Report No. 26841-MA Malaysia Firm Competitiveness, Investment Climate, and Growth Based on the Productivity and Investment Climate Survey conducted between December 2002 and May 2003 with Reference Period of 19992001 June 30, 2005 Poverty Reduction, Economic Management, and Financial Sector Unit (PREM) East Asia and Pacific Region Document of the World Bank MALAYSIA: FIRMCOMPETITIVENESS. INVESTMENTCLIMATE AND GROWTH TABLE OF CONTENTS Executive Summary ......................................................................................................................... i Chapter 1: FirmPerformanceandInvestmentClimate inthe ManufacturingSector ......................... 1 A. Productivity Performance at the National Level.................................................................... 4 B Productivity Perfonnance inManufacturing: Comparisons across Sectors andwith Korea 8 C. Malaysia's Investment Climate.............................................................................................. . 16 D. Determinants ofFirmPerfonnance........................................................................................ 26 E. Conclusions .............................................................................................................................. 38 Chapter 2: FirmPerformanceinthe Services Sector ................................................................................ 39 A. Performanceat the National Level...................................................................................................... B. Industry Performance.......................................................................................................................... 41 47 50 D Investment Climate............................................................................................................................. C. FirmPerformance................................................................................................................................ . E. Skills Shortage..................................................................................................................................... 57 53 F. Innovation Readiness........................................................................................................................... . 62 H. DeterminantsofFirmPerformance..................................................................................................... G Regulatory Regime............................................................................................................................. 64 I.Conclusions.......................................................................................................................................... 71 76 Chapter 3: ImprovingSkills ......................................................................................................................... A. The GapinSkills for Work: The Macro Evidence............................................................................. 78 80 B. The GapinSkills for Work: The Micro Evidence.............................................................................. 89 D. Conclusions andPolicy Implications..................................................................................... C. Determinants o f Skills Match................................................................................................. 100 112 Chapter 4: StrengtheningTechnologicalCatch-Up ................................................................................... B. How DoFirmsMoveUpthe Technology Spectrum?......................................................................... A. International Comparisons:StockTaking of ExistingMeasuresofTechnological Performance......117 118 123 D. Firm-Institution Collaboration ........................................................................................................... C. What are the Determinants of Technological Activities amongMalaysian Firms?............................ 132 136 F. SourcesofTechnica1Knowledge ....................................................................................................... E. FiscalIncentives ................................................................................................................................. 138 139 G IPRs .................................................................................................................................................... 141 . H. Conclusions:What shouldthe PolicyThrustbe? .............................................................................. 143 Bibliography............................................................................................................................................ 146 Appendices Appendix -Sampling Characteristics........................................................................................................ 151 Appendix 2 - Chapter 2 ........................................................................................................................... Appendix 1-Chapter 1............................................................................................................................ 154 176 Appendix 3 -Chapter 3 ........................................................................................................................... 198 Appendix 4 -Chapter 4 ........................................................................................................................... 210 ListofTables Table 1.1Growth Accounting Data. 1960-2000 ...................................................................................... 5 Table 1.2ISIC IndustryCodes ................................................................................................................. 13 Table 1.3 Number o fDays to Obtain a License/Permit/Approval from Specific Agencies ..................... Table 1.4 Number o f Days to Process Application (Including Payment) for Incentives.......................... 21 22 Table 1.5: Correlations among Measures o f FirmPerformance .............................................................. 27 Table 1.6: Correlates of FirmPerformance ............................................................................................. 28 Table 1.7: Benefit from Reducing Skill Shortages .................................................................................. 35 Table 1.8: Increases inAverage Wages as Skills Expand ....................................................................... Table 2.1: The Performance of the Services Sector Lags behind the Manufacturing Sector ...................37 45 Table 2.2: The Services Sector inMalaysia i s Growing Well but there I s the Potential for 45 Table 2.3: Business-Support Services Have GrownFaster Than the other Services ............................... Improvement ....................................................................................................................... 50 Table 2.4: FirmPerformance inBusiness-Support Services and Manufacturing, 2001 .......................... Table 2.5: Sales and Export Performance o fthe Business-Support Services Sectors inthe PICs ...........51 Table 2.6: Comparisons of Firms' Concerns about the InvestmentClimate ........................................... 52 54 Table 2.7: Percentage o f Services Sector FirmsIdentifying the Indicated Problem as 55 Table 2.8: Training and Services Sector Firms......................................................................................... One o f the Top Three Obstacles to Doing Business inMalaysia........................................ 60 Table 2.9: Wage Premiums to Current and Previous Training, Controlling for Firm Fixed Effects ...................................................................................................................... 61 Table 2.10: SelectedBusiness-Support Services FirmsReport Same Level o f Technological 62 Table 2.1 1:Key Drivers o f Innovation (Adaptation and Creation) ......................................................... Activities as Manufacturing Firms ........................................................................................ 64 Table 2.12 Malaysia Appears to Have One of the Most Restrictive Regimes for Services Overall .................................................................................................................. 68 Table 2.13 Measures o f Discriminatory Restrictions Against ForeignFirms inServices are 69 Table 2.14: Higher Restrictions Translate to Higher Price Effects inMalaysia ........................................... Higher inMalaysia ............................................................................................................. 69 Table 2.15 Qualitative Comparison o fBusiness-Support Services Regulation inMalaysia, Taiwan, Korea and Chile .................................................................................................... 70 72 Table 2.17: Determinants o f FirmPerformance ...................................................................................... Table 2.16: Correlates o fFirmPerformance ........................................................................................... 75 Table 2.18 Domestic Firms Tend to Benefit from the Presence o f Foreign Firms .................................. 76 Table 3.1: Stock o f Educational Attainment o f the Total Population Aged 25 and Over......................... 83 83 Table 3.4: Returns to Education, Spline SpecificationControlling for FirmFixedEffects .................... Table 3.3: InternationalTest Scores in Science in Selected Countries..................................................... Table 3.2: Scientists and Engineers inR&D(per million people)............................................................ 88 91 Table 3.5: Wage Premiums to Training, Controlling for FirmFixedEffects........................................... 93 Table 3.6: Shlls Workers Lack Most ...................................................................................................... 95 Table 3.7: Skills Mismatch ........................................................................................................................ 98 Table 3.8: EducationMismatch- Level o f Education ............................................................................. Table 3.10: Unemployed Graduates Registered under the Training and Attachment Scheme.................103 Table 3.9: EducationMismatch: Relevant Education............................................................................... 99 99 Table 3.11: Skills Content o f Education .................................................................................................. 99 Table 3.12: Wage Premiums to Training by Content, Controlling for FirmFixedEffects ...................... Table 3.13: Training Incidence Over Time inMalaysianManufacturing ................................................ 105 106 112 Table 3.15: Matching the Excess Demand for Skilled Labor inMalaysia ............................................... Table 3.14: Quality of Firms' Three Most Used Skills Development Institutes ..................................... 114 Table 4.1:Matrix o f Competitive Interactions between China and the Rest o f the East Asian Region inExports Markets ....................................................................................... 122 Table 4.2: In-House Skilled Resources Matter for Technology Adoption .............................................. Table 4.3 :What Drives Technology Adaptation within Firms? .............................................................. 133 135 135 Table 4.5 :What i s the Explanation for Technology Creation? ................................................................ Table 4.4: FirmsAre Not Skipping Stages .............................................................................................. 136 Table 4.6: Satisfaction Rating of RTIusers (percentage) ........................................................................ 137 Table 4.7: Sources of Technological Innovations .................................................................................... 140 Table 4.8: Strengthening the IPRRegime Would Induce R&D Active Firmsto D o More .................... 142 ListofFigures Figure 1.1 Sources o f Growth inMalaysia. 1960-2000............................................................................ 6 Figure 1.2 International Comparisons of Growth and Productivity.......................................................... 6 8 Figure 1.4 Malaysia's Manufacturing GrowthinInternationalPerspective............................................. Figure 1.3 Future Growth Scenarios. 2003-2020 ..................................................................................... 10 Figure 1.5 Manufacturing Productivity Growth inMalaysia.................................................................... 10 12 Figure 1.7 Correlates o f Productivity Growth inthe 1990s...................................................................... Figure 1.6 Sectoral ManufacturingProductivity Growth inMalaysia...................................................... 14 Figure 1.8 Comparing Sectoral Productivity Growth inMalaysia and Korea.......................................... 15 Figure 1.9 Malaysia Fares Well inInternationalComparisons o fthe Investment 18 Figure 1.10 Firms' Concerns about the Investment Climate .................................................................... Climate Based on PICS Surveys ......................................................................................... 20 21 Figure 1.12 Are Labor Regulations an Obstacle to Firms?....................................................................... Figure 1.11Why are FirmsOverstaffedor Understaffed? ....................................................................... 21 Figure 1.13 Malaysia's Higher Education Stocks Lagbehind Its Level o f Development........................ 23 Figure 1.14 InternationalPerspectives on Regulatory Burdens inMalaysia............................................ 25 Figure 1.15 Regional Differences inthe Investment Climate .................................................................. 33 Figure 1.16 FirmPerformance and the Investment Climate..................................................................... 34 Figure 1.17 Benefit from Relaxing Skill Shortages and Investment Climate Perceptions on Slulls .......................................................................................................... 36 Figure 2.1 Employment Share o f Services Sector: Selected Countries, 1981 and 1999 ......................... 42 Figure 2.2: Share of Services Sector inGDP in 1981 and 1999............................................................... 43 Figure 2.3: Share o f Services Sector inGDP Relative to GDP per Capita............................................... 44 Figure 2.5: Cross-Country Comparisons o f Labor Productivity o f the Services Sector ........................... Figure 2.4: Labor Productivity o f the Services Sector: Selected Countries in 1984 and 1999.................43 44 45 Figure 2.7: Value of Services Export in Selected Countries in 1981and 2001 ....................................... Figure 2.6: Labor Productivity o f All Sectors in Selected Countries in 1984 and 1999........................... 48 48 49 Figure 2.10: Investment Climate -Results from Open-Ended Question ................................................ Figure 2.9: IndustryPerformances inServices Sector, 1981 and 2001 ................................................... Figure 2.8: Value o f Services Import in Selected Countries in 1981and 2001 ....................................... 53 Figure 2.11: Investment Climate -Results from Closed Question .......................................................... Figure 2.12: Open-Ended Question - Sample o f All Domestic Firmsthat Responded (175) ..................54 56 Figure 2.13: Open-Ended Question -Sample o f All Firmswith Positive Foreign Equity that Responded (53) .......................................................................................................... 56 Figure 2.14: EducationalAttainment of Workers inthe Business-Support Services i s High Figure 2.15: Time Taken to Fill a Vacancy inthe SelectedBusiness-Support Services Sector ...............58 Compared to the Manufacturing Sector ........................................................................... 58 Figure 3.1: Completion Rates at Different Levels o f Education, 2000..................................................... 84 85 Figure 3.3 : Malaysia Performs Poorly inTertiary Education................................................................... Figure 3.2: Malaysia Has PerformedWell in Secondary Education: ....................................................... Figure 3.4: Malaysia Shows a Persistent Gap inTertiary Enrollment Rates in2000 ............................... 85 86 Figure 3.5 Unemployment Duration Population with University Education (Labor Force Survey) ........................................................................................................... 87 Figure 3.6: MeanLogHourly Wage by Years o fFormal Education ....................................................... Figure 3.8: The Extent o f EnglishSlulls Mismatch.................................................................................. Figure 3.7: The Premium for Tertiary Education I s Higher inMalaysia Than inOECD Countries ........92 95 97 Figure 3.9 Employer-ProvidedTraining Incidence inMalaysia............................................................... 106 Figure 3.10: Plant-Level Hire and QuitRates by Occupation and FirmSize, 2001................................. 107 Figure 3.11: Share of Plants UsingSlulls Development Institutes by Sector........................................... 112 Figure 4.1: Innovation Outputs: Deviation from "Expected" Values ...................................................... 121 123 Figure 4.3: FirmsReporting R&DActivity .............................................................................................. Figure 4.2: Mediumand HighTechnology Products inDirectly Threatened Exports ............................ 125 Figure 4.4: The Technology Spectrum inMalaysia ................................................................................. 127 Figure 4.5: What are Firms (not) Doing? ................................................................................................. 128 Figure 4.6: Technological Activity Among Firms.................................................................................... 128 128 Figure 4.8: Size, Exports, and Technology Performance ......................................................................... Figure 4.7: Which Sectors Have More Inventors?.................................................................................... 129 Figure 4.9: R&D and LicensinginMalaysia............................................................................................ 130 Figure 4.10: Determinants o f Technological Activities............................................................................ 132 Figure 4.11: Usage o f RTIsby Inventors ................................................................................................. 137 Figure 4.12: InventorsUsingFiscal Incentives ........................................................................................ 139 ListofBoxes Box 1.1: Recent Policy Responses to Improve the InvestmentClimate................................................... 3 Box 2.1: What i s Services? ....................................................................................................................... 41 Box 2.2: Costs o f Business Services inMalaysia are Relatively High .................................................... 46 80 Box 3.2: The MalaysianEducation System.............................................................................................. Box 3.1: Measuring Skills for Work ........................................................................................................ 102 Box 3.3: Ten Years o f HRDFinMalaysia (1993-2003) .......................................................................... 110 Box 3.4: Recent Government Measures to Enhance HumanResource Development ............................. 111 Box 4.1:Methodology Underlying the GCR............................................................................................ 119 Box 4.2: Technology Audit o f Firms inMalaysia: Where D o Firms Lie?............................................. 131 Box 4.3: Initiatives Under the SecondNational Science and Technology Policy.................................... 144 ACKNOWLEDGEMENTS This report was prepared by Ejaz Ghani and Albert G. Zeufack (Co-Task Leaders), Ana Margarida Fernandes, Hiau Looi Kee, Aart C. Kraay, Manjula M. Luthria, and Charles C. Udomsaph. The peer reviewers were Barry Bosworth, Indermit Gill, h i n d Subramanian, and Shahid Yusuf. Albert G. Zeufack took the lead in the design and implementation o f the Productivity and Investment Climate Survey. Production o f the report was supported by Hedwig Abbey. Editorial assistance was providedby EmilyEvershed. The preparation of this report benefited greatly from the assistance and cooperation provided by the Government o f Malaysia. In particular, the authors would like to thank officials in the Economic Planning Unit (EPU), the Ministry o f Finance, the Ministry o f Trade and Industry, the Malaysian Industrial Development Authority, Bank Negara Malaysia, the Ministry o f Human Resources, and the Department of Statistics for their generous help inproviding data and explanations to the mission. A World Bank mission visited Malaysia from June 9 to June 18, 2003. The diagnostics o fthe firm level data was carried outjointly with the EPU staff duringthis period. This report is part of the advisory services being provided by the World Bank at the request o f the Government o f Malaysia. The first phase o f advisory services includes (i)the firm-level surveyand(ii) report on firmcompetitiveness. The second the phase i s aimed at implementing the findings of this report and will include (i) TA to reform skills and technology institutes, and (ii) a detailed study on services as a source o f growth. EXECUTIVESUMMARY 1. The objective of this report is to explore firm competitiveness, the investment climate and growth in Malaysia. The analysis is grounded in data from the Malaysia Productivity and Investment Climate Survey (PICS), a stratified random survey of 1,151 firms conducted by the Malaysian Department of Statistics and the Economic Planning Unit from December 2002 to M a y 2003 in collaboration with the World Bank. The objective o f the survey was to identify the key constraints to competitiveness as perceivedby the firms inthe manufacturing and selected business support services sectors. A total o f 902 firms in the manufacturing sector and 249 firms in selected business support services sector were surveyed. The core questionnaire on the investment climate covered issues related to the business environment. As Malaysia will need to shift from an investment led growth strategy to a productivity led growth strategy, the survey also probed firm's perceptions o f what i s holding back productivity growth. Firms were asked to assess the skills and technology programs, the two pillars o f productivity enhancing programs in Malaysia. Finally, as the services sector is expected to play a larger role in supporting future growth, business support services were also asked to assess the investment climate. The findings o f this report served as inputs into the mid- term review o f the EighthMalaysia Plan, completed in October 2003. It i s noted that the timing o f the survey, which coincided with the invasion o f Iraq and the outbreak o f SARS, would have strongly influenced the perception o f firms with respect to the macroeconomic environment and future growth. In addition, given that the survey was completed inMay 2003, it certainly does not capture the full impact o f the measures undertakenby the Government o f Malaysia (GoM) in 2003 andthe positive sentiment o f investors that has followed the measures. 2. The firm-level analysis pointed to key concerns regarding the regulatory burden, skills shortages and weak innovation capabilities which could be addressed by the following recommendations: (i) to carry out a detailed assessment o f the regulatory environment and make changes to reduce the regulatory burden; (ii)to accelerate tertiary education; (iii) to rebalance the economic and social objectives o f the education policy; (iv) to further reduce restrictions on the import o f skills; and (v) to scale up appropriate skills and technology programs. These measures are summarized inthe matrix attached to this Executive Summary. If they are implemented, Malaysia could achieve its Vision 2020 and make the transition from growth based on factor accumulation to growth based on higherproductivity. 3. The Government of Malaysia is acutely aware of concerns regarding the investment climate and skills shortages. Inthe past years it has taken active steps to reduce the regulatory burdens and streamline the business environment, with the objective o f raising investment and growth. It has also taken steps to increase the supply o f skilled workers and enhance the employability o f the human resources. On M a y 21, 2003, the Government o f Malaysia announced a four-pronged stimulus package intended to promote private investment, strengthen i competitiveness, develop new sources o f growth, and enhance the effectiveness o fpublic service delivery. In order to promote private investment, the Government has announced a variety o f incentives for small and medium enterprises and has liberalized previous restrictions on foreign investment, particularly in the manufacturing sector. Most significantly, the Government has announced the elimination o f restrictions on the fraction o f manufacturing firms' equity that can be held by non-Malaysians inthe case o f new ventures or the expansion o f existing ventures. It has also significantly relaxed restrictions on the employment of expatriates inthe manufacturing sector. 4. The impactof these actions is reflectedinthe recoveryindomestic investmentas well as FDI more recently. Both the applications for FDI in the manufacturing sector and FDI measured through the cash balance o f payments showed significant increases in the second half o f 2003. Inaddition, applications for domestic investment inthe manufacturing sector doubled in 2003 compared to the level recordedinthe previous year. 5. This report is structured as follows. The first two chapters provide an assessment o f the investment climate. The last two chapters examine how productivity enhancing programs-skills and technology-have performed inMalaysia. The diagnosis i s based on the survey data as well as on cross-country comparisons. Chapter 1 reports on the investment climate and firm performance in the manufacturing sector. Chapter 2 presents the findings on the investment climate and firm performance inthe services sector. Chapter 3 examines skills performance, and the contribution o f training and education to skills performance. Chapter 4 provides a diagnosis o f the technological capability o f firms. This report also identifies areas that require further work (e.g., carrying out a detailed assessment o f key elements o f the regulatory environment; undertaking a larger firm level survey to achieve a better assessment o f the services sector as a source o f growth). The InvestmentClimate 6. A strong investment climate is needed to provide firms with incentives to invest, innovate, and grow.A good investment climate consists o f an environment inwhich firms can realize their productive potential to grow, create jobs, and provide the goods and services demanded by consumers at home and around the world. This requires a sound macroeconomic environment, the availability o f high quality physical and financial infrastructure, skills and technology, the existence o f a legal and regulatory framework that promotes competition, and a governance environment that overcomes bureaucratic inefficiencies and enables firms to access the factors that they need in order to grow. The PICS extensively probed firms' perceptions o f the investment climate in which they operate. This report focuses on key aspects o f the investment climate as identified by firms inMalaysia. 7. Malaysia's investment climate compares favorably with the Asian countries in which PICSshavebeenimplemented. Malaysia's investment climate is rated substantially better than that o f even the most dynamic parts o f China. Firms were asked to rate 17 different dimensions o f the investment climate as a constraint to doing business. For every one o f the 17 categories, firms in China are more likely than firms in Malaysia to rate the factor as a serious constraint. As another example, the number o f days required to clear customs inMalaysia is three days, but the time requiredi s over twice as long in China (Figure 1). The same i s true when Malaysia i s 11 compared with other indicators o f the investment climate in countries that have comparable data from the PICS. ~ ~ ~ Figure1: MalaysiaFaresWell inInternationalComparisonsofthe InvestmentClimateBasedonPICS Surveys Malaysia Guangzhou (China) Shanghai (China) China (Overall) India (Overall) Gujarat (India) Maharashtra (India) Bangladesh Pakistan 0 2 4 6 8 10 12 14 16 Days to Clear Customs (Average of Imports and Exports) 35 1 'Tax Rates 30 i 8E" 'Electricity 'RegUncert. 2 5 - 'Ik%%Instability 'TaxAdmin / 'Customs eg 20 I 9 ne Yransport Labor Re/ 15 5 10 15 20 25 30 Malaysia Source: Investment Climate Survey StandardTables, World Bank. 8. But not all is well with Malaysia's investment climate. If we look closely at firms' responses, a consistent pattern of concerns emerges. Firms are concerned about (i)shortages of skilled workers, and (ii) regulatory burdens, as key adverse features of the investment climate. Consider, for example, labor market regulations. Firms note that the difficulty in hiring local workers, the regulations for hiring foreign workers and skill shortages are the reasons why they are understaffed. Firms face considerable uncertainty regarding the length o f time required to complete bureaucratic procedures. For example, the time required to obtain a license from the landoffice is more uncertain than that requiredto obtain animport permit. ... 111 Figure 2: FirmPerformance and the Investment Climate: Percentage of firms identifying skills and regulation as obstacles to their operations Exporters versus Non-Exporters Foreign versus Domestic Firms 90% 80% - 80% 70% - 70% 60% - 60% 50% - 50% 40% - 40% 30% - 30% 20% - 20% 10% - 10% 0% I I 0Yo Exporters Non-Exporters FDI Domestically-Owned 10Skills W Regulation1 10Skills RegulationI 'ource:Malaysia PICS, 2002 9. In addition, for the services sector, international comparisons with nearly 40 countries reinforce the firm-level finding that there i s a relatively higher cost resulting from these regulatory burdens in Malaysia than that seen in most other countries. According to a WTO services trade restrictiveness index (constructed by assigning subjective weights and values that estimate the restrictiveness o f the country's trading regime for services based on the number and severity o f restrictions), for most services industries, restrictions in Malaysia are greater than the averages for Asia, Latin America, and OECD countries. The cross-country evidence as well as the firm-level survey data all point inthe same direction-to concerns about regulatory burdens as having a negative impact on the investment climate inMalaysia. 10. Regulatory impediments are especially harmful in Malaysia because they tend to be felt most by the best performing firms. The best performing firms in the manufacturing sector- large scale firms, exporters, or firms with FDI-voice strong concerns about regulation as an obstacle to their operations. Large firms, which have a better performance in terms o f value added per worker (VAL) and total factor productivity (TFP), identify regulation as a very important obstacle to their operations relative to smaller firms. In fact, 90 percent o f large firms complain about regulation. Inparticular, large firms face a heavy regulatory burden interms o f customs and labor regulations. Similarly, exporter firms, which also perform better in terms o f VAL and TFP, are more concerned with regulation than non-exporter firms (see Figure 2). Firmswith FDI, which have significantly higher VAL and TFP, also consider regulation to be a greater obstacle to their operations than i s the case with domestic firms. Firms in the most dynamic regions (those inPeninsular Malaysia, especially inthe south and west, which perform much better than those on the east coast, in Sabah and in Sarawak) voice more concerns about regulation. Finally, irrespective o f size, export orientation, or level o f FDI, firms identify skill shortages as a main obstacle to doing business-a concern that i s reinforced by the findings throughout this report. iv 11.Firms identifythe investmentclimatein the selectedbusiness support services sector to be more unfavorable for doing businessthan the investmentclimate in the manufacturing sector. Evidence shows that business support service firms that are constrained to have 30 percent or less foreign ownership are less productive, while firms that operate in less restrictive industriesare moreproductive. 12. Firms that use high technology show greater productivity.Evidence from firm-level data shows that firms with newer capital equipment are more productive. For example, firms with a larger fraction o f their machinery that i s computer-controlledhave higher VAL and TFP. 13. Policymakers face trade-offs between scale and efficiency. Firm size measured by firm assets (machinery and equipment) i s positively associated with firm performance. Although one cannot establish a strong causal relationship between firm size and productivity, there i s some preliminaryevidence that large firms inMalaysia tend to be more productive than smaller ones, as they benefit from economies o f scale, face lower transaction costs, and are better able to access skills and technology programs. These results do not sit well with the goal o f the Eighth Malaysia Plan which i s to expand the programs targeted at the SMEs. Malaysia i s implementing a large number o f productivity-enhancing public programs that have been nominally targeted towards the SMEs but are not extensively used by them (see Table 1). This raises a number o f issues. An important one i s whether skills development andtechnology support institutions target SMEs more aggresively, even if this imposes a cost in terms o f efficiency. The Government should adopt a clear objective against which the skills and technology development institutions will be evaluated. Table 1: SME Use of Skills DevelopmentandTechnology Support Institutionsand Incentivesis Low Percentage o f firms that report using the institutions Skills Development Technology Support Technology Institutions Institutions Incentives SMEs 14.7 percent 14.5 percent 1.2percent Large Firms 39.1 percent 17.4 percent 11.4percent Source: Malaysia PICS, 2002. Skills andEducation 14. Shortage of skills is a pervasive problem among the firms surveyed. Irrespective o f location, industry or firm characteristics, the large majority o f firms identify skill shortages as a "severe" or a "very severe" problem. The average time taken to fill a vacancy for a skilled technician takes longer than in China and a few other Asian countries where PICS have been carried out. When firms were asked about the possible causes o f skills shortages, around 70 percent o f managers surveyed identified the insufficient supply of university graduates as the most important reason. The complaints o f the firms regarding skills shortages are found to be consistent with the analysis on returns to education, returns to training, and trends in unemployment. V Figure3: MeanLogHourly Wage by Years of FormalEducation . 2.75 2.5 4 p . 2 5 3 B- 1F 2 l.15 1.5 1.25 , 0 3I 6 9I 10l 11l 12/ 15 161 I 1 Years o f formal education 21 Source: Malaysia PICS, 2002. 15. Firms place a high premium on workers with tertiary education, as evidencedby their much higher wages compared to workers with less education. The return for tertiary education is nearly 18 percent versus 9.5 percent for secondary education and only 4.5 percent for primary education. The premium for tertiary education is higher in Malaysia compared to OECD countries. Figure 3 presents the results o f a semi-parametric regression o f log hourly wages on formal education. The curve connects the average o f the predictedvalues o f log hourly wages for each year o f formal education. There is a significant increase inthe slope for primary, secondary, and tertiary education. 16. The rate of return on training is also large: 10 percent in terms of higher wages for manufacturingworkers with any kind of training. Returns also increase with more training. Workers with training only from their current employers receive the lowest premium (7 percent), those with training only from their previous employers receive higher premiums (11 percent), and those with training from both their current and their previous employers receive the highest premiums (15 percent). The returns to training also differ by level o f workers' education. By industry, the training premiums are concentrated in food processing, rubber and plastics, household electrical appliances, and automobile parts. 17. Malaysia'shigher educationrates are lower than the normfor itsincomelevel. Figure 4 below shows the relationship between real per capita GDP (on the horizontal axis) and the fraction o f the population with a higher education (on the vertical axis). Data for comparator countries refer to 2000, while data for Malaysia are shown for 1980, 1990, and 2000. Inall three periods, Malaysia's higher education rates are lower than the norm for its income level, as indicated by the simple line o f best fit through this graph. Although the gap has narrowed over time, the current fraction o f Malaysia's population completing higher education of 6 percent is still below that o f Thailand or Chile (at nearly 10 percent). 18. Firms are concerned about the quality of Malaysian-educated workers, especially professionals. Managers were asked inthe survey to rate Malaysian professionals against their vi foreign counterparts. Twenty-ninepercent o f managers believe that foreign trained professionals performed better than Malaysian trained professionals. Workers were also asked inthe survey to do a self-assessment and identify important areas o f deficiency infirm-specific skills and general skills. On firm-specific skills, almost half o f the workers surveyed reported English language proficiency as the skill that they lacked most. Information and technology skills emerged as the second most needed skill. Lack o f adequate professionalhechnical skills ranked third. On general skills, workers identified lack o f communication skills to do their job and/or adapt to changes inlabor market conditions. Giventhat general skills are usually better provided through the education system and that firms have little incentive to provide training ingeneral skills, the workers survey also highlightsdeficiencies inthe education system. Figure 4: Malaysia's Higher Education StocksLagBehindIts Level of Development Korea .. .. . Philippines rn I) 0- 1 G 7 8 9 10 13 log(Rea1 Per Capita GDP in2000) Source: Barro and Lee (2000), Penn World Tables. 19. There i s an educationmismatch,especially at the tertiary level. Because of the shortage o f university graduates, firms are forced to hire workers with a diploma to do the job o f a graduate. The shortage in tertiary education graduates has contributed to sub-optimal hiring policies and loss o f productivity at the plant level. The education background mismatch i s also noticed from the qualifications o f unemployed graduates registered for training schemes. Around 40 percent of the unemployed are in areas that are not o f interest to manufacturing activity. 20. Evidence suggests that increases in the skills of the workforce can have substantial benefits. Fromthe standpoint o findividual firms that are unable to hire as manyskilledworkers as they would like, the benefit in terms o f higher sales i s estimated to be on the order of 10 percent. The industries in which an increase in skilled employment would have the greatest benefits in terms o f increased sales also have larger proportions o f firms reporting concerns about skill shortages. Moreover, to the extent that an expansion inthe supply o f skilled workers would allow Malaysia to expand into sectors that are intensive in skilled workers, there are also likely to be substantial increases inaverage earnings. vii 21. Although Malaysia has a world class skills training infrastructure, the number of firms that use these facilities is low. A large number o f firms reported that the availability o f skills training institutes and the existence o f the HumanResource Development Fund (HRDF), a levy- grant scheme for the retraining and skills upgrading o f employees administered by the Pembangunan Sumber Manusia Berhad (PSMB), was critical to their decision to train. Firms believe that they would provide more training if the processes for training were made more efficient. But firms, especially the SMIs, seldom use the skills training programs. However, the plants that use skills development institutes rate them very high. From the PICS survey, more than 75 percent o f managers rank the top three institutes they use as o f "good quality," and around 20 percent believe they are o f "very good" quality. Innovation Readiness 22. Malaysian firms have benefited from technology diffusion. With a backdrop o f openness to trade and FDI, Malaysians have been technologically active in terms o f undertaking technology adoption and adaptation. The PICS data show that nearly 40 percent o f firms report making some improvements to their products or processes, and over 60 percent o f firms have upgraded machinery and equipment inthe last two years. 23. But there is room for expanding technological diffusion among firms. The data also show that nearly 70 percent o f firms have not introduced a substantially new technology and over 50 percent have not been able to enter new markets on the basis o f quality or cost improvements. The use o fnew or computerizedtechnology varies very widely among firms, and inter-sectoral variations in technological activity are also quite notable. Non-exporters and SMEs lag behind exporters and large firms intechnological activity. While this i s to be expected, understandingthe extent o f linkagesbetween large multinational corporations (MNCs) and small domestic firms i s being taken seriously in Malaysia and should continue to be the focus o f investment andtechnology policies. 24. In-house skilled resources, fiscal incentives, and inter-firm collaboration influence technology diffusion. The empirical analysis o f the data shows that humanresources dedicated to undertaking technological activity in a firm, the use of fiscal incentives by the firm, and the level o f inter-firmcollaboration are significant determinants of the level o f technology adoption and adaptation at the firm level. Hence, as a matter of public policy, attention to the quality and availability o f skills in the labor force, and efforts to address any regulatory or institutional constraints preventing the take-up o f fiscal incentives by firms (outlined in the policy matrix), are proposed as priority items for policy work. However, it i s not clear what role exists for public policy in facilitating firm-firm collaboration; hence, this i s identified as a role for the private sector to assume. 25. The move from technology diffusion to technology generation is also overdue. After controlling for income level, Malaysia exhibits a "deficit" in R&D expenditures as well as scientific publications and patents - measures o f inputs and outputs o f technology generation, respectively. Further evidence o f this i s provided by the findings o f a qualitative technology audit conducted on a sample o f 81 Malaysian firms to complement the data emerging from the 'There maybe sector-specific reasons driving this variation. viii PICS survey. These industry case studies corroborate the PICS finding that most Malaysian firms are reactive (Le., firms that recognize the need to keep up with technology, but lack skills andcapabilities and are slow inresponding) rather than creative (Le., knowledge-intensive firms with hlly developed capabilities that are able to redefine technology frontiers, challenge existing business models, and create new markets). 26. Collaboration with technology institutions matters for technology generation. The empirical analysis shows that collaboration with research or technology institutions (RTIs) increases the probability that a firm will be able to produce technological innovations. The incidence o f such collaboration i s quite low in Malaysia, and firms surveyed under the PICS reported some constraints to greater collaboration which should receive attention in the follow- up diagnostic work that is intended to produce specific recommendations for improving the design and delivery of services from RTIs. Tentatively, a review o f the governance structures o f RTIs, the nature o f financial incentives for collaboration on the part o f RTIs, and the outreach programs for improving awareness among firms are identifiedas starting points. 27. Learning-by-hiringand the improved appropriability of investments are identified as important to technology creating firms. Firms surveyed under the PICS have identified new hires in a firm as a significant source o f technical knowledge. This also conforms to the emerging consensus in the broader innovation literature which identifies "tacit" (rather than "coded") knowledge as the key to building the innovative capacity o f firms. Hence, some attention to increasing the international mobility o f labor intechnology-intensive sectors may be worth considering. The increased certainty o f the contractual environment through an improved intellectual property rights regime (the regime i s WTO-compliant but the efficiency o f the Patent Office needs to be enhanced: some items are suggested in the policy matrix at the end o f this summary) would provide the needed impetus for undertaking risky investments in innovation andwould also reinforce learningfrom themobility o f labor. ISSUES FOR CONSIDERATION 28. The matrix that follows (Table 2), identifies the problems to be considered and suggests measures that could address the concerns o f the firms. The Malaysian authorities are already addressing these concerns. The key areas o fpolicy challenges are the following: Strengthening the investment climate (e.g., reducing the high cost o f the regulatory burden) 0 Accelerating tertiary education 0 Rebalancingthe economic and social objectives o f the education policies 0 Reducingthe restrictions on the import o fprofessional skills 0 Scalingup the relevant skills andtechnology programs. ix 6iY 3 2 3 - El C L BY rd eB8 6 t 0 L B C C 3 v) 3 9 3 U a8 Y 3 4M x e. e * e . . e e e e . . e . . a U8 d I 'w 8 . e . e . . . . I :#8 a s 0 . 0 0 0 . 0 0 *0 1. FIRMPERFORMANCEAND INVESMENTCLIMATEIN THE MANUFACTURINGSECTOR The recent financial crisis clearly demonstrated the need for resilience, both economically and socially. I t served to remind us that we can no longer depend onpastformulas for ourfuture success. During the Eighth Malaysia Planperiod, we will be faced with even greater challenges from globalization and liberalization as well as the rapid development of information and communications technology. We will have to enhance the competitiveness of the economy, strengthen our economic resilience and improve our total factor productivity. Dr.MahathirBinMohamad Prime Minister, Malaysia EighthMalaysiaPlan2001-2005 April 23,200 1 1.1 This chapter explores the links between firm performance and the investment climate in Malaysia. The chapter begins by looking at past trends in economy-wide and manufacturing productivity growth. This evidence indicates that Malaysia's past rapid growth has been driven primarily by factor accumulation, combined with modest rates o f total factor productivity (TFP) growth on the order o f about 1percent per year. While these modest rates o f TFP growth are comparable to those o f other high-performing East Asian economies, they may not be sufficient for sustained long-run growth. A simple projection exercise suggests that over the next 20 years increases in the quality o f the workforce and higher rates o f productivity growthare requiredto support fbture growth. 1.2 A strong investment climate is needed to provide firms with incentives to invest, innovate, and grow. A strong investment climate combines four inter-related components: the first i s a sound macroeconomic environment, the second i s the availability o f high quality physical and financial infrastructure, skills and technology, the third is the existence o f a legal and regulatory framework that promotes competition, and the fourth is a governance environment that overcomes bureaucratic inefficiencies and enables firms to access the factors that they need inorder to grow. 1.3 This chapter next documents the evidence on perceptions o f the investment climate in Malaysia based on the data from the Malaysia Productivity and Investment Climate Survey (PICS). The PICS i s a stratified random survey o f 1,151 firms conducted by the Malaysian Department of Statistics and the Economic Planning Unit from December 2002 to M a y 2003 in collaboration with the World Bank. The objective o f the survey was to identify the key constraints to competitiveness as perceived by the firms in the manufacturing and selected business support services sectors. A total o f 902 firms inthe manufacturing sector and 249 firms 1 in the selected business support services sector were surveyed. The findings of this survey served as inputs into the mid-term review of the Eighth Malaysia Plan, completed in October 2003. Given the timing o f this survey, which was completed inM a y 2003, it certainly does not capture the full impact o f the measures undertakenby the Government o f Malaysia in 2003 and the positive sentiment o f investors that has followed the measures. It is also noted that the timing o f the survey, which coincided with the invasion o f Iraq and the outbreak o f SARS, would have strongly influenced the perception of firms with respect to macroeconomic uncertainty. 1.4 The firms' views captured by the PICS focus on three areas o f concern: (i) shortages the o f skilled workers, (ii) the overall macroeconomic environment, and (iii) regulatory burdens the as key adverse features o f the investment climate. International comparisons based on a variety of other sources also point inthe same directions. 1.5 Firm performance differs sharply across regions and types of firms. The fourth section o f this chapter investigates differences in firm performance using data from the PICS. Using a variety o f measures o f firm performance, the evidence shows that firms in Peninsular Malaysia, especially in the south and west, perform much better than those on the east coast, in Sabah, and in Sarawak. Inaddition, older firms, larger firms, exporting firms, and firms with FDIall tendto havebetterperformances inavariety o fdimensions. 1.6 These differences in performance across types of firms can be linked to differences in the perceptionsof the investmentclimate. The worst performing regions are also perceived by firms as having the worst investment climate, with macroeconomic uncertainty' as a key obstacle to doing business. Incontrast, the key obstacle for firms inthe most dynamic regions i s heavy regulation, and the same is true for large firms, exporter firms and firms with FDI. Finally, irrespective o f location, industry or characteristics, firms point to skill shortages as a major problem faced in their operations. This suggests that the benefits o f easing regulatory burdens may be especially large, as these burdens are disproportionately borne by the best- performing firms inthe economy. 1.7 The Governmentof Malaysia is acutely aware of concernsregardingthe investment climate and skills shortages. Inthe past years it has taken active steps to reduce the regulatory burdens and streamline the business environment, with the objective o f raising investment and growth. It has also taken steps to increase the supply o f skilled workers and enhance the employability o f the human resources. On May 21, 2003, the Government o f Malaysia announced a four-pronged stimulus package intended to promote private investment, strengthen competitiveness, develop new sources o f growth, and enhance the effectiveness o f public service delivery (Box 1.1). In order to promote private investment, the Government o f Malaysia has announced a variety o f incentives for small and medium enterprises and has liberalized previous restrictions on foreign investment, particularly in the manufacturing sector. Most significantly, the Government has announced the elimination o f restrictions on the fraction o f manufacturing firms' equity that can be held by non-Malaysians, inthe case of new ventures or the expansiono f existing ventures. It has also significantly relaxed restrictions on the employment o f expatriates inthe manufacturing sector. 'Thetiming of the survey, which coincided with the invasion of Iraqand the outbreak of SARS in2003, would have strongly influenced the perception of fmwith respect to macroeconomic uncertainty. 2 3 1.8 It must be emphasized that the results from the PICS are valid only for the reference period of the survey. The impact of actions undertakenby the Government has generatedpositiveresultsand is reflectedin the recoveryin domestic investmentas well as FDI more recently. Both the applications for FDI in the manufacturing sector and FDI measured through the cash balance o f payments showed significant increases in the second half o f 2003. Inaddition, applications for domestic investmentinthe manufacturing sector doubled in 2003 compared with the previous year. A. PRODUCTIVITYPERFORMANCEAT THENATIONAL LEVEL 1.9 Although Malaysia has enjoyed rapid growth over the past 40 years averaging about 4 percent per capita annually, total factor productivity (TFP) growth has averaged only 1percent per year. These modest rates o f aggregate TFP growth are similar to those o f other rapidly growing East Asian economies. Malaysia can probably continue to grow rapidly for some time with these rates o f TFP growth, but inthe long runfaster productivity growth will be necessary to support sustained rapid growth. Sources of Past Growth 1.10 Over the past 40 years Malaysia has witnessed impressive growth inper capita GDP and commensurate improvements in a variety o f development indicators. However, as is the case for the other strong-performing East-Asian economies, factor accumulation has been the predominant source o f growth while improvements inTFP have been relatively modest. 1.11 This can be seen from the data shown inTable 1.1,which provide information onrealper capita GDP,capital accumulation, demographic change, and improvements inlabor force quality over the period 1960-2000. Duringthis time, per capita incomes increased almost five-fold, to their current levels o f around 10,000 purchasing power-adjusted 1996 U S dollars in 2000. This growth has been accompanied by very rapid capital accumulation, averaging 5.7 percent per year in per capita terms. There have also been sharp increases in the numbers and skills o f the workforce. The working-age share o f the population aged 15-64 has increased from 51 percent to 61 percent, and the average number o f years o f schooling o f the working-age population has increased dramatically from around 2.3 years in 1960 to 8 years in2000. 1.12 Increases in capital and in the numbers and quality of workers have contributed substantially to past growth. A simple growth accounting exercise, which is described in detail in Appendix 1A, and i s summarized in Figure 1.1, suggests that the bulk o f past growth has been driven by factor accumulation, and that productivity growth has been quite modest, moderating from 1.2 percent per year during 1960-1980 to 0.5 percent per year during 1980- 2000. While the exact estimates o f TFP growth that can be obtained from an exercise such as this can vary with the assumptions used to construct them, the available evidence appears to be 4 consistent with fairly modest and possible declining TFP growth for Malaysia over the past 40 years.2 Table 1.1: GrowthAccounting Data, 1960-2000 - Levels AverageAnnual Growth 1960 1980 2000 1960-1980 1980-2000 GrowthAccounting Data GDP Per Capita 2,119 4,876 9919 4.0% 3.9% Capital Stock Per Capita 2,119 6,733 20,994 5.6% 5.7% Workers Per Capita 0.35 0.38 0.41 0.5% 0.4% WorlungAge Pop.Share 0.51 0.57 0.61 0.5% 0.3% Participation Rate 0.67 0.67 0.68 0.0% 0.1O/O Years of Schooling 2.3 4.5 8.0 3.2% 3.4% Contributions to Per Capita GDP Growth Capital Stock Per Capita 1.7% 1.7% Workers Per Capita 0.3% 0.3% Schooling 0.7% 1.5% TFP 1.2% 0.5% Note: See Appendix 1.A for variable descriptions and data definitions. Source: Appendix 1.A. 0 These estimates o f TFP growth net out the effects o f increases in schooling, which are often included as part of TFP in Malaysian estimate^.^ In order to make comparisons with many Malaysian estimates, the contribution o f schooling to growth will need to be added. From Figure 1.1, this i s roughly on the order o f 1percent per year over the entire period. 0 An important factor drivingthe reduction o f TFP growth inthe 1980-2000 period relative to the 1960-1980 period i s the acceleration o f growth inhuman capital. Duringthe 1990s there have been sharp increases in school enrollment, especially in higher education, which have led to especially rapid growth inschooling duringthis latter period. 0 Incontrast with the previouspoint, the effects ofthe financial crisis inEast Asia in1997-1998 have little effect on productivity growth averaged over a 20-year period. Although estimated productivity growth dipped sharply during the crisis, excluding these years does not change the conclusionthat TFP growth i s lower inthe latter 20-year period relative to the former. Interms ofinternational comparisons, Malaysia's productivity growthisnotunprecedented, as shownin Figure 1.2. Malaysia's average productivity growth o f around 1percent per year i s slightly slower than that inKorea, Singapore and Thailand, and slightly faster than that inIndonesia. It i s also interesting to note that Malaysia's productivity growth i s very similar to that o f several industrialized economies such as the United States, Canada, and the UnitedKingdom. 'Theslowdownin measured productivity growth inpart reflects faster growth inhuman and physical capital, which may inthe future contribute to higher productivity growth. See, for example, Malaysia National Productivity Corporation (NPC). The NPC's estimate o f growth o f TFP for the period 1981-2000 is 2.8 percent. The higher growth rate couldbe attributed to the inclusion of schooling inthe measurement o f TFP as well as the difference inmethodology usedby the NPC and the World Bank. 5 Figure 1.1: Sources of Growth inMalaysia, 1960-2000 4.0% 3.0% ., . . . . . . . . . ................... . . . . . . . . . 2.0% 1.0% 0.O% 19601980 19802000 ICapital,StockPerCapita KlWorkers PerCapita 0Schooling 0TFP I Source: See Appendix l.A. Figure 1.2: International Comparisonsof Growth and Productivity Average Annual TFP 3 - Growth, 1960-2000 . 2.5 China - . 2 - .Taiwan ...... . . 1.5 . - 9Singapore 8 . ' ThailamlSouth 1 - .. . .!ndia Korea . % . . S$L&gn . Malaysia . # 0.5 - ='Bf$Qadesh Indonesia . m . ...... . I .. I V n 1 I I I I -2 -1 . 0 m,. Philippine2 AverageAnnual Grow% in Real GD8 4 .. -0.5 -, .... Per Worker, 1960-2000 I -1 - Notes: This graph shows the relationshp between real GDP growth per worker and total factor productivity growth, for a sample o f 73 countries over the period 1960-2000. These estimates were constructed by Bosworth and Collins (2003) usinga methodology similar to the one described inAppendix 1.A. Source: Bosworth and Collins (2003). 6 Implications for Future Growth 1.13 There are naturallimits to growth based on factor accumulation. While aggregate growth accounting exercises such as the one above are necessarily crude, they can be useful for thinking about future long-run growth prospects. What are the implications for future growth given that the past history o f rapid growth was driven primarily by factor accumulation? Consider, for example, years o f schooling. Inview o f past trends, it is reasonable to think that over the next 20 years until 2020, average educational attainment can rise to levels observed in highlyindustrialized such as the United States @e,, an average o f around 12 years o f schooling per person). Similarly, available demographic projections suggest that the upward trend in the working-age population share is not yet exhausted, and that the share o f the population aged 15- 64 will peak at 68 percent by 2020, up from the current level o f 61 percent. Incontrast, labor force participation rates are already high at near 70 percent o f the working-age population and are unlikely to increase further. 1.14 Increases in productivitygrowthwill have to be a more importantsource of growth inper capita incomesover the long run.What will the future growth rates be under alternative scenarios for investment rates and productivity growth? Four scenarios are considered below. We consider a "low investment" scenario inwhich investmentrates remain at their current low levels o f around 20 percent of GDP, while in the "high investment scenario" we assume that investment rates rise to their average during the first half o f the 1990s, at 27 p e r ~ e n t .We ~ also consider a "low TFP growth" scenario o f 0.5 percent per year (corresponding to the average we observed in Malaysia during the 1980s and 199Os), and a "high TFP growth" scenario of 1.5 percent per year (corresponding to Malaysia's average duringthe period 1960-80). 1.15 Figure 1.3, which summarizes the results o fthis exercise, highlights several implications: 0 Even under conservative assumptions about investment and productivity growth, there remains scope for considerable per capita GDP growth inMalaysia. Looking across all scenarios, per capita GDP growth projections range from a very respectable 3 percent per year to 5 percent per year. 0 Not surprisingly, higher investment rates and higher TFP growth rates can contribute to faster growth. The more important point is the relative importance o f capital accumulation and TFP growth in these scenarios. Consider, for example, the difference between the low-TFP growth and high-TFP growth scenario in the last two sets o f columns, which correspond to growth o f 3.6 percent in the former, and 5 percent in the latter, during 2010-2020. In contrast, the difference between the low-investment and high-investment scenarios i s much smaller: 4.6 percent versus 5 percent growth over the same period. This reflects the assumptionunderlying these scenarios to the effect that the return to capital decreases as capital accumulates, so that progressively capital accumulation contributes less and less to growth. Thus in the long run, increases in productivity growthwill be a more important source o f growth inper capita incomes. These investment rates are expressed in purchasing-power-parity adjusted terms, as reported in the Penn World Tables Version 6.1. Domestic currency-denominated investment rates are roughly 10 percentage points higher, reflecting the Penn World Tables (PWT) estimate that capital goods prices are closer to world levels in Malaysia relative to the other goods and services that comprise GDP. 7 0 It is worth noting that the optimistic estimates o f TFP growth are very much at the high Figure 1.3: FutureGrowth Scenarios, 2003-2020 iI 6% Ic] 2003-2010 I 5% ~2011-2020 4% 3% 2% 1% 0% I I 1/"=2[1%, Ifl=20%, Ifl=27%, Ifl=27%, TFPG=[1.5% TFPG=1.5% TFPG=[1.5% TFPG=1.5% Note: This graph shows fbture per capita real GDP growth scenarios for the indicated periods and assumptions. Source: Appendix 1.A. end o f those observed inhistorical experience. InFigure 1.2 we have seen that the vast majority o f countries have estimates o f long-run TFP growth under 1.5 percent. Only a few exceptions stand out, including Taiwan and Ireland with estimated productivity growth rates o f slightly more than 2 percent5 1.16 Projections such as these emphasize an important point: namely, that in the long run, improvements in productivity growth will have profound implications for future growth prospects in Malaysia. At this level o f aggregation, TFP growth reflects much more than improvements in technical efficiency at the firm level. It also captures improvements in the investment climate that create the opportunities and incentives for firms to innovate and grow. Understandingthe strengths and weaknesses o f Malaysia's investment climate, and how it has an impact on firm performance, i s therefore an important ingredient in designing strategies for future growth. B. PRODUCTIVITYPERFORMANCEINMANUFACTURING: COMPARISONS ACROSS SECTORS AND WITH KOREA 1.17 Despite the rapid growth in production, overall productivity growth in Malaysian manufacturing has been modest over the past 20 years, similar to that o fthe economy as a whole. There are stark differences across industries in productivity performance, with some industries consistently doing well (electrical machinery and non-industrial chemicals), but these are also a worrying number o f sectors consistently having negative estimated productivity growth China's very high estimates o f TFP growth inthis figure should be treated with some caution. GDP growth rates inChina are likely to beoverstated, andinadditionat least some ofthe measuredproductivity gains reflect one-time benefits starting from a very low base, owing to dramatic reforms during the 1980sand 1990s. 8 (transport equipment, garments, iron and steel production). Sectors with rapid productivity growth are also expanding as a share o f total manufacturing, suggesting a reasonably efficient allocation o f resources in manufacturing. Average productivity growth for manufacturing i s comparable to that in Korea, but with similar large differences across sectors in their relative performance. 1.18 Over the past 30 years Malaysia's manufacturing sector has seen extraordinarily rapid growth, both inabsolute terms and relative inrelation to the rest o f the economy. Manufacturing production has grown at an average annual rate o f almost 11 percent per year since 1970, substantially faster thanreal GDP growth o f 6.6 percent duringthe same period. As a result, the manufacturing sector has expanded dramatically as a share o f GDP, from 12 percent in 1970 to 30 percent in 2000 (Figure 1.4). Malaysia's manufacturing share o f GDP is similar to that o f Korea and Thailand, but these countries and several others in the region have manufacturing shares that are extremely high, giventheir level o f development. 1.19 Growth in manufacturing has been driven primarily by factor accumulation, with improvements in TFP playing only a small role. Estimates suggest that TFP growth in manufacturing has been quite similar to that o f the economy as a whole, averaging around 1 percent per year. Figure 1.5 summarizes the results o f these calculations for the manufacturing sector as a whole. These estimates o f manufacturing TFP growth are based on manufacturing input and output data from 1968-97, reported in various issues o f the Malaysia Statistical Yearbook, The details o f the econometric techniques usedto estimate TFP growth are described inAppendix 1B. 1.20 A reasonable range of manufacturing TFP growth net of improvements in labor quality is from 0.5 percent per year to about 1 percent per year. As shown in Figure 1.5, manufacturing output growth surged from 8.9 percent during the 1980s to 13.6 percent during the 1990s. The key feature o f this growth acceleration was a sharp increase in investment and capital formation: the growth rate o f the capital stock increased from 13.4 percent to 18.8 percent duringthese two periods. As a result, there was little increase in estimated TFP growth during the two periods, with productivity growth averaging just under 2 percent per year. Moreover, unlike the estimates o f TFP growth for the whole economy reported inthe previous section, these estimates do not reflect the rapid growth in skills o f the workforce in manufacturing. The portion o f the workforce inmanufacturing with 12 years o f schooling (average skills) was 10.2 percent in2001 and 9.6 percent in2002. It is not unreasonable to guess that the skills profile o f workers in manufacturing has evolved in a way that is very similar to the rest o f the economy. Since growth in human capital has contributed about 1.5 percentage points o f growth to the economy as a whole, this means that a reasonable range o f manufacturing TFP growth net of improvements in labor quality is from 0.5 percent per year to about 1percent per year. 1.21 Interestingly, there is not much evidence of a slowdown in manufacturing productivity growth over the period 1980-1997. This contrasts with the evidence o f a slowdown in aggregate TFP growth during the same period. We do not have the data required (e.g., capital stock) to measure TFP growth in services directly. However, Chapter 2 provides some evidence that the value added per worker in services grew more slowly than that in the manufacturingsector over the last two decades. 9 Figure 1.4: Malaysia's Manufacturing GrowthInInternationalPerspective 45% 1 Share of 40% P anufacturing Sector inGDP,China 35% - Thailand Korea . I 30% - .. AMalaysia 1970-2000 25% - Japan 20% - .. Canada 15% - . 10% - .. 9 Greece . 8 . 1 . 5% - . m . ... 1. 0% , Source: WorldBank S I M A Database, PennWorldTables. Figure 1.5: ManufacturingProductivity Growth inMalaysia: Average Growth of Output, Inputs andTFP inMalaysia 1980-1997 0.20 0.15 0.10 0.05 0.00 198G1990 1990.1997 -0.05 0OutputGrowth Labor Growth HMaterials Growth UJCapital Growth HTFP Growth (gr. acc.) llTFPGrowth(OLS) Source: Appendix 1.B. 10 1.22 The aggregate figures for manufacturing as a whole disguise a great deal o f variation in productivity growth across sectors within manufacturing. This i s highlightedin the two panels o f Figure 1.6. The upper panel graphs average productivity growth in the 1980s (on the horizontal axis) against average productivity growth in the 1990s (on the vertical axis), for 28 manufacturing sectors. The lower panel plots TFP growth inthe 1990s (on the horizontal axis) against real output growth (on the vertical axis) for the same sectors. Manufacturing sectors are identifiedby their three-digit ISIC codes (a list o f these codes is provided in Table 1.2). This graph has a number o f striking features: 0 There is a great variation across industries in productivity performance, ranging from -6 percent (Transport Equipment in the 1990s) to 16 percent (Pottery, China and Earthenware Products in the 1990s). It i s likely that at least some o f this variation in productivity performance reflects a variety o f data weaknesses, some o f which are discussed inAppendix 1B. Nevertheless, the evidence does suggest that the modest TFP growth for manufacturing as a whole reflects a combination o f strong performance in some sectors and weak performance inother sectors. 0 There is much less variation in productivity performance over time. Sectors with strong productivity growth, such as Electrical Machinery, Non-Industrial Chemicals, and Rubber Products had strong TFP growth in the 1980s as well as the 1990s, while poor performers such as Transport Equipment and Garments stand out for their slow TFP growth in both periods. The few sectors that show sharp changes in TFP performance over time are also those for which we are less confident in the data quality, and these changes are less likely to reflect true underlying large changes in productivity performance (Footwear, Pottery, etc., andNon-Metallic Minerals). 0 Sectors with faster productivity growth have also, on average, grown more rapidly during the 199Os, as shown in the bottom panel o f Figure 1.6. A core o f seven expanding sectors has had rapid TFP growth and also output growth rates faster than average growth in manufacturing. These sectors include Electric Machinery, Non- Electric Machinery, Rubber, and Plastics. Incontrast, a group o f eight declining sectors have seen both negative TFP growth and also an output growth that i s slower than the average for manufacturing. These sectors include Transport Equipment, Garments, and Wood Products. Overall, the positive correlation between growth and productivity growth suggests a reasonably efficient allocation o f resources in the manufacturing sector.6 Measurement error in output growth will tend to exaggerate the correlation between output growth and (residual) TFP growth. We do, however, find that a similar pattern can be seen using growth in employment rather than growth inoutput. 11 Figure 1.6: SectoralManufacturingProductivity Growth inMalaysia TFP GrowthVariesWidely Across IndustriesandOver Time 0.18 0.16 m361 0.14 m372 0.12 0.1 0.08 i 0.06 0.04 =354 0.02 0 -0.02 -0.04 8342 m324 -0.06 -0.08 0.18 - .Nonelec.mach. mE1ec-mach- . 0.16 - .354 'Plastics Rubb.er 0.14 .323 - Avg. Manuf.Output Growth = 13.6 Autoparts 0.12 - .369 ,371 =362 m372.361 0.1 Jnd. Chem. . moth. Chem. t 353 0.06 . 0.04 - Furnit. ,Garm nts mWO?341 0.02 - .385 .3 0 - .324 -0.02 - -0.04 - -0.06 - m342 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Avg. TFP Growth (OLS) in MalaysianIndustries 1990-1997 Source: Appendix 1.B. 12 Table 1.2: ISIC IndustryCodes 311 FOODPRODUCTS (311) 354 MISC.PETROLEUMAND COALPRODUCTS(354) 313 BEVERAGES(313) 355 RUBBERPRODUCTS(355) 314 TOBACCO (314) 356 PLASTIC PRODUCTS(356) 321 TEXTILES (321) 361 POTTERY, CHINA, EARTHENWARE(361) 322 WEARING APPAREL, EXCEPTFOOTWEAR (322) 362 GLASSAND PRODUCTS (362) 323 LEATHERPRODUCTS(323) 369 OTHERNON-METALLIC MINERALPROD.(369) 324 FOOTWEAR,EXCEPTRUBBEROR PLASTIC (324) 371 IRONAND STEEL (371) 331 WOOD PRODUCTS,EXCEPTFURNITURE (331) 312 NON-FERROUS METALS(372) 332 FURNITURE,EXCEPTMETAL(332) 381 FABRICATED METALPRODUCTS(381) 341 PAPERAND PRODUCTS(341) 382 MACHINERY,EXCEPT ELECTRICAL(382) 342 PRINTINGAND PUBLISHING(342) 383 MACHINERYELECTRIC (383) 351 INDUSTRIALCHEMICALS (351) 384 TRANSPORT EQUIPMENT(384) 352 OTHER CHEMICALS (352) 385 PROFESSIONAL& SCIENTIFICEQUIPM.(385) 353 PETROLEUMREFINERIES(353) 390 OTHERMANUFACTUREDPRODUCTS(390) 1.23 What explains these large differences in productivity performance across industries? Throughout the remainder o f this report, our objective is to link firm performance with the investment climate, skills, and technology use, based on firm-level data from the Malaysia PICS. Here we briefly illustrate how differences in TFP growth at the sectoral level are correlated with concerns about the investment climate, skills, andtechnology use. 1.24 Concerns about skill shortages are highest in the sectors with the weakest productivity performance, while regulatory burdens appear to be felt most heavily by firms inindustries with rapid productivity growth. Figure 1.7 summarizes some o f these patterns at the broad sectoral level. The upper panel focuses on three groups o f concerns reported by firms: skill shortages, macroeconomic uncertainty, and regulatory burdens. Inthis graph we divide the nine industries covered in the PICS into three groups according to their productivity growth inthe 1990s. The industries with high productivity growth are Chemicals, Electronics, and Rubber and Plastics, while Furniture, Machinery, and FoodProcessing comprise the group with mediumproductivity growth. Finally Textiles, Garments, and Auto Parts stand out as sectors with the lowest aggregate TFP growth during the 1990s. We then report the fraction o f firms inthe PICS within each o f the three performance groupings that identified (i) shortage, (ii) skills macroeconomic uncertainty' or (iii) regulatory impediments as their top concerns about the investment climate. Although, as we will discuss in more detail later, sectoral differences in the perceptions o f the investment climate are modest, it is nevertheless interesting that concerns about skill shortages are highest in sectors with the weakest productivity performance, while regulatory burdens appear to be felt most heavily by firms inindustrieswith rapid productivity growth. 1.25 Firms with stronger innovative capability have higher productivity growth. In the lower panel we also examine data on innovation among firms inthese three groups o f industries. We find that 44 percent o f firms inindustries with rapid TFP growth adopt new technologies and adapt them to their own needs. Incontrast, inthe industries with moderate and low TFP growth, 35 percent o f firms do so. This i s the first suggestive piece o f evidence o f the importance o f specific innovative activity for manufacturing productivity growth. In Chapter 4 o f this report we will returnto this issue ingreater detail. 13 Figure1.7 :Correlates ofProductivityGrowthinthe 1990s InvestmentClimateConcerns Differ across PerformanceCategories 80% 70% 60% 50% 40% 30% 20% 10% 0% HighTFP Growth MediumTFP Growth LowTFP Growth 1 uskills Economic Uncertainty mRegulation 1 Note :See footnote 1. FirmsinIndustrieswith RapidProductivityGrowth also InnovateMore HighTFP Medium LowTFP Growth TFP Growth Growth Note: See the discussion in this section for a description of the categories in the top panel, and Chapter 4 o f this report for a definition of adapters o ftechnology. Sources;Appendix l.B and Malaysia PICS. 14 1.26 We conclude this section by briefly comparing Malaysia's manufacturing productivity performance with that o f Korea. Like Malaysia, Korea experienced double-digit real manufacturing output growth in the 1980s and the 1990s. But a number o f studies have documented that Korea has also had a quite unremarkable TFP growth, on the order o f 1 to 2 percent per year. We show this inFigure 1.8, which plots productivity growth inthe 1990s for manufacturing sub-sectors inKorea andMalaysia. Inthis figure, points above the 45-degree line correspond to sectors in which Korea has enjoyed faster productivity growth than Malaysia, while the opposite is true for points below the 45-degree line. 1.27 Although there are large differences in the rates o f TFP growth between Korea and Malaysia in individual manufacturing sectors, overall manufacturing TFP growth in the two countries i s virtually identical over the past 20 years, at 1.6 percent per year for both countries. These figures include improvements in the educational attainment o f the manufacturing workforce. As discussed previously, we did not have the data required to accurately adjust for this, either in Malaysia or Korea. However, Malaysia has seen faster growth in years o f schooling than Korea for the economy as a whole. Between 1980 and 2000, the average total number o f years o f education inMalaysia increased from 4.5 to 8.5 at an annual average rate o f 3 percent per year. Incontrast, average years o f schooling rose from 6.8 to 10.5 inKorea, by an average rate o f 2.2 percent per year. Thus, if anything, the limiteddata we draw on here suggest that Korea has had a slightly faster manufacturing TFP growth than Malaysia, and in neither country has this growth ratebeenparticularly impressive. Figure 1.8: Comparing SectoralProductivity Growth inMalaysia and Korea 0.18 0.16 5M - 0.14 - 21 v) 0.12 - 3 9ah 0.1 - 0.08 - #In\ f i g E T 'I 0.06 - 3, 0.04 - 6=F 0.02 - 0 - .361 E9 -0.02 - -0.04 - 4 -0.06 - -0.08 - -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Avg. TFP Growth (OLS) in MalaysianIndustries 1990-1997 Source: Appendix 1B. 15 C. MALAYSIA'S INVESTMENT CLIMATE 1.28 Data from the PICS show that firms rank shortage o f skills, macroeconomic uncertainty, and regulatory burdens as their top three concerns. These firm-level perspectives about skills and regulatory burdens are validated with other cross-country indicators. This underscores the importance and the likely future benefit o f Malaysia's economic stimulus package announced in May 2003, which has targeted a number o f these regulatory burdens as well as expansions in incentives for education. It should be noted that the timing o f the survey, which coincided with the invasion o f Iraqand the outbreak o f SARS, would have strongly influenced the perception o f firms with respect to macroeconomic uncertainty. 1.29 Total private investment and inward FDI over the past four years have been modest relative to the first part o f the 1990s. The slow recovery o f aggregate investment to pre-crisis levels in the short run i s likely to be due to substantial excess capacity installed during the investment boom o f the 1990s. Over the medium term, the incentives for future private investment are likely to be enhanced by continued regulatory reform as well as the increased availability o f highly skilledworkers. How Do Malaysian Firms Perceive theInvestment Climate? 1.30 A good investment climate consists of an environment in which firms can realize their productive potential to grow, create jobs, and provide the goods and services demanded by consumers at home and around the world. This requires a sound macroeconomic environment, the availability o f high quality physical and financial infrastructure, skills and technology, the existence o f a legal and regulatory framework that promotes competition, and a governance environment that overcomes bureaucratic inefficiencies and enables firms to access the factors they need to grow. The Malaysia PICS was administered to 902 manufacturing firms in2002, and extensively probed firms' perceptions o f the investment climate in which they operate. Elements o f the same core questionnaire have also been administered to firms in China and India, as well as several other countries worldwide, and similar surveys will soon be carried out in Thailand, Indonesia and the Philippines. This set o f surveys provides a rich body o f information about firms' perceptions, and how they vary across industries, regions, and countries. Here we provide an overview o f firms' responses. Later in the report we will delve in more detail into firms' key concerns, especially in the area o f skills shortages, to which Chapter 3 o f this report i s devoted. 1.31 Appendix 1C provides details on a number o f key questions regarding the investment climate inthe survey, and these responses are tabulated inTables lC.l-1C.3. Table l C . l reports the national average response, disaggregated responses by region in Malaysia, and comparisons with Guangzhou and Shanghai in China. Table 1C.2 reports responses to the same set o f questions, but disaggregating Malaysian responses by type o f firm, and Table 1C.3 does the same by industry. For each disaggregation and each question, an asterisk ("*") beside the figure indicates that the average response in the subcategory o f interest i s statistically significantly different from the average for Malaysia as a whole. 1.32 A quick look at the internationalcomparisonsinthese tables shows that, invirtually all dimensions,Malaysia enjoys an investment climate that is substantiallybetter than that 16 of the Asian countrieswhere PICS surveys havebeen completed. The upper panel o f Figure 1.9 shows the number o f days required to clear customs, where Malaysia (three days) ranks substantially better than all comparator countries. The upper panel focuses on differences in firms' perceptions o f the investment climate inMalaysia and China. Firms were asked to rate 17 different dimensions of the investment climates as constraints to doing business, The figure shows the percent o f firms that identified each constraint as "severe" or "very severe" inthe two countries. It is striking that for every one o f the 17 categories, firms in China were more likely to rate the factor as a serious constraint than firms inM a l a y ~ i a .Corporate tax rates inMalaysia ~ are also relatively low at 28 percent, which i s on a par with Korea at 28 percent and lower than in China at 33 percent. 1.33 However,we should not conclude from these simple regionalcomparisonsthat all is well with Malaysia's investment climate. Looking more closely at firms' responses, a consistent pattern o f concerns emerges. Figure 1-10summarizes the main responses o f interest, Inthe upper panel we show firms' responses to an open-ended question in the survey, where Malaysian firms were invited to identify the three biggest obstacles that they face in doing business. The responses were coded into 21 different areas, and we rank these according to the frequency with which firms identify each one as a top area o f concern. The graph shows the seven highest-ranked concerns. The lower panel provides the same information, but i s based on a "closed" question inwhich firms were invited to rate 18 factors according to the severity o f the constraint to doingbusiness. 0 Shortages of Skilled Workers. Almost 50 percent o f firms reported inadequate worker skills as a top obstacle to business in the open-ended question, while a quarter o f firms rated the skills and education o f workers as a "major obstacle" in the closed question. We will see later inthis chapter that concerns about skills are pervasive across industries, and are especiallypronouncedinthe most dynamic areas ofPeninsular Malaysia. 0 Dissatisfaction with the Overall Economic Situation. Firms' concerns about the overall economic environment can be seen inthe large proportion o f firms which identify "weak demand for their products" or "stiff import competition" as factors adversely affecting their business (inthe open-ended question), and "macroeconomic uncertainty'" as a key concern (inthe closed question). 0 Regulatory Burdens. A variety o f firm responses point to concerns about the regulatory and policy environment as a constraint to doing business. On the open-ended question, many firms identified "tax regulations," "labor regulations," and "bureaucratic burdens" as main concerns. On the closed question, "economic policy uncertainty," "labor regulations," and "customs regulations" were also consistently identified as severe obstacles to doing business. Responses to other questions inthe survey suggest that these concerns about regulatory burdens appear to reflect on firms' interactions with all levels o f government. To some extent this may reflect cultural or other differences in how firms react to questions about the problems they face. Nevertheless, the ordering between Malaysia and China i s striking. 17 Figure1.9: MalaysiaFaresWellinInternationalComparisonsofthe Investment ClimateBased onPICSSurveys Malaysia Guangzhou (China) Shanghai (China) China (Overall) 4 India (Overall) 1 Gujarat (India) Maharashtra (India) Bangladesh 1 Pakistan I 0 2 4 6 8 10 12 14 16 Days to Clear Customs (Average of Imports and Exports) PercentageofFirmsRankingthe Constraintsas "Severe" or "Very Severe" 'TaxRates 'Electricity 'RegUncert. - 'pills acro Instability ''TaxAdmin / cOrruPtiO?Co weg/F` of ante 45-DegreeLine Customs `Transport 1Labor Ref 'Monopolies/ 5 10 15 20 25 30 Malaysia Source:InvestmentClimate Survey StandardTables, World Bank. 18 1.34 Further details on Malaysian firms' perceptions o f regulatoryburdens can be gained from the PICS. Consider, for example, labor market regulations. Firms were asked if they were overstaffed or understaffed, and the reasons why. Figure 1.11summarizes their responses. Firms are overstaffed because in the face o f the current downturn inthe Malaysian economy, they are expecting an upturn in sales. Only a small percentage o f firms point to regulations on firing workers as a reason for their overstaffing. Incontrast, however, firms that are understaffed cite the difficulty in hiring local workers, the regulations in hiring foreign workers and skill shortages. This finding reinforces the aforementioned obstacles to firms o f regulatory burdens and shortage o f skilled workers. Focusing on specific labor regulations, Figure 1.12 shows that a large majority o f Malaysian firms find that dealing with the hiring procedures for foreign workers i s a major problem, whereas only 30 percent or less o f firms find that dealing with other aspects o f labor regulation is problematic. Actions have been taken by the Government o f Malaysia to alleviate this problem by increasing the number o f work permits issued for foreign workers in the manufacturing sectors from 271,892 in 2000 to 375,631 in 2003, an average annual growth rate o f 11.4 percent. Inorder to overcome the shortage o f skilled workers, foreign workers with 5 years o f working experience in Malaysia were eligible to be certified as skilled workers by the National Vocational Training Council (NVTC) and the Construction Industry Development Board (CIDB) and their employment can be extended for another five years to a maximumo f 10 years. 1.35 Tables 1.3 and 1.4 provide another perspective on the bureaucratic burdens that firms face. Table 1.3 shows the number o f days taken by firms in recent years to obtain licenses, permits or approvals from different agencies. Because different agencies are involved in procedures o f varying complexity, there are substantial differences across agencies in the average number o f days. More interesting, however, i s the coefficient o f variation in the third column -- this summarizes the uncertainty regarding the length o f time required to obtain a permit, and thus captures one dimension o f regulatory burden. According to the firms' experience, the time required to obtain an import permit i s less uncertain, whereas the time required to obtain a license from the land office is more uncertain. Table 1.4 shows similar information for a variety o f incentives. Although relatively few firms responded to this question, their responses again show significant variation across incentives interms o f the predictability of the time required to process an application. It is, however, acknowledged that the Malaysian Government has accepted the task o f reviewing regulatory burdens as one o f its top priorities with commitments coming from the highest level o f the Government. A whole set o f initiatives introduced by the Government summarized i s inBox 1.1. 19 Figure1.10: Firms' Concerns about the Investment Climate Resultsfrom Open-EndedQuestion Skilled Labor Shortage c Import Competition 1 1 Tax RegulatiodighTaxes Insufficient Demand Lack o f Business Support Labor Regulations Bureaucratic Burden 0 10 20 30 40 50 Percent ofFirmsIdentifyingIndicatedProblemas One of Top Three Concerns 1 ResultsfromClosedQuestion I I 1 S p l l s and education of work Macroeconomic v o l a t i h h l Economic policy uncertai- l 1 PercentofFirmsIdentifyingIndicatedPro as "Severe" or "Very Severe" I 1 I Source:MalaysiaPICS, as describedinAppendix l.C. 20 Figure1.11: Why Are Firm Overstaffedor Understaffed? Overstaffed Understaffed 1 70% 7 I 90% 60% 80% i 70% 4- 60% aF 50% 40% $e 50% fiE 4 40% 30% 3e 30% 20% : a E 20% 10% 0 10% 4- 0% %I I 0% $: Laws/regul.hiring Difliculty Shortage of Lawslregul.firing Anticipation of workers upturn i n sales foreignworkers employinglocal skilledworkers workers Figure1.12: Are LaborRegulationsanObstacle to Firms? 1 70% 60% )3 a 50% 40% 11 30% 20% 10% *s 0% Hlrlngprocedures Hiringprocedures Layoff procedures Limits on Inflexible salary for local workers for foreign temporary hiring scale for skilled workers workers Source: Malaysia PICS, as described inAppendix 1.C. Table 1.3: Number ofDaysto ObtainaLicense/Permit/Approvalfrom SpecificAgencies Avg. St. Dev. Coeff. Variation N.Observ. Foreign Investment Committee (FIC) 19 28 1.5 51 Commercial Vehicles Licensing Board (LPKP) 20 34 1.7 268 ImmigrationDepartment 25 34 1.4 415 Land Office 46 88 1.9 214 Fire and Rescue Department 12 18 1.5 423 Approval for Construction 79 99 1.2 266 Import Permit 23 24 1.o 282 Operating License 32 40 1.3 441 Source: Malaysia PICS as described inAppendix 1C. 21 Table 1.4: Number ofDaysto ProcessApplication(Including Payment) for Incentives St. Coeff. Avg. Dev. Variation N.Observ. Double Deduction for Promotion of Exports 66 86 1.3 49 Double Deduction for Promotion of MalaysianBrands 117 148 1.3 12 Tax Exemptionon Value of Increased Exports 20 16 0.8 44 Tax Exemptionfor Malaysian Intern. Trading Company (MITC) 20 10 0..5 7 Double Deductionof Export Credit Insurance Premiums 26 24 0.9 32 SingleDeductionfor Quality Certification 46 32 0.7 14 SingleDeductionfor Registrationof Patents 22 13 0.6 3 Single Deduction for HotelAccommodation 14 17 1.2 12 IndustrialBuilding Allowance (IBA) 61 89 1.5 28 Deductionof Cost of DevelopingWebsites 27 24 0.9 9 Tax Incentives for Offshore Trading via Websites 17 12 0.7 3 Source: MalaysiaPICS, as describedinAppendix 1.C. Broader International Comparisons of the Investment Climate 1.36 International comparisons support and complement the evidence from the firm level survey about key concerns inthe investment climate inMalaysia. Inlight o f the Asian financial crisis o f 1997, the slowdown in economic growth in2001-02, the Iraq invasion and the outbreak o f SARS, it i s not very surprisingthat so many firms point to a negative overall macroeconomic environment as one o f their key concerns. 1.37 Malaysian firms' concerns about skill shortages are consistent with international comparisons of educational attainment, as summarized in Figure 1.13. This graph shows the relationship between real per capita GDP (on the horizontal axis) and the fraction o f the population with a higher education (on the vertical axis), for over 100 developing and industrial countries for which data are available. Data for comparator countries refer to 2000, while data for Malaysia i s shown for 1980, 1990, and 2000. In all three periods, Malaysia's higher education rates are lower than the norm for its income level, as indicated by the simple line o f best fit through this graph. Although the gap has narrowed over time, the current fraction o f Malaysia's population in 2000 completing higher education o f 6 percent is still below that o f Thailand or Chile (at nearly 10 percent). It is, however, noted that the aggressive efforts undertaken by the Malaysian Government to increase the number o f people with a higher education and with skills, and to improve the quality o f humanresources, would have narrowed the gap currently. This is reflected by the increase in the output o f skilled and semi skilled human resources as well as first degree graduates in both public and private higher education institutions between 2000 and 2003. The output of skilled and semi-skilled humanresources has 22 increased from 44,486 to 60,193. Similarly, the output o f first degree graduates from the public higher education institutions has increased from 37,878 to 39,064, while the output o f graduates from the private higher education institutions has also increased, from 48,115 to 59,43 1, during the period. Figure 1.13: Malaysia's Higher Education Stocks Lag behind I t s Level of Development . >Age 25 With Higher . Korea 16 . B . Edikddion Complete, 2000 Phillppines - 13 -- I i .T h a=i(b. d / n w.. "1 ..*: PercenG P Malaysia 2000 . I I 6 7 8 9 10 11 Log(Rea1 Per Capita GDP in2000) Source: Barro and Lee (2000)' PennWorld Tables. 1.38 While Malaysia does well in the assessments of overall government effectiveness (ranking in the top quartile, and rating in a way similar to Korea), it rates lower in the assessments o f regulatory burdens (falling below the top quartile, and with a lower score than Korea). Figure 1.14 provides evidence supporting firms' concerns about regulatory burdens. The two panels o f this figure contrast Malaysia's international performance on measures o f regulatory burdens versus other dimensions o f economic governance. The upper panel shows Malaysia's ranking on two perceptions-based measures constructed by Kaufmann, Kraay, and Mastmzzi (2003). The two measures capture perceptions o f overall government effectiveness (on the horizontal axis) versus concerns about regulatory burdens (on the vertical axis). These measures are composite indices built up from a large number o f different sources o f perceptions data on governance, and cover over 170 industrial and developing countries.8 The graph also shows a horizontal and a vertical line indicating the 75th Percentile o f Regulatory Quality and Details on the methodology and data sourcesunderlying this indicator canbe found at www.worldbank.org/wbi/govemance/govdata2002.See also Appendix Table El for a lists o f all components comprising the composite Regulatory Quality Index. 23 Government Effectiveness. Malaysia i s shown as a large triangle, and selected comparator countries are also labeled. These distinctions between government effectiveness on the one hand, and perceptions o f over-regulation on the other hand, also persist inthe individual sources underlyingthe indicators inthis figure. 1.39 The lower panel o f Figure 1.14 repeats this comparative exercise, but instead uses two "factual" as opposed to "perceptions-based" measures o f the investment climate, as reported by Djankov et al. (2002, 2003). These authors have compiled data on the number o f days required to start a business across countries, as well as the number o f days required to resolve a typical contract dispute, namely, collecting on a bad check. These measures are based on consultation of legal statues and extensive interviews with local lawyers, cover about 100 developing and industrial countries, and refer to 2002.' hterms o f dispute resolution, Malaysia perfoms quite well, again ranking inthe top quartile o f all countries covered by this source, and its ranking o f just under 100 days is again comparable to that o f Korea. In contrast, it scores lower in the particular dimension o f regulation measured by this source, namely, restrictions on the entry o f new firms. Malaysia's 56 days to start a business places it well within the second quartile o f all countries, and i s longer than Korea's 36 days, as o f 2002. In short, this figure suggests that firms' concerns about regulatory obstacles inthe business environment in the PICS do appear to be capturing some important features of the investment climate in Malaysia that are verified by other data sources. 1.40 Insummary, boththe firm-level and cross-country evidence point inthe same direction - to concerns about skills, the macroeconomic environment, and regulatory burdens as negatively affecting the investment climate in Malaysia. The Government o f Malaysia i s well aware o f these areas o f concern and i s taking measures to address them. Details onthe methodology underlyingthis indicator can be foundat http://rm.worldbank.org/doingbusiness/default.aspx. 24 Figure 1.14: International Perspectives on RegulatoryBurdensinMalaysia ' .= I I I ' STNGAPORE 1.50 'C3$&%&!%% STATES _-Qf~-~a~$_-----_-__-__-_-_ 75th Percentde 1.00 _ _ _ --E ..;I.. -.: I ) .,I r-- ..E&MEA- .- ------------ Government I Effectiveness ... ....,........-,..: . 060 ':"WYSIA I +,"i'H+ ,,I. '11, I , I , ' . . . I , -.+, 5 I I m~-o.5Id~ I , ,:gllrbbr* ESIA CHINA I I ' -1.00 , I 'I I I I 9 . + I I I .I -1.50 I . I I .I. -2.00- I I I I -2.50- II 75th Percentile ofGE=0.81 -3.00- I I 140- IBad25th Percentde Check = 120 Days I I ,201 I DayBoh Start a Business b I I I * O i IT$DIA I I . I I I VIET.Ni).M 60' .kfA.hySIA PHIEIPPINES :CHINA I r-:TR.A~LAND' I I I 40- -- - ------ -,IKOR.EAr-- I . I ---I- -r---_- -- ~ ------__ -_ --- --- I 20- I I SIFGqPORE j ! IRELAND: ' Start25thPercentde DaysBusiness = 30 0 - Source: Kaufmann,Kraay andMastruzzi (2003), Djankov et al. (2002,2003). 25 D. DETERMINANTS FIRMPERFORMANCE OF 1.41 This section explores the links among firm characteristics, performance and the investment climate. Larger firms, exporting firms, firms with FDI and firms using more computerized machinery exhibit a stronger performance than other firms, measured by value added and total factor productivity, Firms on the East Coast, and in Sabah and Sarawak, are less capital-intensive and have lower labor productivity and total factor productivity than do firms in the central region. 1.42 There are important differences inthe investment climate facing by the better performing firms and that facing the worse performing firms. The East Coast, Sabah and Sarawak are also perceived by firms as having a less attractive investment climate, with economic uncertainty' as a key obstacle to doing business. Incontrast, the key obstacle for firms in the most dynamic regions is heavy regulation. And, irrespective o f location, industry or characteristics, firms point to skill shortages as a major problem in their operations. Large firms, exporter firms and firms with FDI are more strongly concerned than other firms about regulation as an obstacle to their operations. 1.43 Finally, the potentialbenefits to be gained from a reduction inskill shortages are clear: the costs to firms o f skill shortages represent, on average, 11percent o f sales across industries. Measuring Firm Performance 1.44 To promote more rapid productivity growth it i s important to understand the factors that shape firm performance and the ways in which the investment climate differs for firms with different characteristics. A variety o f measures can be used to characterize firm performance. This chapter focuses on four measures o f firm performance calculated with the data from the PICS (these measures are listed below). Details on how the measures are constructed are discussed inAppendix 1D. 0 Value added per worker (VAL) measures the productivity o f a single factor o f production: labor. 0 Total factor productivity (TFP) is a multi-factor productivity measure that represents the efficiency o f the firm in transforming inputs into output. TFP captures technology, managerialquality, andgovernment policies, among other factors. 0 Profit margin(PM)providesinformation on the financial strength o fthe firm. 0 Sales growth (SG) indicates which firms expand and which contract." 1.45 These four performance measures are consistent with one another, but they capture different dimensions o f firm performance. It is important to examine how consistent the measures are in classifying firms as good or bad performers. Table 1.5 reports the correlations 10 Ifmarkets are relatively free from distortions (e.g., there are no obstacles to entry, exit and growth of firms) and the fm that grow are those most efficient, sales growth is also a good indicator o f productivity. 26 among performance measures within industries and regions. All measures are positively correlated, but the magnitude o f the correlation varies substantially. VAL and TFP exhibit the highest correlation among performance measures with a coefficient o f 0.65. The correlation between VAL or TFP and SG i s very low, the coefficients smaller than 0.1. PM exhibits significant correlation coefficients with respect to VAL and TFP o f around 0.4. These findings suggest that while the four performance measures are consistent with one another, they are sufficiently different to make it informative to pursue the analysis o f the determinants o f performance for all four measures. Table 1.5: Correlations amongMeasuresof FirmPerformance Sales Net Profit VAL TFP Number o f firms=3 77 Growth Margin p/worker Lev. Pet. Sales Growth 1 Net Profit Margin 0.107""" 1 Value Added per worker 0.039 0.404""" 1 TFP Lev. Pet. 0.054 0.366""" 0.65 1 *** 1 Note: ***indicates significance at the 5 percent confidence level. The correlations are calculated using the residuals from a regression of each performance measure on 2-digit industry andregion effects. The number o f firms excludes outliers insales growth, value added per worker and net profit margins. Correlates of Firm Performance 1.46 Firms differ considerably in terms o f their performance. At any point intime, there are firms that are growing, others that are contracting, firms with high productivity and firms with low productivity. The focus in this section is to understand how different types o f firm characteristics account for the observed heterogeneity observed in firm performance. A regression framework i s used to relate the four measures o f firm performance in 2001 to a set o f firm characteristics: age, size, export status, foreign ownership status, and measures o f technology use, industry affiliation and regional location.l1The patterns o f association between firm-level characteristics and performance are presented inTable 1.6. 1.47 Although the findings from the regressions in Table 1.6 are informative and intuitive, one should be cautious intheir interpretation. The regressions potentially suffer from a problem o f reverse causality: some firm characteristics whose effect on performance is estimated may themselves be affected by performance. For example, foreign ownership may have a positive effect on firm performance, but concurrently, foreign ownership may be influenced by firm performance, (Le., foreigners invest in andbuy only the best performing domestic firms). Given the cross-sectional nature o f the PICS, this reverse causality problem cannot be addressed l1The exact empirical specification and the measurement of fmcharacteristics are described inAppendix l.D. The regression evidence is based on roughly half o f the full sample o f fm,and is limited by data availability. Results may therefore not be fully representative o f the full sample. 27 convincingly. Therefore, the findings from the regression analysis need to be interpreted as indicative o f a correlation between performance and characteristics but not o f a causal relationship. Table 1.6: Correlatesof Firm Performance Correlates of Plant Performance Regressors VA TFP Sales Net Profit plworker Lev. Pet. Growth Margin (1) (2) (3) (4) North Region(Penang, Kedah) -0.271*** -0.093 0.007 -0.036 South Region(Johor) -0.296*** -0.084 -0.025 -0.032 East Coast (Terengannu) -0.975*** -0.481*** -0.124 0.022 Sabah -0.367** -0.103 0.021 -0.027 Sarawak -0.209 -0.112 0.036 0.004 Firm Age 0.002 0.013 -0.007 0.003 (0.015) (0.008) (0.005) (0.003) FirmAge Squared -0.0004 -0.0003** 0.0001 -0.0001 (0.0003) (0.0002) (0.0001) (0.001) Firm Size (capital) 0.139*** 0.051**' 0.003 0.009*** (0.02) (0.011) (0.008) (0.004) Exporter Dummy (more than 10%) 0.062 0.039 -0.028 0.035*' (0.099) (0.048) (0.03) (0.018) ForeignOwnership Dummy 0.138*** 0.121*** -0.007 -0.011 (0.098) (0.047) (0.03) (0.018) % Computer-Controlled Machinery 0.001 0.0003 -0.0003 -0.0001 (0.001) (0.0007) (0.0004) (0.0003) CapitalVintage (% machineryunder 5 years) -0.00002 -0.00001 0.001*** 0.00007 (0.001) (0.0006) (0.0005) (0.0003) IndustryEffects2-digit Yes Yes Yes Yes N. observations 401 401 487 487 R-squared 0.371 0.862 0.076 0.103 Note: Estimation is done by OLS. Robust standard errors are in parentheses.*** and ** indicate significance at 5% and 10% levels, respectively.All regressionsinclude a constant. The standard errors of the region effects are not reported.. 1.48 The regressions may suffer from another potential problem: the correlation among determinants o f performance (co-linearity). Such correlation makes it difficult to isolate the effect o f a particular characteristic on firm performance when all other correlated characteristics are also included in the regression. In our sample, the firm characteristics included as determinants o f performance are significantly correlated among themselves. Hence, we complement the analysis o f the regressions with all o f the characteristics included by a series of partial regressions o f performance are one firm characteristic at a time, within industries and regions. The results from the partial regressions are in line with those shown in Table 1.6. However, given the co-linearity problem, the results from the partial regressions tend to be stronger and more significant. For example, the effect o f exports on firm performance is much stronger when the exporter dummy is the only characteristic included inthe regression thanwhen 28 other characteristics, such as firm size, are included. Exporters are generally larger than non- exporters, and therefore the inclusion o f firm size reduces the sign and significance o f the exporter variable. 1.49 A final area o f caution is the comparison of performance across industries that can be problematic for productivity measures such as VAL and TFP. For example, the comparison o f VAL across industries will simply reflect differences in the industries' capital intensity. Thus, we do not focus on cross-industry comparisons o f productivity. However, we include in the regressions the industry affiliation o f each firm. The rationale behind this i s to prevent the results from being driven by a different distribution o f firm characteristics across industries. Although the coefficients o f the industry dummy variables are not reported in Table 1.6, we highlightthe main industry differences inPM and SG, which are more comparable. Relative to firms in the food industry, firms in the rubber/plastics industry have a 6.5 percent lower SG, whereas firms in the machinery/equipment industry have a 9.3 percent higher SG. Relative to the food industry, firms in the machinery/equipment industry also have a 6 percent higher PM, whereas firms inthe furniture industry have a PM that i s lower by 7.8 percent. For firms in all other industries, the SG and PM do not differ significantly from those o f firms in the food industry. 1.50 Firmsperformworse in all regions other than the central region, as shown in Table 1.6. Some interesting regional differences infirm performance emerge from the regressions. In particular, the regions o f the East Coast and Sabah perform poorly in terms o f VAL. VAL for firms in the East Coast i s on average 38 percent lower than VAL for firms inthe central region, This lower VAL is not simply due to a lower capital intensity o f production, since TFP is also significantly lower for firms in those regions. TFP for firms in the East Coast region i s on average 62 percent lower than that in the central region.I2 The regional divergence in performance i s obtained within industries and is therefore not driven by a different industry composition across regions. Infact, the regional divergence inperformance i s found also within product categories (Le., 5-digit industries in the Malaysian classification).' This means that firms in even very similar lines o f business perform worse if they are located in the East Coast region and to a lesser extent in Sabah and Sarawak. For SG and PM, the differences in performance across regions are much less pronounced. 1.51 Firm age i s weakly associated with performance. The following interesting non-linear relationship between VAL and TFP and firm age is found: firm performance increases with age but at a decreasing rate. For a 9-year-old firm (25th percentile o f the age distribution), TFP increases with every additional year by 1percent whereas for a 19-year-old firm (75th percentile 12 The logarithm of VAL i s the explained variable in column (1). With all else constant, ln(VAL East Coast) - In(VAL Central Region) = In(VAL East Coast/ VAL Central Region) = -0.975. This means that VAL East Coast / VAL Central Region= exp(-0.975) = 0.377. Hence, the value o f 38 percent mentionedinthe text. The 62 percent difference for TFP inthe East Coast relative to the central region i s obtained similarly. l3 thenegativeeffect fortheEastCoastissignificant intheTFPregressionsinTable 1.6. However, inaTFP Only regression o f industry and region effects only, the three negative coefficients for the East Coast, Sabah and Sarawak are significant. This difference infindings is driven by differences inthe distribution o f f i size across regions. In fact, adding firm sue to the TFP regression reduces the significance o f the effects for Sabah and Sarawak, as these regions have relatively more small f i . 29 o f the age distribution), TFP increases with every additional year by 0.7 percent. For firms more than 43 years old, performance decreaseswith age.14 1.52 Firm size measured by firm assets (machinery and equipment) is positively associatedwith firm performance. Firmswith large capital stocks exhibit significantly higher VAL: a one standard deviation increase infirm size is associated on average with a 31.7 percent higher VAL.15 The TFP o f large firms is also higher than that o f smaller firms: a one standard deviation increase in firm size i s associated with an 11.6 percent increase in firm TFP. Finally, large firms also grow faster and have significantly higher profit margins.16 The finding on size andproductivity performance shouldbe interpretedcautiously. First, the correlation betweensize and TFP may be spurious as TFP i s estimated using revenues as the measure o f output. Since large firms are likely to have market power and to charge higher prices, this will be reflected in higher revenues and hence inhigher measured TFP. Second, our cross-sectional results cannot disentangle the direction o f causality between size and performance: it is possible that firms grow because they are more efficient rather than the reverse. Third, the production fbnctions generating the TFP measures exhibit decreasing returns to scale across industries. Hence, the positive effect o f size on TFP presented in Table 1.6 needs to be counteracted with a negative effect o f size on efficiency owing to the absence o f economies o f scale. Although one needs to be cautious about drawing policy implications from the estimated effect o f size onperformance, it i s likely that large firms derive benefits from other sources, such as beingbetter able to cover fixed costs, and having easier access to productivity-enhancing activities such as the use o f training andtechnology institutes (see Chapters 3 and 4). 1.53 Exporter firms perform better than non-exporter firms according to all measures but SG. The regressions relate performance to a firm's integration in international markets by including a dummy for exporter firms. Controlling for all other characteristics, exporter firms have, on average, 6 percent higher VAL than non-exporter firms. Although exporter firms are generally more capital-intensive, they also exhibit on average a 4 percent higher TFP than non- exporter firms.17 However, the performance o f Malaysian exporter firms is generally not significantly stronger than that o f non-exporter firms. Therefore, these findings are weaker than those obtained for manufacturing firms in other countries (see, for example, Bernard and Jensen [1999] for the United States, Isgut [2001] for Colombia, and Kraay [1999] for China).l8>l9The l4 increaseofTFPwithfmageholdsfor almosttheentiresample of401firmsinthe regressionanalysis, since The only 1.5 percent o f those firms are more than 43 years old. l5BothTFP and fmsize are inlogarithms, sothe coefficientonsize indicatesthe percent change inTFPper 1 percent change in fm size (with all else constant). The figure o f 1.4 percent mentioned inthe text i s obtained as 0.139* 10percent. 16Itis possible that P M is infact capturing capital intensity, because firms with large capital stocks relativeto sales would have high payments to capital that is high profitability. Using a different measure o f profitability (profitslcapital stock), the effect o f fmsize on profitability i s reversed: largerfirmshave smaller profit rates. l7 the exporterdummyvariable withallelseconstant, in(VALexporters) -ln(VALnon-exporters) =ln(VAL For exporters/ VAL non-exporters) = 0.062. This means that VAL exporters / VAL non-exporters = exp(0.062) = 1.064. This is the value o f 6.4 percent mentionedinthe text. The 4 percent difference for TFP of exporter firms relative to non-exporter firms is obtained similarly and so is the 12.9 percent difference for TFP of firms with FDI relative to domestic firms. ''These authors finda significantly stronger performance o f exporter firms when compared to non-exporter firms. l9 W e should note that the TFP o f exporter firms is significantly higher than that o f non-exporters in the partial regression where the exporter dummy, industry effects and region effects are included. Once some o f the other 30 stronger performance o f exporter firms may be due to their exposure to fierce competition in international markets and to more advanced technology and the more stringent demand for quality by their foreign buyers. Over time, however, this "knowledge" and productivity advantage o f exporter firms i s transmitted and difhsed as they interact with other firms. This type o f spillover is more likely to occur when large numbers o f exporters are present and could justify the lack o f significance o f the exporter firm performance advantage. Infact, this may be relevant for Malaysia given the large numbers o f exporters across industries and the fact that the participation o f Malaysian firms in export markets for manufacturing products has been strong for more than a decade. 1.54 Partially or fully foreign-owned firms have a significantlybetter performance than domestic firms in terms of VAL and TFP. Inparticular, firms with FDI have on average a 12.9 percent higher TFP than domestic firms. This result is consistent with the findings for other countries (see, for example, Aitken and Harrison [19991 for Venezuela). Firms with FDI have lower sales growth and lower profit marginsbut the differences relative to domestic firms are not significant. 1.55 Firms in which a larger fraction of the machinery is computer-controlled have higher VAL and TFP. Technologically more advanced firms are expected to exhibit a better performance. Two measures o f technology used by firms are included in the regressions: the fraction o f firm machinery that is computer-controlled and the fraction that i s less than five years old. Where a larger fraction o f the machinery i s computer-controlled, the firms are found to have higher VAL and TFP. For example, a firm with an average percent o f machinery controlled by computers (43 percent) has a TFP that is higher by 1 percent than a firm with no machinery controlled by computers.20 This effect may seem small, but it should be noted that in the partial regressions o f TFP that include only this measure o f technology use, the increase in TFP due to an increase from no computer-controlled machinery to the average level is 4 percent. Firms that have more modem machinery grow at a significantly faster rate but have profit margins that are similar to those o f firms with older machinery. Inthe partial regressions o f TFP on the vintage o f capital, the correlations are stronger. But the inclusion o f other firm characteristics correlated with the use o f modern machinery by firms dissipates the effect on TFP. The analysis o f technology use and adoption byMalaysian firms i s pursuedinChapter 4. Investment Climate and Firm Performance 1.56 This section examines the role of the investment climate in an understanding of differences in firm performance. Our methodology consists o f splitting the sample o f firms into groups according to each o f the characteristics that affect performance as noted in the previous section and identifylng the differences in the investment climate faced by firms in the different'groups. Namely, we group firms according to their location (region), their size, their export status and their FDIstatus. The main idea o f the approach is to identifythe differences in characteristics correlated with the exporter dummy (namely size) are included in the regression, the significance of the exporter dummy disappears. 20 The average of 43 percent represents the average for fm with some positive percent o f their machinery controlled by computers. 31 the investment climate faced by good performers versus that faced by under-performers. This approach was followed as an alternative to the estimation o f regressions o f performance on investment climate perceptions, as the latter might suffer from reverse causation problems. For example, firms with low sales and low productivity are very likely to "blame" the investment climate for their bad year (e.g., to identify "insufficient demand" for their products as a main obstacle to their operations). ' 1.57 InFigure 1.15, we showthe mainconcerns voicedbyfirms acrossregionsto relate them to regional divergence in performance.21 The differences in concerns across regions are very clear (see also previous section). Skill shortages are a concern for firms in all regions. The percentage o f firms pointing to skills as a major obstacle i s relatively similar across regions. Only in Sabah, a region with weaker performance, do the skill shortages appear to be a minor concern. Macroeconomic uncertainty i s much more important as an obstacle to operations for firms inthe weaker performing regions than it i s for firms inthe stronger performing regions. In fact, more than 54 percent o f firms in the East Coast region, and in Sabah and Sarawak point to insufficient demand for their products and competition from imports as main concerns, whereas much smaller percentages point to these concerns in the other regions. Finally, the burden o f regulation i s an obstacle with more relevance for firms in the regions with a better performance (where about 70 percent o f firms find that the regulatory burdenhurts their operations) and also for firms inthe East Coast. The relevance o f the separate items different items that add up to the regulation concerns differs across those regions, however. In the East Coast, lack o f business support systems is the major concern, whereas in the better performing central, north and south regions tax regulations and labor regulations are more important.22 1.58 Figure 1.16 relates the differences infirm performance to the investment climate.23 0 Large firms were found to have better performance interms o f VAL and TFP. From the top panel o f Figure 1.16, it i s clear that large firms are much more likely to point to regulation as a very important obstacle to their operations relative to smaller firms. In fact, 90 percent of large firms complain about regulation. Large firms face a particularly heavy regulatoryburden interms o f customs and labor regulations. 0 Exporter firms also perform better interms o fVAL andTFP. The middle panel o fFigure 1.16 shows that exporter firms are more concerned with regulation than non- exporter firms: 78 percent o f exporter firms point to regulation as a main obstacle to business, whereas 64 percent o f non-exporter firms do Given the nature o f their 21 This Figure is constructed using the answers from an open-ended question inwhich firms were asked to list the three main obstacles they faced indoing businessinMalaysia. The detailed items making up each o f the three types of obstacles (skills, macroeconomic uncertainty, regulation) are shown inthe previous section. 22 Some complementary evidence supporting the idea that regions with weaker fm performance are characterized by a weaker investment climate lies inthe fact that, when asked to identify the regionwith worst investment climate inMalaysia, a large majority of firms named the East Coast, the region with worst performance in the regression analysis. 23 These Figures are constructed using the answers from a closed question in which fm were asked to classify each o f items interms of the degree of obstacle it constitutes to their operations. The detailed items making up each of the three types o f obstacles (skills, macroeconomicuncertainty, and regulation) are shown inthe previous section. 24 Inpart this may simply reflect the fact that exporters naturally face more regulatory requirements, as they needto deal with customs declarations and other regulations. 32 operations, exporter firms voice particularly strong concerns about customs regulations. Furthermore, we also find that exporter firms report larger numbers o f days spent dealing with inspections by public institutions. Macroeconomic uncertainty is a more important concern for exporter than for non-exporter firms, but the difference is very small. 0 Finally, firms with FDIhave significantly higher VAL and TFP. The last panel inFigure 1.16 suggests that firms with FDI are more likely to consider regulation as an obstacle to their operationsthan domestic firms. Butthe operations o f domestic firms are hampered more by macroeconomic volatility than the operations o f firms with FDI. A related findingis that firms with FDIhave easier accessto finance thandomestic firms. 1.59 Regardless o f location, industry or firm characteristics, firms point to skill shortages as a main obstacle to doing business, but there are important differences in the perceptions of the investment climate for the better performing firms relative to the worse performing firms. Firms in the worse performing regions point to economic uncertainty as the main obstacle to doing business, while the main obstacle for firms in the better performing regions i s regulation. In addition, the best performing firms, large firms, exporters or firms with FDI voice strong concerns about regulation as an obstacle to their operations. This last findingi s important, since it shows that regulatory impediments are especially harmfklbecause they tend to be felt most by the best-performing firms inthe economy. Figure1.15: RegionalDifferencesinthe InvestmentClimate 70% 60% 50% 40% 30% 20% 10% 0% Central Region North Regton South Reeon East Coast Sabah Sarawak 0Slulls 0Macroeconomic Uncertainty IRegulation Note: See footnote 1. 33 Figure1.16: FirmPerformanceandthe Investment Climate Large versus SmallFirms 100% 1 90% 80% 70% 60% 50% 40"/0 30% 20% 10% 0% 1 Large Small/Medium 0Skills I3Macroeconomic Uncertamty Regulation ExportersversusNon-Exporters Foreignversus DomesticFirms 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% I Exoorters Non-Exoorters F D I DomesticallyOwned 0Slulls 0Macroeconomlc Uncertamty HRegulauon 0S M s 0MacroeconomicUncertainty Regulation Note: See footnote 1. Skill Shortages 1.60 Shortageof skills appears to be a pervasiveproblemamong the firms surveyed and Chapter 3 will analyze this problem in depth. In order to emphasize the importance of these shortages, we briefly develop here two sets o f estimates which quantifythe output costs of skill shortages. We first provide "micro" evidence based on the estimates o f firms' production functions. We then provide "macro" estimates based on assumptions about trends in economy- wide wages. 1.61 Consider first the "micro" estimate^.'^ In the absence of skill shortages, a firm that maximizes profits hires skilled workers until the marginal product o f a skilled worker i s equal to the workers' payments (the skilledwage). Ifthe firm is constrained, i.e. (ifit cannot hire as many skilled workers as desired owing to skill shortages), the marginal product of skilled workers is larger than the skilled wage. This difference can be viewed as the potential benefit to the 25A detaileddescriptionofthe frameworkis providedinAppendix 1.D. 34 constrained firm interms o f increased saledprofits ifskill shortages are reduced.26 We have data from the PICS on average wages for skilled workers across industries. Also, we can calculate the marginal product o f skilledworkers using firm-level production data and the production function parameter estimates used in the TFP calculations (described in Appendix 1D). Finally, we can identify the constrained firms as those having a skill mix (their number o f skilled workers divided by their total number o f workers) that i s lower than the optimal skill mix for their industry.27We then consider the following experiment: supposethat skills shortages are reduced (e.g., due to large numbers o f college graduates entering the labor market) so that each constrained firm can increase the number o f skilled workers to approximate it to the optimal skill mix forits industry. The benefit from this increase in skilled workers i s calculated for each constrained firm. The total benefit for an industry is calculated as the sum o f the individual firms' benefits. This total benefit i s then divided by the sum o f current sales o f those constrained firms and the results are presented in Table 1.7. Overall, the potential benefits from relaxing the skillconstraints are largeinmost industries, averaging 11percent o fsales. Table 1.7: BenefitfromReducingSkill Shortages Industry Benefit from reducing skill shortages as Yo sales 15=Food Processing 4.01 17=Texules 14.29 18=Garments 9.64 24=Chemicals 10.86 25=Rubber/Plastics 8.95 29=Mach./Equipm. 17.68 32=Electronics 5.44 34= Auto. Parts 16-01 36=Furnit./Oth.Man. 10.85 1.62 In Figure 1.17, we plot the sales gains from relaxing the skill constraints against the proportion o f firms in each industry that indicated skill shortages as a major obstacle to their operations. The industries where an increase in firms' skilled employment would have the greatest benefits in terms o f increased sales also have larger proportions o f firms reporting concerns about skill shortages2' 1.63 One shortcoming o f these simple calculations is that if they focus only on the costs to individual firms o f skill shortages, and ignores other benefits o f expanding the supply o f skilled workers. These "macro" benefits consist of resulting expansions of sectors that employ skilledworkers. Since Malaysia trades a lot with the rest o f the world, increases inthe fraction 26 More precisely, this difference represents the benefit derived from being able to hire one additional skilled worker. 27 The optimal skill mix is calculated for each industrybased on the production function parameter estimates (see Appendix 1.D for details). 28 Figure 1.17 i s based on the answers to the closed question. A very similar graph would be obtained using the answers to the open-ended question. 35 o f skilled workers in Malaysia are likely to encourage production and exports in sectors that employ a lot o f skilled workers. This shift to skill-intensive industries is likely to be accompanied by increases in average wages, as more and more workers earn higher wages commensurate with their improved skills. This increase in average wages can serve as a measureof the potentialbenefitsof improvementsinthe supply of skills inMalaysia. Figure 1.17: Benefit fromRelaxing Skill Shortages and the Investment Climate Perceptionson Skills 1 36 1 MackJEquip. .Auto-parts 34 .Garments .Textiles .Electronics . Chemicals Rubberdastics .Furniture 2o 181' .Food 1 I I , I I 1 1 3 5 7 9 11 13 15 17 19 YOSales Gain from Reduction in Skill Shortages 1.64 How large will this benefit be in terms of higher wages? The answer, o f course, depends on three factors: (i) the fraction o f the workforce that is skilled, (ii) current premium the o f skilled wages over unskilled wages, and (iii) extent to which this premium falls as the the overall skills o f the workforce increase. The first two figures can be obtained from the PICS. In manufacturing, roughly 30 percent o f the workforce is skilled (i.e., managers, professionals, and skilled production workers), while in selected business support services roughly half of the workers are skilled. We therefore take a simple average o f 40 percent o f the workforce that is skilled. The survey also tells us that skilled wages are roughly double unskilled wages in manufacturing, and inservices this skill premiumi s more than fourfold. We again take a simple average o f a threefold premium o f skilled wages over unskilled wages. The third itemi s difficult to quantify precisely, but the past experience o f other East Asian economies suggests that the decline in the wages o f skilled workers relative to unskilled workers has been modest, even as the supply o f skilled workers has increased dramatically. In Singapore, for example, there were roughly 50 workers with no university education for every university-educated worker in the mid-1960s. By the early 1990s, this ratio had fallen to 10 to 1. But despite this dramatic 36 increase inthe education o f the workforce, the relative wage o f university-educated workers fell byonly 20 percent. 1.65 Table 1.8 provides some simple estimates o f the possible order o f magnitude o f the aggregate benefits o f skill upgrading measured inthis way. Inthe first column we show a range o f possible values o f the elasticity o f the skill premium with respect to the relative supply o f skilled workers @e. the percent change in the ratio o f skilled to unskilled wages that accompanies a 1 percent increase in the ratio o f skilled to unskilled workers). Inthe second column we show the percent increase in average wages in manufacturing due to a 20 percent increase in the fraction o f the workforce that i s ~killed.~'Ifthere are no reductions in the skill premium, average wages could increase by as much as 9 percent. This figure i s shown in the first row, and reflects the purely compositional effect that there are now more skilled workers earning higher skilled wages. Of course, it is likely that there will be some declines inthe skill premium as well. As shown in Table 1.8, the historical elasticity o f the skill premium with respect to the relative supply of skilled workers was small in Singapore, Hong Kong SAR, and Taiwan (China), ranging from -10 percent to -20 percent, andwas substantial only inKorea. The other rows o f Table 1.8 show the estimated effect on average wages for different value o f this elasticity consistent with the experience o f these countries. For moderate values o f this elasticity, the increase in average wages falls to 6 to 8 percent, and falls to 2 percent if the elasticity i s equal to the value observed inKorea. Table 1.8: IncreasesinAverageWages as Skills Expand Elasticity of Skill Premium Percent Increase in to Relative Supply of Skilled Average Wages Due to Workers 20 Percent Increase in e nf Wnrkers 0 9% -0.1 8% ,-0.2 6% -0.5 2% MemorandumItem: Average elasticityof skillpremiumwith respect to relative supplyof skilled workers (calculatedover mid- 1960sto mid-1990s) Singapore: -0.22 HongKong SAR: -0.11 Taiwan, China: -0.10 Korea: -0.55 Source: Dataonwages by educationlevel from Hsieh(2002). 1.66 Overall, this evidence suggests that increasesin the skills of the workforce can have substantialbenefits. From the standpoint of individual firms that are unable to hire as many skilled workers as they would like, the benefit interms o f higher sales could be on the order o f *'Thisexercise i s based on the following simple calculation. Let wsand wu denote skilled and unskilled wages, and let I$be the fraction of the workforce that i s skilled. Average wages are therefore w=I$ws+( l-I$)wu . The elasticity o f w with respect toI$i s 4 - 4 (%(I 1 whereEistheelasticityoftheskill - , ] @ w +E) CbWs~W, +(I-$) wu premiumwithrespect to I$, and we have assumedthat there is no change inthe levelo fthe unskilled wage. 37 10 percent. Moreover, to the extent that expansions inthe supply o f skilled workers would allow Malaysia to expand into sectors that are intensive in skilled workers, there are also likely to be substantial increases in average earnings as more and more skilled workers are able to achieve higher earnings. Chapter 3 of this report will discuss skill shortages and policies to alleviate them inmore detail. E. CONCLUSIONS 1.67 Malaysia's past growth has been driven by the accumulation of physical and human capital, with improvements in total factor productivity playing a relatively modest role. For Malaysia's growth to continue well into the future, Malaysia will need to continue to invest in human and physical capital. But given the limits to growth from these sources, Malaysia will also need to find ways to improve total factor productivity growth. 1.68 A good investment climate will be required to provide firms with incentives to invest, innovate, and grow. A good investment climate will also provide individuals with incentives to invest inskills that are valued by dynamic and growing firms. Although Malaysia's investment climate i s good compared to many middle-income countries, it also exhibits key vulnerabilities, namely: (i)shortages of skilled workers, (ii)uncertainty about the overall economic environment, and (iii) regulatory burdens. Ofthese, skill shortages and regulation have a doubly pernicious effect: evidence from firm-level data indicates that the most dynamic firms (firms in strong-performing regions, firms with FDI, and firms that export) are most likely to be adversely affected by these factors. Progress in these three areas should be the focus o f policy efforts in the short to medium run, inorder to provide the foundations for sustained growth into the future. 1.69 The remainder o f this report delves into these vulnerabilities in more detail. First, Chapter 2 provides a diagnostic o f the services sector, where regulatory burdens are felt especially keenly. Chapter 3 then discusses in greater detail the determinants o f skill shortages and outlines policy responses. Chapter 4 investigates factors driving technology adoption and adaptation inMalaysian firms, and evaluates the effectiveness o fpolicy interventions designed to expand technology use. 1.70 This chapter also points to two key areas o f policy relevance that require further investigation. First, although firms identify regulatory burdens (together with skill shortages) as a major concern in the investment climate inMalaysia, this report does not take up this issue in depth. A detailed assessment o f key elements o f the regulatory environment, and the bottlenecks that it creates can be useful in informing future policy moves to ease regulatory burdens. Second, while this chapter has emphasized the importance o f productivity growth for Malaysia's future growth perfonnance, the evidence on the determinants o f productivity growth at the firm level based on the Malaysia PICS remains incomplete. The main issue here i s that this first survey observes firms only at a single point in time. One or more follow-up surveys will be needed to collect data on how firms respond over time to different policies and opportunities, which in turn will shed more light on the effectiveness o f various policy interventions. It i s acknowledged that the Government o f Malaysia has put inplace appropriate programs to address the various concerns. 38 2. FIRMPERFORMANCEINTHE SERVICES SECTOR 2.1 The services sector in Malaysia i s larger than the manufacturing sector. It contributes more than 50 percent to GDP and employs nearly half o f the labor force. In comparison, the manufacturing sector contributes less than 30 percent to GDP and employs 23 percent o f the labor force, As Malaysia makes the transition from a middle income country to a high income country, the significance o f the services sector i s likely to increase further. The EighthMalaysia Plan, 2001-2005, has identified the services sector as a key source o f growth. How efficient i s the services sector in Malaysia? What are the major constraints that are holding back its medium-term development? 2.2 This chapter assesses the performance o f the selected business-support services sector at the firm level, based on a survey o f 249 firms carried out under the Productivity and Investment Climate Survey (PICS).' The analysis o f the firm-level data is further investigated using the national accounts data and cross-country comparisons. , 2.3 The keyfindings are the following: 0 Performance varies within the services sector. While business support services appear to be doing well, other services industries such as wholesale trade, retail trade, andhotels are growing more slowly. 0 The efficiency performance o f the services sector as a whole i s lagging behind the manufacturing sector. 0 Firms in the selected business-support services sector have identified skills shortage and the high cost o f regulation as major constraints that are holding back the growth o f the services sector. 0 The investment climate inthe selected business-support services sector appears to be less healthy than that in the manufacturing sector, with more firms in the selected business-support services sector complaining about a higher level o f bureaucratic burdenrelative to the manufacturing sector. The PICS does not cover banking, retail, tourism, and other services industries. 39 0 Cross-country comparisons suggest that the services sector inMalaysia is not as open to trade as are the agriculture andmanufacturing sectors.' 0 Malaysia i s inthe process o f liberalizing the services sector consistent with the goals o f the Eighth Malaysia Plan and negotiations with the World Trade Organization (WTO). 2.4 The chapter i s organized as follows. Section A presents the performance o f the services sector at the national level, by looking at employment, value added, and labor productivity. The services sector in Malaysia i s benchmarked to that o f some selected countries, as well as other sectors within the Malaysian economy from 1980 to the present. Industry performance within the services sector i s fbrther examined using trade and national accounts statistics in Section B. Section C presents the efficiency indicators at the firm level. InSection D the chapter reports the findings on the investment climate as perceivedby the firms. Section Einvestigates the problem of the skills shortage, a key constraint identified by the firms in the survey, while Section F reviews the innovation readiness o f the services firms. Section G reviews the regulatory regime, and Section H investigates empirically the key drivers o f efficiency performance at the firm level. Section Ipresents conclusions to the chapter. 2.5 The robustness o fthe findings inthis chapter is limitedby several factors: 0 First, measuring efficiency for the services firms is more challenging than measuring it for manufacturing firms, as output, inventory, and material costs are difficult to measure for the services firms. However, this concern may be somewhat overstated. This chapter uses sales growth, along with value added per worker where data are available, and net profit margins as proxies for the firm-level efficiency in the services sector. Inaddition, by carefully isolating the movement o f labor productivity owing to changes incapital per worker, the effects o f various policy variables on firm productivity can be properly estimated despite the fact that firm productivity remains unobservable. 0 Second, the determinants o f firm performance in the services sector are likely to be similar to those in the manufacturing sector. Indeed, this chapter shows that increasing the education o f workers and improving regulatory institutions drive firm performance inthe services sector, as they do inthe manufacturing sector. 0 The firm-level analysis in this chapter is based on a survey o f 249 firms in the following five business-support industries: information technology, communications services, accounting and related services, advertising and marketing, and business logistics. Given their disparate characteristics, the findings based on this sample o f business-support services firms cannot be generalized to other services industries in Malaysia, such as wholesale trade, retail trade, hotels, restaurants, and tourism. The latest economic package announced on May 21, 2003, relaxes some o f the restrictions on foreign ownership on a case-by-case basis. Please refer to http://www.mida.gov.my and Appendix Box 2A.1 in Appendix 2F for a summary o fthe package. 40 A. PERFORMANCEAT THE NATIONAL LEVEL 2.6 The services sector is a major part of the Malaysian economy. hterms of employment, it provides about 50 percent of current employment, while two decades ago it provided about 40 percent3 However, relative to the more advanced countries, the employment share o f the services sector inMalaysia i s low as shown inFigure2.1. Interms of value added, the share o f the services sector inMalaysia is even lower than inneighboring Thailand, as shown inFigure 2.2, and insome countries that have a comparable GDPper capita level, as shown in Figure 2.3. The share o f the private services sector in GDP is only 33 percent (excluding government services). 2.7 The labor productivity of the services sector in Malaysia has improved, but it is still lagging behind that of other countries. When it i s measured by labor productivity interms of value added per unit o f worker in the services sector (in 1995 US$), as seen in Figure 2.4, it i s clear that the productivity of Malaysian services workers has improved by 50 percent, from US$7,000 in 1984 to nearly US$10,500 in 1999. The average annual growth rate o f labor productivity inthe services sector inMalaysia i s about 2.7 percent and around 1.6 percent inthe latter part of the 1980s. However, the cross-country comparison in Figure 2.4 shows that the Malaysian services sector is less productive relative to other countries. Figure 2.5 presents the labor productivity o f Malaysia relative to a sample o f more than 60 countries for the period 1980 to 2001, conditional on the GDP per capita o f countries, and further supports the finding that the labor productivity o f the services sector inMalaysia, although improved, still lags behind that o f countries such as Korea when it was at a comparable level of GDP per capita. For the same time period, the labor productivity for the whole economy of Malaysia, which i s measured as GDP per employment, has improved from US$7,400 to nearly US$12,000, as shown in Figure 2.6. Employment in services i s the proportion o f total employment recorded as working in the services sector. Employees are people who work for a public or private employer and receive remuneration in wages, salary, commission, tips, piece rates, or pay inkind. Services include wholesale and retail trade and restaurants and hotels; transport, storage, and communications; financing, insurance, real estate, and business services; and community, social, and personal services-corresponding to divisions 6-9 (ISIC Revision 2) or tabulation categories G-P (ISIC Revision 3). Data are from World Development Indicators, which gathers information from national and international sources. So it i s possible that different countries have slightly different definitions o f the variables, and may have gone through various revisions indifferent periods. 41 The average annual growth rate o f labor productivity for the whole o f Malaysia is 3.2 percent, and was 2.1 percent inthe latter part o f the 1980s. 2.8 Furthermore, the labor productivity of the services sector is lagging behind other sectors in Malaysia as well, but over time the efficiency performance of the services sector appears to be catching up with the manufacturing sector. Table 2.1 reports the average annual growth rate of value added per worker for the services andmanufacturing sectors over the last two decades (see Appendix Table 2A.11inAppendix 2B for yearly figures). The long-run growth rate in value added per worker in the manufacturing sector i s nearly double the growth rate in the services sector. Nevertheless, cross-country comparisons suggest that the services sector inMalaysia is doing well but that there i s potential for improvement. Table 2.2 reports the average annual growth rate o fvalue added per worker for the services sector for a group o fAsian and European countries. Singapore stands out as the star performer, but Korea and Malaysia are close behindwith long-run growth rates o f value added per worker close to 2.5 percent over the last decade and a half. It i s difficult to infer, using such macro data, just what is driving the high performance o f the services sector in Singapore relative to Malaysia. Among the East Asian countries, Malaysia and Singapore are the only two countries that have consistently maintained highgrowth rates invalue addedperworker inthe services sector. Figure 2.1: Employment Share of Services Sector in Selected Countries in 1981 and 1999 90 80 70 60 Yo 50 40 30 20 10 0 I MYS FIN HKG JPN KOR SGP THA Source: WorldDevelopment Indicators (2003). 42 Figure 2.2: Share of Services Sector in GDP in 1981 and 1999 90 80 70 60 Yo 50 40 I 1 9 8 1 30 1999 20 10 0 MYS FIN HKG JPN KOR SGP THA I Source: World Development Indicators (2003). Figure 2.3: Share of Services Sector in GDP Relative to GDP per Capita SGP coef = .09263646, (robust)se = ,00250225, t = 37.02 I Note: The line corresponds to the predicted share of the services sector inGDP from a regression on logper capita GDP and a constant. Source: WorldDevelopment Indicators (2003). 43 Figure 2.4: Labor Productivity of the Services Sector Selected Countries in 1984 and 1999 Labor Productivityof Service Sector SelectedCountries, 1984 & 1999 100000 90000 80000 70000 3 a 60000 2 50000 1984 40000 1999 30000 20000 10000 0 I MYS FIN GBR JPN KOR SGP THA Source: WorldDevelopment Indicators (2003). Figure 2.5: Cross-Country Comparison of the Labor Productivity of the Services Sector coef = ,86720303,se = .00600154,t = 144.5 2.194351I ...', KOR I I 44 Figure 2.6: Labor Productivityof All Sectorsin Selected Countriesin 1984and 1999 Labor Productivity ofAll Sectors Selected Countries, 1984 & 1999 90000 80000 70000 I 20000 I MYS FIN HKG JPN KOR SGP THA Source: World Development Indicators (2003). Table 2.1: The Performanceof the Services Sector Lagsbehindthe Manufacturing Sector: Average Annual Growth Rate of Value Added Per Worker (YO) 1985-1989 1990-2000 1985-2000 Services 0.45 3-44 2.50 Manufacturing 3-47 5.69 5.OO All Sectors 1.81 3.35 2.87 Source: Yearbook of Statistics Malaysia Table 2.2: The Services Sector inMalaysia is Growing Well but There I s the Potentialfor Improvement Average Annual Growth Rate of Value Added Per Worker (YO) 1985-1989 1990-2000 1985-2000 Malaysia 0.45 3.44 2.50 Japan 2.66 1.35 1.79 Korea 5.05 1.47 2.59 Philippines -0.01 -0.30 -0.20 Singapore 5.08 3.92 4.37 Thailand 7.07 -1.13 1.60 Finland 1.24 1.32 1.30 France 1.70 0.38 0.79 Norway 0.98 1.30 1.20 UnitedKingdom 0.97 1.03 1.01 ~~ Sources: Yearbook of Statistics Malaysia; WorldDevelopment Indicators (2003). 45 46 B. INDUSTRY PERFORMANCE 2.9 In the absence of easily available international time series data on the performance of different services industries in different countries (which is available for the manufacturing sector), international finance statistics and national accounts data can be utilized to provide some overall impressions of the performance of different services sector industries across countries and time. First, international finance statistics from the World Development Indicators 2003 are utilized for cross-industry, cross-country, and time series comparisons o f the services sector. Next, Malaysian national accounts data will be employed. The advantage o f rising national accounts data is that they give better detailed coverage of several services sector industries over time. However, data on capital stock for the services sector are also not available and therefore TFP growth cannot be estimated. Also, the business- support industries covered by the PICS (information technology, communications services, accounting and related services, advertising and marketing, and business logistics) cannot be neatly mapped into the national accounts disaggregation o f the services sector. Business-support services in the national accounts data come under two categories in Malaysia: transport, storage and communications; and finance, insurance, real estate, and business services. One other services sector category covered in this section is wholesale and retail sales, hotels, and restaurants. 2.10 While the export of Malaysian services has been booming in the past 20 years, the import of services into Malaysia is growing as well, leading to a narrowing trade deficit in the services sector overall. In 1981 the total value o f private commercial services exports in Malaysia was about US$1.2 billion. The volume swells more than 10 times and approaches US$14.3 billion in 2001. The average growth rate o f services export in Malaysia i s about 12.3 percent per annum, second only to Ireland, which has the fastest growth inthe sample countries analyzed, with an impressive 14.3 percent per year (see Figure 2.7). Figure 2.8 presents the value o f commercial services imports o f Malaysia, which increases from US$2.8 billion in 1981 to US$16.5 billion in 2001. The average annual growth rate o f services imports during this period i s 8.9 percent, as shown inFigure 2.8. While the growth rate o f services imports is less than that o f exports, Malaysia i s still running a US$2.2 billion trade deficit in commercial services, which i s slightly over 1percent o f the total value added o f the services sector. On the other hand, Hong Kong and Singapore in 2001 are running trade surpluses in commercial services o fUS$17.1 billion andUS$5.8 billion, respectively. 2.11 The private commercial services sector is firther disaggregated into three main industry groups: (i) Transport Services-covers alltransport services (sea, air, land,internalwaterway, space, and pipeline) performed by residents o f one economy for those o f another and involving the carriage o f passengers, movement of goods (freight), rental o f carriers with crew, and related support and auxiliary services. Excluded are freight insurance, which is included in insurance services; goods procured in ports by nonresident carriers and repairs o f transport equipment, which are included in goods; repairs o f railway facilities, harbors, and airfield facilities, which are included in construction services; and rental o f carriers without crew, which i s included in other services. 47 Figure2.7: Value of Services ExportinSelectedCountriesin 1981and2001 50 M rA 40 30 1981 .B.- - - 0 20 m2001 a lo 0 MYS FIN HKG IRL KOR SGP THA Source: World Development Indicators (2003). Figure 2.8: Value of Services Import in Selected Countries in 1981and 2001 40 35 30 2 ;20 25 3 15 z I O 5 0 MYS FIN HKG IRL KOR SGP THA Source: World Development Indicators (2003). (ii) TravelServices-coversgoodsandservicesacquiredfromaneconomybytravelersin that economy for their own use during visits o f less than one year for business or personal purposes. Travel services include the goods and services consumed by travelers, such as lodging andmeals andtransport (withinthe economy visited). (iii) Other Business Services - includes insurance and financial services, international telecommunications, postal and courier services, computer data, news-related service transactions between residents and nonresidents, construction services, royalties and license fees, miscellaneous business, professional and technical services, and personal, cultural and recreational services. 2.12 Among the three main industry groups in services, the importance of transport services has diminished,while the importanceof travel and business services has increased. 48 Figure 2.9 presents the composition o f services exports and imports in 1981 and 2001 according to the three main industry groups o f the selected countries. In 1981, the most important industry group within the services sector in Malaysia in terms o f trade was transport services, when the share o f transport services in services export and import was about 40 percent and 50 percent, respectively. This i s not unique to Malaysia, as transport services was the most important industry within the services sector for many other countries, as shown in the top two panels o f Figure 2.9. A different picture emerges in2001 as the importance o f travel and business services increases over time. Currently, travel services has the largest share in terms o f services sector export while business services dominates the imports o f services in Malaysia, as seen in the bottom two panels o f Figure 2.9. This is a widespread phenomenon where the increasingly important role o f business services intrade for many countries is observed. Conpsition ofCommercial Service Exportin1981 ConpsitionofCommercialService Inportin1981 I MYS FIN IRL JPN KOR SGP THA I I MYS FIN IRL JPN KOR S c f THA MTransport .Travel OBusmess Services BTmspoR .Travel OBusmess Services CompositionofCommercial Service Exportin2001 I ConpositionofCoon+al Service Inport inLOO1 MYS FR'I IRL JPN KOR SG? THA &!Transport .Travel OBusiness Services I MY3 FIN IRL JPN KOR SGP THA lPTransport .Travel OBusmess Services Source: WorldDevelopment Indicators (2003). 2.13 The national accounts data also suggest that business-support services have grown faster than to the other services. In the absence o f TFP data, growth rate in value added per worker is used as an indicator o f sectoral efficiency. Table 2.3 reports the average annual growth rate o f value added per worker for different categories o f services in Malaysia for the period 1985-2000 (see Appendix Table 2A.10 in Appendix 2B for yearly figures). Transport, storage, and communications were the fastest growers at 4.5 percent per year over the last one and a half decade. Wholesale, retail trade, and hotels was the slowest, although its performance has improved dramatically. Finance, insurance, and business services has improved over time and is somewhere in between the other two categories. The annual growth rate in value added per worker in transport, storage, and communications matches the growth rate o f value added per worker inthe manufacturingsector, but this i s not true for the other services sectors. 49 Table2.3: Business-SupportServices Have GrownFasterThanthe Other Services: Average Annual GrowthRateofValue AddedPer Worker (YO) 1985-1989 1990-2000 1985-2000 Transportation,Storage, and Communications 4.06 4.71 4.5 1 Wholesale andRetailTrade, Hotels, andRestaurants -3.31 4.26 1.89 Finance,Insurance,Real Estate,andBusinessServices 1.34 3.33 2.71 SourceBuku TahunanPerangkaan Malaysia (Yearbook of Statistics Malaysia). C. FIRMPERFORMANCE 2.14 Firm characteristics. The PICS survey was carried out on 249 firms belonging to the selected business-support services group. This group includes information technology, communications services, accounting and related services, advertising and marketing, and business logistics. The firm level data were collected for the years 1999,2000, and2001. 2.15 These firms are diverse in terms of location, age, and ownership structure. More than 70 percent o f the firms are from the regions o f Selangor, Kuala Lumpur, and Melaka, while the rest o f the regions make up the remaining 30 percent o f the ample.^ Interms o f the age o f sample firms, new entrant firms are 2 years old in2003, and old mature firms are more than 100 years old. The average age o f the firms inthe sample i s 20.5 years old, while the median firm is about 18 years old. Only 6 percent o f the sample firms are less than 5 years old, and they are mainly in the software publishing, consultancy, and supply industry. The majority o f the industries sampled have not had any new entrants in the past 5 yearsa5 In terms o f ownership structure, domestic-owned firms that have no foreign equity make up 80 percent o f the sample, and 20 percent o f the sample firms, largely concentrated inthe advertising industry, have some foreign equity.6 Among firms that have some foreign equity, 37 percent have no more than 30 percent foreign investments, while foreigners have majority control over only 7 percent o f the sampled firms. The proportion o f joint venture firms in the sample is relatively small (16 percent), and the majority o f domestic firms do not enter into any partnerships with foreigners. 2.16 Business-support services firms appear to be doing well. Firm performance is measured using several indicators, including sales growth andnet profit margin (see Chapter l).' As a large number o f firms didnot report material and inventory costs, TFP levellgrowth cannot estimated. Table 2.4 compares the average performance in 2001 for the business-support services firms with the manufacturing sector, although any conclusions drawn from such a comparison will need to be interpreted carefully, as the sample for the services sector is limited to business-support services and does not include consumer services. Firm performance as Table 2A.1in Appendix 2A shows the distributionof firms by industryandregions. 5Table 2.A2 inAppendix 2A presentsthe distributionofthe sample by industryandage group. We have ownershipinformationon only 240 firms inthe sample. 7See Appendix 1D for a discussionof the details onhowthe measures are constructed. 50 measured by average sales growth i s high for the business-support services sector compared to the manufacturing sector, although the average net profit margin is lower. The average sales growth o f firms in the services sector i s high and positive, unlike the manufacturing firms that turned out to be more cyclical during the post-crisis period.' It is worth noting that the manufacturing firms generated a positive net profit margin despite the contraction in sales growth, which suggests some degree o f firm restructuring during the post-crisis period. These firms should do well as sales growth expands. Table 2.4: FirmPerformance inBusiness-Sumort Services andManufacturing. 2001 ___ Services Manufacturing Efficiency Indicators (Percent): Sales Growth 6.2 -1.5 Net Profit Margin 7.3 11.6 Memo Items: Assets Per Worker (Ringgit) 198,694 132,186 Average Years of Schooling Per Worker 11.7 9.3 Note: Business-support services included in the PICS sample included the following five industries: information technology, communication services, accounting and related services, and business logistics. See Appendix for description of the industries at the 4-digit ISIC level.. Source : Productivity and Investment Climate Survey 2002. 2.17 Business-support services firms are characterized by a relatively higher level of education and assets per worker. The average years o f schooling per worker in the services sector, as measured at the firm-level, is nearly 12 years compared to 9 years for the manufacturing sector. On average, 50 percent o f workers in services have completed secondary education and have some tertiary education. Average assets per worker are 50 percent higher in business-support services than in the manufacturing sector. It would appear that the superior performance o f the business-support services firms i s to some extent being driven by the high level o f education and total assetsper worker relative to the manufacturing sector. 2.18 Table 2.5 presents the average annual total sales o f the firms in the sample by industry and year, Overall, the growth rate for business services i s quite high at 16 percent. The comparison of average sales growth o f firms in each o f the five business-support services industries shows that information technology firms have outperformed the other services, with * The depreciation o f the ringgit may also be driving up the profit margins inthe manufacturing sector, given their highlevelo fexports. 51 firms in data processing, web-based firms, and the software-related industry growing rapidly over the period 1999-2001 (see Appendix Table 2A.6 in Appendix 2A for average annual total sales by industry at the 4-digit ISIC level). Table 2.5 also reports the export performance o f the services sectors. Firms in business logistics have maintained a high level o f export activity throughout the period. The average growth o f exports has outpaced the average growth in sales in several industries, although this is inlarge part a result of starting from a low base in 1999, (see Appendix Table 2A.9 in Appendix 2A for the percentage share o f export by industry at the 4-digit ISIC level). Table 2.5: Sales and Export Performance of the Business-Support Services Sectors in the PICS Average Total Sales (Thousands of RM$) Annual Growth (YO) Average 1999 2000 2001 1999-2001 InformationTechnology 66,568 52,027 49,983 97,695 45.8 Communication Services 1,334,663 1,137,284 1,344,15 1 1,522,553 15.7 Accounting & Related Services 10,841 10,562 10,340 11,621 5.1 Advertising & Marketing 54,235 48,171 57,188 57,346 9.5 Business Logistics 167,490 158,448 173,096 170,925 4.0 Average 326,759 281,298 326,951 372,028 16.0 Average Total Exports (Thousands of RM$) Annual Growth (YO) Average 1999 2000 2001 1999-2001 InformationTechnology 19,208 8,259 7,642 41,724 219.2 Communication Services 16,317 12,656 17,447 18,849 22.9 Accounting & Related Services 65 68 71 55 -8.7 Advertising & Marketing 2,710 212 232 7,686 1,611.8 Business Logistics 88,658 82,693 91,889 91,393 5.3 Average 25.392 20.777 23.456 31,941 370.1 Source: Productivity andInvestmentClimate Survey2002. 52 D. INVESTMENT CLIMATE 2.19 As in the manufacturing sector, firms in the selected business-support services sector find the shortage of skilledlabor to be a major problem. Firmswere asked to identify the biggest obstacles to doing business in Malaysia, and the results from the open-ended and closed questions are presented in Figures 2.10, and 2.11, respectively. Similar to the sample o f manufacturing firms, responses to the open-ended question show that 33 percent o f selected business-support services sector firms identify "skilled labor shortages" as one o f their top three concerns, by far the most common problem identified. Incontrast, the result that diverges from the manufacturing sector i s the response o f business-support services sector firms to the closed question with regard to labor.. "Skills and education o f workers" i s rated as "severe" or "very severe" by a relatively smaller percentage o f firms in the business-support services sector (12 percent), suggesting overall satisfaction with the quality o f current employees. Thus, while the firms find the quantity of skilled labor to be a problem, they may be less concerned about the quality o f skilled workers. Evidence o f these points is presented in the next section on skills shortage. 2.20 As with the responses of manufacturing firms, a large proportion of firms in the selected business-supportservices sector express dissatisfactionwith the overall economic situation and the regulatory burden. Some 24 percent o f firms indicate that macroeconomic instability i s a severe problem, and 22 percent o f firms have problems with insufficient demand for their services. It i s noted that the timing o f the survey, which coincided with the invasion o f Iraq and the outbreak o f SARS, would have strongly influenced the perception o f firms with respect to macroeconomic uncertainty. Another major constraint to better performance identified by the selected business-support services sector firms is the regulatory burden, as evidenced by several concerns relating to governance, such as economic policy uncertainty (23 percent) and the bureaucratic burden(2 1percent). Figure2.10: Investment Climate-Resultsfrom Open-Ended Question I I - 1 Insufficient Demand Bureaucratic Burden Lack o f Business Support Tax RegulationdHigh Taxes InadequateAccess to Credit ~l Official Corruption I I ~ 0 5 10 15 20 25 30 35 PercentofFirmsIdentifying IndicatedProblemas One ofTop Three Concerns Source: Productivityand InvestmentClimate Survey 2002. 53 Figure2.11: InvestmentClimate-Resultsfrom ClosedQuestion Macroeconomic Instability L 4 `. Economic Policy Uncertainty Cost o f Financing Tax Rates Corruption Access to Domestic Credit Anticompetitive Practices Skills and Education o f Workers 0 5 10 15 20 25 Percent of Firms Identifying Indicated Problem as "Severe" or "Very Severe" Source:Productivity and Investment Climate Survey 2002. 2.21 The key obstacles highlighted by the firms are the same as those in the manufacturing sector: skills shortage, dissatisfaction with the overall economic situation, and the regulatory burden. As seen in Table 2.6, shortages o f skilled workers is the most common constraint to doing business identified by firms in both the selected business-support services sector and the manufacturing sector. Similarly, a significant proportion o f firms inboth sectors express dissatisfaction with the overall economic situation prevailing at the time o f the survey. The bureaucratic burden as a constraint to doing business is more widespread in the selected business-support services sector than inthe manufacturing sector (21versus 13 percent, respectively). This finding i s consistent with cross-sectoral and cross-country comparisons of the regulatory regimes, which suggest services to be more restrictive relative to manufacturing and agriculture. Table2.6: Comparisonof Firms' Concernsaboutthe InvestmentClimate Results of Open-Ended Question (Percent) Services Manufacturing Shortages of Skilled Workers 33 44 Dissatisfaction with Overall Economic Situation 22 21 Bureaucratic Burden 21 13 Source: Productivity and Investment Climate Survey 2002. Note: See footnote 1. 54 2.22 Table 2.7 reports the responses to the open-ended investment climate question by different industries o f the services sector: information technology, communications services, accounting and related services, advertising and marketing, and business logistics. The table shows that the biggest complaint o f skills shortage comes from the advertising and marketing firms. Nearly all firms complain about the overall economic situation, with stiff import competition and macroeconomic uncertainty expressed as concerns. More firms in accounting and advertising complain about the bureaucratic burden. The number o f complaints about the bureaucratic burden i s low ininformationtechnology andbusiness logistics. 2.23 Figures 2.12 and 2.13 report the responses o f domestic and foreign-owned firms to the open-ended investment climate question. Both groups have similar concerns: the skilled labor shortage, insufficient demand, and the bureaucratic burden. The skilled labor shortage i s by far the greatest obstacle to doing business for both domestic and foreign firms. A lower percentage o f foreign firms find insufficient demand to be a problem, most likely because o f a higher percentage o f exporters among them. The bureaucratic burdeni s also identified as a concern by a large percentage o f both domestic and foreign firms. Ownership regulation shows up as a concern for the foreign firms surveyed. 2.24 A strong investment climate is needed if firms are to invest and grow in the services sector. This calls for action on several fronts: eliminating the skills shortage through education and training; lowering the highcost o f the regulatory burden; and liberalizing the restrictions on the inflow o f professionals in the services sector. These reforms would help strengthen the investment climate inthe services sector. It i s acknowledged that the Government o f Malaysia has already introduced several initiatives to address these issues. The next section examines the symptoms and manifestations o f the skilled labor shortage in the business-supporting services sector. Table 2.7: Percentageof Services Sector FirmsIdentifyingthe IndicatedProblem as One ofthe Top Three Obstaclesto DoingBusinessinMalaysia Information Communicationsand Accounting Advertising Business Technology Services Related Services Marketing and Logistics Skilled Labor Shortage 37% 33% 36% 71% 29% Dissatisfactionwith OverallEconomic 19% 22% 26% 24% 24% Bureaucratic Burden 11% 22% 30% 41% 16% Number o fFirms 27 9 86 17 89 Source: Productivity andInvestment Climate Survey 2002. 55 Figure 2.12: Open-Ended Question Sample of All Domestic FirmsThat Responded(175) - Skilled Labor Shortage Insufficient Demand Lack of BusinessSupportServices Bureaucratic Burden Tax Regulationsand/or High Taxes InadequateAccess To Credit Official Corruption High Collateral Requirements Regulationsfor Startinga Business Competitionfrom Imports 0 5 10 15 20 25 30 35 Percent ofFirmsIdentifyingthe IndicatedProblemas One of the Top Three Concerns Source: ProductivityandInvestmentClimateSurvey 2002. Figure 2.13: Open-Ended Question- Sampleof All Firmswith PositiveForeign Equity That Responded(53) I I I I I 1 Skilled Labor Shortage I I Bureaucratic Burden m Tax Regulationsand/or High Taxes m Insufficient Demand Ownership Regulations Lack of Business Support Services Official Corruption High CollateralRequirements 0 5 10 15 20 25 30 35 Percent ofFirmsIdentifying the IndicatedProblemas One of the Top Three Concerns Source:Productivityand Investment ClimateSurvey 2002. 56 E. SKILLS SHORTAGE 2.25 This section describes and investigates the symptoms and manifestations of the skilled labor shortage in the selected business-supportservices sector, usingdata from the PICS, and also examines the possible recourse through increasing formal training. The analysis o f the investment climate inthe previous section as well as Chapter 1clearly shows that the shortage o f skilled labor i s one o f the top three concerns facing services and manufacturing firms inMalaysia. About one-third o f the firms cited the skills shortage to be o f concern. The text below i s a detailed discussion and analysis o f the skills shortage in the selected business- support services sector. With a larger sample o f firms, Chapter 3 i s dedicated to the investigation of the skills shortage inthe manufacturing sector. 2.26 Data from the workers survey support the firm-level findings, established in the earlier section on firm performance, to the effect that the educational attainment of workers in the selected business-support services is higher than in the manufacturing sector, reflectinghigher skill requirementsfor workers in these specific services industries. Figure 2.14 shows the distribution o f the educational attainment o f workers in the business- support services is to the right o f the distribution for workers in manufacturing. The average years o f formal education completed per worker i s 13.4 years in business-support services compared to 10.4 years for manufacturing. The percentage o fworkers with at least a highschool diploma is 57 percent in business-support services compared to 18 percent for manufacturing. The percentage o f workers with a college degree is 34 percent in business-support services compared to 7 percent in manufacturing. The business-support services sector also hires more graduates who have studied abroad (13 percent for business-support services and 6 percent for manufacturing). 2.27 Consequently, the lengthy average time taken to fill a vacancy is reflective of the shortages of skilled labor shortages in the selected business-supportservices sector. Figure 2.15 shows that the time taken to fill a vacancy for professionals inthe selected business-support services sector i s 11.6 weeks compared to 5.6 weeks inthe manufacturing sector. The time taken i s also longer at the low end of skills. Although this reflects the skills shortage, it also suggests that the managers in the business-support services firms are more selective in their hiring decisions, since workers are the key input in the production process. These results suggest that services firms cannot compromise on the skill level o f workers (for example, the workers must have an accounting/architecture/engineering degree to be an accountant/architect/engineer). Therefore, vacancies are filled only after qualified workers with satisfactory skills are identified, further reconciling the discrepancy regarding the quantity and quality o f skilled workers that is observed in the responses o f firms to the open-ended and closed investment climate questions. The long time taken to fill a vacancy in the services sector strengthens the argument for increasing the import o f professional skills. (Chapter 3 takes up this policy discussion in greater detail.) 57 Figure2.14: EducationalAttainmentof Workers inthe Business-SupportServices is High Comparedto the ManufacturingSector DistributionofEducationalAttainment Years of F o d Source.PICSManufacturingandSemces W o r h Survey Figure2.15: TimeTakento Filla Vacancyinthe SelectedBusiness-SupportServices Sector Skills Shortagesin Malaysia: Time Taken to Filla Vacancy (Weeks) 12 10 8 6 4 2 0 Services Manufacturing 1W ProfessionalBSkilled TechnicianRUnskilled Production Workdrs 58 2.28 In-house, or on-the-job, training incidence among firms in the selected business- support services sector is high, and wage analysis suggests that these firms are offering general rather than specifictraining. Table 2.8 shows that 51percent o f firms inthe business- support services sector run formal training programs. The percentage o f large firms offering training is even higher at 79 percent. General training, as opposed to specific training, is the type of on-the-job training that, once acquired, i s equally useful (that is, it can enhance productivity) in another firm. Table 2.9 reports the results of the regression of log hourly wages on several individual and firm-specific determinants o f wages, including training.' The coefficient on the variable "Training received in current$rm" i s insignificant, implying that the wages paid to workers who are currently receiving training are not statistically significantly different from the wages paid to workers who are not receiving training. Conversely, the coefficient on the variable "Training received in previousfirm" i s significant andpositive, implyingthat firms pay 6.8 percent higher wages to workers who received training in a previous firm than to those who have not received training ina previous firm. 2.29 Given the general nature of the skills that are required in the selected business- support services sector, the poaching of trained workers is a distinct possibility as is evidenced by the wage premiums paid to workers who have received training from previousemployers. Employers do not have any incentive to fundthe training o f a wider set o f general skills because such on-the-job training increases the employability o f the individual to other firms and heightens the probability that the individual is likely to be hired away. This i s reflected inthe regression results that suggest that firms do not pay higher wages to workers who are currently receiving on-the-job training, but do pay higher wages to workers who have received training from previous employers. Further evidence o f the unwillingness o f firms to pay for general training i s the observation that almost a third o f the business-support services firms hire fresh graduates from public vocational training institution, as seen inTable 2.8. lo 2.30 The use of outside formal training, or off-the-job training, such as Skill DevelopmentInstitutions (SDIs), can increasethe supply of skilledworkers and also avoid the disincentive presented by the possibility of poaching. Compared to in-house training, outside training i s modest. As seen inTable 2.8, 22 percent o f services firms report using Skills Development Institutes. The use by small and medium-size enterprises (SMEs), firms with less than 150 employees, i s lower at 18 percent. However, more than 91 percent o f firms that do use SDIs rank the top three institutes as at least of "fairly good quality," and o f those 91 percent, more than 23 percent believe that they are o f "very good quality." The explanation for this incongruity between the very small number o f users and the very high satisfaction expressed by users may be found by examining the number o f days it takes to receive refunds from the HRDF. On average, the HRDF takes 27 days to process reimbursement claims for the typical services firm, and an even longer 36 days for large firms, as seen inTable 2.8.. This is much longer than the 18-day target set by the Pembangunan Sumber Manusia Berhad (PSMB). Asked whether the See Item A in Appendix Chapter 3 for a description o f the specification o f the returns to training equation and detailed discussion o fthe econometric methodology for the results presentedinTable 2.8. lo These results are incontrast to the specific training that characterizes on-the-job training offered inmanufacturing firms. In the manufacturing sector, firms are willing and do pay higher wages to workers who are currently receiving on-the-job training invery fm-specific skills. Chapter 3 discusses these results inconsiderable detail. 59 firmwould train more workers ifthe HRDFwere more efficient, 76percent said yes; the share of large firms saying yes i s even higher at 83 percent. These results support the need to improve the efficiency o f the HRDF and also the scaling-up of SDIs in an effort to mitigate the constraining impact of the skills shortage on the economy. This same inefficiency of the HRDF i s mirrored inthe manufacturing sector, which Chapter 3 discusses in great detail." Table 2.8: Training andServices Sector Firms Employer-ProvidedTraininginthe Business-Support Services Sector Didyour establishmentrunformal trainingprograms?(YO) 51 S M I 45 Large 79 Didyou hirefreshgraduatesfrompublic vocationaltraininginstitutions?(%) 28 S M I 24 Large 43 Have you sent your workers for training inan SDI inthe pastthreeyears? (YO) 22 S M I 18 Large 38 How long, on average, didittakeHRDFto processyour claimsfor reimbursementin 27 2001?(days) S M I 24 Large 36 Would you train moreworkers ifHRDF was more efficient?(%) 76 S M I 74 Large 83 Source: Productivity and Investment Climate Survey 2002. 11 See Chapter 3 for the historicalbackground of the HDRF. 60 Table 2.9: Wage Premiumsto Current andPreviousTraining,Controllingfor FirmFixedEffects Dependentvariable: Loghourly wage Training receivedin current firm -0.0004 (0.026) Training receivedin previousfirm 0.068** (0.027) Degree 0.763*** (0.084) Diploma 0.456*** (0.082) Upper secondary 0.232*** (0.079) Lower secondary 0.103 (0.078) Experience 0.057*** (0.005) Experience squared -0.001*** (0.0001) Tenure 0.017*** (0.005) Tenure squared -0.0001 (0.0002) Female -0.061*** (0.021) Married 0.034 (0.023) Manager 0.497*** (0.048) Professional 0.359*** (0.046) Skilled production worker 0.240*** (0.046) Non-production worker 0.064 (0.044) Apprentice 0.143 (0.115) Constant 1.118*** (0.089) Number o f observations 2,125 Number of firms 249 Adjusted R 0.428 Note : "Training received in current firm "=1 if yes to Question C34 and =O otherwise. "Training received in previous firm "=1 if yes to Question W35 and =O otherwise. The omitted educational attainment comparison group i s "primary or less years of schooling'I. The omitted occupation group i s "unskilledproduction worker 'I. i,..i , ,*denotes significanceatthe 1%,5%, and 10%level, respectively. Source: Productivity and Investment Climate Survey 2002. 61 F.INNOVATION READINESS 2.3 1 Business-supportservices firms report the same level of technological activities as the manufacturing firms. What is the level of technological effort at the firm level in the selected business-support services sector? How productive are the technological activities inthe business-support services sector? Table 2.10 shows that a large proportion o f the business- support service firms report upgrading machinery and equipment in the last two years, entering new markets, and introducing new technology. As inthe manufacturing sector, business-support services firms appear to be good adopters and adapters. There i s room for further productivity growth from strengthening technological activities at the adoption and adaptation levels.12 Data also show that very few business-support services firms are creators (see Chapter 4 for a detailed discussion on technology). That is, few firms report filing any patents, or more pertinent to the business-support services sector, copyrighting protected materials. Table2.10: SelectedBusiness-SupportServices FirmsReportSame Levelof Technological Activities as ManufacturingFirms Firms Reporting Technological Activities (Percentage) Total Services Manufacturing Upgraded machinery and equipment in the last 2 years 66.5 60.4 Entered new markets due to process or service improvements inquality or cost 36.3 47.5 Filed any patents/utility models or copyright protected materials 6.2 10.6 Developed a major new service line 22.7 28.3 Upgraded an existing service line 49.6 49.4 Introduced new technology that has substantially changed the way that the main service i s provided 31.1 28.8 Number of Observations 248 899 Source: Productivity and Investment Climate Survey 2002. l2A fmis classified as an adopter if it updates machinery or introduces a substantially new technology only. A fmis classified as an adapter ifit adopts (as previously defined) andupgrades a major product line (product improvements) or enters new markets due to quality or cost improvements (process improvements). A fm i s a creator if it adopts and adapts (as previously defined) as well as obtains copyrights or patents. See Chapter 4 for a more detailed discussion o f these classifications. 62 2.32 This section now examines three factors that can drive anation's ability to inn~vate.'~ 0 First, the extent to which there i s a common innovation structure, which includes support for basic research. 0 Second, the extent to which firm-linkages promotes innovation-based competition. 0 Third, the extent to which there are linkages between the innovation infrastructure (e.g., research institutes, universities) and industry clusters that allow the resources broadly available for innovation to flow to their most competitive use. The PICS provides data to assess the ways in which these three factors are influencing the productivity o f technological activities inthe services sector inMalaysia. 2.33 Few firms generate "ideas." Ideas can be measured either by the level o f Research and Development (R&D) activity or by copyrights and patents being filed by firms. As shown in Table 2.11, the level o f R&D activities i s low in Malaysia, with only 12 and 20 percent of business-support services and manufacturing firms, respectively, reporting R&D activities. The number o f firms filing copyrights or patents - an output measure o f ideas - is also low. The cumulative stock of technological knowledge i s also low as measured by both the proportion o f the adult population completing tertiary education and the percentage o f engineers and scientists inthe adult population(see Chapter 3). 2.34 Malaysia performswell on firm-firmlinkagesthat promote technology transfer, but performsless well on the linkages betweeninnovation infrastructure and industry clusters. Linkages between innovation infrastructure and industry clusters are likely to be acting as a drag on the productivity o f Malaysia's innovation infrastructure. As seen in Table 2.11, the most common type o f collaboration on innovation i s firm-firm linkages, with 41 and 37 percent o f business-support services and manufacturing firms, respectively, developing and adapting technology in this manner. Incontrast, linkages between the innovation infiastructure and firm clusters are quite weak inboth sectors. Collaboration between firms and universities i s lagging farthest behind, and firm linkages with research institutions, government, and multilateral agencies are also inneed o f further promotion. l3Stem, Scott, Michael Porter, and Jeffrey Furman (2000), "The Determinants o f National Innovative Capacity," NBERWorking PaperNo. 7876. 63 Table 2.11: Key Drivers of Innovation (Adaptation and Creation) (Percentage of firms undertaking technological activities) Services, Manufacturing R&DActivities 12 20 Firm-FirmLinkages 41 37 Linkages between innovation infrastructure and firm clusters: Firm-University 14 10 Firm-Research Institutions 17 21 Firm-Government 15 23 Firm-Multilateral Agencies 15 12 Source: Productivity and Investment Climate Survey 2002. G. REGULATORY REGIME 2.35 In this section, the business regulations and barriers in various business-supporting industries o f the Malaysian services sector are highlighted.14 Cross-country comparisons will then be examined using an index o ftrade restrictiveness inservices. 2.36 Accounting and Taxation Services. Foreign accounting firms may operate inMalaysia through the local affiliates. All accountants must register with the Malaysian Institute o f Accountants (MIA), which has a citizenship or permanent residency requirement, before applying for a license at the Ministry o f Finance. Registration also requires either a local university degree or a membership in at least one o f the 11 recognized overseas professional bodies recognizedby Commonwealth countries. 2.37 Architectural Services. With the approval o f the Board of Architects for operating in Malaysia for a specific project, a foreign architectural firm may act as ajoint-venture participant o f a local firm. However, Malaysian architectural firms may not have foreign architectural firms as registered partners. Foreign architects may not be licensed in Malaysia, which i s necessary for any submission o f architectural plans. Nevertheless, foreign architects are allowed to be managers, shareholders, or employees o f Malaysian firms. l4 boxon A the business incentives of the services sector as well as manufacturing and agriculture inMalaysia is presentedinAppendix 2E to provide a more balanced view. 64 2.38 Engineering Services. Similar to architectural services, foreign engineers must be sponsored by Malaysian firms to operate inMalaysia for any specific project, which i s subject to the approval of the Board o f Engineers. The license i s project-specific. To qualify for the license, a foreign engineer must have at least 10 years o f professional experience in the home country. A foreign engineering firm may establish a commercial presence in Malaysia if all directors and shareholders are Malaysian. Foreign engineering firms may collaborate with a Malaysian firm, but the Malaysian firm i s expected to design, and i s required to submit, the plans. 2.39 Telecommunications. Under the WTO Basic Telecommunications Agreement, Malaysia guarantees market access and national treatment for most basic telecommunications services only through the acquisition o f up to 30 percent o f the shares o f the existing licensed public telecommunication operators, and limits market access commitments to facilities-based providers. For value-added service, foreign suppliers are allowed a maximum o f 30 percent equity. Currently, the Government has allowed foreign equity inthe sector to a maximumo f 61 percent for the first five years, and 49 percent after. However, companies with MSC status are allowed to have full foreign ownership. Since February 1, 2000, the Communications and Multimedia Act has gone into effect, which i s Malaysia's first legislation to address anti- competitive activities. 2.40 Wholesale and Retail Trade. Foreign involvement in the wholesale and retail trades musthave a local incorporation. Foreign equity of 30 percent is allowed with at least 30 percent to be reserved for the indigenous people, the Bumiputeras. There are also minimum capital requirements and restrictions on the number o f expatriates for managerial or technical posts (1 key post per company with a maximum o f 10 posts total). However, recent developments show that increasingly more licenses are issued to foreign-owned hypermarkets, which works to promote liberalization and competition inthis sector. 2.41 Cross-country comparisons of trade restrictiveness in services were also conducted. A services trade restrictiveness index, which is calculated separately for domestic and foreign firms using data on the policy regimes, is presented in Table 2.12 for several ~ountries.'~The value o f the index i s between 0 and 1, with 1being most restrictive. The index i s constructed by assigning subjective weights and values that estimate the restrictiveness o f the country's trading regime for services based on the number and severity o f restrictions. Restrictions that are common to a number o f countries are grouped into restriction categories. Scores are then assigned on the basis o f the stringency and the relative economic cost o f the restriction. The more stringent and the greater the cost, the higher is the score.l6 2.42 The index methodology classifies restrictions in two ways. The first distinguishes whether the restriction effects "establishment" or "ongoing operations": (a) Establishment. Restrictions that affect the ability o f service suppliers to establish a physical outlet ina territory and supply services through those outlets are l5The dataareavailablefromtheProductivityCommissionofAustraliawebsite: http://www.pc.gov.au/research/memoranda/servicesrestrictionl. World Trade Organization, World TradeReport 2003. 65 included inthis category. Restrictions on establishment often include licensing requirements for new firms, restrictions on direct investment inexisting firms, and restrictions on the permanent movement o fpeople.l7 (b) Ongoing operations. Restrictions that effect the operation o f a service supplier after it has entered the market. Restrictions on ongoing operations often include restrictions on firms conducting their core business, the pricing o f services, and the temporary movement o fpeople.l8 The second way a restrictionis classified is whether it is: (c) Non-discriminatory. That is, these type o fregulations restrict domestic and foreign service suppliers equally. (d) Discriminatory. That is, these types o f regulations restrict only foreign or only domestic service suppliers. The domestic index represents restrictions that are applied to domestic firms, and it generally only covers non-discriminatory restrictions. The foreign index i s calculated to measure all the restrictions that hinder foreign firms from entering and operating in a country. It covers both discriminatory and non-discriminatory restrictions. The difference between the foreign and domestic index scores is a measure o f discrimination against foreign firms.'g 2.43 According to the index in Table 2.1220, Malaysia compares favorably with the averages inAsia, LatinAmerica, and OECD countries but appears to imposemore restrictions than other countries in the areas o f banking, distribution and engineering. Furthermore, the measures o f discriminatory restrictions against foreign firms, presented in Table 2.13, are highest in architectural services. These restrictions translate into higher price effects in Malaysia for selected industries as compared to those in some other countries. Table 2.14 shows the available results from studies that estimate the effect o f trade restrictions on price.211t should be noted, 17 See Productivity Commission o f Australia website at http://www.pc.gov.au/research/memoranda/servicesrestriction/ for the list o frestrictions o n establishment specific to each services industry examined. 18 See Productivity Commission o f Australia website at http://www.pc.gov.a~research/memoranda/servicesrestriction/ for the list o f restrictions o n ongoing operations specific to each services industry examined. 19 See Productivity Commission o f Australia website at http://www.pc,gov.au/research/memoranda/servicesrestriction/ for the list o f discriminatory restrictions specific to each services industry examined. 2o The set o f Latin American countries includes: Argentina, Brazil, Chile, Colombia, Mexico, Uruguay, and Venezuela. For the set of Latin American and OECD countries, measures for some countries are not available in some industries. For more specific details, the data are available from the Productivity Commission o f Australia website at: http://www.pc.gov.au/research/memoranda/servicesrestriction/. See also Findlay and Warren (2000). 21 Price effect measures are calculated using econometric models to estimate the effects o f restrictions o n the price of services. For more specific details, the data are available from the Productivity Commission o f Australia website at: http://www.pc.gov.au/research/memoranda/servicesrestriction/. See also Findlay and Warren (2000). 66 however, that these indices were computed in 2000. Since then, Malaysia and other countries have undertakenreforms. Therefore, these quantitative rankings may be subject to change.22 2.44 Table 2.15 presents a summary o f some qualitative comparisons o f the business-support services regulations in Malaysia, Taiwan, Korea, and Chile. It supplements the quantitative indices presented in Table 2.8. It i s clear that Malaysia, relative to the other countries, has a rather restrictive environment against foreign firms and entry. Most industries are subject to the 30 percent foreign ownership restrictions, and most professions are subject to citizenship or residency requirements. The foreign equity restrictions will be investigated in the next section where the determinants of firm performance in the services sector are analyzed. It i s expected that firms that are constrained to have 30 percent or less foreign ownership are less productive, while firms that operate in less restrictive industries in terms o f entry are expected to be more productive. ~~ ~ 22 For a list and description of current restrictions in individual countries, see United States Trade Representative, 2002National TradeEstimate Report on Foreign TradeBarriers, available at: http:llwww.ustr.govlreports/ntel2OO2lindex.htm. 67 Table 2.12: Malaysia Appears to Have One of the Most Restrictive Regimes for Services Overall Services trade restrictiveness index for selected industries, 2000. Accountancy Architectural Banking Distribution Domestic Foreign Domestic Foreign Domestic Foreign Domestic Foreign Index Index Index Index Index Index Index Index Malaysia 0.09 0.51 0.04 0.33 0.27 0.65 0.09 0.40 Asia 0.20 0.48 0.03 0.19 0.08 0.42 0.09 0.26 HongKong 0.20 0.32 0.09 0.22 0.04 0.09 0.03 0.05 India 0.31 0.44 0.02 0.08 0.05 0.60 0.15 0.32 Indonesia 0.00 0.56 0.04 0.30 0.07 0.55 0.09 0.32 Korea 0.24 0.48 0.00 0.19 0.19 0.43 0.26 0.33 Philippines 0.29 0.63 0.05 0.33 0.14 0.53 0.06 0.37 Singapore 0.18 0.41 0.00 0.08 0.11 0.37 0.03 0.07 Thailand 0.19 0.49 0.00 0.12 0.00 0.39 0.06 0.39 LatinAmerica 0.14 0.35 0.05 0.19 0.07 0.29 0.05 0.15 OECD 0.19 0.36 0.08 0.22 0.02 0.11 0.09 0.22 Engineering Legal Maritime Telecommunications Domestic Foreign Domestic Foreign Domestic Foreign Domestic Foreign Index 'Index Index Index Index Index Index Index Malaysia 0.08 0.26 0.13 0.54 0.25 0.52 0.24 0.58 Asia 0.03 0.14 0.10 0.44 0.18 0.51 0.31 0.56 HongKong 0.08 0.13 0.08 0.27 0.09 0.40 0.21 0.21 India 0.00 0.10 0.09 0.40 0.25 0.61 0.39 0.69 Indonesia 0.05 0.24 0.17 0.57 0.21 0.56 0.34 0.67 Korea 0.00 0.12 0.11 0.44 0.28 0.58 0.35 0.68 Philippines 0.00 0.15 0.10 0.54 0.17 0.64 0.13 0.45 Singapore 0.01 0.11 0.08 0.42 0.10 0.21 0.34 0.44 Thailand 0.04 0.11 0.10 0.44 0.13 0.60 0.43 0.79 LatinAmerica 0.02 0.24 0.22 0.49 0.16 0.46 0.25 0.40 O E C D 0.08 0.16 0.20 0.43 0.13 0.37 0.13 0.22 ' ' e omeshc an Oreign IeStriCUveneSSIn ex scores range rom to. ''e lg er e score, e greater are e restnctlons or an economy. N%!e12c:s were com;:ted in2000andconstru:ted based on th: reg&oi r i g e o:th:industrie?in various yea: 1999 for proiessional services, 1997 for banking, 1999 for distribution, 1998 for maritime, and 1998 for telecommunications. Source Productivity Commission of Australia, hrtp://www.pc.gov.au/research/memoranda/servicesrestriction/; World Trade Organization, World Trade Report 2003. 68 Table 2.13: Measuresof DiscriminatoryRestrictionsAgainst ForeignFirmsin ServicesAre Higher inMalaysia Discrimination in services trade restrictiveness (foreign index domestic index) - Accountancy Architectural Banking Distribution Engineering Legal Maritime Telecommunications Malaysia 0.42 0.29 0.37 0.31 0.18 0.41 0.27 0.34 Asia 0.27 0.16 0.34 0.17 0.11 0.34 0.34 0.25 HongKong 0.11 0.13 0.06 0.02 0.05 0.19 0.31 0.00 India 0.14 0.05 0.55 0.17 0.10 0.31 0.35 0.30 Indonesia 0.56 0.26 0.48 0.22 0.19 0.40 0.35 0.33 Korea 0.24 0.19 0.24 0.07 0.12 0.33 0.30 0.33 Philippines 0.34 0.28 0.39 0.31 0.15 0.43 0.47 0.32 Singapore 0.23 0.08 0.27 0.05 0.10 0.34 0.10 0.10 Thailand 0.30 0.12 0.39 0.33 0.07 0.34 0.47 0.36 LatinAmerica 0.24 0.12 0.22 0.09 0.19 0.27 0.30 0.16 OECD 0.17 0.13 0.09 0.14 0.08 0.23 0.23 0.09 Notes 1. The domestic and foreign restrictiveness index scores range from 0 to 1.The higher the score, the greater are the restrictions for an economy. 2. Indices were computed in2000 and constructedbased o n the regulatory regime o f the industries invarious years: 1999 for professional services, 1991 for banking, 1999 for distribution, 1998 for maritime, and 1998 for telecommunication. Source: Productivity Commission o f Australia, hnp://www.pc.gov.au/research/memoranda/servicesrestriction/; World Trade Organization, WorldTmde Repod 2003. Table 2.14: Higher RestrictionsTranslate to Higher PriceEffects inMalaysia Price effects of trade restrictionsfor selected industries, 2000. Engineering Telecommunications Distribution Banking Domestic Foreign Domestic Foreign Domestic Foreign Domestic Foreign Price Effect Price Effect Price Effect Price Effect Price Effect Price Effect Price Effect Price Effect Malaysia 0.05 0.12 0.07 0.16 0.04 0.08 0.22 0.61 Asia 0.02 0.07 0.99 1.83 0.00 0.01 0.07 0.37 HongKong 0.02 0.05 0.01 0.01 0.00 0.00 0.03 0.07 India ... ... 5.61 10.00 ... ... 0.04 0.55 Indonesia 0.03 0.10 0.71 1.38 0.00 0.04 0.05 0.49 Korea ... ... 0.04 0.08 ... ... 0.15 0.37 Philippines ... ... ... I.. 0.21 0.73 0.11 0.47 Singapore 0.01 0.05 0.02 0.03 0.00 0.00 0.08 0.31 Thailand ... ... 0.30 0.55 ... ... 0.00 0.33 LatinAmerica 0.01 0.09 0.08 0.13 0.01 0.01 0.05 0.20 OECD 0.02 0.05 0.02 0.03 0.02 0.03 0.01 0.10 Notes 1. The domestic and foreign restrictiveness index scores range from 0 to 1.The higher the score, the greater are the restrictions for an economy. 2. Indiceswere computed in2000 and constructed based o n the regulatory regime o f the industries invarious years: 1999 for professional services, 1997 for . banking, 1999 for distribution, 1998 for maritime, and 1998 for telecommunications. source Productivity Commissiono f Australia, hrtp://u?r?v.pc.gov.au/research/memoranda/servicesrestriction/; World Trade Organization, WorldTmdeRepor?2003. 69 0 r- H. DETERMINANTSFIRMPERFORMANCE OF 2.45 This section investigates the extent to which firm efficiency in the selected business- support services i s influenced by the level o f education of workers, the training o f workers, export orientation, ownership structure, and new entrants. Several performance indicators are examined, specifically sales, growth rate o f sales, labor productivity (as measured by sales per worker), and firm productivity. The first three indicators are observable and are positively correlated. Firm productivity, on the other hand, is not observable. Given that information on the quantity and prices o f output o f the firms in the survey is unavailable, neither the construction o f TFP measures nor the estimation o f firm productivity i s feasible. Nevertheless, it is still possible to estimate the effects o f the aforementioned factors on firm productivity in a regression analysis framework, which follows. 2.46 There are many hypotheses in the literature regarding the driving forces behind firm performance, which can be summarized as follows: 0 Competition enhances efficiency - entry and exports promote productivity gains (Bernard and Jensen [19991) a Foreign investment enhances technology transfer - restrictions on foreign ownership deter productivity gains (Blomstrom and Sjoholm [19991, Javorick and Spatareanu [20031) 0 Foreign firms or new entrants steal market shares from existing domestic firms - restrictions on foreign ownership or entry may help protect domestic firms (Aitken and Harrison [19991) 0 Physical and human capital investment enhance labor productivity (Mankiw, Romer and Weil [19921). 2.47 The above hypotheses are tested using regression analysis on the sample o f 249 services industry firms inthe survey. The observable performance indicators o f firms are first examined, and then the estimation procedure on unobserved productivity will be discussed. The correlation coefficients between log of sales, growth o f sales, and log of labor productivity are 0.66, 0.19 and 0.27, respectively. Table 2.16 reports the partial correlations o f the different observable performance indicators of the firms with firm size (log assets value), export orientation (share o f exports in sales greater than 5 percent), foreign ownership (positive foreign equity), firm age, and firm age squared, controlling for region and industry fixed effects. The first three columns o f the table are estimated by OLS regressions using the cross-sectional dataset o f firms for 2001, while the last column uses the data for 1999 to 2001 ina panel regression setting, controlling for firm fixed effects. 2.48 It is clear that larger firms tend to perform better in terms o f value o f sales and labor productivity. This could be due to economies o f scale which rewardlarger firms. Similarly, the exporting firms are better performers than the non-exporting firms, and the older firms perform better than the younger firms. While these three factors are robust in expected signs in all 73 columns, they are statistically significant in at least three out o f the four columns reported. On the other hand, while the effect o f foreign ownership on firm performance is robustly positive, it i s only significant in the panel regression. Given that some o f the industries in the business- support services sector are subjected to the 30 percent foreign ownership restriction, a better specification o f this variable i s therefore necessary. Table 2.16: CorrelatesofFirmPerformance Explanatory Variables (1) Log o f sales sales ( O W (FGLS) Firmsize 0.691*** 0.018 0.280*** 0.248 ** * (Log o f assets value ) (0.062) (0.022) (0.072) (0.013) Exporter dummy 0.523** 0.152 0.565** 0.292*** (> 5% sales is exported) (0.245) (0.110) (0.267) (0.060) Foreign ownership dummy 0.173 0.078 0.122 0.115** (positive foreign equity) (0.224) (0.087) (0.273) (0.003) Firmage -0.003 -0.016** -0.032* * -0.025*** ' (0.014) (0.007) (0.015) (0.003) Firmage squared 0.000 0.0001 ** 0.0001 o.ooo* ** (0.000) (0.000) (0.0001) (0.000) Constant 4.930*** -0.142 7.406*** 7.727*** (0.961) (0.348) (1.087) (0.196) Firmeffects no no no random 2 digit industry fixed effects Yes Yes Yes Yes Region fixed effects Yes Yes Yes Yes Autocorrelation no no no 0.6201 Number o f observation 210 206 210 608 R-squared 0.695 0.053 0.270 Heteroskedasticity white correction white correction white correction ves Note: Standard errors in parentheses.*, **,and*** indicatesignificance at 90%, 95%, and 99% confidence levels respectively. 2.49 As previously mentioned, firm productivity is unobservable, and given the lack o f information on price and quantity o f output, construction o f TFP i s also infeasible. Nevertheless, the following specification allows for the estimation of the effect o f various determinants on firm productivity without explicitly estimating firm productivity. 2.50 Specifically, labor productivity (log value o f sales per worker) i s regressed on the log value o f assets per worker. Under the assumption o f a constant return to scale production finction, the residuals o f such a regression capture the unobserved productivity, provided that good controls for firm prices are included. Region and industry fixed effects are used, together with a firm's random effect to control for firm prices. In other words, once the movement o f labor productivity that i s due to prices and capital intensity i s isolated, the remainingpart o f the movement o f labor productivity is attributed to the movement o f firm productivity. Thus, in 72 order to explain the effects o f various determinants on firm productivity when firm productivity i s unobservable, those variables can simply be included in the regression to obtain unbiased estimates. 2.51 Table 2.17 reports the regression results o f the effects o f human capital (log o f average years o f education per worker), export shares o f sales, foreign ownership, training, access to credit, and new firms (aged 5 years or less) on firm productivity. In addition, given that many industries in Malaysia are subject to the 30 percent foreign ownership restriction, the foreign ownership variable is further refined into the following four mutually exclusive categories, each o f which is represented by a dummy variable - fully domestically owned, 30 percent or less foreign owned, 30 to 50 percent foreign equity, andmajority foreign owned: foreign -equity, (0,301 E 2.52 The expectationi s that the constrained firms (that is, firms that have foreign equity o f less than or equal to 30 percent) are less productive, and that the unconstrained firms, especially those that have majority control (that is, firms with foreign equity o f more than 50 percent), are more productive. This hypothesis is tested by comparing the estimated coefficients for the individual dummy variables. 2.53 Giventhat many variables o f interest are firm-specific andtime invariant, the hypotheses can only be tested using random effect estimation, which takes into account both the within and between variations of the variables. The result o f the regression estimation i s presented inTable 2.17, where column (1) shows the random effect panelFGLS estimation and column (2) presents the result o f a cross-section regression based on data in2001 as a robustness check. It is clear that these results are qualitatively very similar to that o f the random effect panel estimates, but given that only cross-section variations of the variables are explored, not all coefficients are precisely estimated when compared to the panel regression result in column (1). (Appendix 2C for the derivation o fthe regression specifications). 2.54 The results presented in Table 2.17 are very similar to those in Table 2.16, which suggests that firm productivity, while unobservable, is highly correlated to the other firm performance indicators. Several results stand out. First, exporting services firms are associated with higher levels of efficiency. Every 10percent increase inthe share o f exports intotal sales i s , associated with an increase in firm efficiency o f 3.3 percent. Second, high levels o f education per worker are associated with high sales per worker. Every 10 percent increase in the average years o f schooling of the workforce i s associated with 2.3 percent higher firm productivity. Third, firms that provide training to workers have a higher level o f sales. Training provided by firms appears to be enhancing the return to the firm, but the return to workers is not statistically different from zero (to be discussed indetail inChapter 3). Fourth, firms with foreign equity are more productive, except when it i s constrained by foreign ownership regulations. Not all firms with foreign equity are more productive than domestic firms. In fact, only firms that are not restricted to have 30 percent foreign ownership turn out to be more productive than otherwise 73 identical domestic firms. Those firms that have 0 to 30 percent foreign ownership are significantly less productive than their domestic counterparts. On the other hand, the efficiency o f unconstrained firms increases with the share o f foreign equity, to the extent that those firms that have majority foreign ownership are the most productive. These findings suggest that the foreign equity restriction has an impact on the productivity o f firms inthe services industries. 2.55 Finally, the effects o f entry competition and the spillover o f foreign firms on firm productivity are investigated. Firm fixed effects are used to control for firm heterogeneity and only the effects o f the presence o f foreign firms and new entrants in the industry on firm unobserved productivity i s estimated, controlling for the capital intensity and export share o f the firms. (See Appendix 2C for more details). Table 2.18 reports the results. 2.56 Table 2.18 shows that domestic firms tend to benefit from the presence o f foreign firms. If sampleisrestrictedtodomesticfirmsthathavenoforeigninvestmentintheirequity,the the regression results suggest that domestic firms benefit significantly from the presence o f foreign firms in the same industry. For the business-support services sector, a 1 percent increase in industry sales that originated from firms with foreign equity increases the productivity o f domestic firms inthe same industryby 3.3 percent. This i s not surprising, since the presence o f foreign firms in the industry may bring in foreign technology and know-how, from which domestic firms may learn through the networking o f personnel and contacts. Such an effect i s also confirmed by a firm survey conducted in Lithuania, United Kingdom, and the United States.23 Finally, ifthe sample i s restricted to firms with FDI, the regression results show that the foreign presence is associated with higher efficiency as well. 2.57 On the other hand, while the estimated coefficient for entry competition on the productivity o f domestic firms i s positive, it is not precisely estimated. Foreign firms, on the other hand, seem to benefit from the presence o f other foreign firms as from well as new entrants inthe industry. For the business-support services sector, a 1percent increase inindustry sales that originated from firms with foreign equity or new entrants increases the productivity o f foreign firms in the same industry by 3.1 percent and 2.0 percent, respectively. Such results suggest that a foreign presence and entry competition may provide win-win scenarios for the business-support services sector. 23 See Smarzynska (2002), Haskel et al. (2001), and Keller and Yeaple (2003). 74 Table 2.17: Determinantsof FirmPerformance DependentVariable: Logof Sales per Worker (1) FGLSPanel (2) OLS Cross Section ExplanatoryVariables Estimates Estimate Logof assets value 0.47"' 0.51*** per worker (0.02) (0.09) Logof averageyears of 0.23*** 0.28 educationper worker (0.07) (0.22) Share of sales for exports 0.33"' 0.48% (0.11) (0.28) Foreign equity 30% or less -0.28** -0.26 (0.08) (0.26) Foreignequitymore than 30% 0.59*** 0.58" to 50% (0.09) (0.28) Foreignequitymorethan 50% 1 . O P 0 . 8 P (0.11) (0.28) Traininginpast three years 0.37** 0.38** (0.06) (0.16) New entrant -0.24' -0.24 (0.13) (0.40) Loan 0.15*** 0.29 (0.04) (0.20) Constant 5.45"** 4.90*** (0.26) (1.18) Firmeffects random NA 2 digt industry fixed effects Yes Yes Regionfaed effects Yes Yes Autocorrelation 0.7297 NA Number of observation 536 184 Heteroskedasticity Yes white correction Note: Standarderrors inparentheses.*, **, and *** indlcate sigmficance at 90%, 95%, and 99% confidencelevels, respectively. 75 Table2.18: DomesticFirmsTendto Benefitfrom the Presenceof ForeignFirms Dependent Variable: Log of Sales per Worker Estimation: Fixed Effect Panel Regressions Explanatory Variables (1) Domestic Firms (2) Foreign Firms Log of assets value 0.19*** 0.23*** per worker (0.06) (0.09) Logof average years of -0.10 1.46 education per worker (0.70) (2.25) Share of sales for exports 3 . 0 Y -0.27 (0.96) (0.72) ForeignPresence 3.28** 3.08*" (1.39) (1.15) EntryCompetition 0.51 2.03*** (0.54) (0.65) Constant 9.32" 4.97 (1.87) (5.71) Firmfaed effects Yes Yes Number of observation 464 116 Note: Standard errors in parentheses.*, **,and *** indcate significance at go%, 95%, and 99% confidence levels, respectively. I.CONCLUSIONS 2.58 The analyses o f firm-level data, national accounts data, investment climate data, and cross-country comparisons highlight a number o f salient features o f the services sector in Malaysia. First, the firm-level indicators o f performance (e.g., sales growth, growth in value added per worker) suggest that the business-support services sector in Malaysia i s doing well. Second, the performance o f the different sectors within the services sector is varied. Business- support services industries have outperformed the other services industries and have done as well as the manufacturing sector. Third, the services sector as a group lags behind the manufacturing sector interms o f efficiency performance. Fourth, firms report that the investment climate inthe business-support services sector is more unfavorable to doing business than the investment 76 climate in the manufacturing sector. A greater percentage o f firms in the business-support services sector complain about the bureaucratic burden. 2.59 The determinants of firm-level performance in the business-support services sector are the same as in the manufacturing sector. Increases in education per worker are associated with improved firm performance. A better regulatory environment goes hand inhandwith better firm performance. However, there are differences as well as illustrated by the longer time it takes to fill a vacancy in the services sector. This may reflect the fact that the quality o f education is more important in the services sector, as i s suggested by more foreign-trained graduates being hiredinthe services sector. 2.60 Strengthening the services sector as a source o f growth will require immediate and medium-term reform on several fronts. First, in line with the goals o f the EighthMalaysia Plan andMalaysia's discussionwith the WTO, the highcost o f the regulatoryburdenfacing firms in the services sector needs to be further reduced in addition to other measures that are already taken. 2.61 Second, given that a large number o f firms complain about a skills shortage, reforming the education system, providing skills training, and liberalizing the restrictions on the inflow o f professionals should be high on the Government's agenda. Some measures have already been introduced to address the above concerns under the M a y 2003 Economic Package o f the Government. 2.62 Third, further analysis is needed to identify specific constraints facing the different services sub-sectors. Another survey will be needed to include more firms from a larger group of the services sector in order to accurately assess the investment climate, supply side constraints, and deficiencies inthe business environment that are likely to hold back the services sector as a source o f growth over the medium term inMalaysia. 2.63 Specific objectives include improving and monitoring the investment climate in the services sector; and also increasing on-the-job and off-the-job training, especially among SMEs. Activities that will help achieve these goals include carrying out periodic firm-level surveys in the services sector; improving and adapting the survey questionnaire for the services sector, especially the questions on firm operations such as material cost; carrying out detailed diagnostics o f the regulatory regime inthe services sector; encouraging firms, especially SMEs, to establish and expand in-house training; and improving efficiency in the disbursement o f the available funds for training. Inter-agency coordination in the services sector can also be improved. 77 3. IMPROVINGSKILLS ...TheEighth Malaysia Plan emphasizes the creation of a strong human resource base to support the development of the knowledge-based economy as well as to enhanceproductivity and competitiveness. EighthMalaysia Plan 2001-2005 3.1 The findings in this chapter are based on the survey o f 902 manufacturing sector firms carried out under the Productivity and Investment Climate Survey (PICS). The PICS included interviews with CEOs, human resource managers, and workers. Employers were asked about their experience in filling vacancies and their experience with deficiencies inthe quality o f their existing workforce. Employees were asked independently about the skills they lack most for doing their jobs as well as about the adequacy of their field of education as it relates to the work they do. The data from this survey offers insightsinto how firms and workers perceivethe quality of education,skills, and trainingprogramsinMalaysia. 3.2 Firms have identifiedthe skills shortage as a key constraint. Irrespective o f location, industry or firm characteristics, the large majority o fthe 902 entrepreneurs inthe manufacturing sector interviewed identified skill shortages as a "severe" or a "very severe" problem.' Around 50 percent o f firms reported inadequate worker skills as a "top obstacle to business," while a quarter o f the firms rated the skills and education o f workers as a "major obstacle". The complaints o f the firms on skills shortage are consistent with the analysis of returns to education, returns to training, and trends in unemployment. Firms place a high premium on workers with tertiary education, signaling a shortage o f skilled workers. The rate o f return on (firm-specific) labor training is also high which suggests a highpremium on skills. The average time taken to fillavacancyforatechniciantakeslongerinMalaysiathaninafewotherAsiancountrieswhere PICS have been carried out. Econometric estimates suggest that the wrong skills mix at the firm level results insevere output losses, with an average firm losing 11percent o f its output. 3.3 Workers have also identifiedimportantareas of skills deficiency. Almost halfo f the workers surveyed reported English language proficiency as the skill that they lack most. Information and technology skills emerged as the second most needed skill. Lack o f adequate professionalltechnical skills rankedthird. A large proportion o f workers lack "general skills" to do their jobs and/or to adapt to changes in labor market conditions. Given that general skills are provided through the education system and firms have little incentive to provide training in general skills relative to firm-specific skills, the workers survey also highlights concerns Finnsinthe business-support services sector also identifiedthe skills shortage as a key problem (see Chapter 2 on the services sector for a discussion). 78 regardingthe education system. One-third o f the human resource managers surveyedrated the performance o f the locally trained professionals lower in comparison with the foreign trained professionals. These findings suggest that the problem o f skill shortage in Malaysia will need to be addressed throughreforms inboththe education systemandthe skilltraining programs. 3.4 Malaysia's completion rate for higher education is substantially lower than the norm for its income level, despite improvements in enrollment. While the Government efforts over the last two decades have paid off in terms o f improving the basic, primary, and secondary education, the country still lags behind in tertiary education. Malaysia i s outperformed in tertiary education by most countries in its income level category such as Chile, Mexico, and Thailand. More effort i s needed to expand higher education to meet the demand o f the economy as Malaysia shifts to products with a higher skill and technology content. The quality o f the education will also need to be improved by better alignment o f the social objectives (e.g., achieving equity) with the economic objectives (e.g., strengthening the competitiveness ofworkers). 3.5 While training incidence is high in some sectors in Malaysia, and firms rate the skills training programs highly, the overall use of training programs by the firms is very low. Firm-level data show that the return to firm-specific training i s very high in Malaysia. Despite this, many firms, especially small firms, are reluctant to make use of training programs. But firms that used training programs rated them highsinthe survey. Scaling up the training programs and removing the restrictions on the inflow o f professionals, along with education reform, should help resolve the problem o f skill shortage inMalaysia. Attracting Malaysians in the diaspora and increasing outreach activities and funding inorder to achieve lifelong learning, which is being undertakenby the Government o f Malaysia will contribute towards alleviating the problem. 3.6 This chapter examines the level, distribution, and quality of the skills and competenciesof workers in Malaysia,as well as the matchbetweenthese qualificationsand those requiredby businesses. The widespread perception o f skills inadequacyrelates infact to two distinct types of disequilibria inthe market for skills. The first is skills shortage (insufficient quantity), and the second is skills mismatch or skills deficiency. This situation has been fuelled by a steady increase in the demand for skilled workers in Malaysia over the last decade. The problems o f skills shortage and skills mismatch could hinder the aspirations of Malaysia to transform itself into a knowledge-based economy. 3.7 The next section reviews the macro evidence to uncover the extent o f the gap in skills performance inMalaysia. The following section uses PICS data to highlight the disequilibria in the market for skills, documenting the intensity o f the shortages and mismatch. This is followed by and examination ofthe key drivers o f skills match including educationand in-service training, and a review of the role of skills development institutes. The last section provides a conclusion and suggeststhe policy implications ofthe findings. 79 A. THEGAPINSKILLS FORWORK: THEMACRO EVIDENCE 3.8 In 1980, only 9.8 percent of the Malaysian population aged 25 and over had completed secondary education. In2000, the number had nearly tripled to 24 percent. This extraordinary achievement mirrors the rapid development o f primary education since the 1970s. Relatively imperfect measures o f the quality o f education also point to good quality primary and secondary education in Malaysia. However, the country still has a large deficit in tertiary education. The share o f its population aged 25 and over that has completed higher education i s about 3 percentage points lower than the norm for the country's income level, despite a tripling in enrollment rates in the past five years. In addition, Malaysia lags behind in its share of engineers and scientists engaged in R&D in the population. Macro data support the perception that there i s a shortage inthe stock and the flow o f skilled labor. (Procedures for measuring skills for work are discussed inBox 3.1.) 80 Assessing the Extent of Skills Shortages 3.9 While the Malaysian Government's efforts over the last two decades have paid off in terms of basic, primary, and secondary education attainment, the country still lags behind all the benchmark countries on tertiary education. To analyze this gap, we use two sets o f macro indicators. The first i s the stock and flow o f humancapital and the gap between its actual andpredictedvalues, given the country's level o f income. The second set o f indicators looks at the duration o f unemployment for graduates in the Malaysian labor market. The skills gap is measured with respect to international benchmarks, including countries that have successfully completed the transition from low-skills manufacturing to high-tech, high-skills dominated manufacturing. The Stock of Human Capital 3.10 The standard educational attainment indicators for Malaysia are higher than these for several Asian and Latin American countries. As shown in Table 3.1,the mean years of schooling for Malaysians aged 25 and over was 4.5 in 1980. By 2000, the country had achieved an average o f 7.9 years, better than Thailand (6.1), China (5.7) and India (4.8) . In the same vein, the share o f the population that hadno schooling hadbeenreducedby more than half (from 34 percent to 14percent) between 1980 and 2000. 3.11 There is room for improvement, especially if Malaysia is to become a knowledge- based economy or is to achieve the status of a developed nation by 2020. The gap between Malaysia and other countries that have to some extent achieved this goal is still significant. In 2000, the average years o f schooling inthe United States were 12.2, almost four years more than Malaysia. The gap was approximately 3 years for Korea and 2 years for Japan. In order to achieve educational attainment levels comparable to these inindustrialized countries (more than lo), Malaysia should further reduce its no schooling rate (14 percent against 0 percent for Japan), andshouldincrease enrollment and completion rates at the tertiary level. 3.12 While primary and secondary education completion rates are good, tertiary education lags behind. As shown in Figure 3.1, Malaysia outperforms the United States and Japan in secondary school completion. Only Korea has a larger share, with 35 percent o f the adult population aged 25 and over having completed secondary school. These results are underpinnedby consistent Government policies. In 1995, the concept o f basic education was revised upwards to include five years o f secondary school. Basic education in Malaysia therefore means receiving 11 years of education. A similar result is evidenced for primary education completionrates. 3.13 The secondpoint, emergingfrom bothFigure 3.1 andTable 3.1,is that Malaysiantertiary education's performance in terms o f number o f graduates produced remains low, although there has been an improvement in enrollment rates in recent years. Malaysia lags behind Thailand *International Benchmarks include: China, India, Japan, Korea, Singapore, Thailand, and the United States. The choice of fm-level benchmarks i s dictated by data availability and includes: Bangladesh, China, India, and Pakistan, where similar surveys were conducted recently. 81 (11.2 percent), Japan (15 percent), and Korea (19.1 percent) as shown in Table 3.1,3 Malaysia must close its gap intertiary education ifit is to succeed inadvancing to the high income group of countries. 3.14 Figure 3.1 also shows that Malaysia has a well-balanced educational transition. The country has tackled the educational challenge in consecutive phases, starting by developing a solid infrastructure for primary education. Indeed, throughout the 1970s, more than 50 percent o f public spending on education was directed to the development and expansion o f primary school education programs. Inthe 1980s, spending on primary education moderated to around one-third o f the total education expenditure as more funds were directed to expanding the secondary education infrastructure. Since 1995, the increased demand for skilled workers has been sustained specifically by university education (see Razali [19991,. Thailand, on the other hand, has developed very competitive primary and tertiary education systems, but secondary education i s deficient (see World Bank [1999]).4 3.15 The percentage of engineers and scientists in the (adult) population is low. An alternative indicator o f the Malaysian gap inthe stock o f skilled labor is the number o f engineers and scientists engaged in R&D. As shown in Table 3.2, Malaysia lags behind all benchmark countries. In 1998, China had twice as many scientists per million as Malaysia, and Korea had 2000 engineers and scientists engaged in R&D per million inhabitants. Still, the gap i s largest betweenMalaysia and Japan, which has more than 5,000 scientists per million. 3.16 The shortage of engineers and scientists may limit the country's innovative potential. Chapter 4 o fthis report shows that most Malaysian firms are still at the adaptation stage o f the technology spectrum and innovation i s very limited. The Flow of Human Capital: High Enrollment rates for Primary and Secondary education, low for tertiary, despite recent improvements 3.17 Inorder to assess the gap inthe flow ofhumancapital, measuredbythe enrollment rates at various levels of education, we regress the net enrollment rate on the log o f per capita GDP anda constant. The regression, runon a large sample of countriesY5is weighted bythe population ineachcountry. The line inFigures 3.2, 3.3, and 3.4 corresponds to the predictednet enrollment rate, given the country's income level. The gap i s measured by the distance between the actual net enrollment rate o f a given country and the norm for its income level (the line). InFigure 3.2, we present results o f the regression for secondary enrollment rates for the year 2000. The figure shows that Malaysia has done well in secondary education, as the country is slightly above the line. See Chapter 1o f this report for a benchmark by level of income. Malaysia's transition i s similar to that o f Korea's or Japan's. "Unbalanced" transition, o n the other hand, is similar to most Latin American countries (see D e Ferranti et al., [2003], Chapter 2). Figures 3.2,3.3 and 3.4, some country names have beenomitted to make the charts more legible. 82 Table 3.1: Stock of EducationalAttainment of the Total PopulationAged 25 and Over Highest level attained Mean Country Year Yearsof N o Primary Second Level Tertiary Schooling Schooling Complete Incomplete Complete Incomplete Complete Incomplete (Percentage of the Donulation aged 25 and over) Malaysia 1960 2.34 58.5 11.5 21.2 2.4 4.8 1.3 0.2 1965 2.67 52.1 13.8 24.8 3.3 4.5 1.2 0.2 1970 3.05 46.4 15.9 26.8 4.2 5.2 1.3 0.2 1975 3.70 37.8 19.7 28.3 6.0 6.6 1.3 0.3 1980 4.49 34.3 23.0 21.4 9.8 10.1 1.2 0.2 1985 4.88 29.9 19.8 25 11.1 12.2 1.7 0.3 1990 5.54 25.6 22.7 21.8 13.9 13.2 2.4 0.4 1995 7.65 16.7 20.6 13 23.5 19.3 5.8 1 2000 7.88 13.9 21.8 13.8 23.6 19.4 6.3 1.2 Korea 1955 3.30 29.7 18.2 43.9 1.5 5.3 1.1 0.4 1960 3.23 56.9 26.2 3.4 5.8 5.1 1.9 0.7 1965 4.43 43.6 33.5 1.7 7.8 9.7 2.7 0.9 1970 4.76 34.5 16.1 22 10.1 11.7 4.1 1.5 1975 5.77 25.2 17.2 22 14.2 14.5 5.1 1.8 1980 6.81 19.7 17.1 17.4 18.7 18.2 6.6 2.3 1985 8.03 15.4 18.1 9.5 25.6 19.7 8.7 3 1990 9.25 11.0 20.9 0.8 35.0 18.9 9.9 3.5 1995 10.09 8.7 17.3 0.9 36.2 15.7 15.6 5.5 2000 10.46 8.0 15.9 0.8 34.5 15.0 19.1 6.7 China 2000 5.74 20.9 15.3 26.6 14.1 21.6 2.3 0.4 India 2000 4.77 44.5 12.4 24.1 6.5 10.9 3.3 1.5 Japan 2000 9.72 0.0 12.9 14.3 17.4 30.5 15.0 9 Singapore 2000 8.12 12.7 16.8 12.1 13.2 35.3 7.2 3.4 Thailand 2000 6.10 17.3 27.3 36 4.1 5.2 11.2 0.1 United States 2000 12.25 1.0 4.5 5.1 21.6 18 30.3 19.8 Source:Batro, RobertJ. andJong-WhaLee, "InternationalData on EducationalAttainment:Updates and Implications,"NBERWorkingPaper No. 7911, Cambridge, MA,September2000. Table 3.2: Scientistsand Engineers inR&D(Per MillionPeople) Cnuntrv --_.. 1992 .. 1993 1994 1995 1996 1997 1998 1999 2000 Malaysia 85 114 92 160 Korea 2033 2642 2233 2193 2245 2009 2160 2319 China 353 351 454 473 387 420 545 Japan 5678 5148 6301 5369 4909 4962 5160 5196 5095 India 133 157 Singapore 2182 2699 2832 2991 3215 4140 Thailand 113 118 74 u s 3729 3730 3863 4099 Source: SIMA, World DevelopmentIndicators(April 2003). 83 Figure 3.1: Completion Rates at Different Levels of Education, 2000 k Malaysia Korea Chma India Japan Singapore Thailand United States 10Prrnarv 0 Secondarv ITertiarv I Source: Barro and Lee [2000]. 3.18 Figure 3.4 shows that, despite its tripling in the last two years, Malaysian tertiary education enrollment rate were about 3 percentage points lower than the norm for its level o f income in 2000. Indeed, in one decade Malaysia had nearly tripled the share o f its population aged 25 and over that had completed higher education (from 2.4 percent in 1990 to 6.3 percent in 2000). 3.19 The results on enrollment rates mirror those on tertiary education completion rates presented in Chapter 1 and earlier in this section and lead to the conclusion that much more effort is needed at the tertiary level for Malaysia to achieve the Eighth Plan's objectives to generate a sufficient supply of adequate skills to meet the growing demand of the economy and specifically, o fthe private sector. The demandfor skilled labor has increased 3.20 The demand for skilled labor has increased consistently, despite the sluggish post-crisis recovery and a slowdown in employment growth.6 Employment growth in manufacturing had contracted by 0.1 percent in 2001, the lowest since the 1997 financial crisis, and the aggregate unemployment rate was slightly higher than the crisis levels (3.5 percent in 2002 versus 3.1 percent in 1998). The Eight plan has rightly anticipated the possibility o f skills a shortage, but only under its highcase scenario of 7.5 percent GDP growth (see the EighthMalaysia Plan, 2001-2005, Chapter 4). 84 Figure3.2: Malaysia Has PerformedWell inSecondaryEducation FittedValuejl 100 0 80 1I3 c ar 60 w i BB 40 m z 20 I 5I 6I 7I 8I 9I 10I 11 I Logof PerCapitaGDP,2mO Source: WorldDevelopment Indicators (April2003). Sample sue 107 countries. Figure 3.3: Malaysia PerformsPoorly on Tertiary Education Fitted values 100 - 80 - United States Finland Korea 60 - 40 - 20 - 0 - I 5I 6I 7I 8I 9I 10 I 11I Log of Per Capita GDP, 1998 Source: World Development Indicators (April2003). Sample size 116countries. 85 Figure 3.4: Malaysia shows a PersistentGap inTertiary EnrollmentRates in2000 Fittedvalues 100 - 8 80 - Korea N United States dk? 60 - c. w1 4 0 1 1 ' 5I 6I 7I 81 9I 10I 11I Log of Per Capita GDP, 2000 Source: WorldDevelopmentIndicators (April 2003). Sample size 116countries. 3.21 At the same time, the unemployment duration for workers with a university-level education has decreased over time, as evidenced by an increase in the percentage o f unemployed persons with tertiary education that find ajob within three months from 52 percent in1980to 62percent in2000.7Figure 3.5 presents the share ofthe unemployedwith a university education in 1980, 1990 and 20008. It appears from this figure that there has been a persistent decline inthe share o f the unemployed engaged in ajob search for more than six months. From nearly 50 percent in 1980, this proportion declined to around 30 percent in2000. Together, these figures suggest that the demand for skilled labor at the aggregate level had been steadily increasinginMalaysia over the past two decades. 3.22 Figure 3.5 shows that the popular perceptions o f increasing unemployment among graduates are not totally supported by the data. Although aggregate unemployment rates were higher in2002 (3.5 percent) than the crisis levels (3.1 percent in 1998), and employment growth inmanufacturinghad contracted by 0.1 percent in2001, the lowest since the crisis, the demand for university educated workers was increasing. The next section o f this chapter uses the PICS data to establish the fact that this positive trend was sustained well after 2000, despite a lower than expected GDP growth. * Basedfigure on Labor Force Surveys data and EPUcalculations. The uses EPUcalculations, from Malaysia Labor Force Surveys data. 86 a Figure3.5: UnemploymentDuration, Populationwith University Education (Labor Force Survey) 60% I 7 50'/o 40% 30% 20% 10% 0% 1930 2000 3.23 The demand for skilled labor has increased. This has arisen through several sources. First, the competition from emerging the lower-wage neighbors i s forcing Malaysia to shift to products with higher skills/technology content. Second, while investment remains sluggish, capacity utilization has increased, suggesting that firms are using the excess or new capacity built up during the crisis years. Using the new equipment requires skilled workers, given the skills-technology complementarities. Last, but not least, Malaysia i s a relatively open economy. Import penetration i s high, and a significant proportion o f firms are at the adapting stage in the technology ladder. Openness to foreign trade and technology i s likely to induce skills-biased technological change by raising the relative price o f skill-intensive products andraising the wage o f skilled workers relative to unskilled worker^.^ Cross-Country Assessments of the Quality of Education 3.24 At the aggregate level, two indicators are often used to assess the quality o f human capital across countries. The first i s the quality o f educational inputs, and the second is results at standardinternational test scores.lo 3.25 The Quality of Educational inputs for Primary and Secondary School'' in Malaysia i s good. The pupil-teacher ratio is better than in most OECD countries, both at the primary and secondary school levels. In 1990, the ratio was 20.4 for primary and 19.1 for secondary; while for Korea it was 35.6 and 25.2 respectively. Only the United States, Japan and Thailand outperformed Malaysia. Similarly, the real government current expenditure per pupil at the secondary school level in PPP-adjusted 1985 international dollars was 869 in 1990, is higher This hasbeenlargely studiedinthe literature, andmost recently by Bermanet al. (1998) andAcemoglu (2003). lo This approach assumes implicitly that the amount ofpublic hnds spent on education is correlatedwithquality, which i s debatable. "TheBarroandLeedatabasethatweusedoesnotprovideinformationontertiary educationforMalaysia andthe most recent data-point i s 1990 (see Appendix). 87 than Korea (671), Thailand (550), and China (375). Only Japan (2,456) and the USA (4181) had higher figures. However, the real government current expenditure per pupil at primary school i s lower in Malaysia than in most comparators in 1990 (see Appendix 3), as more funds were focused towards the expansion o f secondary and tertiary educationinfrastructure at this stage. 3.26 Performance on standardized international tests tends to suggest that Malaysian primary and secondary educationprovides competitive skills in mathematics and science. Using the International Student Achievement in Science Test (TIMSS) data, we find that Malaysian students are well above the average in mathematics and close to the average in science, as seen in Table 3.3. The TIMSS 1999 test was taken by eighth graders (with an average age o f around 14 years) in38 countries around the world. 3.27 Malaysia ranks sixteenth out o f 38 countries inthe mathematics test with a score o f 519. The international average inmathematics is 487. Inscience, Malaysia scores 492, which is four points above the international average and ranks twenty-second out o f 38 participant countries. As shown in Table 3.3, Malaysia has one o f the lowest gaps inperformance between boys and girls. Table 3.3: International Test Scores in Science in Selected Countries T I M S S 1999, gthGrade students Difference (absolute Country Girls Boys value) Asia Malaysia 488 498 9 Singapore 557 578 20 Korea 538 559 21 Japan 543 556 14 HongKong 522 537 14 Thailand 481 484 3 Chinese Taipei 561 578 17 United States 505 524 19 United Kingdom 522 554 32 International Averape 480 495 15 3.28 However, it i s important to note that despite their intuitive appeal, international test scores are a very imperfect measure o f skills quality or deficiency. Scoring highina science test for 14 year old children is no guarantee that they will be productive workers or will operate well on a team. Also, more generally, because o f data limitations, skills are most often measured at the macro level by education attainment. This approach has serious limitations. First, equal investments in education can lead to different quantities o f skills or to skills that differ from their 88 market value. Second, owing to the mismatch, the labor market does not always hlly utilize the available skills. Third, the acquisition -but also the depreciation-- o f skills, continues after school, hence the importance o f employer-provided training (formal and informal), and o f lifelong learning. Finally, the scope for policy relevant discussion is very limited. The next section addresses the issues o f lack o f stock o f skills and education quality using plant-level and worker-level information. B. THEGAPINSKILLSFORWORK: THEMICRO EVIDENCE 3.29 In Malaysia, managers face shortages of skilled labor. In 2001, 27 percent of manufacturing plants had vacancies for professionals; it took longer to fill these vacancies than it took to fill vacancies for technicians and unskilled worker vacancies. As a consequence, managers pay high wage premiums to workers with tertiary education and to those with appropriate training. Inaddition to the shortfall inquantity, managers are also concerned with the quality o f the Malaysian-educated professionals. Twenty-nine percent o f managers perceive foreign trained professionals as better performers than Malaysian trained professionals. This can be a concern because the adequacy o f the supply (in volume and quality) in relation to the demand o f skills i s what ultimately determines employment, productivity and competitiveness o f nations. 3.30 The previous section presented evidence on the shortage o f skilled workers from the supply side. This section uses PICS data to look at the issue from the demand side. We first explore the incidence o f skills shortages andthen analyze skills mismatch. How Do Skills ShortagesAffect thePrivate Sector? 3.3 1 The Malaysian PICS worker survey data provide valuable insights to an understanding o f the market for skills in Malaysia. The incidence o f unemployment among workers is low and o f short duration. Only 26 percent o f workers in the PICS sample" have ever been unemployed during their professional life. For more than 60 percent o f these unlucky ones, the last unemployment stint lasted less than 12 months. Inaddition, only 9 percent o f workers have ever been self-employed. This situation characterizes a tight labor market with an excess demand for skills. Skills shortages are measured here by two indicators: the incidence and intensity o f vacancies o f the professional and sub-professional categories, and the wage premium paid to skilled workers. Wage PremiumsPaidto SkilledWorkers 3.32 Usingdata from the PICS workers survey for manufacturing establishments, we estimate the returns to education. The regression results provide strong evidence that tensions at the high end o f the Malaysian labor market have resulted inhighwage premiumsto workers with tertiary education and to those who have received firm-specific training. This reflects the extent of the l2The PICS Malaysia collects informationfrom 8,590 workers inthe manufacturing sector. 89 skills shortages and the high value managers place on skilled workers. Two specifications are estimated: the spline function and semi-parametric regression^.'^ 3.33 Table 3.4 presents the results for the spline specification for the sample o f manufacturing workers with 0 to 16 years o f education from the PICS survey. The return to tertiary education i s very highrelative to the returns to primary and secondary education. 3.34 Indeed, the return i s smallest for primary education (4.5 percent), large for secondary education (9.5 percent), andhighest for tertiary education (17.7 percent). Also, it i s important to note that all changes inslope are significant at the 0.1 percent level. Finally, anF-test shows that the return to secondary education i s significantly different from the return to tertiary education. Together, these findings suggest considerable convexity in the relationship between log hourly wages and years o f completed schooling. 3.35 The resultsbased on the semi-parametric regression presented inFigure 3.6 are consistent with those reported in Table 3.4. These estimates show clear increases in the slope for primary, secondary, and tertiary education. Figure 3.6 shows the mean log hourly wages, by years o f formal education. The curve connects the average o f the predicted values o f log hourly wages (for each year o f formal education) from a semi-parametric estimation o f log hourly wages on formal education. 3.36 These results are consistent with other empirical evidence inthe literature suggesting that the returnto education is high in Malaysia. Schafgans (2000) reported that men and women in Malaysia all experience increasing returns to higher education. For men in Malaysia, the returns to primary and secondary schooling and above are 4 percent and 17 percent respectively. Moreover, between 1987 and 1993 the wages o f skilled and semi-skilled workers rose by 10 percent annually, those o f managers, technical workers and professionals rose by 7.5 percent, while those o funskilled workers roseby less than 5 percent. 3.37 Figure 3.7 portrays the negative relationship between the rate o f return and completion rate for tertiary education, where the line represents the predicted annual rate o f return to tertiary education given the rate o f completion for tertiary education. l4The premium for tertiary education is higher inMalaysia compared to OECD countries, suggesting that there is scope for increasingthe supply o ftertiary education graduates. As the supply o f skilled workers increases, the shortage will reduce and so will the premium to tertiary education, as has been the case for Canada andNew Zealand. 3.38 Return to training is also high. The return to training i s substantial, as shown inTable 3.5. The first column o f Table 3.5 shows that the estimated premium to workers with any training i s 10 percent, and the coefficient i s very significant. The second column presents the results fi-om the specification o f firm-specific training. The coefficients on the variables are highly significant, suggesting that training, like education, has a positive effect on wages. Workers with training only from their current employers receive the lowest premium (7.2 13See equations (1) and (2) inAppendix 3 for details. l4This result is froma regressiononthe percentage o ftertiary education completed inthe total population and a constant. The sample size for the regression i s 63 countries. Some country names have been omitted to make the figure more legible. 90 percent), those with training only from their previous employers receive higher premiums (11.1 percent), and those with training from both their current and previous employers receive the highest premiums (15.8 percent). Table 3.4: Returnsto Education,Spline SpecificationControllingfor FirmFixedEffects Dependent variable: Loghourly wage Years of formal schooling 0.045**' (0.008) D6*(S-6) 0.050**' (0.010) Dll*(s-11) 0.082'** (0.008) Experience 0.049*** (0.002) Experience squared -0.001*** (0.00004) Constant 0.405*** (0.053) Number of observations 7,847 Numberof fwms 898 F-test [D6*(S-6)=D ll*(S-ll)] 4.15 Prob > F 0.042 R2 0.253 Note: D 6 and D 11are dummy variables for those who have completed at Least 6 and 11 years of schooling, respectively. S i s the years of completed schooling. D 6*(S-6) is an interaction term between the DG dummy and (S-6). D ll*(S-11) is an interaction term between the D 11dummy and (J-11). F-test tests that the two interaction terms are equal to each other. I** denotes significance at the 0.1% level. 91 Figure3.6: MeanLogHourlyWage byYears ofFormalEducation 2.75 2.5 F2.25 -F 2ti0 2 3 i 1.75 1.5 0 3 6 9 10 11 12 15 16 18 21 Years of formal education Sources: MalaysiaPICS 2002, andWorldBank Figure3.7: The Premiumfor TertiaryEducationI s HigherinMalaysiaThaninOECDCountries 9 3 5 - B Pakistan $ 3 0 'g8 - SubSaharanAfrica 4 25 - Europe/Middle East/North Africa* .1 i PanaAustralia * 20 - u8 15 - 3 'iii& lo a Hi&%%&&%ECD Philippines N e w Zealand - I I 1 I I I I I 92 Table3.5: Wage Premiumsto Training,Controllingfor FirmFixedEffects Dependentvariable: Loghourly wage Training O.lOO*** (0.016) Training inpresent firm only 0.072**- (0.019) Training in past firm only 0.11l*** (0.027) Training inpast and present firm 0.158'*' (0.024) Degree 0.795*** 0.794*** (0.032) (0.032) Diploma 0.507*** 0.502*** (0.028) (0.028) Upper secondary 0.262"' 0.261*** (0.020) (0.020) Lower secondary 0.138*** 0.137*** (0.019) (0.019) Experience 0.027*** 0.026m** (0.002) (0.002) Experiencesquared -0.0004**' -0.0004*** (0.00004) (0.00004) Tenure 0.025"' 0.025'-* (0.002) (0.003) Tenure squared -0.0002-* -0.0002*' (0.00008) (0.00008) Female -0.173'** -0.172'" (0.012) (0.012) Married 0.078-** 0.078*** (0.013) (0.013) Manager 0.821*** 0.816"' (0.022) (0.022) Professional 0.656"* 0.651*** (0.028) (0.028) Skilledproductionworker 0.303*" 0.302*'* (0.016) (0.016) Non-productionworker 0.372*** 0.372'x* (0.019) (0.019) Apprentice 0.061 0.068 (0.067) (0.067) Constant 0.733"- 0.736*** (0.027) (0.027) Number of observations 8,066 8,066 Number of fums 897 897 R' 0.459 0.460 Note:"Truining"=I ifyes to "34 or W35 and =O otherwise. "Training inpment on5J"=1 ifyes to W34 and no to W35 and =O otherwise. "Trainingin puston5J"=l if no to W34 and yes to W35 and =O otherwise. "Training inpustandpresent"=l if yes to W34 and W35 and =O otherwise. The omitted **. educational attainment comparison group i s ')Primaryor lessyears gscboobng". The omitted occupation group i s "unski/hdpmductionworker". denotes significance at the 0.1% level and" denotes significance at the 1% level. 93 3.39 It is importantto note that returns to training differ by the level o f education o f workers. Results from the specification o f training by level o f education suggest that the returns from training are higher for workers with tertiary education. Point estimates for trainingpremiums are: 0.17 for degree holders, 0.11 for diploma holders, 0.08 for workers with upper secondary school, 0.10 for workers with lower secondary school, and 0.09 for workers with primary or less education. By industry, the training premiums are concentrated in food processing, rubber and plastics, household electrical appliances, and automobile parts (see Appendix 3). Incidence and intensity of hard-to-fill vacancies of professionals and sub-professional categories 3.40 Managers in Malaysia face shortages o f skilled labor. In 2001, 27 percent o f manufacturing plants had vacancies for professionals; it took longer to fill vacancies for these vacancies than it took to fill technicians andunskilledworker~.'~ . 3.41 Around 70 percent of managers surveyed by the PICS identify the insufficient supply of university graduates as the most important reasonwhen asked about the possible causes of skills shortages. This result confirms the diagnosis inthe previous section. Given the growth prospects and the lags inincreasing the tertiary education completion rate, there i s a high probability that the demand for skilled workers will increase duringthe second halfo f the Eighth Malaysia Plan. 3.42 In addition to the shortfall in quantity, there are concerns about the quality of the Malaysian-educated professionals. This can be a concern because o f the adequacy between the supply (involume and quality) inrelation to the demand o f skills is what ultimately determines employment, productivity and the competitiveness o f nations. The literature on skills and labor market outcomes (see Middleton et al. [1993], and the new generation of endogenous growth models that explicitly takes into account the complementarity between skills and technology, (see Huw and Roberts [2002]), have provided consistent evidence o f the key role adequate skills can play inthe fortunes o f countries. Micro Evidence of Skills Mismatch 3.43 We explore three different approaches in measuring the skills mismatch at the micro level. The first approach i s the mismatch in relation to an optimal level o f skills that would maximize output, the second is the mismatch from a worker's perspective, a self-assessment o f workers skills' adequacy, and the third i s the mismatch inrelation to what employers consider as the desirable skills. The first approach, presented in Chapter 1 (Appendix D) o f this report exploits the production function framework and calculates for each industry the optimal skill mix, that is the optimal combination o f skilled and unskilled labor that would maximize output. The results show that there i s skills mismatch in several industries in the manufacturing sector. l5 The incidence and intensity ofhard-to-fill vacancies is the most commonindicator of skill shortage inthe literature (see Green et al. (1998). 94 The mismatch is highest in machineqdequipment and in the auto processing industries. This chapter focuses onthe last two sets o fmeasures of skills mismatch. A Worker's Perspectiveon Skills Mismatch 3.44 English language proficiency is the skill workers lack the most. Workers were asked to list the three skills they lack most severely indoing their jobs. Almost half o f the workers (47 percent) identified Englishlanguage proficiency as the most severe constraint in doing their job. As for the second most neededskill, workers were divided between information and technology skills, identified as second by 19 percent o f workers, and communication skills, ranked second by 18 percent o fworkers. TechnicaVprofessional skills came only as the third most needed skills. 3.45 Skills that workers lack most are "general skills." Table 3.6 presents the distribution of answers by occupational category for the single most important skill workers lack in doing their job. From Table 3.6, it appears that while Englishlanguage proficiency i s identified by 47 percent o f workers, another non-negligible share (14 percent) identify professional communication skills as lacking. Together, English and professional communication skills are the most needed skills for more than 61% o f workers. Following Becker's (1964) distinction, English and professional communication skills are "general skills" because they are not firm- specific. This may point to the education system as the source o f the problem because it should have provided these skills. When asked about the level o f education where the skills they lack should have been provided, more than half (52 percent) o f the sampled workers identified secondary education as the level at which they should have been taught these skills. Together, these results suggest that a reform of the curriculum at the secondary level aimed at strengthening the provision o f general skills could be beneficial. Table 3.6: Skills Workers LackMost: WhatI s the FirstMostImportantSkill WorkersLackinDoingTheir Job? (Percentage of workers by occupation) I L 5 Lt D U M-btt R b ~ h lSHlw- I k s k U d ~ d u N O I N C O & X hrticr L ~ TotalsampleI 1 E@sh-pmfiky 30.03 28 D4 51.I6 55.5 50.57 50 47.52 2 P l o f e s s i o d c o d a t k n s k 21.73 17.9 11.37 10 83 18.78 8.06 13.9 3 S midskills 7.57 7 98 725 8 97 74 8 9.68 7.98 4 T e r n o r k i n g 4 % 8 32 7.95 7 2 1 3.71 8.45 6.84 5 Leadership skills 5.05 5 9 7 4.49 3.7 4.09 3.23 4.38 8 Tim managementskills 8 . 4 7 98 2.88 2.8 3.03 8.46 3.77 7 Adaptability 22 5 2 s9 1.78 1.78 1.82 1.61 1.9 8 creativity/ innovaticlnsML 8.39 5 9 7 3.12 3 2 3 3.18 1.81 4.08 9 Numelical5kills 1.44 127 0.98 0 7 8 1.74 1.11 10 Pmbhmsolving 1.08 2.17 1.7 1 1.44 4.84 1.44 11 ITskdh 7.12 8 32 4.22 199 3.28 3.23 4.14 1.2 TechinaUpm&sskM].~kilL 4.24 5 2 4 3.18 2 A7 2.43 4.84 3.18 T ~ t d 1 OD 100 100 100 1 00 100 100 3.46 English language deficiency i s present across all occupations, as shown inFigure 3.8. More than 30 percent o f managers and one professional out o f four inMalaysia feels constrained in hidher daily job by a lack of English language proficiency. In addition, skilled production 95 workers stand out, with 50 percent declaring that they lack Englishskills. The extent o f English proficiency o f skilled production workers, nearly as high as that o f unskilled production workers (56 percent) is likely to translate into productivity gaps. Skilled production workers are the backbone o f any manufacturing plant. They operate machinery, read and interpret the manuals and instruct unskilled production workers. Enforcing the English curriculum in secondary schools and as the language o f training intechnical vocational schools would be useful. 3.47 The East Coast exhibits the highest gap in English language proficiency as seen in Figure 3.8 (63 percent o f workers as against 45 percent for the Central Region). The English skills deficiency is likely to be more pronounced at the national level, as the workers interviewed in the PICS are better educated than the average Malaysian worker. Thirty-six percent o f sampled workers have some upper secondary schooling, 28 percent have some lower secondary education, and 16 percent have some primary school education. Around 20 percent o f sampled workers have some tertiary education with more than 12 years schooling (university-level education), higher than the national average (see Barro and Lee [2000.]). 3.48 Almost one-third of the Malaysianworkforcein the manufacturingsector lacks the skills they need to do their job, andthe industrieswith the highest proportion o f workers with skills mismatch are also those identified in Chapter 1 of this report with the highest cost of the mismatch and the highest rate o f managers' complaints about skills shortages. Inaddition, more than one-thirdof workers do not have the skills needed to adapt to labor market changes. From Table 3.7, it appears that 36 percent of workers in the total sample are not properly equipped. It is worth noting that the industries with the highest proportion o f workers with skills mismatch are also those where the cost o f the mismatch i s highest and where managers complain most about skills shortages. Indeed, consistent with the findings o f Chapter 1, auto parts (46 percent) and machinery/equipment (41 percent) are among the industries with the largest skills mismatches. 96 Figure3.8: TheExtentofEnglishSkills Mismatch Share ofworkers rankingEnglishlanguageproficiencyas the first skillthey lackthe mostindoingtheir job 70% 60% 40% 50% 30% 20% 10% 0% Source: Malaysia PICS 2002. 3.49 Lifelong learning is not yet a reality. Workers were asked if they knew where they could acquire these skills they believe they lack. Only 17 percent o f them gave a positive answer. Moreover, only 5 percent of the workers interviewed are currently enrolled in an after work learning program, and there i s some Government hnding in less than 20 percent o f these few cases. The Government could reach out to more people, providing information on the existing facilities and the numerous training opportunities. 3.50 There is an education mismatch at the tertiary level. Because o f the shortage of university graduates, firms are forced to hire workers with a diploma to do the job o f a graduate. As shown in Table 3.8, more than 21 ercent o f workers with a diploma estimate that their current job requires a University Degree!6 Similarly, 18 percent ofhighschool graduates believe they are over-utilized in jobs requiring a diploma. These results suggest that the shortage in tertiary education graduates is leading managers to an under optimal hiring policy. This mismatch could causeproductivity losses at the plant level. 3.51 Fourteen percent of workers do not have the relevant education background for their work. An alternative approach to assess the skills mismatch inthe manufacturing sector is to look at the proportion o f workers whose field o f education is remote from the field needed to do their present jobs. Employees were asked to rate the level o f relevance o f their educational background to the work they were doing. l6In a similar study of 2,460 Dutch graduates, Allen and Velden (2001) find that 14percent o f higher vocational education graduates and 8 percent o f university graduates were working injobs for which they considered a higher level o f education was more appropriate. 97 Table 3.7: Skills Mismatch Share of manufacturing workers by education and industrywho declare not having the skills needed to adapt to labor markets changes (Percentage inparentheses for sample of workers who have studied abroad) Food Processing Textiles Garments Chemicals l=Degree 16% (1l0/o) 15% (8%) 27% (0%) 22% (13%) 22% (9"/0) 2=Diploma 23% (20%) 30% (45%) 14% (0%) 21% (0%) 5% (67%) 3=Upper secondary 31% 35`/o 32% 29% 21Yo 4=Lower secondary 43% 47% 42% 42% 28% 5=Primary 52% 63% 68% 44% 38% Total 36% 43% 38% 37% 24% Rubber & Machinery Plastics 82 Electronics Auto Parts Furniture Equipment l=Degree 14% (9Yo) 14% (13%) 11%(15%) 18% (150/0) 21% (13%) 2=Diploma 19% (5%) 28% (0%) 22% (25%) 30% (33%) 20% (33%) 3=Upper secondary 28% 39% 29% 45% 31% 4=Lower secondary 36% 50% 48% 63% 38% 5=Primary 49% 56% 44% 48% 44% Total 31% 41% 31% 46% 35% Source: Malaysia PICS 2002 3.52 Overall, only 8 percent o f workers think the ideal field o f education to do their job i s the one they possess. As shown in Table 3.9, 14 percent o f workers are employed in areas inwhich the ideal field o f education is completely different from their own. This proportion is equally distributed across ethnic groups. 3.53 Around 40 percentof the unemployedgraduates are in areas that are not of interest to manufacturing activity. The education background mismatch is also evidenced by the qualifications of unemployed graduates registered for training schemes. Table 3.10 gives the distribution between November 2001 and April 2003 by field o f education. It appears that 17 percent have a degree in arts, 8 percent are religious studies graduates, and 14 percent are qualified inother disciplines. The proportion of unemployed graduates is lowest inEconomics and Accounting. The highincidence of Information Communication and Technology (ICT) may be explained by a combination o f factors, including the global downturn inthe IT sector that led to lower employment opportunities and the heterogeneity o f qualifications in the I C T sector, some o fwhich may not match employers' requirements. 98 Table 3.8: EducationMismatch:Levelof Education I What 1 vel of education is more appropriate for your work? Worker's EducationA l=Degree 4=Lower sec. 5=Primary 77.7 1.5 0.4 2=Diploma 14.8 56.7 18.1 6.8 2.4 S=Upper secondary 3.9 16.7 54.8 31.2 16.3 =Lower secondary 2.6 4.1 16.8 44.1 32.3 5=Primary 0.7 0.3 4.2 10.6 36.3 6=Informal 0.2 0.3 0.8 2.3 5.0 17=None (aterate) 0.2 0.3 1.9 3.4 7.3 Some: Malaysia PICS 2002 Table 3.9: EducationMismatch:RelevantEducation Bumiputera.Chinese Indian Others Total l=Degree 16% 11% 10% 50% 13% Source: Malaysia PICS 2002 Table 3.10: UnemployedGraduatesRegistered under the Training andAttachment Scheme Cumulative figures from November 2001to April 2003 Percentage Number of Individuals Engineering 10.2 4,503 Economics 3.9 1,711 Accounting 7.8 3,436 Business Administration 14.3 6,300 Science 8.6 3,796 Arts 16.6 7,284 ICT 16.9 7,441 Religious Studies 8.0 3,526 Others 13.6 5,982 Total 100.0 43.977 ~Source: HR Section EPU 99 MismatchinRelationto What EmployersConsider the DesirableSkills 3.54 Managersare concernedaboutthe quality of Malaysian-educatedprofessionalsand unskilledlabor. Managers were asked, basedon their experience inthe years 1999-2001, to rate the quality o f Malaysian workers in comparison with their foreign counterparts working in the same establishments. Twenty-nine percent o f managers believe that foreign trained professionals perform better than Malaysian trained professionals and seventeen percent believe that Malaysian trained professionals perform better. However, the results are mixed when managers rate Malaysian professionals against their foreign counterparts without distinction as to where they received their education. Thirty-seven percent o f managers rank foreign unskilled workers as better than Malaysian unskilled workers, while 28 percent state the reverse. c. DETERMINANTSSKILLSMATCH OF 3.55 This section examines the key drivers of skill adequacy, including education and in- service training, and explores the interaction between firms and skills development institutes. Education and in-service training are found to be complementary means o f reducing the skills mismatch and increasing the employability o f graduates. Analysis o f the content o f training reveals that firms value specific training. The Government has set up high quality skills development institutions but their use i s low, especially by the SMEs. Estimating a Skills' Production Function 3.56 Inorder to investigate the key drivers of skills matching, we adopt the framework of a simple aggregate skills productionfun~tion.'~Tables 3A.6 and 3A.7 inAppendix 3 report probit regression results o f a skills adequacy (sufficiency) function estimating the probability that a worker i s adequately equipped with the skills needed to perform hisher job. We find that both education and training are key determinants o f skills adequacy. The higher the education level, the greater i s the likelihood o f being adequately skilled. Similarly, having received training and having completed a professional certification program or having attended a Polytechnic, is associated with a higher probability o f being properly equipped. Moreover, firm characteristics matter. Exporting status and plant size both significantly affect the probability o f a worker's being skilled. Together, these results suggest that a comprehensive approach to skills provision should be adopted, that would better integrate the school, the firm and professional training infrastructure. However, these results should be interpreted with caution, There i s a potential endogeneity o f training variables, as workers who have received training are likely to self-assess themselves as skilled. The role o f education and training in providing the workers with skills needed for work needs to be further explored. l7The skills supply function assumes that the level of slulls embodied ina worker at a given period is determined by a vector o f education inputs and a vector o f indices o f prior inputs o f work-based learning, see Green et a1 (2001) and Appendix for the specification. 100 Education Education Policy in Malaysia 3.57 Since independence, Malaysia has viewed education as central to its development strategy. The Malaysian Government envisions developing Malaysia as a center o f educational excellence in the region, increasing the participation rate in higher education to 40 percent by 2020 and making education a revenue generating sub-sector. Box 3.2 briefly describes the Malaysian education system. 3.58 Malaysia has witnessed an extraordinary expansion o f the demand for skilled workers over the last decade. In order to address the issue, the Government has taken measures to increase the supply o f skilled labor. The most significant measures have included expanding the national education and training capacity (both public and private) and providing incentives to increase female labor participation. Among other achievements, it is worth notingthat three new universities and three new polytechnics were opened during the Seventh Malaysia Plan (1996- 2000). In addition, the country has continued to rely on foreign workers, particularly skilled workers, and currently outsources 10 percent o f its demand for skilled workers. 3.59 These actions have resulted in increased accessibility to education and training and enrollment at all levels. But they have yet to translate into lower skills shortages, given the lag in producing new graduates. As the current university intakejoins the labor market inthe next four years, the skills shortage could be reduced. 3.60 The quality o f the output o f the Malaysian education system continues to be an issue o f concern. As shown earlier in this chapter, the skills mismatch i s pervasive in the Malaysian manufacturing sector. The mismatch is likely to stem from the low skills-content o f the education. 101 Estimating the Skills-Content of Education Over Time 3.61 We formally test the contribution of the education system to the supply o f skills for work. The question we ask i s the following: has education improved its skills-content over time? Or, alternatively, has education reduced the inadequacy o f its content with the needs o f the workplace overtime? To answer this question, we follow Freeman and Shettkat (2001) and regress a measure of skills adequacy o f the workforce on education as a sole explanatory variable with age andwith an interactionterm betweenage and education. 3.62 Table 3.11 reports the change in the probability for a iven worker to be adequately skilled for an infinitesimal change in each independentvariable.If The regression results show that an additional year of education increases the likelihood of possessingthe needed skills by 3 percent. The coefficient of the interaction term in column 3 is positive and not significant, suggesting that the education system has not improved its effectiveness in providing skills over time, so that the education of older cohorts increases the probability o f being skilled more than the education does of younger cohort^.'^ This result may imply that the Malaysian education See Table 3A.8 inAppendix 3 for the full results. l9Using the Comparable German-American Structural database (CGAS), Freeman and Shettkat (2001) find a negative sign for the interaction term for both Germany and the United States, suggesting that education and age are substitute ways o f acquiring skills. 102 system has provided older generations with more English skills than it i s now providing to the youth.2o Also, it i s quite plausible that this result captures the effect o f a higher percentage o f workers o fthe older generation who studiedabroad (inEnglish). Table: 3.11: Skills Content of Education Probit regression Dependant variable is a dummy variable equals one ifthe worker i s skilled Marginal Effects (dF/dX) Equation (1) Equation (2) Number ofyears of .0324806 .0314018 Education completed (.0017953)*** (.0023241)*** Worker's Age -.0001715 (.00005GG) Interaction .0000448 Age*Education (.0000566) Notes: Robust standard errors inparentheses. *** Sipficance at 1%level . In-Service Training2` 3.63 Training incidence is high in Malaysia. Figure 3.10 presents the employer provided training incidence for five countries for which plant level data collected by the World Bank is available. Forty-two percent o f manufacturing firms provide some formal training for their employees. This result i s consistent with previous findings inthe literature. Tan andBatra (1995) have shown that, in the early 1990s, the training incidence inMalaysia was at levels comparable with the UnitedStates. 2oEnglishlanguage proficiency is the skill the most indemand. 21 In-service training i s defined here as formal in-house or outside training, that is, a learning event taking placeina classroom setting, inside or outside the firms' premises. 103 3.64 A key explanation o f the high training incidence relative to the few countries with comparable data i s that Malaysian manufacturing firms have engaged in massive technology upgrading over the last decade.22Using panel data from Malaysian manufacturing firms, Tan (2002) finds that the percentage o f firms introducing new technology rose from 53 percent in 1988 to 75 percent in 1997. Similarly, the share o f firms using advanced process IT rose from 48 percent to 83 percent over the same period. As a consequence, the need to operate new machineries/technologies has ledfirms to look for and value firm-specific training. 3.65 Indeed, analysis of the content of training from the PICS worker data suggests that employers pay higher wage premiums to workers with firm-specific training, as shown in Table 3.12. Courses such as instruction in production, information, and management/quality technologies are the most sought after courses. Conversely, there are no premiums to training in more general skills, namely marketing, intellectual property, safety procedures, and language skills. The pattern in these results suggests that improvement in general skills needs to be achieved by the educational system. Indeed, as evidenced in the literature, firms have no incentive to provide their workers with general skills because their transferability favors poaching externalities (see Becker [19601, Acemoglu andPiscke [19981). 3.66 Within Malaysia, large firms provide four times as much formal training as small ones. Figure 3.9 presents formal training incidence in Malaysia by firm size and some firm characteristics, including exporter status and the presence o f foreign equity. This figure shows that while around 80 percent o f large establishments provide training, less than 19 percent o f small ones provide it. Training incidence for medium-size firms (50 percent) i s slightly higher than the average for the sector (42 percent). While this i s a standard result in the literature, the magnitude i s very important inMalaysia and may signal impediments to training for small firms. Figure 3.9 also points to the fact that the percentage o f exporting firms and those with foreign equity that train i s double that o f firms producing exclusively for the domestic market and domestic owned firms. While 52 percent o f exporters provide training, only 31 percent o f non- exporters provide it. The result is more striking for foreign-owned firms with 62 percent against domestic. However, this result likely to capture firm size as most foreign or exporting firms are large. 3.67 In addition, there is preliminary evidence that the formal training incidence for small firms may have decreasedover time in Malaysia. Table 3.13 presents the incidence o f formal training inMalaysian manufacturing by firm size inthe ILTD Survey conducted in 1997 and the PICS survey. It appears that training incidence for large firms remained constant at around 78 percent between 1997 and 2002. However, the proportion o f small firms providing formal training to their employees dropped dramatically compared to the pre-crisis levels. In 1996, around 34 percent o f small firms were trainingtheir workers. By 2001, only 18 percent o f small firms were providing training. While this result i s striking it should be interpreted with caution as the comparison is made from two points in time, using two different data sources. The result may therefore capture differences insample size and structure. Ideally, we should use 22 Another factor could be the existenceo f an excellent skills development infrastructure. W e next present an assessment o f the skills development institutes inthis section. 104 a balanced panel to assess the trend in incidence of firm-level training, but the data are not available. Table3.12: Wage Premiumsto Trainingby Content, Controllingfor FirmFixedEffects Dependent variable: Loghourly wage Productiontechnologies 0.057# (0.020) Marketing 0.021 (0.044) Information technology 0 . 0 8 T (0.029) Management/quality technologies 0.142*# (0.019) Intellectualproperty 0.124 (0.095) Safety procedures 0.006 (0.024) Language skills -0.010 (0.037) Others 0.102*** (0.024) Notes : Covariates: dummies for educational attainment (Degree, Diploma, Upper secondary, Lower secondary, Primary or less), experience, experience squared, tenure, tenure squared, gender, marital status, and dummies for occupation (Manager, Professional, Skilled Production Worker, Unskilled Production Worker, Non-production Worker, Apprentice). *-*, ,and significanceatthe 1%,5%, and 10%levelrespectively. I* * 105 Figure3.9: Employer-ProvidedTrainingIncidenceinMalaysia Share of establishments that train workers 90 80 I 70 x 60 50 40 30 20 1 0 0 Malaysia Small Medium Large Ekporter Non- DomesticForeign E+ orters Source: PICS 2002 Table3.13: TrainingIncidenceOver Time inMalaysianManufacturing 1997 ILTD Survey 2002 PICS Survey Small Firms 34.2 18.8 Medium Size Firms 56.2 50.6 Large Firms 77.8 78.5 Total Sample 58.4 42.0 Number of Observations 603 902 Source: Tan, 1999, PICS. 3.68 Can the modest training performance by small firms be explained by highlabor turnover? The literature on the importance o f quits for firms' training decisions and the possibility o f market failure, inparticular training externalities, i s extensive. See Booth and Zoega (1999) and Booth et al. (2002) for a review. The basic idea that i s widely shared i s that quits discourage training (both general and specific) by firms. Also, worker turnover imposes capital losses on firms that provide specific training. Has HighTurnoverReducedthe Incidenceof Trainingfor SmallFirmsinMalaysia? 3.69 The labor turnover rate is high and policymakers and the private sector alike tend to see it as a key impediment to employer-provided training. This result is confirmed by findings from the worker survey. Indeed, only 26 percent o f the workers in the PICS sample have ever been unemployed duringtheir professional life. Inaddition, for more than 60 percent of these unlucky ones, the last unemployment stint lasted less than 12 months, signaling a tight labor market with potentially significant firm-to-firm transfers. 106 3.70 Worker turnover is highest among unskilledworkers, as shown by worker turnover by occupation inFigure3.10. While the quit rate in2001 was around 10percent for professionals and 12percent for skilled production workers, it was around 40 percent for unskilledproduction workers. The high turnover for unskilled production workers i s puzzling and may point to an issue with labor regulation. Indeed, 16 percent o f unskilled production workers surveyed are o f foreign origin23against only 11percent inthe skilledproduction worker category.24 3.71 The Malaysian legislation concerning foreign unskilled workers is very strict. Foreign unskilledworkers are hired for three years, and should return to their home countries at the end of the contract period unless the employer proves that they have become skilled. The worker survey data confirms the very temporary nature o f the employment contract for foreign workers in Malaysia. Less than 10 percent o f the foreign workers have been in Malaysia for more than 10 years. Moreover, around half o f the sample o f foreign workers has been in Malaysia only since the year 2000. This regulatory constraint may partly explain the high turnover rate inthis category and may have a negative impact on plants' productivity. This could operate by disrupting any learning process that may take longer than three years for unskilled workers. Unskilled employees are sent o f f when they are about to reach the full speed and replaced by new batches o f employers that need more informal training. This would imply adjustment costs for the firm with negative effects on its cost competitiveness. Figure 3.10: Plant-Level Hire and Quit Rates by Occupation and FirmSize, 2001 Hire and Q u i t R a l e s in Manufacturing b y O c c u p a t i o n f o r S M E s a n d L i e Firms 4 5 1 I 3 o i . . b 2 5 B 2 0 1 5 1 0 5 n A I1 M ana gers P r o f e s s l o n a13 S k 111 Unskilled N o " . Employees 1 Prod u clmn Production P m d uetlo n W orkers W arkers W orkers 0 S M E H i r e R a l e , I S M E Q u i t R a l e 0 L a r g e F l r m H i r e R a l e Q L a r p , F l r m Q u ~ t R a l e Source: PICS 2002 3.72 In order to investigate more formally whether quitting explains the low training incidence, we regress the quit rate at the plant level on training variables (dummies for in-house training and outside training), controlling for plant characteristics such as age, size, ownership structure, and a set of technology and innovation variables. Table 3A.9 inAppendix 3 A presents the results o f the regression. We find that quitting i s not always associated with lower training. 23Among foreign unskilled workers, Bangladeshi represent 36 percent. 24This may suggest that the shortage o f skilled workers is coupled with a shortage o funskilled workers. 107 Only training provided outside the establishment premises is positively correlated with quitting in Malaysia. In addition, firm size has a negative and significant coefficient, suggesting that quittingis more o f an issue for smaller firms. Therefore, the question is how to reduce quitting, especially for SMIs? As Becker (1964) would suggest, "Firms can discourage such quits by sharing hiring costs and the return with employees, but they would have less need to discourage them and would be more willing to pay for hiring costs if insurance were provided." Most recently, Stevens (1994), Acemoglu and Pischke (1998), and Zeufack (1999), among others, have shown that imperfect competition inthe labor market may open the way for market failures inthe case ofgeneral training and leadto higherlevels ofemployer-provided training. 3.73 Three suggestions follow from this analysis. The first i s to promote more in-house training by encouraging Malaysian firms to build training facilities in their premises, and encourage training providers to deliver trainingon-site. The second is to lower the costs o f hiring new workers for firms, andthe third is to set up a training insurance system for SME's registered with the HRDF. A possible implementation device could be that HRDF credits back a share o f the training cost to the account o f the SME, if the employer provides evidence that the trainee has quit inthe two years following training. Interaction betweenFirms and the Malaysian Skills Development Infrastructure 3.74 Malaysia has a world class skills development infrastructure. The HumanResource Development Fund (HRDF) training scheme i s one good example o f the quality infrastructure that Malaysia has created. The HRDF training scheme has been a powerful driving force for training and productivity in Malaysia since its inception in 1993 (see De Feranti et al. [2003]). By December 2002, 8,172 employers had registered with the HRDF. In 2002, 722 new employers registered, of which 448 were inmanufacturing and 274 inservices (see Box 3.3).25 3.75 Inthe PICS survey, managers were asked if they are registered with the HRDF. Ifso, they were requested to assess the effectiveness of this scheme. In our sample, 48 percent o f establishments are registered. This highregistrationrate may reflect the fact that the PICS covers only formal establishments with more than 10 employees. Forty-five percent o f manufacturing firms believe that participation to the HRDF has been critical to their decision to train. These results confirm Tan (2002) who finds on a panel o f firms that participating inthe HRDF scheme i s associated with more frequent training. 3.76 HRDF'sperformance. The PICS provides preliminary evidence that the HRDF, which i s critical to firms' decision to train, may have grown slower over time. It takes five weeks on average to process the reimbursement o f claims. This i s slow inrelation to the 18-day target set by the (Pembongunan Sumbero Manusia Berhad (PSMB), the training council created with the HRDF, (see Box 3.3). It is interestingto note that PSMB internal data tracking exhibits different numbers. The average processing time for PSMB sources was 18.7 days in 2000, 18.1 days in 2001, 13.4 days in 2002 and 16.4 days in 2003 (as o f November 7, 2003). This significant gap between firms' estimations andPSMB data may be explained by a difference indefinition o f the processing time. While firms assess total time to obtain reimbursement, PSMB only accounts for 25SeeTable 3A.10 inAppendix 3 for distributionby state. 108 the time to process a properly filled claim application. These are two different concepts, as firms, especially SMIs, most often do not fill out properly their application forms correctly and several supporting documents are missing. As a consequence, they have to go through several iterations before the claim can be processed.26The process has become quite tedious for SMIs to the point where the training schemes managed by the national S M E agency (SMIDEC) have become more attractive to firms. SMIDEC provides integrated business, training and technology development services to SMIs. A better targeting of SMIs inworkshops and clinics organizedby the P S M B as partof its outreach program couldhelp to address this issue. Itis important to note, however, that the PSMB has recently implemented schemes such as PERLA (agreement with training providers) and SBL-PKS (training grant scheme for SMEs) and SBL-Khas (special training assistance schemes) which authorize training providers to claim training fees directly from the PSMB. 3.77 The rate o f use o f the training levy collected by the HRDF in 2002 was only 56 percent (62 percent manufacturing, 37 percent services).27 The rate o f use is even lower for registered SMIs (50 percent). Inaddition, there was a decrease inthe amount o f levy paid by employers in 2002 due to the reduction o f the levy rate from 1percent to 0.5 percent. Although this decrease was not associated with a reduced number o f training places created and the amount o f financial assistance approved in 2002, it highlights the need to attract a larger pool o f employers. Improving the functioning o f HRDFwould be critical for firms' training. Indeed, 84 percent o f firms sampled would provide more training if the process was more efficient. Stepping up outreach programs towards SMIs, simplifying the application procedures, encouraging electronic submission o f applications and processing of claims, and authorizing training providers to help out first time SMIs applicants inthe procedure could be envisaged. 3.78 Overall, firms, especially the SMIs, seldom use the skills development institutions. Managers were asked the extent to which they have used the skills development institutions (SDIs). As shown in Figure 3.11, only 21 percent o f plants sent their workers for training in a skills upgrading institute in the past three years. In2001, while 39 percent o f large firms sent their workers to skills development centers, only 15 percent o f SMIs worked with SDIs in2001, The rate o f use i s even lower for pre-employment training institutions. Only 14 percent o f establishments hiredgraduates from public vocational institutions in2001. 3.79 However, the plants that use skills development institutes rate them very high. Table 3.14 shows the rate o f satisfaction o f managers interviewed. More than 75 percent o f managers rank the top three institutes they use as o f good quality, and around 20 percent believe they are o f very good quality. The Government should scale up the good work o f the SDI, in order to mitigate the impact o f the skills shortage inthe economy. 26We are grateful to HRDFofficials for providing this explanation. 27HRDF Annual Report, 2002. 109 110 3.80 Poor outreach is blamed as the chief reason for low utilization. When asked about the reasons for the absence o f interaction with SDIs, 45 percent o f managers believe that the services offered by SDIs are not relevant to their plants' needs. Another sizable portion (37 percent) do not know of any skills development institute they can refer to. Another equally important reason i s that 37 percent o f managers do not know how to make the first contact with SDIs. Similar findings are uncovered for technology development institutions (TDIs), which are discussed in greater detail inthe next chapter. 111 Figure3.11: ShareofPlantsUsingSkills DevelopmentInstitutesby Sector v ru I I -aI I I I I v) .-CI S m CI S .-v) u S 5VI f3 Q m 6cu L L 0 CI 13 3 ; a: a U Source: PICS 2002 Worker survey Table 3.14: QualityofFirms' Three MostUsedSkills DevelopmentInstitutes Quality of Institute First Most Used Second Most Used Third Most Used (YOof respondents) Very Poor 1.2 0 0 Poor 0.6 4.2 1.8 Fairly Good 75.6 74 79 Very Good 22.6 21.9 19.3 Number of respondents 164 96 57 D. CONCLUSIONSAND POLICYIMPLICATIONS 3.81 Despite the considerable effort made by Malaysia, the gap in skills performance remains. The gap remains in terms o f both skills shortages and skills mismatches across all regions in Malaysia. The gap inskills performance has affected output growth inthe manufacturing sector: the manufacturing sector would have increased its growth by 11percent more if there were no skills shortages. 3.82 Alleviating the effect o f skills deficiencies on the Malaysian economy would require action from the Government and also from the private sector. Inorder to meet the challenges o f globalization and international competitiveness and to achieve Vision 2020, the Government will need to continue to reinforce actions aimed at increasing the supply o f tertiary education graduates. These actions could include further promoting the private sector's role in tertiary education. Scaling up the good work o f SDIs will also be important. Specific actions here could include improving access to and use o f their services, especially by smaller firms. Improvingthe 112 efficiency o f the HRDF could be very beneficial. Also, firther relaxing the constraints on hiring foreign skilled labor could release the pressures in the short run. Attracting Malaysians in the diaspora, and increasing outreach activities and finding in order to promote lifelong learning, which is presently being undertaken by the Government o f Malaysia will contribute toward alleviating the problem. 3.83 Inorder to improve the quality ofthe output ofthe education system, Malaysia should continue to emphasize the needto foster the provision o f Englishlanguage skills and to reinforce the IT and communications skills content o f the secondary and tertiary education curricula. Also, reinforcing public-private sector collaboration in identifying the specific skills needs o f firms and in defining the skills content o f secondary and tertiary education could be critical. Indeed, skills upgrading efforts will not be sustainable unless they are ledby the private sector. 3.84 The results o f the PICS were used as a basis o f discussions with Government officials, representatives o f SDIs including PSMB, and training providers. Table 3.15 summarizes these discussions in a matrix format and offers suggestions on potential ways to reduce the skills shortage and mismatch. 113 U 0 Y e e ee e e e * v, 3 3 e 0 . 0 0 e 4. STRENGTHENINGTECHNOLOGICAL CATCH-UP 4.1 The global economy is undergoing rapid development into a knowledge-based economy where technology, skills, and innovation will be the determinants for enhancing competitiveness and efficiency. Taking cognizance of this, the Government o f Malaysia continues to accord a high priority to innovation-driven development. In fact, the Eighth Malaysia Plan, 2001-2005, identifies the "need for a shift from a growth strategy that is input- driven to one that is knowledge-driven" inorder to enhance the competitiveness o f the economy andstrengthen economic resilience. 4.2 The Eighth Malaysia Plan aims for science and technology (S&T) based development to contribute more than one-third o f the economic growthtarget (of 7.5 percent per annum) by optimizing the application o f new and improved technology, increasing the indigenous innovation capability and providing an enabling environment for technology development by focusing on the following: Adopting an integrated national approach inthe use o f R&D resources in order to ensure more effective and efficient implementation o fresearch and innovationprojects; Accelerating the rate o f commercialization o fR&D findings Enhancing further private sector involvement inand commitments to R&D activities Increasingthe supply o f scientific and technological manpower Acquiring new and imported technologies through, among other means, acquiring equity inforeign companies and forging strategic alliances; Promoting the development o f indigenous S&T capabilities in strategic and key technologies Improving and expanding technical extension services and training to strengthen the technological capability o f SMEs. In2003, Malaysia launchedthe second National Science and Technology policy which is aimed at maximizing the utilization and advancement o f S&T as a tool for sustaining economic development, improving the quality o f life and enhancing national competitiveness. 4.3 To assist in the attainment o f these goals o f the Plan, this chapter addresses the following issues: Where i s Malaysia on the technology-competitiveness spectrum? A r e Malaysian firms technologically active in the adoption and adaptation o f existing technologies? 117 I s Malaysia ready to make the transition to becoming a creator o f technology, or i s this premature? Are Malaysianpolicies appropriate and effective for this transition? 4.4 The analysis in this chapter uses newly collected data from the Productivity and Investment Climate Survey (PICS) o f 902 Malaysian firms in the manufacturing sector on various issues, including technological competitiveness. The survey inquires into the extent o f technological activity within firms, the sources o f technology for firms, and the relationship with other firms/universities/institutions in undertaking technology development. Firms also report their impressions o f the existing technology-support infrastructure as well as the business environment for undertaking technological efforts inMalaysia. 4.5 Because the survey data represent currently a cross-section o f firms at a point in time (rather than a panel dataset with a time series element), it i s important to rely on other sources o f data to obtain a comparative picture for Malaysia, both over time, where possible, and incomparison with competition countries. Fortunately, several innovation-related indices exist and shed light on Malaysia's level o f technological competitiveness at an aggregate level. '. Together, the aggregate data and the PICS dataset help to arrive at a coherent assessment o f Malaysia's technological position. The PICS dataset, o f course, builds on the assessment by pointing to constraints on technological efforts at the firm level. 4.6 The chapter is structured as follows. The next section sets the stage by presenting the macro picture for Malaysia based on the available aggregate data. The chapter then approaches the development o f technological capability from the firm's perspective by outlining a simple framework and diagnostics to empirically support the framework. Then the results o f regression analysis are discussed in order to identify the most important firm-level attributes needed to advance the technological capability o f Malaysian firms. Finally, the chapter puts forth for further discussion issues that are starting points for the design and implementation of policy reforms. A. INTERNATIONAL COMPARISONS: STOCK TAKING OFTHE EXISTINGMEASURES OF TECHNOLOGICAL PERFORMANCE 4.7 Four measures of national innovative performance. In order to gain some insight into Malaysia's technological capabilities, a quick review o f some international comparisons i s useful. International comparisons are based on aggregate level data, and several rankings o f national performance have been attempted. To get a sense o f how Malaysia performs nationally, four popular measures are selected here for discussion: 1. Global Competitiveness Rankings from the World Economic Forum 2. High-Tech Indicators rankingsfrom the Georgia Institute o f Technology 3. Exports 4. Patents. Caution should be exercised ininterpreting cross-country comparisons, especially when the data i s qualitative or based on very small sample sizes, as i s the case with the fnst two indicators discussed inthis section. 118 Global Competitiveness Rankings (GCR) 4.8 These rankings are based on two indices that capture the underlying conditions for sustainable productivity and medium-term growth based on aggregate level institutions, the business environment, and firm-level operations (see Box 4.1 for details). 4.9 Malaysia performs well on micro-level indicators, except on technology competitiveness. Malaysia shows the largest increase inthe microeconomic index byimproving its rank to 26 in 2002 from a rank o f 37 in 2001 (see Table 4a.l in Appendix 4 for selected countries and GCR for ranking o f all countries). Other noteworthy performers are Taiwan, which went from 21 to 16, China which went from 43 to 38, Korea which went from 26 to 23. Meanwhile, Brazil and India worsen in the rankings by going from 30 to 33 and from 36 to 37, respectively. The growth competitiveness index for Malaysia also improves, from 30 to 27. This index comprises a macroeconomic rank which remains the same at 20, a public institution rank which improves from 38 to 33, and a technology rank which worsens from 22 to 26. 4.10 While Malaysia is very well poised to receive global knowledge, local inventive activity i s lagging. The worsening technology rank deserves attention. It comprises an innovation subindex which captures R&D activity in firms and the links with universities - here Malaysia receives a rank of 52, an ICT subindex where Malaysia has a rank o f 32, and a technology transfer subindex based on the level of FDI and importance o f licensing. Here Malaysia receives the highest rank o f 1. The conclusion one draws from this is that while Malaysia i s very well poisedto receive global knowledge, local inventive activity is lagging. 119 High-Tech Indicators (HTI)Index 4.1 1 Performance is weakening on high tech indicators. This index is also a composite o f several attributes captured in sub-indices o f National Orientation, Socioeconomic Infrastructure, Productive Capacity and Technology Infrastructure. Each o f these is explained below briefly, along with Malaysia's performance along each one (see Appendix Table 4A.2). National Orientation (NO): This measures a country's commitment to technology-based development. Malaysia starts out high, but its rank falls over time from 5 to 4 to 15 between 1993, 1996 and 1999. Socioeconomic Infrastructure: This measures the strengtho f the educational system, the mobility o f capital and the encouragement of foreign investment. Malaysia slips again from a place o f 15 to 13 to 21 between 1993,1996 and 1999. Productive Capacity: This measures the capability to produce technology-intensive products. Again, Malaysia slips from 14 to 18 to 21 between 1993, 1996 and 1999. Technolonv Infrastructure: This measures the strength o f the scientific and engineering manpower, the relationship o f R&D to industry, and the ability to make use o f technical knowledge. Malaysia's rank here is low to start with and continues to worsen from 21 to 25 to 27 between 1993, 1996, and 1999. It should also be noted that when the scores are scaled to the United States (Le., score o f 100 for the United States) Japan ranks at 73 and Malaysia at 0.2. Technolonical Standing: This measures the production and exports o f hightech products. Here, Malaysia ranks satisfactorily at 13, 13, and14 over the 1993, 1996, and 1999 period. Once again, the message that emerges is that while Malaysia may be doing well in high tech production and exports (probably due to a high M N C presence), the domestic technological capability is very limited. Patents 4.12 Patents are good outcome measures for national innovation. Patenting activity i s usually interpreted as a sign of success in the creation o f industrial inventions. Inventors in all technological fields -agriculture, health, biotech, and electronics- can choose to apply for patents incountries where they wish to protect their inventions. While several strategic factors influence the patenting decision, the advantage o f usingpatents rather than the frequently usedmeasure o fR&D is that while the latter is a measure o f an input into innovation, the former i s an outcome indicator, measuring successful innovative attainment. In fact, a common deficiency o f most aggregate quantifications o f technological capability i s their tendency to club together input and output data. 4.13 Malaysia has a deficit in outputs of innovation. Because patenting activity and income are expected to be correlated (with causation going both ways) it i s useful to look at Malaysia's performance in patenting, given its income level Figure 4.1 shows the percent deviation from the expected performance for patents as well as scientific publications. Malaysia shows a sizable deficit on bothcounts. 120 Figure4.1: InnovationOutputs: Deviationfrom"Expected" Values InnovationOutputs Patents Scientific Publications `ra 80`s 70`s go \\ 8 \\ I\ III II II p.l I 0 7 - L J/ \\-J \ / 1 /. v I I I I 1960 1970 1980 1990 2000 Scientific Publications----- Patents Source3 Technology-Intensive Exports 4.14 I s there a Chinese threat in technology-intensive exports? Finally, an examination o f Malaysia's technology-intensive export perfonnance in the regional context i s worth consideration. There is growing concern about a competitive threat from China, given the region's dependence on sustained manufactured exports to fuel growth. 4.15 The most useful and practical measure o f export competitiveness i s changes in market shares in foreign markets. Table 4.1 shows the different combinations o f market share changes for China andthe competing countries. 4.16 Presumption of a competitive threat. Where the Chinese market share is declining, as in boxes B and D in Table 4.1, it seems reasonable to assume that China does not pose a threat in the period in question. There i s the presumption o f a "competitive threat" only where the Chinese market share i s growing - with some qualification. The most intense threat i s likely to be in box Cywhere China's market share i s rising and that o f the region i s falling. While this may be suggestive o f causation, it may not always be the case. The second, less intense, threat i s inbox A where both sets o f market shares are rising. If China's share i s rising faster, there i s the presumption that it i s holding back growth in neighboring countries. However, fast rising market shares in exports in China could also provide an opportunity for neighboring countries to become suppliers to China by participating in integrated production systems. 121 Table4.1: Matrix of CompetitiveInteractionsbetweenChinaandthe Restof the EastAsian RegioninExportMarkets Chinese Export Market Shares Rising Falling A B No ostensible competitive threat from China, N o competitive threat from China inperiodunder unless the Chinese growthi s faster, and i s holding consideration. The threat i s reverse, from the downthe growthof regional exports. Itis also region to China. ~ i possible that both sides enjoy higher export growth ~ i ~ ~ because of Chinese entry: integrated production systems covering bothparties become more efficient. Legional xport narket C D hares Possible competitive threat from China, unless No ostensible competitive threat. from China. Both regional market shares were decliningin the parties are losingcompetitive advantage inexport absence of Chinese entry. If exports by China are markets. ~ ~ lundertaken~b l i y enterprises from the region, their fallingmarket shares from home countries do not indicate tme competitive impact: location inChina allow their fading comparative advantage to be extended. 4.17 Malaysian technology-intensive exports stand out. Figure 4.2 shows how countries in the regions stack up against China in terms of medium and high technology products. Notably, Malaysia stands out as the single country inthe region that experiences a rise inthe proportion of technology-intensive exports that may be under direct threat. Of course, it i s not clear whether this threat i s really symptomatic of a zero-sum game where Malaysia's loss i s China's gain, or whether there i s room for win-win outcomes via integrated production systems. 4.18 Whether threat or opportunity - there is work to be done. Ineither case, it points to room for action on Malaysia's part. IfMalaysia loses market share to China in some technology-intensive products, it i s left with no choice but to develop a capability in other technologies. Ifit intends to enter global supply chains by becoming a supplier to China's export producers, it will also need to upgrade its technological capability to compete with world class suppliers o fhightechnology inputs. 122 Figure4.2: MediumandHighTechnologyProductsinDirectlyThreatenedExports 100 , Shares of medium and high technology products in directly threatened exports 90 - 80 - 01990 70 - 12000 GO - 50 - 40 - 30 - Hong K o r e a Taiwan Singapore K~~~ Malaysia Thailand Indonesia Philippines Source: Lall, "Integrated Production Networks in East Asia"Wor1d Bank (2003) B. HOWDOFIRMSMOVE UPTHE TECHNOLOGY SPECTRUM? 4.19 Technological capabilities evolve over time. Despite frequent references to some nations (e.g., South Korea, Taiwan [China]) as "new entrants" on the innovation scene, this notion o f suddenly "arriving on the scene" i s mis-founded. These nations, or, more precisely, firms in these nations, are no strangers to innovation and have been engaged in making product and process improvements for several years. Innovation cannot occur in a vacuum, and as in almost all aspects o f learning, there are stages that pave the way for an evolutionary transformation to take place.2 The central aim o f this section i s to understand the drivers o f this technological transition (TT), to position Malaysia along this TT, and to identify the policy levers that exist to facilitate the transition from one stage to another. In the course of this exposition the links betweentechnology and skills are also highlighted. A Frameworkfor the Technological Transition 4.20 Competitive pressures force firms to adopt new technologies. All firms are in the stage o f adopting one technology or another. When firms are predominantly dependent upon external sources for the supply o f new innovations they are classified as technology adopter^.^ H o w do firms aquire the incentives to adopt newer technologies? The answer lies in the pressures put on them by competitors, domestic or foreign (directly in the form o f competing with them for market share) and by lenders (indirectly by lending on the basis of market performance). These pressures force firms to constantly re-evaluate and improve their production (and other) techniques in search o f efficiency gains. Malaysia has enjoyed a fairly open regime for trade and investment inthe manufacturing sector, ensuringthat the pressure o f a competitive threat i s present for domestic firms. Recent progress in improving corporate governance indicators i s likely to have added to the competitive pressures on firms. 'Naturally, these stages are not water-tight and considerable overlaps exist which are pointedout inthe course o f the discussion below. While it is not uncommon, and perhaps i s even desirable, for firms at this stage to adopt less than world-class technologies, the act o f adopting any `new' technology is commendable and a sign of a movement in the right direction on the TT spectrum. 123 4.21 The global stock of knowledge is a key source of adopted technologies. Once motivated, adopter firms rely on foreign firms, essentially via spillovers, for supplying these innovations. Access to global markets through trade and FDIbecomes a vehicle for the diffusion of technology and i s key inexposing firms to newer products andprocesses. This exposure leads to imitation, and the process o f a technological mimicry - the first step towards any catch- begins, Formal methods o f any technology transfer involving licensing and royalty payments may be less common at the very early stages o f the TT. Very little formal R&D on the part o f firms should be expected, although universities and public research institutions may be engaged inconductingresearchinselected areas. 4.22 Wide technology take-up (diffusion) is not automatic or costless. This should not convey the impression that firms need do nothing but wait passively to adopt technologies - and here i s where the overlaps in these stages o f TT become highlighted. A firm's ability to adopt any technology hinges on its own absorptive capacity, which i s determined, very often, by its own level o f technological effort to stay abreast o f the relevant new advances. The discussion below on the empirical correlates o f technology-adopting firms from the PICS survey data sheds light on this. Inaddition, information asymmetries may retard difhsion and exacerbate the risk associatedwith any new technology adoption. 4.23 Good technical education and training policies have helped Malaysia tap into the global stock of knowledge. Having access to a labor force that possesses general education and basic technical skills is, o f course, a critical determinant o f a firm's absorptive capacity, While some specialized skills may be called for, selective but well designed on-the-job training for workers with secondary education can adequately fill the gap and remove some o f the obstacles faced in the adoption o f technology. Evidence exists from firms in Taiwan, Korea, Colombia, and Mexico as well as Malaysia which supports the view that firm-led training and adoption o f technology go handinhand. Infact, firms that do bothreap the greatest productivity gains over time, and Malaysia has done quite well in terms o f expanding the incidence o f training o f workers as well as improving the functioning o f public training funds (Tan and Batra, 1995), although there i s room for improent. 4.24 Next the process of tinkering begins. In the course o f adopting technology cycle after cycle, workers may begin to recognize opportunities for modifying and making minor improvements. These improvements could be led by a need to take advantage o f local resources, or a need to modify the technology to suit local conditions (environmental, infiastructural, or economic). A natural boost to making the transition to this next stage o f adaptation is usually received when firms try to move from domestic to export markets, or from supplying local markets to becoming suppliers to MNCs. New entrants into export markets typically come in producingrelatively standard products, but at reduced cost - which is derived from low wages as well as a lowered cost o f technology through some tinkering and modification^.^ 4.25 Good secondary education is needed to support this stage. The skills needs become more specialized at this stage; however, it i s unlikely that a large pool o f scientists and It is extremely hard to quantify the role o f export markets or h4NCs in putting pressure on Malaysian firms to modify technologies without conducting detailed case studies. However, the received wisdom has been that competitive participation in export markets creates pressure to improve processes and products. Empirically, using panel data, evidence of learning-by-exporting has been obtained (Hallward-Driemer et al, 2003). 124 engineers can be fully utilized by firms on a fbll-time basis (unlike the Brazil or India model o f over-investment in tertiary education as opposed to secondary education before the adaptation phase has matured). Most tertiary educated workers at this stage o f the economy find suitable employment inresearch institutions or universities, and firms need to call upon them from time to time to assist in making some o f these adaptations to existing technology. Links between firms and some institutions begin to form, although the nature o f the relationship is at arm's- length, where a team o f university professors or research scientists is called in on an as-needed basis. WhereIs Malaysia on the TT? 4.26 Inorder to attempt apositioning ofMalaysian firms onthe TT, we need to take a closer look at the technological activities within firms. Usually, R&D data or patent data are used as measure of technological effort or performance. While R&D statistics are a measure o f formal technological effort within firms, they miss out on many other technology upgrading activities that may take place outside o f their formal R&D departments. In fact, other than for the largest firms, this i s unlikelyto be the main type o f technology activity, as seen inFigure 4.3. Figure4.3: FirmsReportingR&DActivity 90 80 70 60 50 40 30 20 10 0 4.27 Therefore, we use the PICS questionnaire which asks firms to report on several other technology activities to classify them into adoption, adaptation, or creation activities. There i s clearly no formula for this and there are considerable overlaps ineach o f these activities, which makes it difficult to be too precise about these classifications. Yet much could be gained by understanding the determinants o f different types o f activities, so that a clear connection could be made with the framework for TT and a prioritization o f the reforms needed could be facilitated for policymakers. Therefore we construct the following criteria: We classify a firm as an adopter ifit answers "yes" to undertakmg: 1. Updates o fmachinery or 2. Introduces a substantially new technology 125 Butanswers "no" to the questionsbelow. We classify a firm as an adapter if it answers "yes" to undertaking 1. Upgrades ofthe majorproductline (product improvements) or 2. Entersnew markets due to quality or cost improvements (process improvements) Butanswers "no" to filing any IPRs. We classify a firm as a creator ifit answers "yes" to 1. Obtainedpatentsor copyrights(either process or producttechnology). 4.28 Malaysian firms are technologically active. Figure4.4 shows the prevalence o f technology adoption and adaptation within firms. It i s hearteningto note the prevalence o f some technological activity among firms. Nearly 40 percent o f firms report making improvements to products or proce~ses.~ 4.29 Nevertheless, there is room for expanding technological activity in firms. As Figure 4.5 shows, there i s considerable scope for improvement - for example, about 70 percent of firms have not introduced a substantially new technology and over 50 percent have not been able to enter new markets on the basis o f quality or cost improvements. 4.30 The use of new or computerized technology varies widely among firms. Two measures o f technology that lend themselves easily to measurement are the fraction o f firm machinery that is computer-controlled and the fraction o f firm machinery that is less than five years old. Since Chapter 1has found these measures to be linked to the performance o f the firm, some preliminary statistics o f the disparity within the sample are briefly presented. For example, 75 percent o f firms have 0 percent computer-controlled machinery, 16.6 percent o f firms have between 0-40 percent o f their machinery computerized, and 8.2 percent o f firms have between 40-80 percent o f their machinery computerized. Similarly, around 70 percent o f firms report not having any machinery that i s less than five years old, 25 percent o f firms report having 0-40 percent o f their machinery aged less than five years old, and 5 percent o f firms report owning between 40-80 percent of machinery that is no more than five years old. Of course, not all sectors and not all firms within sectors need to have the same level o f new or computerized technologies, and there i s no prescribed target for reducing the level o f dispersion across firms. 4.3 1 Intersectoral variations in technological activity are present. Figure 4.6 shows the sectoral variation in adopter versus adapter firms and shows the chemical and electronics sectors performing better on adaptation, while textiles and garments not doing as much.6 Itshouldbenotedthat there are various degrees ofadaptationthat fmcouldundertake; however, we are not able to disentangle this from this data. An added complication to consider when looking at sectoral variation here is that the nature and pace of change of electrical technology may demand local adaptation whereas textile technology may be relatively standardized and may not needmuch modificationto implement. 126 4.32 Figure 4.7 shows the percentage of inventors by sector. (See Chapter 1 for the correspondence between this variation andvariation inTFP growth across sectors.) 4.33 There is room for improving technology diffusion to non-exporters and SMEs. Figure 4.8 shows that exporters and larger firms (non-SMEs) tend to be more technologically active than non-exporters and SMEse7 This is an entirely expected result but is noteworthy because public policies to bridge the gap between large and small firms, as well as the broader rationale for promoting exports in general, are based on the expected diffusion o f technologies through explicit linkages between firms or demonstration effects. Studying the extent o f linkages between large MNCs and small domestic firms is a subject that i s beingtaken seriously in Malaysia at present and it should continue to be the focus o f investment and technology policies. Figure 4.4: The Technology Spectrum inMalaysia 40 , I 35 30 25 20 15 10 5 0 Adopters Adaptors Inventors An SMEis definedas a fmwith less than 150employees. 127 Figure4.5: What Are Firms(not) Doing? Figure4.6: TechnologicalActivity amongFirms 60 I 0 Stop atAdoption IDo SomeAdaptationAlso Figure4.7: Which Sectors Have MoreInventors? 25 I 128 Figure 4.8: Size, Exports, and Technology Performance N o n - E x p o r t e r E x p o r t e r N o n - S M E s S M E s 10Adoption 0Adaptaaon ICreation 1 4. Can some Malaysian firms move to the next stage? It may be helpful .J note that whereas the move from adoption to adaptation can almost be labeled a natural transition which evolves with time given the right incentive regime and the right supply o f basic education and skills (in which Malaysia fares well) the move to creation requires an intensive and purposeful effort on the part o f firms, very often with a parallel effort on the part o f governments. This i s what Malaysia i s aiming for, as stated in the EighthMalaysia Plan. Box 4.2 reports the findings o f the qualitative technology audit conducted on a small sample o f high- performing Malaysian firms. 4.35 Infact, by some accounts, given its income level, Malaysia should have already made strides in boosting private technological effort. Figure 4.9 shows the negative deviation from the expected value o f R&D (i.e. Malaysia has an R&D "deficit"), whereas in licensing it has a "surplus", (i.e. Malaysia licenses more than the predicted value o f licensing, implying again that it is well plugged into the global stock o f knowledge, but does less R&D than its predicted level o f R~LD).~ Therefore, Malaysia's move to the next stage o f technology readiness may be overdue rather than premature. 4.36 The current threat of losing regional competitiveness in export markets could provide the right backdrop for this transition. The incentive to make this intensive effort on the part o f firms and governments may once again be provided by export markets. Firms that have been present in global markets adapting existing technologies and selling the resulting products at lower cost than their competitors usually find margins eroding as new low wage entrants take away market share. Sustaining a position in global markets would necessitate a leap forward into creating new products and processes, which i s greatly facilitated by the learning as well as the capital accumulation from the adapt-and-export stage. The zero line inthe figure corresponds to the predicated level. The predictions are based on appropriate regressions where the regressors are In(GDP), (In(GDP))2,ln(1abor force), @(labor forces))2, and year dummies. Itis possible that most o fthe `surplus' inlicensingcomes from MNCs transferring technology fromheadquarters to subsidiaries. The data do not allow us to exclude such i n t r a - f m transfers. 129 Figure4.9: R&DandLicensinginMalaysia Innovation Inputs D e c a d e R & D I G D P L l e e n a l n n I G D P //` 9 0 ' 1 0 . 3 3 % 0 . 4 8 % // / 8 0 ' s 0 . 9 0 % 0 . 1 2 % 9 - , 1990 1995 2000 R&D/GDP ----- Licensing/GDP 4.37 There i s no doubt that competition for Malaysia in"standard" goods has intensified over the past years from countries inSouth Asia, and Latin America, as well as in neighboring East Asian economies. The potential threat from the Philippines inelectronics- relatedFDIand exports, and the overall threat from China, discussed inthe previous section, provide the needed backdrop of a credible threat that necessitates the rapid upgrading o f technological capability inMalaysian firms. 4.38 Technological activity in firms will need effective public support. Technological activity, will indeed demandan increase on the part o f firms ininvestment inplant and equipment for state-of-the-art laboratories and testing facilities as well as in-house R&D activity, Given the externalities inknowledge creation arising from the imperfect appropriability o f the returns to such investments, this R&D push i s likely to be compromised unless mechanisms that correct for this market failure exist. Some attention to the design and delivery o f these mechanisms is due in Malaysia - as will become apparent in the next section, which discusses the use o f some of these mechanisms by Malaysian firms. 4.39 Both the quantity and quality of skilled labor will require specialattention. Malaysiahas done well untilnow infocusing on secondary and technical education as well as in- firm training to support the adoption and adaptation activities o f Malaysian firms. However, while technological knowledge at the early stages o f the technology transition is essentially coded or embodied inproducts - hence making its transfer or adoption easier - knowledge at the creation stage will tend to become more tacit and disembodied. When new knowledge is not completely embodied in a product, spillovers from the sale and use o f a product leading to imitation through reverse engineering also become restricted. Highly skilled human capital then becomes central and (more than the mere existence o f skilled labor) its mobility becomes the medium of technology transfer and the fuel for moving a firm up the technological ladder. Malaysia will need to focus on this. The next section, as well as the previous chapter on skills, stresses the importance o f improving the mobility of labor inthe Malaysiancontext. 130 131 c. WHAT ARE THE DETERMINANTSTECHNOLOGICALOF ACTIVITIES AMONG MALAYSIAN FIRMS? 4.40 Using the PICS dataset for 902 firms, this section tries to empirically verify the , importance o f certain firm-level attributes that facilitate technology adoption, adaptation, or creation by firms (see Figure 4.10 for technical summary). 4.41 Determinants of technologicaleffort - identifyingpublic and private roles. Chapter 1 underscores the importance o f technological effort to firm productivity performance. This section now turns to understanding the determinants o f technological activity. For consistency with the earlier section that positioned firms into three broad categories, this section will try to identify key firm-level attributes that increase the probability o f undertaking technology adoption, adaptation and creation activity. We use a probit model to estimate these probabilities. Figure4.10: Determinantsof TechnologicalActivities Adopt Adapt Create Training relationships Contractua relationships Cooperativ:relationships EDUCATION SECTOR Type: EMBODIED b DISEMBODIED Mode: VERTICAL TRANSFERq-b HORIZONTAL TRANSFERS/ CO- LOCATION Actor: PUBLIC R&D b PRIVATE R&D 4.42 In-house skilled resources matter for technology adoption. The regressionresults are quite striking as seen inTable 4.2. Intryingto explain the prevalence o f technological adoption activity among firms, the amount o f dedicated technical and innovation staff stands out as being important. Size as well as an exporter status for a firm also increases the probability significantly, which i s as expected. Interestingly, ownership or collaborative efforts do not drive this decision. The auto parts sector also appears more inclined to undertake more technology adoptionthan other sectors. 132 Table 4.2: In-House SkilledResourcesMatter for TechnologyAdoption Dependent Variable: Technology Adoption Indtcator dF/& Std. Err. z P>lzl exp* 0.118 0.046 2.49 0.013 foreign 0.000 0.001 -0.84 0.403 size 0.000 0.000 3.48 0.001 innstaff 0.017 0.006 2.77 0.006 ffcollab* 0.059 0.044 1.34 0.181 ficollab" 0.026 0.057 0.46 0.647 fucollab" 0.060 0.078 0.77 0.439 textdes* -0.160 0.108 -1.29 0.198 garments * 0.061 0.099 0.62 0.535 chemical* -0.006 0.122 -0.05 0.964 rubber* 0.057 0.081 0.71 0.477 machine* 0.045 0.098 0.46 0.646 electron* -0.124 0.088 -1.3 0.195 autopart* 0.384 0.107 3.12 0.002 food* 0.029 0.084 0.34 0.732 Note: Probit Regression: Probability o f technology adoption = f (exporter, ownership, size, dedicated innovation staff in-house, firm-firm collaboration, firm-university collaboration, firm-institution collaboration, sectoral dummies). Furniture dropped due to collinearity. Number o f obs= 604; Loglikelihood = -361.98262. (*) dF/dx i s for discrete change of dummy variable from 0 to land z and P> Iz I are the test of the underlying coefficient being 0. 4.43 Inorder to increase technology adoption, then, firms needto allocate humanresources to the effort. This means that the availability o f and access to a pool o f such technically qualified staff i s key. The Eighth Plan recognizes this and aims for a 60:40 ratio in science to arts .enrollment at the tertiary level. 4.44 Technology transfer from MNCs cannot be passive. What i s worthy o f mention here are the statistics on technology transfer from an MNC, since this i s commonly understood to be a source o f technology adoption. Only 17 percent o f suppliers to MNCs report a technology transfer from them, with 60 percent o f recipients receiving it through explicit means such as licensing or certification. Notably, 51 percent o f these firms report a need to adapt this transferred technology to local conditions and 65 percent report a need to train the workforce to implement this new technology. This underscores the point made earlier about non-passive technology transfers and the importance o f in-house capacity to absorb and implement this technology, even when it is transferred from a parent company. To facilitate the acquisition o f technology instrategic industries, the allocation for the Technology Acquisition Fund(TAF) will be increased to a 2 5 0 million in the Eighth Plan. The Fund i s available for the purchase o f high-tech equipment and machinery, technology licensing, and the acquisition o f patent rights, prototypes and designs to enhance the transfer o ftechnology to local companies. 4.45 What drives technology adaptation within firms? Inter-firm collaboration and the status o f being an exporter increase the probability o f a firm's undertaking adaptation, as seen in 133 Table 4.3.' Size, foreign ownership, or an in-house staff that i s dedicated to innovation does not drive this technological activity significantly, after other variables have been controlled for." 4.46 But, it is not clear which policies can promote inter-firm collaboration. Given the existence o f inter-firm collaboration, what distinguishes firms that do collaborate from firms that do not? Finding explanatory variables for firm-firm collaboration has been difficult. Interestingly, 63 percent o f collaborators are SMEs, while 60 percent o f collaborating domestic firms are members o f a business association. (sectorally, machinery stands out). No readily available policy lever seems to stand out (empirically tried in the regression framework) as a means o f promoting firm-firm collaboration. It is easy to live with this finding, since the widely researched literature on inter-firm collaboration has thus far not been able to point to any clear firm-level attributes that facilitate collaboration and has attributed most o f such collaboration to idiosyncratic causes such as historical accident, or perhaps the presence o f high social capital in certain geographical or ethnic groups. These idiosyncratic factors dictate whether the level o f interaction between firms i s merely transactional or whether it i s truly collaborative to include joint problem-solving, although policymakers may have little leverage over these factors. Firms are not skipping stages. Finally, what is the explanation for inventive activity in firms? It i s usehl to be reminded o f the importance o f these stages o f technological transition or learning-to- learn phenomena that firms must go through. Table 4.4 shows that without adaptation activity in firms, invention is unlikely to occur. Obviously, only a subset o f adapters makes it to the stage o f becoming inventors. Hence the importance of technology adoption and adaptation activities should not be overlooked inpolicy discussions. 4.47 What is the explanation for technology creation?For invention activity to take place, collaboration with technology institutions matters. Empirically, the key drivers o f inventive activity in firms seem to be firm-institution collaboration as opposedto inter-firm collaboration before, Foreign ownership i s not a significant driver, and in-house skilled resources are no longer sufficient for the undertaking o f enough anymore to undertake invention, as seen inTable 4.5. It i s interesting to note that being an exporter does not increase the chances o f becoming an inventor. This may reinforce again the role o f external markets and associated competitive pressures inthe earlier stages o f the TT by forcing firms to upgrade and tinker with technology. These pressures are not enough to reach the higher end o f the TT. Rather, with the backdrop of open trade and investment regimes, which Malaysia has already followed quite well, a heightenedfocus on domesticpoliciesto boost inventiveoutputis warranted. 4.48 Since it i s this transition to the high end o f the TT that the EighthMalaysia Plan has articulated as a goal o f its technology policy, we investigate the factors that influence firm- institution collaboration, while also exploring the possibility o f increasing the use o f fiscal incentives by more firms. 'Firms were asked if they sought the help of other f i i , or of universities or research institutions in order to develop or adapt technological innovations. If they answered "yes" to any o f these, it constituted technical collaboration with one o fthese entities. lo Interestingly, financial incentives were a significant variable in the regression until size and exporting were introduced into the equation. This may imply that larger f i are availing themselves o f financial incentives more than SMEs. 134 Table 4.3: What DrivesTechnology Adaptationwithin Firms? Dependent Variable: Technology Adaption Indicator dF/dx Std. Err. z P>lzl exp* 0.194 0.049 3.93 0 foreign 0.000 0.001 -0.36 0.717 size 0.000 0.000 1.82 0.069 innstaff 0.014 0.008 1.77 0.077 ffcollab* 0.183 0.043 4.11 0 ficollab* 0.037 0.058 0.63 0.529 fucollab* 0.065 0.078 0.81 0.416 mfincent* 0.148 0.099 1.38 0.166 textiles* -0.321 0.109 -2.54 0.011 garments * -0.215 0.096 -2.15 0.032 chemical* 0.009 0.129 0.07 0.946 rubber* -0.168 0.083 -2.01 0.044 machme* -0.229 0.096 -2.28 0.022 electron* -0.217 0.100 -2.08 0.037 autopart* 0.085 0.124 0.66 0.509 food* -0.151 0.087 -1.72 0.085 N o t e : Probit Regression: Probability o f technology adaptation = f (exporter, ownership, size, dedicated staff in- house, firm-firm collaboration, firm-university collaboration, firm-institution collaboration, fiscal incentive usage, sectoral dummies). Furnituredropped due to collinearity. Number o f obs= 604; LR chi2(16)=76.65 (*) dF/dx is for discrete change o fdummy variable from 0 to 1andz andP> I z I are the test o f the underlying coefficient being 0. Invent Adapt N o Yes No 96.4 3.6 Yes 78.32 21.68 135 Table 4.5: What is the Explanation for Technology Creation? Dependent Variable: Technology Creation Indicator exp* 0.022 0.032 0.67 0.505 foreign 0.000 0.000 -0.16 0.872 size 0.000 0.000 3.5 0 innstaff 0.001 0.002 0.49 0.621 ffcollab* 0.058 0.031 1.92 0.055 ficollab* 0.079 0.044 1.99 0.046 fucollab" -0.046 0.039 -1.04 0.301 mfmcent* 0.081 0.074 1.25 0.212 texdes* -0.062 0.054 -0.9 0.366 garments* -0.040 0.050 -0.72 0.473 chemical* 0.083 0.097 0.99 0.322 rubber* -0.087 0.041 -1.89 0.059 machine* -0.118 0.029 -2.38 0.017 electron* -0.080 0.040 -1.52 0.128 autopart* -0.013 0.069 -0.19 0.85 food* -0.023 0.048 -0.45 0.653 Note:Probit regression: Probability o f technology creation = f (exporter, ownership,size, dedicated staffin- house, firm-firm collaboration, firm-university collaboration, firm-Institution collaboration, fiscal incentive usage, sectoral dummies). Furniture dropped due to collinearity. Number o f obs=604; Loglikelihood = -226.97296. (*) dF/& i s for discrete change o f dummy variable from 0 to 1and z and P> Iz I are the test o f the underlying coefficient being 0. D. FIRM-INSTITUTIONCOLLABORATION 4.49 Users and generators of knowledge need to interact. Recognition o f the fact that scientific progress does not only flow from pure science to applied science, but rather that pure science is enhanced by application (feedback from front line users of technology to researchers i s essential for the refinement o f products and production processes), has led to an interest inthe way research institutions and firms interact with each other. Frequent interactions between the generators and the users o f knowledge, or between basic and applied knowledge, can enhance the productivity o f the institutions and firms equally. 4.50 .Can firm-institution collaboration be facilitated? Some 15 percent o f the surveyed firms report usage o f research or technology institutions (RTIs) - with most inventor- status users being in the chemical and machinery sectors and most non-users in the food, auto parts andfurniture sectors (see Figure 4.11). It should be noted that a firm's use o f institutions is related to the technological characteristics o f the sector, and it is well known that chemical technology lends itself to contractual relationships intechnology development. It i s also possible that there may not be high quality technology institutions present in every sector. Hence, there seems to be considerable scope for scaling up the elements o f this firm-institution interaction. Also, it should be noted that sectors that house the largest share o f inventors are not necessarily 136 the sectors that make the most use o f RTIs. However, the regressions stress that firms that use RTIs aremore likely to engage intechnology creationthanfirms that do not. 4.5 1 Firms inMalaysia havereasonsfor notworkingwith RTIs. Since inventionis correlated with higher RTI usage in Malaysia, promoting such collaborations is key to making the transition the invention stage for private industry. Inthe PICS survey, firms were asked to provide their reasons for not working with an RTI. They reported the lack or relevance o f the services offered by the RTI to the firm's needs as the primary reason. Lack the knowledge o f how to make the first point o f contact with an RTI and lack o f the technical capability in-house to interact with an R T Iwere the secondary reasons for not engaging inany form o f collaborative relationship with these institutions. 4.52 But among users of RTIs, satisfaction was very high. Table 4.6 shows the firms' ranking of these institutions along several attributes and the predominance o f a fairly good to very good rating on all counts. Less than 3 percent o f firms reported using RTI and "having a bad experience" so that they would not go back again to the institution. Figure 4.11: Usage of RTIsby Inventors I IY o o f Inventor F i r m s in F-I C o l l a b s x l Assistance with Modifying R&D in Outreach Ease of Transparency License Application of Process search, purchase exisiting breakthrough of techdogies negotiation technologies technologies Very Poor 2.3 5.6 3.4 4.4 6.3 4.2 5.6 Poor 10.1 10.0 9.1 12.3 18.7 15.5 19.7 Fairly Good 65.2 63.3 71.5 70.7 64.0 66.2 60.6 Very Good 22.5 21.1 15.9 12.3 10.9 14.1 14.1 4.53 The Eighth Plan includes steps to improve firm-institution collaborationby stating that "Business units at RTIs will be reorganized and strengthened to facilitate identification and implementation of market oriented R&D projects through interaction among researchers and the private sector. To generate more R&D projects that can be commercialized, RTIs will be encouraged to place more emphasis o f research related to product and process development for the industries." 137 E. FISCALINCENTIVES 4.54 The rationale for fiscal incentives. Credit constraints reduce the demand by firms for R&D. Credit constraints for R&D activity are exacerbated by the fact that R&D i s risky and traditional bank lendingi s unable to value intangible assets. Further, the public good nature o f knowledge makes the social benefits greater for firms than the private benefits, leading to under-investment and thus adding to the case for government intervention to bridge the gap." 4.55 The most common type o f intervention is through fiscal incentives. Governments have a variety o f fiscal incentives that they can provide to stimulate firms' R&D investments, as inthe case ofMalaysia (double deduction, investmenttax allowance, exemption of duty, capital allowance, industrialbuildingallowance). l2 4.56 Low usage of fiscal incentives by technologically active firms in Malaysia. The use o f fiscal incentives turns out to be a statistically significant variable in explaining invention (as well as adaptation) activity in firms only until the size and exporting variable are introduced explicitly. This i s suggestive o f a strong correlation between size and use of incentives. It i s noteworthy that between 40 and 50 percent o f firms with R&D inMalaysia have never heard of the listed fiscal incentives. Even among inventors, where use o f these fiscal incentives i s high, some sectors have no firms that availed themselves o f much incentive(see Figure 4.12). Such a skewed take-up o f fiscal incentives is hard to explain from the data, and a review o f the qualifying criteria (are they sectorally biased?) for these incentives may need to be considered. 4.57 In addition, there is little correspondence between sectors that house the largest share o f inventors and sectors using fiscal incentives. This points to considerable inter-firm variation within sectors in the use o f fiscal incentives and raises the question o f whether any sectoral preference inawarding incentives i s paying off. 4.58 Complicated process and lack of awareness as deterrents for firms. Notably, 25 percent o f all firms that had heardo f these fiscal incentives chose not to use them because o f the complicated process of application and approvals. Also, 60 percent of SMEs have not heard o f the product, quality, andmarket development schemes designed specifically for SMEs. l1Rates o f return inclusive o f such inter-industry spillover effects, i.e., social rates o f return (SRR), have been estimated by a few investigators. In agricultural R&D projects, the SRR was 150-300 percent greater than the private rate o f return (PRR). Studies for industrial countries have also estimated that a 1 percent increase in spillovers, caused average costs o f other f m to fall by 0.2 percent. Across the major industries that were studied, non-electrical machinery, chemicals, rubber and plastics, and petroleum were found to b e major spillover sources. Their SRRs ranged from four times the PRR in non-elect machinery, to three times the PRR in chemical products, and to twice the PRR inthe other two industries (Bemstein 1989). l2Some grant schemes managed by the Ministry o f Science, Technology and Environment (MOSTE) were not included inthe PICS questionnaire. Follow-up work o n fiscal incentives should take note o f all existing programs. 138 Figure 4.12: Inventors Using Fiscal Incentives 35 30 25 20 15 10 5 0 I IPercentage of Inventors U s i n g Double Deduction Incentive 1 4.59 Among users, a high ranking of fiscal incentives. Among users, satisfaction i s high. Onaverage, 80 percent ofusers found the scheme to beofmoderate-to-critical importance to the firm's technological activity. Among SMEs, users report even higher satisfaction, with between 90-100 percent reporting the use of the scheme to be o f moderate-critical importance to the firm. 4.60 The Eighth Plan states that it "will review fiscal incentives in order to increase private sector participation in R&D activities. The various financial grant schemes for the private sector will also be expanded." The review will include process-related details or outreach and awareness for SMEs, to increase the uptake o f these financial incentives. 4.61 Sources of innovation and the Intellectual Property Rights (IPR) regime. Two previously mentioned issues also deserve some empirical attention: the importance o f disembodied sources o f technical knowledge and the importance of the PR regime. Since these issues do not vary by firms, incorporating them into the above regression fi-amework i s not possible. Some simple statistics based on firm's answers to related questions are used to shed light onthese two issues. F. SOURCESOFTECHNICALKNOWLEDGE 4.62 Learning-by-hiring is ranked as an important source of innovation. Firms in Malaysia were asked to list their first, second, and third most important sources o f technological learning (see Table 4.7). While embodied sources of knowledge via purchases ofmachinery and equipment feature as the main mode o f technological acquisition, learningthrough the hiring o f keypersonnel shows up as a significant channel as well. 4.63 Closing the technology gap with advanced countries on the basis o f innovation will require substantial augmentation o fthe local pool of researchers, entrepreneurs and business service providers. Attracting workers who have trained in the leading research centers abroad 139 and have acquired experience as well as personal contacts inkey overseas clusters i s seen as an expeditious route for technological catch-up. This is largely because technological advances remain localized and production processes are rising in complexity, increasing the share o f "tacit" knowledge that i s not easily communicable through blueprints or specifications. The movement of skilled labor or dense networks that promote contacts between skilled labor across regional or national boundaries therefore becomes central to the discussion on building innovative economies. The most The second The third most important most important important Embodiedinmachinery 48.0 30.7 16.5 Hiringkeypersonnel 10.6 18.2 17.4 Licensing 3.2 4.1 4.3 Transferfromparentcompany 10.6 4.7 3.9 Cooperationwith clients 8.3 13.3 11.4 Cooperationwith suppliers 5.O 10.9 10.8 4.64 Migration is not just brain-drain. Traditionally, the migration o f skilled labor from developing to developed nations has been interpreted as "brain-drain," implying a dead- weight loss to the developing country after years o f public spending on the education o f these individuals. Countries such as Korea and Taiwan, however, have been very successful in attracting back their US-trained and experienced scientists, engineers and entrepreneurs, and this has no doubt been a boost for innovation in these countries. Of course, the ability to attract migrants back to their homelandi s inlarge part a function o f the advancement inliving standards that these countries made and were able to offer to attract skilled emigrants. The last decades o f the twentieth century provided some dramatic examples of the importance o f return migration as a stimulus to innovation (Fontana and Cannel, 2001).13 4.65 Nations that are unable to attract their skilled migrant labor home may still be able to benefit from the role o f diasporic networks. Diasporic networks by emigrants also stimulate capital flows, because emigrants are better placed to evaluate investment opportunities and to retain contacts in their home countries to facilitate this process, as well as to encourage foreigners to invest in their countries o f origin by helping FDI sources find trustworthy and competent partners. As Saxenian (2002) observes, "The scarce resource in this business environment i s the ability to locate foreign partners quickly and manage complex business relationships across cultural and linguistic boundaries. This i s particularly a challenge in high tech industries in which products, markets, and technologies are continually being redefined - andwhere product cycles are routinely shorter than nine months." l3For example, Ireland's software industries benefited from returning expatriates who went abroad duringthe 1980s and 1990s to meet a chronic U S shortfall in IT workers. Ireland constituted, along with the United Kingdom, the largest EU source o f emigration to the United States, with California as a principal destination. In the 1990s, however, net immigration has been positive and over half are returning emigrants: 25 percent o f male emigrants with completed tertiary education returned during the 1990s. Over 40 percent o f Irishgraduates under the age o f 40 have worked abroad for at least one year. 140 4.66 Several success stories abound. Infact, both Korea and Taiwan leveraged their scientists and engineers abroad very well to bridge the knowledge gaps that existed between the United States and Korea. US-based Korean engineers were designated "R&D outposts" who advised the Korean government on technical and business trends and helped design policies to take advantage of them. The Chinese Institute o f Engineers, founded in 1979 by a small group o f Taiwanese engineers working in Silicon Valley, organizes an annual seminar in collaboration with its counterpart organization in Taiwan (China) and provides consultative services to the Government o f Taiwan (China). This also resonates with the Indian IT success story, where immigrants have not returned to India but have facilitated outsourcing o f service development to their homeland. Most U S subsidiaries in India are headed and staffed by employees o f Indian origin. They have not only been a source o f knowledge for their home countries, but they have also been a source of contacts within their countries for US firms and for links to foreign capital markets, filling ingaps indomestic credit markets (Arora, Gambardella, Tonissi 2001).14 4.67 Malaysia may be able to reap these gains also. Malaysia may serve its technological deepening well by paying close attention to these episodic occurrences and deciding whether there is a systemic lesson to be learned here. Chapter 2 on skills highlights some rigidities faced by firms inhiringworkers, especially foreign workers, which may blunt the benefits o f the learning-by-hiring channel. A more aggressive and focused approach to tap into Malaysian scientists and engineers abroad i s likely to pay off, 4.68 Again, the EighthPlan has included this issue by stating that the "program for attracting Malaysian scientists and engineers residing overseas will be enhanced with a revised incentive package. Inaddition, measures will be taken to appoint Malaysian experts in S&T related areas` who are working overseas to undertake short term assignments inMalaysia." G. IPRs 4.69 Why is the IPR regime important? Inmany ways the arguments for the protection o f physical property apply to intellectual property as well: exclusive but limited rights o f ownership need to be awarded to encourage investments in the creation and improvement o f physical or intellectual property. Since knowledge possesses the additional characteristic o f imperfect appropriability (unlike physical property, knowledge cannot be "owned" exclusively), investments in the creation o f knowledge are believed to be curtailed without the help o f instruments that can correct this inherent market failure. Thus copyrights protect the rights o f authors (books, music, and software), trademark registration protects trade logos and symbols, and patents protect inventions with industrialapplicability (products as well as processes). l5 For technology development, patents are usually the most relevant instrument. l4 An analogous story is evident inthe field of economics. Latinacademics occupy faculty positions inmost o fthe major U S universities and serve as a source o f inside information onpromisingpotential graduate students, contacts for flows of foreign academics to visit LatinAmerica, and they keep the finger on the pulse o fU S academic thinking, After a period o fprofessorship inmajor universities, many academics have returned to anchor or to have joint appointments with departments inMexico, Chile, Argentina and Brazil. 15 It should be noted that copyrights only protect the expression o f the idea or concept from plagiarism, not the underlying idea itself from being applied. Patents protect the underlying idea or invention from being copied. Copyrights are therefore not seen as hampering the application or diffusion o f ideas contained inbooks, software, or 141 4.70 Improving contractual certainty could boost licensing and private-public collaboration. Some econometric studies have shown a positive effect o f IPR on R&D as well as on technology transfers inthe form o f FDIor licensing to domestic firms. IPRprotection can also have a positive impact on technological collaboration between different parties due to the improved contractual certainty. For example, the passage o f the 1980 Patents and Trademark Amendments Act (the Bayh-Dole Act) in the United States allowed academic research institutions to retain title to (and thus patent and license) inventions made with government funding, ensuring that a particular institution had the incentive to seek licensing opportunities. After the passage of the Bayh-Dole Act, technology transfer offices at these institutions became an important vehicle for ensuringthat research results reached the marketplace.16 This may be o f relevance to Malaysia in the context o f promoting private-public collaboration in undertaking technical effort. 4.71 Also, ifP R s are well defined and portable (Le., a worker can take all or part o f the rights to the invention with hidher when he/she switches employers), firms are likely to compete fiercely for skilled labor andworkers are likely to move in favor o f more competitive terms. The mobility o f labor i s a key driver o f innovation at the advanced stages and has been discussed earlier inthis chapter, as well as inthe chapter on skills. 4.72 FirmsforecastincreasedR&D effortwith improvedIPRs. When firms inthe PICS data are dividedinto those doing R&D and those not doing any R&D, a majority o fthe former report that they would increase R&D efforts ifthe IPRregime was strengthened (see Table 4.8). Among non-R&D performing firms, strengthening the IPRregime would probably not induce them to undertake it. Table 4.8: Strengtheningthe IPRRegimeWouldInduce R&DActive Firmsto Do More Would do moreR&DifIPR stronger Do anyR&D currently? Yes No Yes 55.1 44.9 No 26.0 75.0 4.73 However, the welfare implications of improved IPRs are hard to draw. ll__le strengtheningIPRs may increase R&D in firms (dynamic benefits), it would also impose a cost (static costs) in the form of raising the cost o f technology access for producers and would also result in higher prices for consumers. Hence, the political economy repercussions o f stronger IPRswill haveto becarefully managedwithinthe current WTO obligations. music. This is the rationale behind copyrights lasting a very long time (50 or more years after the author's death) andpatents lasting only around 20 years, and i s frequently used as the basis for advocating the adoption o f copyright law earlier inthe development process than patent law. l6The level of sophistication of technology transfer offices varies greatly from one institution to another, but examples from the United States document that such institutions can have tremendous success. 142 H. CONCLUSIONS:WHAT SHOULDTHEPOLICYTHRUST BE? 4.74 Overall, Malaysian technology policy is on the right track in assisting domestic firms in upgrading their technological capabilities (see Box 4.3). The implementation and delivery o f some public support mechanisms meant to boost private technological effort may deserve additional attention. Exploring ways to improve the effectiveness o f some o f the public support measures will complement the technological efforts o f the private sector that are necessary if they are to compete in sophisticated markets - and achieve the right balance between a private and public focus. The Eighth Malaysia Plan and the Economic Package o f May 2003 have already identifiedthese issues and the empirical investigation conducted on the PICS dataset highlightsthe following elements.I7 4.75 Malaysian firms are technologically active interms o f adopting new technologies as well as in making some modifications to technologies. Yet there is scope for expanding such activities, since wide inter-firm and inter-sectoral variation exists in the level o f technological effort. Scope also exists for facilitating technology diffusion from large firms to SMEs and from exporting to non-exporting firms. In order to facilitate wider technology adoption, significant improvements in the availability o f skills, as well as in the environment for attracting skilled labor are likely to be key, giventhe skill-biased nature o f technical change. A preliminarylook at the services data shows that mostfirms are technologically active and some form ofinter- fmcollaboration(likelytobetheoutsourcing oftechnologydevelopment) iskeytotechnologicaleffortinthe services sector. New workers are relied uponto serve as sources o f innovationinthe services sector also (see Chapter 2). Furtherwork is neededinthe services sector. 143 144 4.76 To promote a move to the next level o f the technological transition (Le,, adaptation o f technology), a better understanding o f the quality and determinants o f inter-firm collaboration will needto be attempted. Attention to alleviating the regulatory or institutional constraints faced by some firms in availing themselves of fiscal incentives would also be usehl as will development o f skills for research and development. 4.77 Inorder to facilitate highperformingMalaysian firms inundertakingtechnology creation activities (where Malaysia exhibits a "deficit" controlling for level o f income), some attention to the quality o f technology support institutions and constraints to collaboration between these institutions and private firms is well deserved. A detailed look at the governance o f these institutions is being considered as a follow-up to the empirical findings o f this report. Improvements inthe mobility o f skilled labor as well as improvements in the appropriability o f technology investments are also likely to assist such public-private collaboration. Issuesfor Consideration Issues Suggested Improvements 1. Quality andMobility of Labor 1. Focusonimprovingthe stock, quality, andmobility of labor force (PromoteLearning-by- skills to undertaketechnology effort Hiring) Considerreduction inrestrictions for hiringhighly slulled foreign labor intechnology intensivesectors 2. Incentives 1. Review definition ofR&D(is it too narrow?) (Improve outreachand 2. Improve proceduresfor fast-track clearance of applicationswithout delivery of Fiscal Incentives) diluting integrity 3. Review compensationstyle-currently reimbursementrather than seed capital 4. Review early-stageVC prospects 3. Institutions 1. Improve outreachprogramsto buildawareness (PromoteFirm-Institution 2. Considerprivatization of some RTIs Collaboration) 3. Allow governmentRTIsto keep remuneration from private sector collaboration 4. Governancereview: Boardcomposition of RTIsreviewed 5. Improve contractualcertainty through improvementsinIPR administration 145 Bibliography Acemoglu, Daron. 2003. "Patterns of Skill Premia." Review of Economic Studies. Forthcoming. Aitken, Brianand Ann E.Harrison. 1999. "Do Domestic FirmsBenefitfrom Direct ForeignInvestment? Evidence from Venezuela." American Economic Review. 89: 605-618. Allen, J. and R. V. D. Velden. 2001. "Education Mismatches versus Skill Mismatches: Effects on Wages, Job Satisfaction, and On-the-Job Search." Oxford Economic Papers. 3: 434-452. Arora, Ashish, Alfonso Ganbardella, and Salvatore Torrisi (2001). "Inthe Footsteps of Silicon Valley? Indian and Iris Software inthe InternationalDivision of Labour". Stanford Institutefor Economic Policy Research. Barro, Robert and Jong-Wha Lee. 2000. "International Data on Educational Attainment: Updates and Implications." NBERWorking PaperNo. 7911, Cambridge, MA. Becker, G. S. (1964). Human Capital. NBER, New York. Berman E. et al. 1998. "Implications of Skill-Biased Technological Change: International Evidence." Quarterly Journal of Economics. 113(4): 1245-1280. Bernard, Andrew and J. Bradford Jensen. 1999. "Exceptional Export Performance: Cause, Effect, or Both?," Journal of International Economics. 47: 1-25. Bessant, J., Francis, D.,Meredith,S., Kaplinsky, R and Brown, S. (2001). "Developing manufacturing Agility inSMEs". International Journal of Technology Management, 22 (1,2,3), 28-54 Burpitt, W.J.and Rondinelli, D.A.(2000). Blomstrom, Magnus and Fredrik Sjoholm. 1999. "Technology Transfer and Spillovers: Does Local Participation with MultinationalsMatter?" European Economic Review. 43: 915-923, Booth, Alison L and Gylfi Zoega. 1999. "Do Quits Cause Under-training?" Oxford Economic Papers. 5l(2): 374. Booth, Alison L, ,Yu-FuChen, and Gylfi Zoega. 2002. "Hiring and Firing: A Tale of Two Thresholds." Journal of Labor Economics. 20(2): 217-248. Bosworth, Barry and Susan Collins. 2003. "The Empirics of Growth: An Update." Manuscript, Brookings Institution and GeorgetownUniversity. Burnside, Craig, Martin Eichenbaum, and Sergio Rebelo. 1995. "Capacity Utilization and Returns to Scale." NBER Macroeconomics Annual. Cohen, Daniel and Marcel0 Soto. 2001. "Growth and Human Capital: Good Data, Good Results." OECD Technical Paper 179. D e Ferranti et al. 2003. Closing the Gap in Education and Technology. World Bank, Latin American and Caribbean Region. D e Fontenay, C. and Carmel, E. (2001) "Israeli's Silicon Wadi: The Forces BehindCluster Formation", 146 Working Paper N.40, Stanford Institutefor Economic Policy Research (SIEPR), Stanford University, StanfordCA. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2002. "The Regulation o f Entry." Quarterly Journal of Economics. 117: 1-37. Easterly, William and Ross Levine. 2001. "It's Not Factor Accumulation: Stylized Facts About Growth Models." WorldBank Economic Review. 15(2): 177-219. Fernald, John and Susanto Basu. 1995. "Are Apparent Productive Spillovers a Figmento f Specification Error?" Journal of Monetary Economics. 36: 165-188. Findlay, C. and Warren T. (eds.) 2000.. Impediments to Trade in Services: Measurements and Policy Implications. Routledge, London andNewYork. Freeman, R. and R. Shettkat. 2001. "Slulls Compression, Wage Differentials, and Employment: Germany vs. the US." Oxford Economic Papers. 3: 582-603. Ghosh, Swati R. and Aart Kraay. 2000. "Measuring Growth inTotal Factor Productivity." PREM Note 42, The World Bank. Gill, Indermit Singh and Amit Dar. 1997. Constraints and innovationsinreform o fvocational education and training : country study summaries and issues briefs Washington, D.C., U.S.A. : World Bank ; Geneva, International Labour Office. Gill, I. H.Tan. 1998. "Malaysia: Meeting the Demand for Skilled Workers in a Rapidly Growing and Economy." World Bank. Green, F., D. Ashton, and A. Feltstead. 2001. "Estimating the Determinants of Supply of Computing, Problem Solving, Communication, Social, and Teamworking Skills." Oxford Economic Papers, 3: 406-433. Green, F., S. Machin, and D. Wilkinson. 1998. "The Meaning and the Determinants of Slulls Shortages". Oxford Bulletin of Economics and Statistics. Griliches, Zvi, Ernst Berndt, Timothy Bresnahan, and Marilyn Manser. 1992. Output Measurement in the Service Sectors. Studies inIncome and Wealth 56, NBER. Hall, Robert and Charles Jones. 1999. "Why Do Some Countries Produce So Much More Output Per Worker Than Others?" Quarterly Journal of Economics. 114(1):83-116. Hallward-Driemeier, M; Iarossi, G., and Sokoloff, K (2002), "Exports and ManufacturingProductivity in East Asia: A Comparative Analysis with Firm-LevelData," NBER WorkingPaper No. 8894. Haskel, J. and C. Martin. 1993. "The Causes of Skills Shortages inBritain." Oxford Economic Papers. 451573-588. Haskel, J., S. Pereira, and M. Slaughter. 2001. "Does Inward Foreign Direct Investment Boost the Productivity of Domestic Firms?" Paper presentedat the NBER Summer Institute. 147 Hsieh, Chang-Tai. (2002) "What Explains the IndustrialRevolution inEast Asia? Evidence from Factor Markets." American Economic Review. 92(3):502-526. Hungerford, T. and G. Solon. 1987. "Sheepslun Effects in the Returns to Education." Review of Economics and Statistics. 69: 175-177. Huw, L. and J. Roberts. 2002. "Twin Engines of Growth: Skdls and Technology as Equal Partners in BalancedGrowth." Isgut, Alberto. 2001. "What's DifferentAbout Exporters? Evidence fkom Colombian Manufacturing." Journal of Development Studies. 37: 57-82. Javorcik, Beata S. and Mariana Spatareanu. 2003. "To Share or Not To Share: Does Local Participation Matter for Spillovers fkom ForeignDirect Investment?" World Bank Worlung Paper. Jeffrey I. Bernstein & M.IshaqNadiri, (1989). "Research and Developmentand Intraindusty Spillovers: An Empirical Application ofDynamic Duality," NBER WorkingPapers 2002, NationalBureauof Economic Research, Inc. Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. 2003, "Governance Matters 111: Governance Indicators for 1996-2002." Manuscript, The World Bank. (www worldbank.org/wbi/governance/govdata2002). . Keller, Wolfgang and Stephen R. Yeaple. 2003. "Multinational Enterprises, International Trade, and Productivity Growth: Firm Level Evidence from the United States," NBER Working Paper # 9504. Kraay, Aart. 1999. "Exports and Economic Performance: Evidence from a Panel o f Chinese Enterprises." RevueD 'Economie du Development. 1-2: 183-207. Lederman, Daniel, andWilliam F.Maloney.(2003). "R&D and Development." The WorldBank Policy Research WorkingPaper No. 3024. Levinsohn, James and Ami1Petrin. 2002. "Estimating Production Functions UsingInputsto Control for Unobservables." Review of Economic Studies, Forthcoming. MalaysiaNational Productivity Council (2000) Productivity Report 2000. (http://dominoapp.npc.org.my/publications .nsf). Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. "A Contribution to the Empirics o f Economic Growth." Quarterly Journal of Economics. 107(2): 407-437. Mattoo, Aaditya, Randeep Rathindran, and h i n d Subramanian. 2001. "Measuring Services Trade Liberalization and Its Impact on Economic Growth." World Bank Policy Research Working Paper #2655. Middleton, J., A. Ziderman, and A. Adams. 1993. "Slulls for Productivity: Vocational Education and Training inDeveloping Countries." World Bank. PermWorld Tables Version 6.1. 148 Prime Minister's Office of Malaysia. 2003. "New Strategies Towards Stimulating the Nation's Economic Growth" (May 21, 2000) and "Government Further Liberalizes Equity and Expatriate Employment Policies," Ministryof InternationalTrade and IndustryNews Release (June 17,2000). Psacharopoulous, G. and H.A. Patrinos. 2002. "Returns to Investment inEducation: A Further Update." World Bank Policy ResearchWorking Paper 2881. Razali, N. 1999. "Human Resource Development in Malaysia: A Study of the Government's Role." ASEAN Economic Bulletin. 16(3):307-29. Saxenian, Annalee (2002). "Transnational Communities and the Evolution of Global Production Networks: The Cases of Taiwan, China and India", Industry and Innovation, Special Issue on Global ProductionNetworks, Fall 2002. Schady, N.T. 2003. "Convexity and Sheepskm Effects inthe Human Capital Earnings Function: Recent Evidence for FilipinoMen." Oxford Bulletin of Economic and Statistics. 65(2): 171-196. Schafgans, M. 2000. "Gender Wage Differences in Malaysia: Parametric and Semiparametric Estimation." Journal of Development Economics. 63(2):35 1-78. Smarzynska, Beata. 2002. "Does Foreign Direct Investment Increase the Productivity o f Domestic Firms? In Search of Spillovers through Backward Linkages." World Bank Policy Research Working Paper #2923. Stasz, C. 2001. "Assessing Skills for Work." Oxford Economic Papers. 53(3): 385-405. Tan, H. 2002. "Do Training Levies Work?Malaysia HRDFand its Effect on FirmLevel Productivity". Tan, H.W. and G. Batra. (1995). Enterprise training in developing countries: Incidence,Productivity effects andpolicy implications. Washington, D.C.:The World Bank. Tan, H.and Batra. 1995. "Enterprise Training inDeveloping Countries". World Bank, Washington DC. Triplett, Jack E. and Barry P. Bosworth. 2000. "Productivity in the Services Sector". Brookings Institution. UNDO, Industrial Statistics Database. United States Trade Representative. 2002. 2002 National Trade Estimate Report on Foreign Trade Barriers. Available at . World Bank. 1999. "Thailand Secondary Education for Employment", Report No. 22660-TH. World Bank. 2000. "Labor Markets inthe East Asian Financial Crisis." World Bank. World Development Indicators. 2003. World Trade Organization. 2002. TradePolicy Review, Malaysia 2001. 149 Young, Alwyn. 1995. "The Tyranny of Numbers: Confrontingthe Statistical Realities of the East AsianGrowth Experience." Quarterly Journal of Economics. 110: 641-680. Zeufack, A. G. 1999. "Employer ProvidedTrainingunder Oligopolistic LaborMarkets." DECRG, The WorldBank. 150 APPENDIX: SampleCharacteristics Malaysia PICS: Sampling Design The technical discussions with the Department o f Statistics (DOS) led to the following design. For the coverage, the survey will cover establishments inMalaysia engaged inthe following 10 manufacturing and 5 services activities: 1) Manufacturing 1. FoodProcessing 2. Textiles 3. Garments 4. Pharmaceuticals 5. Plastic/Rubber 6. Machinery and Equipment 7. Domestic Appliances 8. Electronics 9. Automobile 10. Furniture ii) Services 1. Information Technology 2. Telecommunications 3. Accounting/ Engineering/ Architecture 4. Advertising 5, Transportation and Related Services Since a substantial portion o f the information required i s not relevant for very small establishments, these establishments are excluded from the survey. For the manufacturing sector, only establishments with more than 10 employees are covered. For the services sector two employment cut- offs are used. Only establishments with more than 20 employees are covered for Transportation and Related Services and Accounting, Engineering and Architecture services, while the same employment cut-off as for the manufacturing sector i s appliedto the rest. Geographically, the survey covers 6 regions. Within each region, towns or areas to be covered are selected based on a concentration o f establishments as follows: 151 Appendix Table 1: SpatialCoverageof the Survey Region State TodConcentration Area PeninsularMalaysia (1) North Pulau Pinang Pulau Pinang Kedah Kulim (2) Middle Kuala Lumpur Kuala Lumpur Selangor Kelang Petaling Melaka Melaka Tengah (3) South Johor Johor Bharu BatuPahat Muar (4) East Terengganu Kuala Terengganu Kemaman East Malaysia (5) Sabah Sabah Kota Kinabalu (6) Sarawak Sarawak Kuching However, for the services sector, the eastern part o f Peninsular Malaysia i s excluded since the number o f establishments there i s negligible. The sampling frame consists o f the listings of establishments from the Department o f Statistics. This sampling frame was last updated duringthe Economic Census 2000. The Sample design used for the survey i s a single stage stratified systematic sampling design. The stratification variables are industry and region. Based on the coverage requirements, the total number of establishments in the frame and the number sampled i s summarized inTable 2 below: 152 Appendix Table 2: NumberofEstablishmentsinthe Frameandinthe Sampleby Sector No. of establishments Sector Frame Sample Manufacturing 2,841 1,000 Services 1,175 Total 4,985 1,300 The total sample size for each sector is based on practical considerations, particularly the time constraints of canvassing lengthy questionnaires within a short duration. Within each sector, the total sample size i s distributedto the substrata based on proportional allocation, as follows: where nijk = sample size for industryi, regionjand area k NUk total number of establishmentsinindustryi, regionjand area k = IZ = total sample size for the sector N=total number o f establishments inthe sector To select the sample, for each sector, establishments within each industry,region and area combination are arranged according to the value o f output. Selection i s then carried out independently for each sub-stratum based on a linear systematic method. 153 Appendix- Chapter 1 Appendix 1 Chapter 1 - A. Methodologyfor the AggregateGrowth AccountingExercise The aggregate growth accounting exercise i s based on the following Cobb-Douglas production function.relating GDP in constant 1996 dollars at purchasing power parity (Y) to inputs o f physical capital (K) and human capital (H), as well as an index of total factor productivity, A: (Al) Y = A . K a *H1-a We assume that the share o f physical capital in production i s 0.3. This i s in the range o f common estimates used inthe cross-country growth-accounting literature (see, for example, Bosworth and Collins [2003]). For a more detailed description o fthis growth accountingexercise, see Ghosh andKraay (2000). Estimates o f the physical capital stock are constructed by cumulating past investment flows and assuming a constant depreciation rate o f 10percent per year: (A2) Kt+l = 0.9 .Kt +It where Ii s investment in constant 1996 dollars at PPP. The initial capital-output ratio in 1960 i s assumed to be 1. Variations inthis assumption have only minor effects on the overall results. Human capital i s measured as the schooling-adjusted number o f workers inthe economy: (A3) H = L D P.eo.'" - - where L i s the total population, D i s the fraction of the population aged 15-64, P i s the labor force participation rate, and S i s the number o f years o f schooling. L.D.P i s simply the labor force, and the last term captures the assumption that each additional year o f schooling raises the productivity o f a worker by 10 percent. This assumption on the returns to schooling i s commonly used and is loosely based on a range o f microeconomic estimates of schooling returns. See Hall and Jones (1999) for a justification o f thisparticularvalue. Data on real GDP, investment, and population are from the Penn World Tables Version 6.1. Demographic data are from the World Bank's SIMA database. Data on labor force participation rates are from the International Labor Organization's Yearbook of Labor Statistics. Data on years o f schooling refer to the average total number of years o f schooling completed by the population aged 15 and above, and are taken from Barro and Lee (2000). An alternative source of data on schooling i s Cohen and Soto (2001), who report average years of schooling o f 3.2 in 1960 and 9.3 in2000 (as opposed to 2.4 and 8 for the same years in Barro and Lee). Both sets o f estimates imply similarly rapid rates o f growth o f schooling, o f 2.7 percent and 3.1 percent per year, respectively. However, the Cohen-Soto estimates show a decline over time inthe growth rate o f schooling\ while the Barro-Lee data do not. 154 Appendix - Chapter 1 B. Methodology for SectoralProductivityEstimates This appendix describes indetail the methodology and the data usedto construct the measures o f productivity (TFP) growth for Malaysian industries. The methodology i s taken from Basu and Fernald (1995). Estimation Methodology For each industry, we assume that real gross output Qit i s produced according to a production function o f the Cobb-Douglas form with Hicks-neutraltechnological progress: Qit = Ai, (Kit (Lit)"" (Mit)*iM by combining the industry's real capital stock Kit, total employment Lit and real materials Mi,. Ai, = Ai, egt't is an industry-specific index of technology growing at rate git. The parameters aiK, ai^, and aMare the elasticities of gross output with respect to the factors o f production, and we allow these to vary across industries. The sum of the elasticities, yi= aiK+aiL+ ai^ measures the extent o f returns to scale: ify>l (yl, which can vary over time. In this case, the share o f revenues going to each factor o fproductioni s smaller than under perfect competition, i.e.: One can show that the each input's share in total costs i s related to its corresponding revenue share: p...s.. =yi cijt, whereciLt = Wage payments inindustryi at time t IJt IJt Costs in industry i at time t and cxt and cMt defined 155 Appendix Chapter 1 - analogously.' Substituting this expression into Equation (B2) and using Equation (B4), we obtain the main empirical specification: where GCit = (CiKt GKit + CiLt GLit + CiMt GMit>is a cost-weighted average of growth rates of inputs, * * ' Treating g as the unobservable residual and estimating Equation (B5) by OLS, we can retrieve the degree o f returns to scale for each industry,as well as TFP growth as a residual. However, it i s likely that the returns to scale estimated from Equation (B5) by OLS are biased since one does not account for an endogeneity problem: firms in an industrymay respond to higher than average TFP growth by increasing their input usage. A possible solution to this problem i s to use instrumental variables estimation, provided that suitable instruments for the composite input GCi, are available, In fact, we experimented unsuccessfully with this type o f estimation, but failed to find instruments with sufficient strength to explain the endogenous variables. Rather than report spurious results based on weak instruments, we simply report the OLS results, recognizing that they may result in overestimates o f the extent o f returns to scale, and underestimates o f TFP growth.2 Data Sourcesand Definitions We use data on output and inputs for Malaysian industries fi-om the UNDO Industrial Statistics Database at the 3-digit level o f the ISIC classification. The results for Korea are based on similar data sources and methods, and are described inmore detail inWorld Bank (2000). The mainproductionvariables are defined as follows: (a) Real Output (Q) i s approximated using sales in local currency units (lcu) deflated by an output price deflator at the 3-digit ISIC level.3 The output price deflator i s implicit in the difference in the growth rates o f nominal productionand the real production index.4 (b) Real materials use is measured as materials in local currency deflated by a materials price deflator at the 3-digit ISIC level. The materials price deflator i s constructed usingan input-output table for Malaysia in 1987 and the 3-digit output price deflators mentioned in a).' The formula used is: 'Summing the output elasticities inEquation (B3) gives ~ K + ~ ~ + ~ M = ~ ( s K + s L + sSo,.Y=~(SK+SL+SM) M ) and therefore, -=(sK+sL+sM)= Y (PKK PLL Prn) - costs + + - P PQ revenues , Now, usingthis expression together with the definition o f revenue share for inputj: sj = Pjj - Pjj costs costs Y revenues --.costs revenues = c j .revenues =c. .-. so P.S.J ==y.c.J ,u ' The instruments we tried to use were one and two lags o f cost-weighted average input growth, current, one and two period lagged changes inoil prices. When obtaining TFP growth for the manufacturing sector as a whole, real output and inputs are defined as the sum o f the variables for individual industries here described. For 3-digit industries leather (323) andpetroleum derivatives (354), IPI data are not available. The output price deflators for these industries are approximated by those of relatively similar industries. The output price index o f leather i s approximated by that o f textiles (321) and the output price o fpetroleum derivatives i s approximated by that o f petroleum refining (353). The source for the input output tables for Malaysia in 1987 i s "Malaysia Input-OutputTables 1987" Department o f Statistics, Malaysia (pp. 157-161). Only a subset o f the columns and rows inthe input-output matrix were considered: those corresponding to the 3-digit manufacturing industries. The weights for each industry were 156 Appendix-Chapter 1 PM, = PjtWij IJ 28 where w1i s the input-output coefficient that represents how much output of industryj is j=1 used as an input into the production o f industryi. (c) Employment is measured by the total number o f employees (production and non-production) in the industry. (d) The capital stock is obtained by the perpetual inventory method. Nominal investment flows in local currency are available for all 3-digit industries fi-om UNID0.6These flows are deflated by the national accounts deflator for total investment to generate real investment flows.' The capital stock for industry i and year t i s obtained by cumulating real flows o f investment (IiJnet o f depreciation (6) according to the following formula: Kit =(1-8)*Kit-l Iit + We assume a depreciation rate o f 10 percent for all industries. Information on the initial capital stock o f industryiis neededto apply this formula.8 We follow Young (1995) inobtaining the initial capital stock for each industryaccording to the following formula: K. =-(h+S) Ii0where his the average growth rate o f real investment flows for industryiinthe lo first 10years o f data.g (e) Inputshares inrevenue (for growth accounting TFP): - labor revenue share: i s obtained as the ratio o f total wage payments lcu to total sales lcu. - materials revenue share: i s obtained as the ratio o f materials lcuto total sales lcu. - capital revenue share: i s defined as 1minus the labor and materials shares. For any given year t, the actual sL, sM, sK used inEquation (B3) are calculated as the average between the revenue shares inyear t and the revenue shares inyear t-I. (f) Cost shares (for econometric estimates ofTFP growth): constructedbased on t h s restricted subset o f columns and rows. Some o f our 3-digit industries o f interest show up inthe input-output table aggregated with one another under aunique industry name. Inthose cases we proceed as follows. Suppose 3-digit industries a andb show up as aggregatedunder industry c inthe input-output table. Then, for any other industry i,the weight coefficient that i s given to industry c is calculated as the input o f industry c used byindustry iI total inputsusedbyindustryi.Then an allocationrule is usedto determine the weight coefficient representingthe use by industry iof inputs from industries a and b: a `s weight coefficient will be proportional to the weight of a intotal sales a +b in 1987 and similarly for b. For some industries and years, data o n investment are missing or are equal to 0. To obtain a continuous series of investment flows per industry, we linearly interpolate the missing values. The aggregate investment deflator i s obtained as the ratio o f gross fixed capital formation incurrent prices to gross *fixedThecapital formation inconstant prices and transformed into a 1990 base year. year 1968 is the first year o f available data for most industries with some exceptions. For industry 361 (ceramics) the initial year i s 1969, for industries 385 and 390 (professional equipment and other manufacturing), the initial year i s 1973. The explanation for this formula is to assume that there were infinite periods prior to period 0 when real investment grew at the same rate hbut no data are available for those periods. Assuming that his representative o f -+I the growth of real investment prior to period 0, it is possible to show that the initial capital stock i s given by: Kio=Iio = Ii01-6 * (-)= 1-8 <1.This condition holds ifh>Oandi necessary to ensure that the s s=o l+h h+S l+h infmite geometric series converges. 157 Appendix- Chapter 1 To construct total costs, we sumthe total wage payments lcu, materials lcu and we need to impute a price to capital to construct the total cost o f capital. For simplicity, we assume that the price per unit o f capital (or rental rate) i s 15 percent for all industries and years (the sum o f a 10 percent depreciation rate and a 5 percent average annual interest rate). The total cost o f capital (or flow o f services) in constant prices i s obtained as 0.15. K,t. We transform this cost o f capital into current prices by multiplying it by the investmentdeflator fkom the WorldBank Development Indicators. - labor cost share share: i s obtained as the ratio o f total wage payments lcuto total costs - materials share: i s obtained as the ratio o f materials lcuto total costs - capital share: is obtained as the ratio o f the cost o f capital incurrent prices to total costs For any given year t, the actual CL, CM, CK used inEquation (B5) are calculated as the average between the cost shares inyear t and the cost shares inyear t-1. Table 1B.1shows the distribution o f sales and employment across 3-digit industries in 1999. Appendix Table 1B.1: Distributionof Sales andEmployment acrossIndustries,1999 313 Beverages 0.52% 0.32% 314 Tobacco 0.48% 0.77% 321 Textiles 1.82% 2.97% 322 Apparel 1.18% 4.53% 323 Leather products 0.04% 0.15% 324 Footwear 0.03% 0.11% 331 Wood products 3.36% 9.11% 332 Furniture 1.15% 3.31% 341 Paper 1.02% 1.72% 342 Printing 1.08% 2.62% 351 Industrial chemicals 6.14% 1.18% 352 Other chemicals 1.36% 1.35% 353 Petroleumrefineries 3.56% 0.28% 354 Petroleumderivatives 0.18% 0.09% 355 Rubber products 2.50% 4.91% 356 Plastics 2.20% 5.71% 361 Ceramics 0.10% 0.47% 362 Glass 0.71% 0.58% 369 Nonmetallic minerals 1.50% 2.50% 371 Iron and steel 2.38% 1.64% 372 Nonferrous metals 0.99% 0.57% 381 Metal products 2.73% 4.45% 382 Nonelectrical machinery 6.50% 4.79% 383 Electricalmachinery 39.29% 32.14% 384 Transport equipment 3.99% 3.48% 385 Professional equipment . . 0.87% 1.49% 390 Other manufacturing 0.39% 1.08% Source: Malaysia StatisticalYearbook data (source for U N D O data). 158 Appendix Chapter 1 - C.Investment ClimateIndicatorsfromthe MalaysiaPICSSurvey andExternalIndicators Our analysis relies on investment climate indicators calculated for Malaysia as a whole, but also on indicators disaggregated across regions, 2-digit industries and types o f firms (small-medium-large, exporters-nonexporters and FDI-nonFDI firms). The description o f the indicatorsbelow i s valid for either type o f disaggregation o f the data. For all the quantitative indicators (most inparts b and c), we restrict our analysis to observations that are within 3 standard deviations o f the mean for Malaysia as a whole. Perceptions of theInvestment Climate by Firms 0 Closedquestion Part I- question V.1: we calculate the number o f firms that respond 3 or 4 (severe or very severe obstacle) to each o f the items .1-18 in the question. We divide this number o f firms by a common denominator to obtain the percentage of firms complaining about each type o f obstacle. The common denominator i s the maximumnumber o f non-missingresponses among items 1-18. 0 Open -ended question Part I- question V.2: for each o f the 21 items on the list (constructed post-survey), we calculate the percentage o f firms that pointed it out as either the main obstacle, the second main obstacle or the third main obstacle to doing business inMalaysia. Infrastructure Indicators Part I- question VI.14 a): for each firm the number o f power outages per month i s multiplied by 12 to generate the total number o f power outages in 2001; these values are averaged across firms to generate the frequency o f power outages in2001. Part I- question VI.15: the average o f this variable indicates the percentage o f production lost due to power outages in2001. Part I- question VI.16: the proportion o f firms answering yes to the question indicates the percentage o f f i r m s that own a generator. Part I- question VI.13: the average o f this variable indicates the percentage o f production that i s lost in shipment. Part I- question VI.6 a)-c): the average o f these variables indicates the number o f days needed to obtain telephone, water and electricity connections. Regulatory and Administrative Burden Indicators Part I- question VI.9: the average o f this variable indicates the percentage o f time that senior management spent dealing with regulations in2001. Part I- question VI.7 1st column: we calculate a total number o f days spent dealing with inspections (summing the number o f days spent dealing with visits from the tax inspectorate, labor and social security, fire andrescue and local) and average it across firms. Part I- question VI.7 2nd column: we sum the costs o f fines from the tax inspectorate, labor and social security, fire and rescue and local, divide this value by the firm's sales in 2001 and average the ratio to obtain the value o f fines from inspections as a percentage o f sales in2001. 159 Appendix- Chapter 1 Part I- question VII.5.y.b. 1st item: the average of this variable indicates the average number o f days needed to clear customs for exports by Malaysian firms. 2nd item: the average of this variable indicates the longest number o f days neededto clear customs for exports by Malaysian firms. Part I- question VII.6.y.a. 1st item: the average of this variable indicates the average number o f days needed to clear customs for imports by Malaysian firms. 2nd item: the average o f this variable indicates the longest number of days needed to clear customs for imports by Malaysian firms. GovernanceIndicators Part I- question V.6: we calculate the proportion of firms that respond 1 or 2 (fully disagree or disagree in most cases) to the question to obtain the percentage of firms that do not have confidence in the judiciary. Part I- question V.5.y.a.: the average o f this variable indicates the percentage o f payment disputes resolvedby firms incourts in2000-2001. Indicators on the Regulatory Burdenfor Opening a New Busin,essin 2001 Part I- question VI.2.: the averages of cells in the lst, 3rd and 5th columns o f the table indicates the average number o f licenses/permits/approvals needed to start a business in 2001 by the federal, state and local governments. The averages of cells in the 2nd, 4th and 6th columns indicates the average time to obtain those documents from the federal, state and local governments. In calculating these averages we restrict the sample to firms that were created in 1997 or after (these younger firms have a better perceptiono f the burdenrequired to open a new firm). Reasonsfor Firms Being Overstaffed or Understaffed Part I- questions IV.4.a. and IV.4.b.: we calculate the percentage o f firms that answer Yes to each o f the items a-e inquestion IV.4.a. and to each o f the items a-d in question IV.4.b. Specific Labor Regulations Part I- question IV.8: we calculate the percentage of firms that respond 1-4 (representing an obstacle) to each of the items 1-5 inthe question. Time to Get Licenses, Permits or Approvalsfrom Specific Agencies Part I- question VI.5.: the average and standard deviation o f the number o f days to get licenses from each o f items 1-5 are calculated excluding values of 0 and the top 1percent o f values. Timeto Get Licenses, Permits or Approvalsfrom Specific Agencies Part I- question V1.6.: the average and standard deviations of the number o f days to get licenses from each o f items 4-6 are calculated excluding values o f 0 and the top 1percent o f values. Timeto Process Applicationsfor Export Incentives Part I- question VIII.4.: the average and standard deviations of the number of days to process applications for each of the items 1-11 are calculated excluding values of 0 and the top 1 percent o f values. 160 L B3 Y 8 I x r E E-r kh E P 9 a c 0 B I Appendix Chapter 1 - D.Firm-LevelPerformanceIndicators Production Data The analysis of firmperformance is based on a sub-sample of manufacturing firms from the PICS survey for which the values of major variables are not considered to be outliers and for which data on all production variables are available. The original total number o f firms in the survey before selection i s 902. We define an observation (firm) to be an outlier for variable X, if its value i s larger than the mean of X in the corresponding 2-digit industry by more than 3 standard deviations of X (inthe industry)or ifits value i s smaller than the industrymean by more than 3 standard deviations. The variables for which this outlier rule i s applied are:" output-labor ratio, capital-labor ratio, materials-labor ratio, labor share in revenue, materials share in revenue." For the inputrevenue shares, two additional rules are applied: (i) firms with labor or all materials revenue shares larger than 1are classified as outliers; (ii) all firms with materials shares smaller than 0.1 are classified as outliers. These outliers are dropped from the calculation of industry means and standard deviations required to apply the 3 standard deviations rule above. Applyingall outlier rules, the sample size is reducedby229 firms. The mainproductionvariables are defined as follows: output for years 2001, 2000, 1999 i s given by total sales in the table o f Part IIA - question IX.1. materials for years 2001, 2000, 1999 are given by direct material cost inthe table o f Part IIA-question IX.1. skilled labor for years 2001, 2000, 1999 i s given by the sum o f the number of management, professionals and skilled production workers in the table o f Part IlB - question X.2. growth insales and value addedper worker. unskilledlabor for years 2001,2000, 1999 i s givenby the sum o f the number o f unskilled productionworkers and non-production workers inthe table of Part IlB- question X.2. '* capital stocks for years 2001, 2000, 1999 are given by net book value o f machinery and equipment inthe table ofPart IIA-question IX.13. Inputrevenue shares usedto determine the outliers are definedas: labor revenue share: i s calculated as the sum of total wages and salaries paid to management, professionals and skilled production workers in the table o f Part III3 - question X.8, plus the sum of total wages and salaries paid to unslulled production '10The definition o f each o f the variables entering these ratios and o f the shares i s provided immediately below. Outliers are identified for the ratios and shares in each of the 3 sample years 1999,2000, and 2001 when data permit. l2Whenever one o f the subcategories of skilled or unskilled workers i s missing we assume that the firm does not employ that type o f worker and we replace that missingvalue by a 0. 164 Appendix Chapter 1 - workers and non-production workers in the table o f Part IIB- question X.8, divided by the total sales in2001 from the table o f Part IIA - question IX.1. l3 0 materials revenue share: i s calculated as direct material cost divided by the total sales in 1999, 2000 and 2001, both from the table o f Part IIA- question IX.1. Total costs are defined as the sum o f the following items inPart IIA- question IX.1: direct material cost + electricity + others (in consumption o f energy) + wages and salaries + allowances, bonuses and other benefits + transport costs + interest charges and financial fees + selling and general administration expenses +other costs.'4 Measures of Performance The four performance measures used in the analysis are sales growth, value added per worker, profit margins and TFP obtained by two different methods. The first 3 measures are defined as follows: (i) growthisthelogarithmicchangeinsalesbetween2000and2001 Sales (ii) addedperworker istheratioofvalueadded(defined assalesminusdirectmaterial Value cost) to total employment defined as the sum o f skilled and unskilled workers. (iii) marginsarecalculatedastheratioofthedifferencebetweensalesandtotalcoststo Profit sales. The TFP measures rely on the following model. For firm iinindustryj, the technology i s described by a Cobb-Douglas production function with Hicks-neutral technical change (inlogarithmic form): where output q i s produced combining the capital stock k, raw materials m, skilled labor s and unskilled labor u. v i s residual firmproductivity and can be decomposed as follows: (D2) vk =wk +E; where w i s a component o f firm productivity that i s known to the firm manager and possibly affects input choices but i s unknown to the econometrician and E i s a random shock to output/productivity that i s realized after input choices are made (therefore it i s not correlated with input choices.'' We obtain estimates o f the parameters(a, y,#) in Equation (Dl) estimated P, separately across industries using ordinary least squares (OLS) techniques. The OLS production l3Whenever one o f the subcategories o f wages o f skilled or wages o f unskilled workers is missing we assume that the fmdoes not employ that type o f worker and replace that missing value o f wages by a 0. l4Some strong assumptions are made inthe calculation o f costs inorder to maximize the number o f f m s for which profit margins can be calculated: any cost item for which there i s a missing value i s replaced by a 0. lSMore specifically, the following equalities are verified for the conditional expected values: E[Eit/ si,]=O, E[Eit / ~it]=O, E[~it/ mit]=O andE[Eit / kit]=O. 165 Appendix Chapter 1 - function parameters are presented in Appendix Table D.1.16But, firms that are more productive will hire more labor and use more materials in order to produce more. OLS estimates o f the production function parameters do not take into account the endogeneity o f input choices, thus they are biased. As an alternative to OLS estimation, we follow Levinsohn and Petrin (2002) to obtain production function parameters that correct for the endogeneity o f input choices. The basic specification i s given by Equations (Dl) and (D2). The main idea i s that firms choose their variable inputs (slulled labor, unskilled labor and materials) with a knowledge o f their productivity w whereas capital i s a quasi-fixed inputthat i s costly to adjust. The demand for variable inputs by a firm can be derived from profit maximization and depends on capital k and on productivity@ . Inparticular, the demand for materials i s given by mit =m(c-oit,kit).", Under fairly generaltechnical conditions(described indetailinLevinsohn and Petrin, it i s possible to invert the materials demand function and express the unobserved (to the econometrician) firm productivity as a function o f two variables (materials and capital) that are observable: wit =w(mit,kit) . The crucial idea behind this estimation method i s to use this proxy for productivity to control for the endogeneity o f input choices with respect to productivity. The estimation proceeds intwo stages. Stage 1: Replacing the proxy equation for productivity into Equation (D2) and this equation itself into Equation (Dl) we obtain: (D3) qit ="emit +Pakit +y-sit +q3.uit+o(mit,kit)+Eit or The function h(.) groups all the terms on materials and capital. Its functional form i s unknown but we can approximate it either (i) a polynomial in materials and capital or (ii) using usinglocally weighted least squares as follows: l6For production h c t i o n estimation, the data for 1999, 2000 and 2001for all f m s inthe industryare included (as long as the firms are not outliers and have no missing productionvariables). However, inthe analysis o fperformance across Malaysian firms we focus on the measures o fperformance in2001only. "Forsimplicityofnotation, wedropthesuperscript j intherestofthe section, butthe estimationis performed for each industryj separately. '*Ideally, the function m(.) should vary across years (Le., m, (wit, kit) to account, for example, for changes inthe price o f materials, other inputs or output. This would be a way (though an imperfect one) to account for these prices for which data are not available. However, inpractical terms, given the small number ofobservations available for eachindividual year and industry inthe MalaysianPICS dataset, we choose to consider a function m(.) that does not vary across years. 166 Appendix Chapter 1 - (i) Addingthe terms o f a 4' degree polynomial inmand kto substitute for h(.) inEquation(D4), it can be estimated by OLS to provide consistent parameter estimates for shlled and unskilled labor. (ii) theexpectationofbothsidesofEquation(D4),conditionalonmandkgives: Taking N o w E[h(mit,kit)/mit,kit]=h(mitykit)and since sit i s uncorrelated with any o f the inputs E[sit / mit,kit]=0 .So Equation(D5)canberewritten as: Taking the difference between Equation (D4) and Equation (D5'), one obtains: This equation with "transformed regressors" can be estimated by OLS (with no constant term) if one obtains estimates for the conditional expected values. These conditional expected values are approximated by quadratic locally weighted least squares regressions of ylt, sit and ui, on (mit, kit).'9It shouldbe notedthat one also obtains an estimate for the unknown function h(mit,kit)that i s used in Stage 2. Stage 2: An additional assumption i s required to obtain the coefficients on materials and on capital. We assume that productivity i s serially correlated: it follows a Markov process wit = E[wit /wit-l]+ vi,. Also, the main identification assumption in Stage 2 i s that capital i s slow to adjust: it may adjust to the expected part o f productivity conditional on lagged productivity '[wit /wit-,], but it does not adjust to the unexpected shock yit. So, one can derive the following moment condition that identifies the coefficient on capital: 2o (D7) E[yit sit /kit]=0 + A separate moment condition is needed to identify the coefficient on materials. A condition for materials exactly parallel to that for capital cannot be used since materials are correlated with productivity (both the expected and unexpected parts): E[vit+sit/mit]# 0.21 l 9The quadratic locally weighted least squares procedure i s described inLevinsohn and Petrin (2002). Basically, weighted least squares estimation i s usedto construct predictions of yit, sit, ui,given (mi,, kit) including as regressorsthe terms from a 2ndorder polynomial in(mit,kit).For any data point (mit*,kit*), for which one needs an estimate of the 3 conditional expected values, the regression gives greater weights to the observations closest to the point (mi,*, kit*).A consistent estimator for the conditional expected value is the intercept from this locally weighted regression. + The definition of it that E[Eit / kit]=0. So E[$it+it/kit]=0. 2o E[$it Eit / kif]=E[$it / kit]+E[Eit / k,J. The main identification assumption says that E[$it /kit] =0. implies 21 Although E[Eit / qt]i s 0, E[$i, / kit]is differentfrom 0. 167 Appendix Chapter 1 - But one can derive a moment condition for laggedmaterials. Materials inyear t-1 are chosenby the firmwithout knowledge o f the productivity shocks realizedonly inyear t, so: (D8) E[y, +sit =0 The residual yit sit in the moment conditions i s obtained replacing productivitywitby + Iits Markovprocess and switching sides for some o f the terms inEquation (Dl): (D9) yit sit =qit -f sit - uit-a mit - kit -fi[wit / wit-l] + 6 P It should be noted that in Equation (D9) we replaced the coefficients on slulled and unskilled labor by their estimated values from Stage 1 and we included an estimate for the expected value o f productivity conditional on lagged productivity. This conditional expected value i s a function o f wit-l. W e call it g(w.11-1 ) but its functional form i s unknown. W e approximate this g(.) function by locally weighted least squares o f an estimate o f wit on an estimate o fqt-,,22 Sample analogs for the two moment conditions in Equations (D8) and (D9) are obtained for all firms and a general method of moments (GMM) criterion function i s constructed. The estimates for the coefficients on materials and capital are those making the sample analogs o f the moment conditions as close to 0 as possible:23 An iterative procedure needs to be used to minimize this function starting from some candidate initial parameters (11and p (e.g, those obtained by OLS). The production function parameters obtained by this estimation method are presented in Appendix Table D.2. The OLS coefficients on variable inputs are expected to be upward biased since they do not control for endogeneity. Generally, the coefficients on most inputs in Appendix Table D.2 are lower than those inAppendix Table D.1.24 For both OLS and Levinsohn and Petrin (2002), the TFP measures are obtained as the residual from Equation (Dl) (the difference between output and inputs weighted by the estimated production function parameters). Since TFP measures obtained by OLS are likely to be biased, 22Specifically, the estimate for the expected productivity conditional on lagged productivity i s givenby an LWLS regression o f an estimate for ait givenby(wit 4 sit) = yit -f sit - uit -a * 6 *.mit /3 *.kit - * . .4 on an estimate for q t - 1givenby hit-l= h(mit-l, kit-l)-a .mit-l - .kit-l Note that both o f these P * estimates depend on the parameters o f interest inStage 2: and p. 01 23The sample analogs for the moment conditions are summed across firms i(the index iinthe 1St summation symbol). For each firm the moment conditions are summed from the 2ndyear o f available data on since the procedure uses lagged inputs (the index t inthe 2ndsummation symbol). N o moment condition canbe computed inthe lst year of available data for a fm. 24Although one expects an upwardbias inthe OLS coefficients on skilled and unskilled labor and materials, theoretically, one can expect either an upward or downward bias inthe OLS capital coefficient (see Levinsohn and Petrin). 168 Appendix Chapter 1 - we rely in the chapter on the analysis o f TFP measures obtained by the Levinsohn and Petrin method. Data on Correlates of Performance In the regressions analyzing the determinants of firm performance, several firm characteristics are used: Region dummies are constructed based on the region code indicated for each firm as its location (6 regions: central region, north region, south region, East Coast, Sabah, Sarawak). Industry dummies are constructed based on a 2-digit classification (9 industries: food, textiles, garments, chemicals, rubber and plastics, machinery and equipment, electronics, auto-parts, wood and furniture). Age is defined as 2001minus the response to Part I- question I. 1. Size i s measured by the logarithm o f the firmcapital stock (defined above). The exporter dummy i s defined to be equal to 1 if in Part IIA - question VIII.9 the percentages given for sales exported directly and sales exported indirectly in2001 sum to more than 10percent o f total sales. The FDIdummy i s defined to be equal to 1if inPart I- question 1.5.1the percentage o f the firm owned by the private sector foreign i s any positive number. Technology variables o computer controlledmachinery is the percentage given inPart I-question 111.7 o vintage o f capital is the percentage o f machinery and equipment of the firm that i s less than 5 years old given inPart I- question 111.6 Determinants of Firm Performance The regression specification used to analyze the determinants o f firm performance i s the following: (D11) Pirj - -6,+6,.agei +S,.size, +6,.expi+6,.FDI, +6, 'comp, +66.Kvinti+I'+I'+&~ where p ii s a measure o f firmperformance in 2001 and the determinants o f firmperformance are firm age age, , firm size size, ,firm export status exp, ,foreign ownership of the firm FDI,, two measures o f technology use comp, ,Kvint, , I'are 2-digit industry effects and I'are region effects. The regressions are estimated by OLS and standard errors that are robust to heteroskedasticity are computed. 169 Appendix Chapter 1 - As mentioned in the main text for the chapter, we also conduct an analysis o f partial regressions that include only one firm characteristic at a time together with industry and region fixed effects. For example, when firm size i s the only characteristic included, the partial regression i s as follows: (D12) p i =6, + A size, e +I'+I'+vi Comment on the Analysis of Determinants of Firm Performance a Although we do not report the results, we also estimated equation (D11) using TFP obtained from the OLS parameter estimates as a measure o f firm performance. The results are generally very close to those reportedinthe main text obtained for TFP usingthe parameters from Levinsohn and Petrin (LP) estimation, except for the effect o f firm size. Larger firms have lower TPFOLS but have significantly higher TFPLP. The LP production function parameter estimates on capital are generally smaller than the OLS estimates. Therefore, TFPLP tends to be larger than TFPOLS, especially for firms with large capital stocks. Hence, the reverse effect o f size i s obtained for TFPOLS relative to TFPLP. a In Appendix Table D.3 we present the correlation among firm characteristics affecting performance for the sample o f 401 firms used for the estimation o f the regressions o f value added per worker and TFP. Firm size i s significantly positively correlated with the exporter dummy, the foreign ownership dummy, the percentage o fcomputer-controlledmachinery and the vintage o f the capital stock. Costs of Skill Shortages The framework o f the production function and profit maximization by firms can be used in thinking about the problem of the shortage of skillshkilled workers. A firm makes its input choices inorder to maximize profits as follows: (D13) Maxs,,,,,,[P.AKaMPSYU4-r.K-P, .M-w, ~S-W, .U] where Q=AK"MPSYU4 represents output produced according to the production function whose estimation we discussed before, P i s the price at which the firm sells its output, r is a rental rate paid on capital, PMi s the price o f materials, WS i s the wage paid to a skilled worker and wu i s the wage paid to an unskilledworker. Ifthe firm faces no constraints, it hires skilled workers untilthe first order optimization condition i s ~atisfied.~' S i s chosen so that: (D14) MarginalProduct o f S = y P Q/ S = w Ifthe firmis constrained, that is, ifitcannot hire as many skilled workers as desired, this optimization condition i s not verified. For the (smaller than desired) number o f skilled workers that it i s feasible for the firm to hire, the condition i s infact: 25This occurs when the derivative of the profit function with respectto S equals 0. The interpretation is that the fmhires more skilled workers as long as the marginal product of each additional skilled worker is larger than or equal to the wage the fmneeds to pay h i d h e r . 170 Appendix - Chapter 1 (D15) Marginal Product o f S= y P Q / S > ws The difference between the marginal product o f skilled workers and the corresponding wage represents the cost (in terms o f outputhales lost) to the firm o f not being able to hire one additional skilled worker. Alternatively, one can think o f this difference as the benefit from reducing the skill shortages. The optimal skill mix for an industry i s derived from profit maximization, combining Equation (D14)with the corresponding first order condition for unskilled labor? (D16) (-)*= S Y'WU s+u yaw, +@'W, Replacing y and @ by the values from the production function estimation, the optimal skill mix can be computed for eachindustry. We analyze the effect o f the following experiment. Suppose that skill shortages are reduced (e.g., owing to large numbers o f college graduates entering the labor market) and constrained firms can hire more skilled workers inorder to approach the optimal skill mix intheir industry.The firms that are constrained interms o f the number o f skilled workers they can hire are defined as those having a skill mix -smaller than the optimal skill mix intheir industry. S s+u For each constrained firm, the benefit in terms o f increased sales from increasing its number o f skilled workers and moving closer to the optimal skill mix i s given by: where the first term inbrackets inthe product represents the additional benefit per skilled worker and the second term in brackets represents the increase in the number o f skilled workers that causes the skill mix o f the firm to become equal to the optimal skillmix. Using the value o f yfrom production function estimation, Equation (D17) can be calculated for each constrained firm. We sum the product in Equation (D17) for all o f the constrained firms in each industry and divide the sum by the sum o f the sales o f the constrained firms in each industry, and call this ratio A.27To obtain the final potential benefits in terms of increased sales from reducing skills shortages we divide the ratio A by 2. The rationale for this final step i s as follows. The graph below represents the marginal product o f skilled labor as a function of the number o f skilled workers. For simplicity it i s assumed to be a linear function. Also inthe graphi s a horizontal line representing the skilled wage ws. 26 That first order condition i s given by: MarginalProduct o f U= @ P Q/ u= w , 27 Itshould be noted that Equation (D17) takes a negative value for some o f the constrained firms. W e disregardthese negative values and focus on constrained firms for which the benefits o f reducing the skill shortages are positive. 171 Appendix Chapter 1 - Mg. Prod. ofs We denote by S the current number of slulled workers of a constrainedfirm andby Snew the number of skilled workers that the firm can hire under the experiment describedbefore. The benefit to the firm from this reduction in skill shortages (the product inEquation (D17)) i s given by the area of the square a.b.c.d. (thisrepresents additional output). We are able to calculate the area of this square for each firm and then sum across firms in an industry.However, the benefit to the firm interms of additional output from a reduction in skill shortages is given by the shaded triangle b.a.d. Ifthe marginal product of skilled labor was truly linear the area of b.a.d. would be equal to half the area of a.b.c.d. Therefore, as an approximation, we assume that the marginal product is linear and we divide the product in Equation (D17) calculated for each firm by 2 to give the benefit from relaxing the skill constraint. AppendixTableD.l: ProductionFunctionParameters Estimatedby OLS N Coeff. Coeff. Coeff. Coeff. R. scale Industry skilled unskilled labor labor materials capital 15=FoodProcessing 312 0.046 0.034 0.827 0.083 0.990 17=Textiles 38 0.116 0.125 0.573 0.160 0.973 18=Garments 98 0.061 0.178 0.601 0.136 0.977 24=Chemicals 32 0.280 0.002 0.695 0.024 1.001 25=Rubber/Plastics 363 0.177 0.096 0.654 0.099 1.026 29=Mach./Equipm. 110 0.174 0.109 0.647 0.090 1.020 32=Electronics 96 0.122 0.072 0.739 0.048 0.981 34= Auto. Parts 53 0.178 0.218 0.597 0.064 1.056 36=Fumit./Oth.Man. 150 0.185 0.135 0.662 0.048 1.030 Notes: Nindicates the number of firm-years includedinthe estimation. The hypothesis of constant returns to scale cannot be rejectedinall industries. 172 Appendix Chapter 1 - Appendix TableD.2: ProductionFunctionParametersEstimatedbyLevinsohnandPetrin(2002) Methods N Coeff. Coeff. Coeff. Coeff. R. scale Industry skilled unskilled labor labor materials capital 15=Food Processing 311 0.033 0.031 0.620 0.070 0.753 17=Textiles 38 0.252 0.149 0.560 0.060 1.020 18=Gments 97 0.115 0.192 0.620 0.120 1.046 24=Chemicals 32 0.226 0.056 0.560 0.210 1.052 25=Rubber/Plastics 360 0.157 0.094 0.660 0.110 1.021 29=Mach./Equipm. 110 0.068 0.027 0.580 0.040 0.715 32=Electronics 94 0.101 0.049 0.720 0.040 0.910 34= Auto. Parts 53 0.170 0.198 0.640 0.084 1.091 36=Furnit./Oth.Man. 148 0.106 0.085 0.560 0.140 0.891 Notes: Nindicates the number of firm-years includedinthe estimation. Appendix TableD.3: CorrelationamongFirmCharacteristics Capital Plant Age Plant Size Exporter Foreign %Comput. Capital Stand. O m . Contr. Intensity Dummy Dummy Mach. Vintage Dev. Capital Intensity 1 Plant Age -0.252*** 1 8.80 Plant Size (total assets) 0.780*** -0.117*** 1 2.28 Exporter Dummy 0.167*** -0.081 0.3 13*** 1 Foreign Ownership Dummy 0.061 -0.05 0.236*** 0.274*** 1 %Computer-Controlled Mach. 0.206*** 0.01 0.296*** 0.079 0.064 1 30.51 Capital Vintage (% mach. under 5 years old) 0.144*** -0.093 0.212*** 0.132*** 0.038 0.154*** 1 29.59 Note: *** indicates significance at the 5% confidence level. These correlations are calculated for the sample of 401 firms usedinthe regressionsof VAL, TFPOLS, TFPLP. 173 Appendix Chapter 1 - Appendix TableEl:Listsof Components Comprisingthe CompositeRegulatoryQuality Index. Code Table RweamtatiueSow- DRI w Reguldjons -i5pott.s: A 2% reduction in export wlume as a resllt of a vrorseringin export regulaticns or restrictions[such as export limits) during any 12month period,Mh respect to the level at thetimeofthe assessment. Reguldjons -lnpottsAZ% reductioninimport volume asa resuR of a wrsening in import regulaticns or restrictions[such as import quotas) during ary 12-month period, vith respectto the level at thetime of the assessment. Reguldjons - DherBusiness: Pn increase inother regulatorybudens, withrespectto the levelat the CwnersA@ofBushes ty M+Residents: A I-point inwe-e ona scale from "0" to "I in legal 0" restrictionson omership of business bv mn-residents durins any 12month period. CwnershQ ofguRies ty Non-ReeHents:A 1.point increaseon a scalefrom "W to "IO" in legal restridionson omership of eguitiesby non-residentsdurincl my12-morth period. EIU I% UnMr comwtitive practices Price controls Discriminatcrvtanffs Excessive protections HER A12 RegAdion GovernmentIntervention WacleBrices Trade Foreiclninvestment Bankim PRS A15 InvedmentPmf#e. lndudesthe risk to operations (scared from 0to 4,increasing in risk1taxation [scoredfrom 0 to r),repatriation[scored from Oto 3); repatridion(xored from 0 to 3) and labor costs wo [scared from 0 to 3. Thev all look at the governmentsattitudet m r d s investment. AI8 Tax ERectiveness Howefliciemlthe comtrys tax collection system is. The rules may be dear and transparent, but vhether they are enfcrced consistently. This fwior looks at the reldive effectiveness too of cornorateand perscnal,indrect and direct taxation. LegMdion: Pn assessment of whether the necesiaty business l a w are inplace, End Mether there any outstandnggaps. This indLdest k extent to vhichthe countrys legislationis compatiblewith, and resnectedbv. other comtries' l e a l svstems. H m - m a t i u esome3 BPS Informationonthe law EndregAdions iseasyto obtain Interpretatims ofthe l a mand regulationsare consistentand predictable Unpredictabilityof changes of regulations Howproblematicare labor regulationsfor the growth of your business. Howproblematicare tax reguationsforthe grovrth of yarr business. Howproblematicare custom Endtrack regllationsforthe grovrth of your blsiness. CP14 A? Competitiveendrorment Factorand productsmarkets Trade policy EBRD R Price liberalisation Trade 8 foreign exchmge system CompetitionpAicy Commerdai LawExtensiveness Commerdal LawEffectiveness FinancialRegllationx e~ensiveness Financialregllations: effectiveness 174 Appendix Chapter 1 - Appendix Table El (cont.) Cock Table Cmcept W a r d Nm--atiue Solrcles GCS A11 Administrativeregulationsareburdemome Tax system is disrtortiomry Import barriersasobstacleto govdh Commition in local market is limitd It iseasyto start company Mimonopolypdicyislaxandineffective Clustersare frequmt Endronmentd regulationshurt competitiveness Cost of twiffs imposedon bmines Government subsidies keep urrcompditiveindustriesalive artificidly A17 T k exchance ratemliw of wur ccuntrvhindersthe mmnetitivemss of enternrises Prdectionim inyour counttyneqativelyaffectst k conduct of businessin your country Competitionleqisldioninyour muntw doesnot prevent unfair compdition Pricecontrolsaffect pricinqof prodctsin most industries Leqd requlationof financid institutims isinadequatefor financial stability Foreiqnfinancial institutionsdo not have mess to the domestic market k e s s t o locd m ~ i t amarketsisrestrictedfw fordun comnmies l k e s sto fordqncapital marketsis restrictd for d mestic compmies Finarrcialinstitutions'traraparenw is nd vridelydevelopedin wur ccuntry Customs'authoritiesdo rrot facilitatethe efficienttramit of qoocts T k lwal framework isdetrimentalto your courtrvscomwtitiveness Foreiqninvestorsarefree to =quire control in domestic companies Public sector contracts are suflicientlv onento foreiun bidders Red ~lers~nal taws we nondidortionarv Red corporatetaxes are non dstotlionwy Bankiw reqllation dclesnot hinder competitivenes 175 Appendix Chapter 2 - Appendix 2 Chapter2- A. FirmSurvey Characteristicsof the Firms There are a total o f 249 firms surveyed in the project.' These firms are diverse in terms o f industries coverage, ownership structure, foreign influence, age, size and locations. We break down the locations of the firms into 5 regions. More than 70 percent o f the firms are from the region o f Selangor, Kuala Lumpur and Melaka, while the rest o f the regions make up the remaining 30 percent o f the sample. Onthe other hand, while these firms spread among 28 service industries, nearly 30 percent o f them are in the architectural and engineering industry, follow by the freight transport by road industrywhich hosts 14 percent o f the sample firms. Other important industries in the sample in terms o f the count o f firms include activities o f other transport agencies, accounting, the book-keeping and auditing industry,and the software publishing, consultancy and supply industry. Table 2A.1 details the distribution o f firms by industries and regions, Table 2A.2 shows the distribution o f firms by industry and age group, and Table 2A.3 gives the distributionby industryand foreign equity. Appendix Table 2A.1: Distributionof Firmsby IndustryandRegion Region SIC Industry descriptions Selangor, Penang & Johor Sabah Sarawak Total Code KL& Kedah Melaka 45 10 Construction site preparation 0 0 1 0 0 1 6010 Railways 1 0 0 0 0 1 6021 Other scheduled passanger landtransport 4 0 0 0 0 4 6023 Frieghttransportbyroad 18 8 4 2 2 34 6110 Sea& coastal water transport 10 0 0 1 1 12 6112 1 0 0 0 0 1 6210 Scheduled air transport 1 0 0 0 0 1 6220 Nonscheduledair transport 1 0 0 1 1 3 6301 Cargo handling 2 2 1 0 1 6 6302 Storage &warehousing 0 1 0 0 0 1 6303 Other supportingtransportactivities 1 1 2 0 0 4 6309 Activities o fother transportagencies 15 2 1 2 1 21 6412 Courier services 3 0 0 0 1 4 6420 Telecommunications 11 0 0 0 0 11 7112 Rentingo fwater transportequipment 0 0 1 0 0 1 7121 Rentingofagricultural machinery andequipment 2 0 0 0 0 2 7123 Rentingofoffice machinery& equipment 0 1 0 0 0 1 7210 Hardwareconsultancy 5 0 0 0 0 5 7220 Softwarepublishing, consultancy& supply 18 0 0 0 0 18 7230 Dataprocessing 2 0 0 0 0 2 7240 Data activities &on-line distributionofelectronic content 1 0 0 0 0 1 7250 Maintainance &repair o foffice, accounting & computingmachinery 1 0 0 0 0 1 7412 Accounting, book-keeping & auditingactivities; tax consultancy 17 1 1 0 2 21 7421 Architectual &engineering activities &related technicalconsultancy 50 5 3 5 9 72 7422 Technical testing & analysis 0 0 0 1 0 1 7430 Advertising 16 0 0 0 0 16 8322 3 0 0 1 0 4 Total I1 183 21 14 13 18 249 One firmi s droppedbecause it belongs to the manufacturingsector. 176 Appendix- Chapter 2 AppendixTable 2A.2: Distributionof Firmsby IndustryandAge Group Age Groups SIC Industrydescriptions 5 years or 6 to 10 11 to 15 16 to 20 21 years or Total less years years years more ~Code 45 10 Construction site preparation 0 1 0 0 0 1 6010 Railways 0 0 0 0 1 1 6021 Otherscheduledpassenger landtransport 0 3 0 0 1 4 6023 Freight transport by road 1 3 11 4 15 3f 6110 Sea & coastal water transport 3 3 1 1 4 1; 6112 0 0 0 0 1 1 6210 Scheduled air transport 0 0 0 0 1 1 6220 Nonscheduled air transport 0 2 0 0 1 2 6301 Cargo handling 0 0 1 0 5 t 6302 Storage &warehousing 0 0 0 0 1 1 6303 Other supporting transport activities 1 1 0 1 1 4 6309 Activities ofother transport agencies 0 4 5 6 6 21 6412 Courier services 0 1 0 2 1 4 6420 Telecommunications 1 7 0 1 2 11 7112 Rentingof water transportequipment 0 0 1 0 0 1 7121 Rentingofagricultural machinery andequipment 0 0 1 1 0 1 7123 Rentingofoffice machinery &equipment 0 0 0 0 1 1 7210 Hardware consultancy 1 1 1 1 1 1 7220 Software publishing, consultancy & supply 5 4 5 2 2 1; 7230 Dataprocessing 2 0 0 0 0 1 7240 Data activities & on-line distribution ofelectronic content 0 1 0 0 0 1 7250 Arquitectual & repair of office, accounting & computing machinery 0 0 1 0 0 1 7412 Accounting, book-keeping & auditing activities; tax consultancy 0 0 0 3 18 21 7421 Maintenance & engineeringactivities & relatedtechnical consultancy 1 9 14 16 32 7: 7422 Technical testing & analysis 0 0 0 1 0 1 7430 Advertising 1 1 3 4 7 1f 8322 0 0 0 - 1 3 1 Total 16 41 44 44 104 245 177 Appendix -Chapter 2 Appendix Table 2A.3: Distributionof Firmsby ForeignEquity ForeignEquity Opercent 30percentor lesmorethan30percMorethan 50perrent Total to 50percent 4510 Constructionsitepreparation 1 0 0 0 1 6010 Railways 0 0 0 1 1 6021 Other scheduledpassengerlandtransport 3 0 0 0 3 6023 Freighttransportbyroad 29 3 2 0 34 6110 Sea& coastalwater transport 10 2 0 0 12 6112 1 0 0 0 1 6210 Scheduledairtransport 1 0 0 0 1 6220 Nonscheduledair transport 3 0 0 0 3 6301 Cargohandling 3 2 0 0 5 6302 Storage& warehousing 1 0 0 0~ 1 1 6303 Othersupportingtransportactivities 3 1 0 0 4 6309 Activitiesofothertransportagencies 17 1 3 0 21 6412 Courierservices 2 1 0 1 4 6420 Telecommunications 7 1 0 3 11 7112 Rentingofwatertransportequipment 1 0 0 0 1 7121 Rentingofagriculturalmachineryandequipment ~ 0 0 1 1 2 7123 Rentingofoftice machinery& equipment 1 0 0 0 1 7210 Hardwareconsultancy 1 0 1 2 4 14 0 0 3 17 1 0 0 1 2 1 0 0 0 1 1 0 0 0 1 19 1 1 0 21 63 2 3 0 68 1 0 0 0 1 3 4 3 5 15 8322 4 0 0 01 4 Total I 191 18 14 171 240 Interms of the distribution of domestic equity among different ethnic groups, 44 percent of the firms in the sample are dominated by Chinese, 43 percent by Malays, and only 3 firms are dominated by Indians.' While Malays have the greatest presence in industries such as activities o f other transport agencies (75 percent), and architectural and engineering activities (49 percent), Chinese play a more significant role in industries such as accounting, book-keeping and auditing activities (65 percent), and software publishing, consultancy and supply (53 percent). Table 2A.4 presents the distribution o f firms by industryand the majority ethnic indomestic equity. We have domestic ownership information for only 220 firms inthe sample. 178 Appendix Chapter2 - Appendix Table2A.4: Distributionof FirmsbyMajority Domestic Equity Majority DomesticEquity aSICb Malays Chinese Indians Others Total y Industrv descriotions Constructionsitepreparation 6010 Railways 6021 Otherscheduledpassengerlandtransport 6023 Freighttransportbyroad 6110 Sea& coastalwater transport 6112 6210 Scheduledairtransport 6220 Nonscheduledair transport 0 0 0 1 1 6301 Cargohandling 2 3 1 0 6 6302 Storage&warehousing 0 1 0 0 1 6303 Othersupportingtransportactivities 1 0 0 0 1 6309 Activities ofothertransportagencies 15 4 0 1 20 6412 Courierservices 1 1 0 1 3 6420 Telecommunications 3 4 0 0 7 7112 Rentingofwatertransportequipment 1 0 0 0 1 7121 Rentingof agriculturalmachinelyandequipment 1 0 0 0 1 7123 Rentingofoffice machinery&equipment 1 0 0 0 1 7210 Hardwareconsultancy 1 1 1 1 4 7220 Softwarepublishing, consultancy& supply 4 8 0 3 12 7230 Dataprocessing 1 0 0 0 1 7240 Dataactivities&on-line distributionof electroniccontent 1 0 0 0 1 7250 Maintenance&repair ofoffice, accounting& computingmachinely 0 1 0 0 1 7412 Accounting,book-keeping& auditingactivities; tax consultancy 4 13 0 3 2c 7421 Architectural& engineeringactivities&related technicalconsultancy 33 30 1 4 65 7422 Technicaltesting& analysis 0 0 0 1 1 7430 Advertising 6 3 0 5 11 0 4 0 01 I 1 ~ Total 95 97 3 25 22c Appendix Table 2A.5: Majority Domestic Equity Malays Chinese Indians Others Total 0 0% 76 84 1 17 178 .I 30% or less 7 5 1 4 17 .-M morethan 30% to 50% 8 4 0 1 13 More than 50% ELu 3 0 1 1 5 Total 94 93 3 23 213 179 Appendix- Chapter 2 Performance of the Firms Table 2A.6 presents the average annual sales o f the firms in the sample by industry and year. With an average value o f sales o f more than 7 billion ringgit, firms inthe airlines industry(6210) are far larger than an average firm in any other industry. Firms in the telecommunications industry (6420) are also relatively large. On the other hand, it i s the data processing industry that has registered the most impressive growth records (7230), with a growth rate o f 115 percent over the past three years, followed by the web-based firms (7240) and the software relatedindustry(7220). AppendixTable 2A.6: Average Annual Sales by Industry andYears, 1999-2001 ISIC Total Sales (Millions of RM$) Growth ('10) Code Industrv descriotions Average! 1999 ZOO0 ZOO1 1999-2001 4510 Constructionsitepreparation 1.081 1.16 0.97 1.12 -4.2 6010 Railways 236.27 233.00 233.80 242.00 3.8 6021 Other scheduledpassenger landtransport 86.18 71.15 88.01 99.39 33.4 6023 Freight transportbyroad 25.07 23.53 26.16 25.53 8.2 6110 Sea & coastal water transport 409.43 421.90 417.90 388.50 -8.2 6112 100.88 78.23 91.31 133.10 53.1 6210 Scheduledair transport 7586.33 7152.00 7888.00 7719.00 7.6 6220 Nonscheduledair transport 59.37 47.85 61.22 69.03 36.6 6301 Cargohandling 19.06 17.99 21.00 18.20 1.1 6302 Storage & warehousing 4.16 4.60 4.24 3.66 -22.8 6303 Other supportingtransportactivities 175.47 171.40 145.60 209.40 20.0 6309 Activities of other transportagencies 14.81 14.75 16.06 13.62 -7.9 6412 Courier services 98.49 79.52 97.25 118.70 40.1 6420 Telecommunications 1112.53 947.60 1121.00 1269.00 29.2 7112 Rentingof water transportequipment 7121 Rentingof agriculturalmachineryand equipment 6.47 9.32 5.12 4.99 -62.4 7123 Rentingof office machinery & equipment 3.01 3.14 2.89 3.00 -4.5 7210 Hardwareconsultancy 26.19 24.16 22.46 31.95 27.9 7220 Softwarepublishing, consultancy & supply 71.36 56.01 46.66 111.40 68.8 7230 Dataprocessing 43.33 22.82 35.18 71.98 114.9 7240 Dataactivities & on-line distributionof electronic content 20.45 11.61 20.62 29.13 92.0 7250 Maintenance& repairof office, accounting & computingmachinery 2.33 2.34 2.40 2.26 -3.7 ,7412 Accounting,book-keeping& auditingactivities;tax consultancy 27.65 28.47 28.40 26.07 -8.8 7421 Architectural & engineeringactivities & relatedtechnical consultanc 7.73 7.26 6.82 9.09 22.4 7422 Technical testing & analysis 4.38 3.84 4.36 4.93 24.9 7430 Advertising 46.40 40.93 49.77 48.48 16.9 8322 1.001 0.95 0.99 1.061 10.6 Average 391.901 364.44 401.47 409.791 11.7 Owing to the lack o f other measurements, we use labor productivity, which i s the value o f sales per worker, to quantify firms' perf~rmance.~Without controlling for assets per worker, the most productive firms are inthe transport and telecommunications industries, with total sales per worker more than FW$2 million per year. In addition to the firms in the transport industries, IT related and architectural firms have the highest growth interms o f labor productivity. Overall, the service sector as a whole has an impressive labor productivity growth o f 23.8 percent from 1999 to 2001, as shown inTable 2A.7. We will control for assetsper worker inthe regression to contain the influence of capital stock on the labor productivity o fthe firms. 180 Appendix Chapter2 - Appendix Table 2A.7: Labor Productivity Growth, 1999-2001 Total S ides Per Worker (RMVOOO) Growth(YO) Average 1999 2000 2001 1999-2001 44.86 52.90 38.77 42.91 -20.9 46.46 44.80 46.18 48.40 7.7 66.1: 50.43 66.52 81.44 47.9 16023 Freighttransportbyroad 177.45 161.60 188.44 182.42 12.1 6110 Sea & coastalwater transport 1361.62 1406.25 1257.53 1421.07 1.o 6112 2894.62 2114.36 2767.10 3802.41 58.7 ~ 6210 Scheduled air transport 393.95 377.86 406.11 398.01 5.2 6220 Nonscheduledair transport 904.18 834.37 949.12 929.05 10.7 6301 Cargohandling 145.86 145.77 162.26 129.55 -11.8 6302 Storage & warehousing 55.75 55.37 58.89 53.0C -4.4 6303 Other supporting transport activities 152.82 166.80 153.94 137.75 -19.1 6309 Activities of other transportagencies 467.85 502.08 484.03 417.57 -18.4 6412 Courier services 200.41 166.09 201.27 233.88 34.2 6420 Telecommunications 2243.36 1714.65 2355.87 2659.56 43.9 7112 Rentingof water transportequipment 7121 Rentingof agriculturalmachineryandequipment 171.42 227.29 136.92 150.05 -41.5 7123 Rentingof office machinery &equipment 35.01 31.36 36.18 37.45 17.8 7210 Hardwareconsultancy 229.0; 227.69 208.75 250.6C 9.6 7220 Softwarepublishing,consultancy & supply 1317.7; 1627.61 765.19 1560.51 -4.2 7230 Dataprocessing 52.36 58.19 30.77 68.1; 15.8 7240 Dataactivities & on-linedistributionof electronic content 189.58 134.95 206.23 227.55 52.2 7250 Maintenance&repair of office, accounting & computing machinery 140.48 146.27 114.07 161.05 9.7 7412 Accounting, book-keeping& auditingactivities; tax consultancy 73.0; 72.49 72.47 74.11 2.2 7421 Architectural & engineeringactivities & relatedtechnical consultanc 186.0; 154.49 141.17 262.4( 53.0 7422 Technicaltesting& analysis 104.15 91.42 103.85 117.3; 24.9 7430 Advertising 676.42 597.16 687.77 744.5; 22.1 65.3; 69.92 64.07 61.96 -12.1 Average 476.7; 432.01 450.13 548.12 23.8 By regressing sales per worker on the assets per worker of firms, we obtain the conditional labor productivity, which i s orthogonal to the movement of assets per worker. Table 2A.8 presents the top 11 industriesplus accounting industry, according to conditional labor productivity. The telecommunications industrystands out as the mostproductiveindustry,followedbythe software publishingindustry. Finally, we look at the export orientation o f the firms inthe sample as shown inTable 2A.9. An average firm in the sample exports 14.4 percent o f its sales. However, most o f the exporting firms are concentrated inthe transportation and IT relatedindustries. 181 Appendix -Chapter 2 Appendix Table 2A.8: ConditionalLabor Productivityfor Top Industries Conditional 743 Advertising 677961 592104 370 630 Activities o f other transport agencies 621 Scheduled air transport 393994 202585 8263 641 Courier services 200413 188281 622 Nonscheduled air transport 904182 185451 31 721 Hardware consultancy 230925 169131 2 Note: Conditional sales per worker = Sales per worker - 0.23"Assets per worker when we regress sales per worker on assetsper worker of the firms. All values are inMalaysianRinggit. 182 Appendix Chapter 2 - Appendix Table 2A.9: ExportOrientationfor Industries, 1999-2001 SIC .- . Percentage of Sales for Export Aversae 1999 2000 200; 601ORailways 0.0 0.0 0.0 0. 602lother scheduledpassenger land transport 0.0 0.0 0.0 0. 6023Freighttransport byroad 7.8 8.1 7.7 7. 611@ea &coastalwater transport 33.1 34.1 32.3 33. 6112 72.7 66.0 71.0 81. 621()Scheduledair transport 50.0 50.0 50.0 50. 622jNonscheduled air transport 12.6 11.0 6.7 20. 6301Cargo handling 0.0 0.0 0.0 0. 6302Storage& warehousing 0.0 0.0 0.0 0. 63030thersupportingtransport activities 49.0 50.0 48.5 48. 6309Activities of other transportagencies 34.2 34.7 32.9 35 6412Courierservices 32.5 32.5 32.5 32 642()Telecommunications 2.2 7112Rentingofwater 3.8 2.4 6. transport equipment 7121Rentingof agricultural machineryand equipment 0.0 0.0 0.0 0. 7123Rentingof office machinery& equipment 0.0 721()Hardware consultancy 0.0 0.0 0. 8.8 7.5 3.0 15 722@oftware publishing,consultancy& supply 8.5 9.5 7.3 8. 7230Dataprocessing 26.5 7240Dataactivities 28.7 26.0 33 &on-linedistributionofelectronic content 0.0 0.0 0.0 0. 7250Maintenance & repair of office, accounting& computingmachin 26.0 7412Accounting, y 15.7 4.0 17 book-keeping& auditingactivities; tax consultancy 0.6 0.4 7421Architectural 0.7 0. &engineeringactivities&relatedtechnical consul1 ICY 0.9 1.o 0.9 0. 7422Technical testing& analysis 0.0 743Wdvertising 0.0 0.0 0 1.2 0.3 0.4 2 8333. 0.0 0.0 0.0 0.C Average 14.4 14.3 13.0 15.7 183 Appendix-Chapter2 B. Value Added Per Worker by Services IndustriesandEconomicSectors, 1984-2000. Table 2A.10 and Figure 2A.1present value added for worker for selected service industries, while Table 2A.11and Figure2A.2 compare value addedper worker inservices and inmanufacturing. Appendix Table 2A.10: Value Added Per Worker in Selected Services, 1984-2000 Value Added Per Worker in Selected Services (1997 Ringgit) Transportation, Wholesale and Retail Finance, Insurance, Storage, and Trade, Hotels, and Real Estate, and Communication Restaurants Business Services 1984 14,267 8,071 25,229 1985 14,859 7,534 25,606 1986 15,497 6,322 24,953 1987 15,965 6,313 26,681 1988 16,898 6,529 28,771 1989 17,406 6,719 26,763 1990 18,169 7,231 30,074 1991 19,360 8,152 31,301 1992 19,880 8,923 32,147 1993 20,119 9,752 32,107 1994 21,229 10,386 33,247 1995 24,644 10,784 35,573 1996 24,235 10,317 35,983 1997 24,876 10,957 36,313 1998 24,896 10,338 37,406 1999 26,804 10,647 35,902 2000 28,600 10,455 38,071 Source:Buku TahtlnanPerangkaanMahysiQearbook of Statistics Malaysia) 184 Appendix - Chapter 2 Appendix Figure2A.1: Value Added Per Worker for Selected Services Industries, 1984-2000 (in1978 Ringgit) 40,000 35,000 - 30,000 - - 25,000- 4 5,0004 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 +Transport, Storage, & Communicatron +Wholesale & Retall Trade, Hotels, & Restaurants -A-Fmance, Insurance, Real Estate, &Business Services Appendix Table 2A.11 Value Added Per Worker in Services and Manufacturing, 1984-2000 ValuxkLMPerWorker (1997 Ringgit) ServicesManufacturin811Sectas 1984 9,684 13,325 10,376 1985 9,586 13,167 10,161 1986 9,187 14,074 10,139 1987 9,402 14,919 10,349 1988 9,810 15,950 10,892 1989 9,882 15,751 11,332 1990 10,356 16,009 11,865 1991 11,108 16,535 12,502 1992 11,727 16,387 13,087 1993 12,289 17,408 13,604 1994 13,083 18,414 14,465 1995 13,853 22,348 15,732 1996 13,607 23,369 15,551 1997 14,179 25,104 16,417 1998 14,152 22,890 15,229 1999 14,252 24,894 15,719 2000 14,266 28,198 16,140 Source:Buku TahunanPerangkaan Mahyfia (Yearbook of Statistics Malaysia). 185 Appendix Chapter 2 - Appendix Figure2A.2: Value Added Per Worker in Services and Manufacturing, 1984-2000 (in 1978 Ringgit) 30,000 25,000 20,000 15,000 10,000 5,000 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 +Services +Manufacturing +All Sectors C.RegressionModel There are many hypotheses in the literature regarding the driving forces behind firms' performance, which couldbe summarized inthe following listing: Competition enhances efficiency - entry and exports promote productivity gains (Bernard and Jensen [19991) 0 Foreign investment enhances technology transfer - restrictions on foreign ownership deter productivity gains (Blomstrom and Sjoholm [19991, Javorick and Spatareanu[2003]) 0 Foreign firms or new entrants steal market shares fkom existing domestic firms - restrictions on foreign ownership or entry may helpprotect domestic firms (Aitkenand Harrison [1999]) 0 Physical and human capital investment enhance labor productivity (Mankiw, Romer and Weil [1992]). We checkthe above hypothesesusingaregression analysis usingthe sample of service industryfirms we have using the following framework. Assuming that the production function o f the firm could be presented by a constant returns to scale Cobb-Douglas function o f labor and capital input, with a capital elasticity, aK,and a Hick's neutraltechnology level, A: 186 Appendix-Chapter2 -=Ait(-) Yit Kit aK Lit Lit In each period t, we do not observe firm i's output quantity, Y, but only its total revenue, we could nevertheless try to infer the productivity o f firm i: In-P.Y. =lnPit +InA, +aKIn7 Kit Lit L i t Let the price level o f firm ifollow the structure below, which i s a summation o f industry and regional fixed effects and a classical regression errors: lnPit =aj+a,+zit. Then the log level o frevenue per worker could be presentedby fixed effect regressions: InPitYit =ai a, InA, + + +aK In7 Kit +tit, L i t L i t To study the hypotheses that human capital, entry, foreign ownership, export orientation and the training o f workers matter for firms' productivity, we assume that these factors affect productivity in a linear fashion, so that the log level o f firmproductivity i s a linear function o f such characteristics: lnA, =a+PHhuman-capital, +PEXexport-share, +P,foreign-equity, +P,training, +PLloani +PENentrantit +uit Together with the industry and region fixed effects, we could test the hypotheses that export orientation, foreign equity, training, access to loan and entry matter to the firms' performance. Assuming that vit = zit +uit is a well behaved classical regression error, then: l P.Yi,i=a aj+a, aKIn-Kit P,human_capital, n + + + +PEXexport-share, Lit +p,foreign Lit -equity, +p,trainingi +PLloani +PENentrantit +tit. Ifexporting firms, or foreign firms, or firms that provide training, or firms that have access to loans, or firms that are new enhnts, are the better firms, then we expect PEX,PF,PT,PL,PEN to be positive. But since there are competinghypotheses about whether a foreign presence or new entrant helps or hurts domestic firms, the signs and magnitudes o f the coefficients are a priori an empirical question. In addition, given that many industries in Malaysia are subject to the 30 percent foreign ownership restriction, we further refine the variable foreign equity into the following categories, each which i s representedby a dummy variable: 187 Appendix - Chapter 2 foreign -equity, E (0,301 W e expected the constrained firms, which are those that have foreign equity that i s less than or equal to 30 percent to be less productive, which the unconstrained firms, especially those that have majority control, (that is, foreign equity of more than 50 percent) to be more productive. We test this hypothesis by comparing the estimated coefficients from the individual dummy variables. Given that many o f the variables we are interested in are firm specific and time invariant, we could only test the hypotheses using random effect estimation, which takes into account both the within and between variations o f the variables. The result of the regression estimation i s presented in Table 10, where column (1) shows the random effect panel FGLS estimation. Finally, as a robustness check, the result of a cross-section regression based on data in 2001 i s presented in column (2) of Table 10. It i s clear that these results are qualitatively very similar to that o f the random effect panel estimates, but given that we only explore cross-section variations o f the variables, not all coefficients are precisely estimated when compared to the panelregressionresult incolumn (1). The second part o f our regression analysis i s to further investigate the effects o f entry competition and the spillover o f foreign firms on firm's productivity inthe sample. Specifically, to test the policy questions on the positive spillover effect of foreign firms and the productivity enhancing effect o f entry competition, we ran the following regression: In-=ai PitYit +a, In-+y,,Kit export-share, +y,foreip-presence, Lit +y,entry-Competition, Lit +uit where foreign presence i s the share o f industry sales that originated from firms that have some positive foreign equity, weighted by the share of foreign equity in the firms. Variable entry competition i s the share o f industry sales that originated from new entrants to the industry. In controlling for firm fixed effects, capital intensity and export share, if there is any positive spillover from the presence of foreign firms in the industry through networking, information sharing or technology transfer, we expect yF to be positive and significant. Similarly, ifentry competition enhances productivity, we expect yENto be positive. ~~ Insome specifications not reported, we also include the age of the firms as a control variable; this didnot change the qualitative results o f the current specification. 188 Appendix - Chapter 2 D. Trade Performance of ServiceIndustries' Export of Malaysia's service has beenboomingin the past 20 years. In1981, the total value of private commercial services exports in Malaysia i s about US$1.2 billion. The volume swells more than 10 times and approaches US$14.3 billion in 2001. The average growth rate o f Malaysia's service export i s about 12.3 percent per annum, second only to Ireland, which has the fastest growth inthe sample countries with an impressive 14.3 percent per year (see Figure 2A.3). Appendix Figure 2A.3 Value of Service Export Selected Countries, 1981 2001 - 45 40 $ 35 3 30 % .-=25 HKG =0 20 15 --+e-- IRL 3 10 +KOR 5 0 +SGP -THA Source: World DevelopmentIndicators(2003). Import of services into Malaysia i s growing as well. Overall, the sector i s still running a narrowing trade deficit. On the other hand, the value o f commercial services imports o f Malaysia increases from US$2.8 billion in 1981 to US$16.5 billion in 2001. The average annual growth rate o f services imports during this period i s 8.9 percent, as shown in Figure 2A.4. While the growth rate o f services imports i s less than that o f exports, Malaysia i s still running a US$2.2 billion trade deficit in commercial services, which i s slightly over 1percent o f the total value added o f the service sector. On the other hand, Hong Kong and Singapore in 2001 are runninga trade surplus in commercial services of US$17.1 billion and USS5.8 billion, respectively. Unlike the manufacturing sector, there is no easily available time-series international data source o n the performance o f different service industries indifferent countries. Inorder to have cross-industry, cross-country and time series comparisons of the service sector, we rely o n the following international finance data which may still provide us with some bigpicture impressions. 189 Appendix -Chapter2 Appendix Figure2A.4 Value of Service Import SelectedCountries, 1981 2001 - 40 +MYS 3 35 30 +FrN cr 25 HKG O c 20 .I 0 15 -h.- .-a I 10 +KOR 5 0 -e-SGP +THA Source: World Development Indicators(2003). Among the three mainindustrygroups in services, the importanceof transport services has been declining. We further break down the private commercial services sector into three main industry groups: transport services, travel services, and other business services. Figure 2A.5 presents the share of transport services in the services exports of the selected countries. It i s clear that while this i s a major industryinmost countries, its importancehas been declining over time inMalaysia. In1981, the share of transport services inservice exports i s about 40 percent; it i s only slightlyhigher than 21 percent in2000. AppendixFigure2A.5 Share of Transport in Commercial Service Exports 1 Selected Countries, 1981 -2001 70 60 YO 40 30 20 ! l o lo 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Source:World Development Indicators (2003). 190 Appendix Chapter 2 - Figure 2A.6 presents the share of transport services in commercial services imports of the selected countries.6 Similar to the trends in export shares, the share of transport services in commercial services imports for Malaysia has been declining, fiom a peak of 50 percent in 1981 to 35 percent in 2001, with a slight recovery since 1997. Such a trend i s not unique to Malaysia. InIreland, the transport sector went through a dramatic decline, from46 percent in 1981to 7 percent in2001, Overall, this sector r;i i s still runninga trade deficit of close to US$3 billion in2001. Appendix Figure2A.6 I Share of Transport inCommercial Service Imports Selected Countries, 1981 2001 - 70 60 50 40 30 20 10 0 +THA Source: World Development Indicators (2003). Tourism hasbecome a mainforeign exchange earner for the economy. Figure 2A.7 presents the share of travel services in the commercial services exports of selected c~untries.~ For most countries the share has been relatively stable. The share of travel services inMalaysia's service exports hit a low of 20 percent in 1998 during the financial crisis, but it has since recovered to nearly 50 percent of the commercial services exports of the country.8 Currently, only Mexico, the Philippines and Thailand rely more on travel services than Malaysia inthe sample counties. 6 Transport services (percent of commercial services exports) covers all transport services (sea, air, land, internal waterway, space, and pipeline) performed by residents o f one economy for those o f another and involving the carriage o f passengers, movement of goods (freight), rental of carriers with crew, and related support and auxiliary services. Excluded are freight insurance, which i s included in insurance services; goods procured in ports by nonresident carriers and repairs o f transport equipment, which are included in goods; repairs o f railway facilities, harbors, and airfield facilities, which are included in construction services; and rental of carriers without crew, which i s includedinother services. 7 Travel services (percent o f commercial service exports) covers goods and services acquired from an economy by travelers in that economy for their own use during visits of less than one year for business or personal purposes. Travel services include the goods and services consumed by travelers, such as lodging and meals and transport (withinthe economy visited). Regarding the rise inthe tourism share inMalaysia from 20 percent to 50 percent, one reason could be due to the nearly 50 percent depreciation o f the ringgit after the 1997198 currency crisis which made Malaysia a cheaper place to visit, especially for tourists from China, who have since become one o f the major groups to visit Malaysia. 191 Appendix- Chapter 2 Appendix Figure 2A.l Share of Travel in Commercial Services Exports Selected Countries, 1981 - 2001 80 +MYS 70 60 +FIN 50 IRL % 40 --X--- JPN 30 +KOR 20 +SGP 10 +THA 0 1 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Source: World DevelopmentIndicators(2003). Figure 2A.8 presents the share o f travel services in the import o f commercial services for the selected countries. It i s evident that the share has been declining inMalaysia, fkom a historic peak of 36 percent in 1987 to 16 percent in 2001. Overall this industry has been cumulating trade surpluses, which in2001are about US$4.2billion. Appendix Figure 2A.8 1 Share of Travel in Commercial Services Imports Selected Countries, 1981 2001 - 40 +MYS 35 30 +FIN 25 IRL. 20 --e JPN 15 +I+KOR 10 +SGP 5 +THA 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 ~ Source:World Development Indicators (2003). 192 Appendix -Chapter2 Figure 2A.9 presents the share o f business services in the commercial services exports o f the selected countries. Figure 2A.10 shows the share o f business services in the imports o f commercial services for the selected countries. AppendixFigure2A.9 Share of Business Services in Commercial Services Exports Selected Countries, 1981 - 2001 90 +MYS 80 70 +FIN 60 I F U --$+ JPN 40 30 +KOR 20 +SGP 10 +THA 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Source:World Development Indicators (2003). AppendixFigure2A.10 Selected Countries, 1981 2001 - ~ 90 80 +MYS 70 +FIN 60 IRL 50 -.+dr- 40 JPN 30 ++KOR 20 -o- SGP 10 +THA 0 I 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 I Source:World Development Indicators (2003). 193 Appendix Chapter2 - E.SummaryofDirectandIndirectTax IncentivesbySectors 194 Appendix Chapter 2 - ManufacturingSector (4) Pioneer Status or Investment tax Allowance (b)IndustrialAdjustment Allowance (c) InfrastructureAllowance (d) Reinvestment Allowance (e) Industrial BuildingAllowance (f)Double deductiononexpensesfor the promotionofexports (8) Double deduction for promotion o fMalaysianbrandnames (h)Doubledeductionincentivefor researchanddevelopment 6) Double deductionincentives on freight charges (4) Doublededuction incentives for training (k)Doubledeductiononexport credit insurance premium (1) Single deductiononpre-operating training expenses (m) Single deduction on expenses incurred for obtaining quality systems and halal certification and accreditation (n)Tax exemptiononthe value ofincreasedexports (0)Accelerated Capital Allowance -- use o f environmentalprotection equipment - upon use o f computers andinformation technology assets expiry o freinvestment allowance only for companies that reinvest inpromotedproducts (p) Dutyexemptiononmachinery, equipment, raw materials & components for: --- research producing manufactured goods activity -training activity environmentalprotection Source: http://www.us-asean,org/rvlalaysia/business-guide/ 195 Appendix- Chapter 2 196 Appendix-Chapter 2 F.May2003 EconomicStimulusPackage Appendix Box2A.1 Source:The Star Online, The Star, May 22,2003, available at : c h i 2 = 0 . 0 0 0 0 Log l i k e l i h o o d = -5324.823 Pseudo R2 = 0.0330 z and P S I a r e t h e t e s t of t h e u n d e r l y i n g c o e f f i c i e n t being 0 ~ ~ 2 ) . dprobit w i 9 educ educsq exper expersq train trainprev vocatrain,robust P r o b i t e s t i m a t e s Number o f obs = 6609 Wald c h i 2 (7) = 373.61 P r o b > c h i 2 = 0.0000 Log l i k e l i h o o d = -4103.4524 Pseudo R2 = 0.0467 educ .0306319 .0094469 3 . 2 5 0 . 0 0 1 10.3969 .012116 .049148 e d u c s q -.0001681 .0004463 - 0 . 3 8 0 . 7 0 6 1 1 7 . 8 3 7 -.001043 .000707 exper .0045478 ,0017987 2.53 0 . 0 1 1 17.8133 .001022 .008073 e x p e r s q -.0001256 .0000383 - 3 . 2 8 0 . 0 0 1 4 4 3 . 4 9 1 - . 0 0 0 2 0 1 -.000051 train* .0295045 .0140231 2 . 0 9 0 . 0 3 7 .316689 .00202 .056989 trainp-v* .0579468 .0164999 3 . 4 3 0 . 0 0 1 .182932 .025608 .090286 vocatr-a* .0932263 ,0172766 5 . 1 3 0.000 .144954 .059365 .127088 205 Appendix-Chapter 3 Appendix Table 3A.7: Estimation of a Skills Supply Function: Probit Estimates Controllingfor Establishment Characteristics ..dprobit wi9 educ educsqexper expersqtenuretrain trainprev vocatrain emplsizeexporter forgneqt eq5_less,robust P r o b i t e s t i m a t e s Number of obs = 2377 Wald c h i 2 (12) = 88.04 P r o b > c h i 2 = 0.0000 Log likelihood = -1349.9829 Pseudo R2 = 0.0339 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II R o b u s t w i g dF/dx S t d . E r r . z P s l z l x-bar [: 95% C . I . 1 - - - - - - - - - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - educ .0123529 .0152525 0 . 8 1 0 . 4 1 8 10.979 -.017542 .042247 educsq .0003082 .0006847 0.45 0.653 129.633 -.001034 .00165 e x p e r .0042411 .00297 1.43 0.153 16.4809 -.00158 .010062 e x p e r s q -.0000799 .0000655 -1.22 0.223 3 7 6 . 8 1 1 -.000208 .000048 t e n u r e IIIII -.0033036 .0018024 -1.83 0.067 5.61759 -.006836 .000229 t r a i n * [ -.0029822 .0211972 -0.14 0 . 8 8 8 .414809 -.044528 .038564 trainp-v*l .OS3999 .0230145 2.27 0.023 .231384 . o o a a g i .099107 vocatr-n*l .0812385 .0238868 3.18 0 . 0 0 1 .169962 ,128056 emplsize I .0000529 .0000194 2 . 7 2 0 . 0 0 7 270.584 .. ooowo1o2i s1 .oooogi exporter*l .os46658 .0242983 2 . 3 1 0 . 0 2 1 .79974a .007042 . i o 2 2 9 f o r g n e q t -.0000477 .0002345 - 0 . 2 0 0.839 5 5 . 5 9 -.000507 .000412 eq5-less II -.0002391 .0003057 - 0 . 7 8 0.434 37.3589 -.000838 .00036 - - - - - - - - - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - obs. P .725284 p r e d . P II .7332936 (at x-bar) 206 Appendix-Chapter 3 Appendix Table 3A.8: Skills Content of Education . dprobit w i 9 educ age P r o b i t e s t i m a t e s Number o f o b s = 8396 LR c h i 2 (2) = 366.22 P r o b > c h i 2 = 0.0000 Log likelihood = -5315.7695 Pseudo R2 = 0.0333 . dprobit w i 9 educ agedu P r o b i t e s t i m a t e s Number of obs = 8396 LR c h i 2 ( 2 ) = 366.75 P r o b > c h i 2 = 0.0000 Log l i k e l i h o o d = -5315.505 Pseudo R2 = 0.0333 207 Appendix-Chapter 3 Appendix Table 3A.9: RegressionAnalysis of Quitting and Training . x i : r e g q u i t 2 tif t o f age kpc i s 0 i i i 1 7 1 - i i i 1 7 1 3 > Inwage2 foreign l n s i z e i . i s i c 2 i . r e g i o n ; i.ic2 -Iisic2-15-36 ( n a t u r a l l y coded; - 1 i s i c 2 - 1 5 o m i t t e d ) iregion .i s -Iregion-1-6 ( n a t u r a l l y coded; -1region-1 o m i t t e d ) Number of obs = 596 F ( 34, 561) = 2 . 7 9 P r o b > F = 0 . 0 0 0 0 R- s q u a r e d = 0.1444 Ad] R - s q u a r e d = 0.0926 R o o t MSE = .24864 - - - - - - - - - - - - - + - - - - - - - - - - - - - - - - - - - - - - - ------ ........................ t i f -.0338108 .0287305 - 1 . 1 8 0.240 - .0902434 .0226217 t o f .oaa3833 . 0 2 7 4 1 3 i 3.22 0.001 .0345385 .1422281 age -.0031046 .0011447 - 2 . 7 1 0.007 -.005353 -.0008562 kpc , 0 4 9 2 4 0 1 . 0 2 3 7 0 6 1 2.08 0 . 0 3 8 .0026766 .0958036 i s 0 .0310973 0 . 6 1 0.542 ---..0980861 0420865 .0800762 i i i l 7 1 -..0189948 0394878 .0298332 -1.32 0.186 .0191106 i i i 1 7 2 -.008584 .0285146 -0.30 0.763 .0645925 .0474245 i i i 1 7 3 .0703073 .0350409 2 . 0 1 0.045 .0014799 .1391347 i i i 1 7 4 .0485016 .0287633 1 . 6 9 0.092 0079953 . l o 4 9 9 8 5 i i i 1 7 5 .0281654 .0298199 0.94 0.345 0304069 .0867377 i i i 1 7 6 -.0449977 .027372 - 1 . 6 4 0 . 1 0 1 .0987618 .0087663 i i i 1 7 7 .0478235 .0280178 1 . 7 1 0. 088 . l o 2 8 5 6 1 i i i 1 7 8 .0124161 .0347331 0.36 0 . 7 2 1 -----....0072091 0558067 .080639 i i i 1 7 9 .1329722 .0510559 2 . 6 0 0.009 .0326882 .2332563 i i i 1 7 1 0 - . l o 9 9 8 7 4 .o489165 - 2 . 2 5 0.025 -.2060692 -.0139057 i i i l 7 1 1 ,1071364 ,0438893 2 . 4 4 0 . 0 1 5 .020929 .1933438 i i i 1 7 1 2 -.0667765 .0377928 - 1 . 7 7 0 . 0 7 8 - .1410092 .0074562 i i i 1 7 1 3 -.0289449 .0398939 - 0 . 7 3 0.468 -.1073045 .0494147 lnwage2 -.0219015 .0163618 - 1 . 3 4 0 . 1 8 1 -.0540393 .0102363 f o r e i g n .0102006 .0247015 0 . 4 1 0.680 - . 0383181 .0587193 -Iisic2-17 I l n s i z e IIIIIIIIIIIIIIIIIIIII -.0344005 . 0 1 1 6 4 1 -2.96 0.003 -.0572658 -.0115352 - .0324593 -Iisic2-18 I .0870026 .0608195 1 . 4 3 0.153 -.1042966 .2064644 -Iisic2-24 I -.0243014 .0407265 -0.60 0 . 5 5 1 .0556937 -Iisic2-25 I -.0706433 .0637869 -1.11 0.269 - .1959336 .054647 -Iisic2-29 I .0721673 .0318369 2.27 0.024 .1347014 .0201394 .0447262 0 . 4 5 0.653 - ..0096333 0677118 . l o 7 9 9 0 6 -Iisic2-32 -Iisic2-34 I I .0982919 .0475494 2 . 0 7 0 . 0 3 9 .0048953 .1916885 -Iisic2-36 I -.0295983 .0613823 - 0 . 4 8 0 . 6 3 0 - .1501654 -.0858278 .0909689 -Iregion-2 I -.0016009 .042881 - 0 . 0 4 0 . 9 7 0 -.0887253 .082626 -Iregion-3 I -.0292232 .0302932 - 0 . 9 6 0.335 -.0541356 .0302788 -Iregion-4 I .0001529 .027639 0 . 0 1 0 . 9 9 6 .0544415 -Iregion-5 I -.1590577 .0717853 - 2 . 2 2 0.027 -.3000585 -.0180569 -.2626516 -Iregion-6 I -.1307074 .0671744 -1.95 0.052 .0012367 -.0671314 .0552518 - 1 . 2 2 0.225 - .1 7 5 6 5 7 1 .0413943 -cons I .5869184 .172386 3 . 4 0 0 . 0 0 1 .2483175 .9255194 208 Appendix-Chapter3 Appendix Table 3A.10: Number of EmployersRegisteredat the HRDFby State (January-December 2002) Manufacturing Services Number of Percentage Number of Percentage registered firms registered firms Perlis I I 1 0.4 Terrenganu 11 2.5 6 2.2 Kelantan 4 0.9 2 0.7 Sabah 28 6.2 11 4.0 Labuan 1 0.2 0 0 I Total Sarawak 20 4.5 17 6.2 448 100 274 100 209 Appendix Chapter 4 - Appendix 4 Chapter 4 - Appendix Table 4a.l: GlobalCompetitivenessReportRankings Growth ComDetitiveness Index Microeconomic Combetitiveness Index 2002 2001 2002 2001 Malaysia 27 30 26 37 United States 1 2 1 2 Finland 2 1 2 1 Taiwan 3 7 16 21 Chile 20 27 31 29 Korea 21 23 23 26 Thailand 31 33 35 38 China 33 39 38 43 Brazil 46 44 33 30 India 48 57 37 36 Sozlrce: World Economic Forum (2002-2003). N O SE TI PC TS 210 d 1 I u t