Report No. 1 6207-EGT Arab Republic of Egypt Country Economic Memorandum Working Papers Annex (In Four Volumes) Volume III March 15, 1997 World Bank Resident Mission In Egypt Middle East and North Africa Region Document of the World Bank Vice President Kemal Dervi§ Director Inder Sud Division Chief/Manager Khalid Ikram Staff Member Chang-Po Yang ANNEX A: EGYPT'S FREE ZONES ............................................... 1 ANNEX B: THE LONG TERM DEMOGRAPHIC TREND AND ITS ECONOMIC IMPLICATIONS ................................... 13 ANNEX C: GROWTH AND SAVINGS IN EGYPT: LESSONS FROM THE EMPIRICAL GROWTH LITERATURE ..... 35 ANNEX D: PENSION REFORM, CONTRACTUAL SAVING, AND CAPITAL MARKETS ............................................... 50 ANNEX E: NATURAL RESOURCE DEPLETION AND SUSTAINABLE GROWTH ............................................... 81 ANNEX A Egypt's Free Zones Introduction 1. This report on Egypt's Free Zones is intended to review briefly the current status of Free Zones in Egypt, assess the operations and management of such zones to meet their intended objectives, and recommend changes that could help such zones meet full potential. Recommendations made in this report will be specific, practical and "how to" in nature. 2. The preparation of this report has been based on several sources of information including a two-week field visit to Egypt. In the field, a desk review of: (i) Egypt's investment laws and regulations; and (ii) relevant documents on the problems and impediments to would-be exporters, was carried out. The preparation of this report has also been benefited from information obtained during the visit to Free Zones, and meetings with government officials, private investors, local and national chambers of commerce, donor representatives, and those in universities and research institutions in Egypt. In particular, the meetings with investors in the Free Zone have provided inside rationales of why they chose Egypt's Free Zones and how they are currently working out! 3. The objectives of establishing free trade zones or export processing zones (EPZ) vary from country to country. But most important of all are to: (i) create jobs to increase employment opportunities; (ii) develop trade to enhance export earnings; (iii) attract foreign capital to increase industrial investments; and (iv) transfer technologies and know-how. For the newly industrialized economies (NIE) in the Pacific Rim, the transfer of high-tech, bio-science and knowledge intensive industries and the mobilization of venture capital are more important than the job creation, because of labor shortage at home, high domestic saving rate, and large foreign reserve accumulation. While in others, the creation of job opportunities and the infusion of foreign capital could be the main objectives of establishing EPZ's. But, almost all the EPZ's were established in the past to attract investors, by providing special laws and regulations to give both substantive benefits and very simplified one-window, one-stop procedures. In Egypt, the performance evaluation of the Free Zones will be addressed according to the objectives, for which they have been established, namely, job creation to increase employment opportunities and export promotion. 4. The analytical approach used to derive the recommendations advanced for immediate actions in this report follows a very simple process. Firstly, current activities in the Free Zones, if cannot contribute significantly to the achievement of the objectives, should be suspended; secondly, efforts and resources should be directed toward those Free Zones, which have a reasonable chance to succeed; thirdly, problems or impediments to the would-be investors should be resolved; and fourthly, an aggressive promotion campaign must be mounted. In addition to the immediate action 2 Working Papers Annex: Egypt's Free Zones recommendations, this report will also venture mid-term and long-term recommendations for the Free Zones and Export Promotions. 5. This report includes six sections. Section II covers Current Status of Free Zones in Egypt; Section III, Performance of Egypt's Free Zones; Section IV, Free Zones and Export Promotion; Section V, Lessons Learnt From Other Countries; and Section VI, Recommendations for Actions. Current Status of Egypt's Free Zones 6. In Egypt, besides Free Zones, there are several special types of land allocated for industrial/residential development. These special zones include New Communities and New Industrial Zones. New Communities have been located on non-agricultural desert lands to be an independent urban centers, which were designed to attract population away from the existing larger urban centers. There are four major examples of New Communities: Tenth of Ramadan, Sixth of October, Sadat City and Borg el Arab, which are administered by the Ministry of New Communities. These Communities have focused their efforts on attracting industrial/residential developments, creating job opportunities, and attracting residents as well. To encourage investors/developers, the Communities offered subsidized land with modem infrastructure and a 10-year tax holiday for all development. 7. The New Industrial Zones have been designed exclusively for industrial purposes and located adjacent to the established urban areas, where the migrant labor force is expected to come. The New Industrial Zones are administered by the Govemorate Investment Approval Committee, chaired by the Governor, and offer the same investment incentives as the New Communities. 8. While the intents for these special zones deserve to be credited, but the old habits die hard. The performance of these special zones so far has not reached their full potential as they are still under utilized. However, in the rest of this report, unless relevant to the assessment of Free Zones, these special zones will not be discussed further. It is to be noted, however, that the Free Zones are administered by the General Authority for Investment (GAFI), under the umbrella of the Minister of Economy (Article 46, Investment Law No. 230/1989). It is conceivable that the integration and coordination of these zones, while under separate administrative responsibilities could be a problem, particularly in Egypt. 9. The creation of Egypt's Free Zones dated back to Law No. 43/1974 The System on the Investment of Arab and Foreign Funds and Free Zones. The Law No. 43 was repealed by Law No. 230/1989 Investment Law, which becomes the basic law for investment projects in Egypt in the Inland Investment and the Investment in the Free Zones. In particular, Chapter Three of the Law No. 230/1989 Investment System in the Free Zones (Article 28 through Article 45) gives detailed rules and regulations on the administration and management of investment projects in the Free Zone. The application Working Papers Annex: Egypi s Free Zones 3 procedures is simple (one page) and fast (one week) (Attachment 1.0 Application for Preliminary Approval on Free Zones Projects) and the benefits offered are substantive and attractive (Attachment 2.0 Free Zone Project Publications by GAFI). 10. Under the current Investment Law No. 230, there are three types of Free Zones: (i) Public Free Zones, (ii) Private Free Zones, and (iii) Free Zones covering an entire city. The Public Free Zones are for the establishment of such projects as may be licensed under the provision s of Law No. 230; the Private Free Zones are established and confined to one project if the nature thereof shall so require; and the establishment of a Free Zone covering an entire city shall be effected by law. 11. GAFI has provided some data on Public Free Zones in operations: Cairo Nasr City, Port Said, Alexandria, Suez, Ismailia, and Damietta. The three tables obtained from GAFI are: (i) by zone, and by project approved, in operation, and under construction, as of June 30, 1996 (Table 1.0); (ii) by type of activities (Industrial, Storage and Service) as of June 30, 1996 (Table 2.0); and (iii) annual approvals of projects by year and by zone from 1975 through 1995 (Table 3.0). 12. From these tables, several statistics emerge: (i) more than half of the projects approved are for storage and warehouses; (ii) total registered capital is about LE 5.2 billion and the estimated project cost is about LE 13.8 billion; and (iii) total jobs expected to be created are about 54,000. In addition, based on meetings with Free Zones officials and investors in Port Said, Alexandria and Cairo Nasr City, most of the industrial projects approved are labor-intensive, low-capital cost, and low-skills in textile/garment industries, employing mostly female workers. The meetings also revealed the project ownership mix: Egyptians, 55%, other Arab nationals, 8%, European, 31%, USA, 4%, and other 2%, reflecting possibly a substantial amount of domestic capital invested in the Free Zones. 13. No data, however, are available to estimate the cost per job created, nor the productivity per worker. The data provided by GAFI do not include temporary and seasonal labor, which according to officials from Port Said, could be as high as 50% of the total permanent labor. In addition, some subcontracting to local firms has been done by the textiles/garment projects in the Free Zones, indicating the existence of some backward linkage to the domestic industries to create local jobs. Performance Assessment of Egypt's Free Zones 14. Egypt's Free Zones have offered very comparative advantages no less than those offered in other countries. Not only Egypt offers simplified one-stop, one-window service coupled with substantive benefits, but also these zones are geographically located in the strategic positions within the region. Egypt does have an abundance of cheap labor, eager, able and actively seeking work. 4 Working Papers Annex: Egypt's Free Zones 15. Government officials, time and again, have indicated that the main objective of the Free Zones is job creation to increase employment. The industrial projects so far established in the Zones, particularly, textile and garment industries, have met the objective. In addition to jobs created in the Free Zones, the subcontracting arrangement also creates local domestic jobs. Although data are not available to estimate the numbers of jobs created overall, yet the visits to the project factories revealed that the number of jobs created per project could be in the thousands. 16. Egypt is in a particularly favorable position to attract apparel projects. The experiences of the current apparel project owners all have high rating on the labor force and working conditions. In addition, Egypt is not currently quota restricted by the international Multi-Fabric Arrangement (MFA). Although EPZ experiences in other countries do show that the apparel industry could someday become a sun-set industry, yet Egypt is still some years away to reach that date. 17. The service projects in the Free Zones have also created low-capital cost, labor- intensive jobs, besides offering opportunities for foreign currency earnings. According to data provided by GAFI, the number of jobs created per service project is about 100, as compared to 25 per storage project. 18. The storage and warehouse projects approved in the Zones have so far created a relatively small number of jobs as compared to, say, apparel projects. It is clear that the Government's objective to increase employment opportunities through the storage projects in the Free Zones will not be achieved. The storage projects do occupy a large space of land in the Zones. Whether they are cost effective or not is yet to be analyzed. But a large portion of land in the Zone allocated for storage/warehouses does produce negative effect to attract industrial projects, say, apparel industry. This is particularly so when the warehouses project is on a piece of land, allocated as a parking lot for cars waiting to be sold, with unknown delivery date, while the cars are gathering dust in the meantime. 19. What is needed most is to assist those Egyptian-owned industries in the Zones to sell overseas. For most industrial projects in the Zones, the foreigner-owned projects do have international market network, and produce relatively higher value products than the ones owned by the Egyptians. The Egyptian-owned projects either sell low-priced items to such discount stores as K-mart and Wal-Mart or produce mostly for domestic market. For instance, the Exact Co., which is fully owned by the Egyptians in the Alexandria Free Zones, currently produces more than 90% of its batteries for domestic market. If Egypt is to develop export and increase export earnings, assistance to these Egyptian-owned projects to penetrate into international market is urgently required. In light of the Egyptian economy consisting of about 50% small to medium scale enterprises, how to design and provide such as'sistance to them is critical for the attainment of export-led growth economy. Working Papers Annex: Egypt's Free Zones 5 20. The projects approved so far in the Free Zones have attracted relatively small amount of foreign capital. About half of the projects in the Free Zones are owned by Egyptian nationals. Presumably, these projects are funded by domestic capital. While the other half of the projects in the Zones are mostly in the labor-intensive, low-capital cost apparel projects, implying a relatively small amount of foreign capital invested in the Free Zones. How to attract foreign capital to invest in the Free Zones and thereby to induce more direct foreign investment in Egypt as a whole is a crucial factor in the growth of Egyptian economy, no matter whether the economy is an export-led growth or not. 21. Egypt's Free Zones do provide an opportunity for export earning. Gross export earnings include: (i) wages, salaries and remuneration paid to Egyptian workers in the Zones; (ii) employer's share contribution for social securities; (iii) payments for services provided in the Zones; (iv) rent and annual fee paid to GAFI; and (v) subcontract payments to domestic firms outside the Free Zones. To achieve the greatest gross export earnings, goods produced in the Free Zones including those owned by the Egyptian Zones should be exported as much as possible. 22. On the subject of know-how transfer in the Free Zones, the projects approved so far do not offer much help. The storage and service projects do not have much know how content, while the industrial projects in the Zone are mostly low-skills projects. Most of the Egyptian workers in the Free Zones are female, low-skills, repetitive and non- managerial positions, which have very little absorptive capacity for know-how transfer. Nevertheless, the subcontracting arrangement by some of the apparel projects do offer opportunities to upgrade domestic industrial standards in productivity and quality, notwithstanding still in a very limited scale. This arrangement should be encouraged and promoted to strengthen backward linkage, which, if attains its maximum potential to integrate with local economic activities, could have a growth pole effect on the neighborhood of the Free Zones. 23. The owners of projects particularly those in the apparel industry in the Free Zones have indicated overall satisfaction in choosing the projects established in the Zones. The reasons given include: (i) easy in project application; (ii) adequate support received from the Zones authorities; (iii) one-stop, one-window service; (iv) substantive tax benefits offered; (v) plenty of relatively cheap but very diligent work force, mostly female workers; and (vi) much more freedom in the hiring and disciplining of workers in the Free Zones than in Inland Projects. Some of the project owners also expressed: (i) difficulties encountered in obtaining work permits for foreign experts, which are controlled and issued by the Ministry of Labor; and (ii) relatively high shipping cost as compared to projects in, say, Dubai's Free Zones. Egypt's Free Zones do offer a very simplified one-window, one-stop procedure and a very substantial benefit package. but if the cheap labor and tax advantages cannot cover the transportation costs, then the substantive benefit will not be there to attract foreign investors to the Free Zones. And, if the work permit for foreign experts needed for the projects is difficult or time-consuming 6 Working Papers Annex: Egypt's Free Zones to obtain, then the prospective new foreign investors could be discouraged to come. Steps need to be taken to address these concerns. Egypt's Free Zones and Export Promotion 24. Egypt's export has suffered from low quality product and needs image building and enhancement. The Free Zones should be treated as a coastal gateway to the international market. By attracting foreign investors to the Free Zones, they bring with them technology, management, quality standard and marketing skills. "Made in Egypt" by these foreign investors in the Zones will be a means to build Egypt's project image abroad. The truth is that the image building is only a step, and most important of all is that the quality product with competitive price internationally is the only sure way to be able to penetrate into the international market for a chance to succeed. The Free Zones, however, could offer such an opportunity for the first step for image-building and enhancement. 25. It is important for Egypt to promote joint-venture and licensing opportunities for local firms and, through such twinning arrangement, the local firms could have on-the-job training to acquire know-how and access foreign markets. Egypt's Free Zones offers a very simplified procedure with substantive benefit package, and could be a grooming ground to promote joint-venture and licensing opportunities through these joint venture opportunities. It is hoped that the domestic industrial capacity could be upgraded to produce quality product with competitive price internationally and, thereby facilitating Egypt's export promotion programs. In Pacific Rim nations, where EPZ's are reasonably successful, almost by intention and design, any prospective foreign direct investors while visiting the countries, are always led to "tour" EPZ's to personally witness the "project in action." The host wants to impress upon the visitors with the FDI friendly environment, like a Western "Industrial Park" airlift and placed in the host country. In Egypt's case, the Coca-Cola plant and the Delta Textile Egypt, both in Nasr City Free Zone, are good examples to give such friendly impressions. Egypt's Free Zones, properly upgraded and maintained, could play such a role. 26. To improve product quality and increase productivity, Egypt needs to import appropriate technology, which, by the way, will require foreign currency. Although Egypt's Free Zones, to date, have not yet shown substantial export earning, they do provide an opportunity for foreign currency earning, which could be used for technology importation and thereby facilitating export promotion. Asian experience in the '70s did show that EPZ was indeed an effective instrument for export earning. 27. Although still in a very limited scale, some subcontracting arrangements to the local firms by the foreign investors in the Free Zones are taking place. The foreign investors insist on acceptable quality standard of goods provided by the local firms, and thereby forcing the local firms to upgrade their industrial standard and capacities. Egypt should encourage and actively promote such sub-contracting arrangements. It is a right step in the right direction to help export promotion by forcing the local firms to upgrade Working Papers Annex: Egypt 's Free Zones 7 production standard. Egypt's Free Zones by attracting foreign investors to come in and encouraging subcontracting arrangements will undoubtedly facilitate export promotion programs. Lessons Learnt From Other Countries 28. As of 1990, there were 45 countries in the world offering more than 145 EPZs, while by 1970, only 10 countries offered a total of 24 EPZs. The pace of growth in offering EPZs has slowed down, due to keen competition. Recently, however, because of government economic reforms and inefficiency in the public ownership and management of infrastructure projects, the private sector participation in EPZ development has taken a much more important role than before, particularly in Asia through the Build, Operate and Transfer (BOT) approach. Despite this new trend, past experience of other countries with EPZs is still of valuable lessons. This report recognizes its limitations of not being able to offer necessary and sufficient lessons to assure a successful operation of EPZs, but the report attempts to list briefly the commonly learnt "lessons", while keeping in mind the stage and conditions of the country's development, political, economic, social and cultural. What follows is a list of lessons from countries with EPZs: * Almost all EPZs in the world offer substantive benefit package and require a very simple one-stop, one-window application procedure; * The rules and regulations must be transparent and predictable, particularly in regards to the right and obligation of investors in the EPZ; - The office of EPZ administration should be professionally managed, competent and efficient; * Most of the successful EPZs, particularly in Asia, started with a concentrated, manageable small program and expanded the program gradually after gaining experience: Taiwan started with only one EPZ in 1966 and then added two more in 1969, only after the first EPZ was fully occupied; Dominican Republic had one EPZ in 1969 and 16 by the end of 1990; Mauritius had one in 1970 and 6 EPZs by 1990; * Most of the EPZs offered by the developing countries in the past started with labor-intensive, low-capital cost, low-skills apparel projects; * Most successful EPZs in Asia had, in the past, received strong and effective government service support in promoting EPZs to attract foreign investment; * Experience of other countries with successful EPZs, indicates that the labor force in these countries all have adequate level of literacy and numeracy to receive 8 Working Papers Annex: Egypt's Free Zones relevant skills training as needed for the effective operations of a variety of equipment/machinery; * Almost, by necessity, the successful EPZs are all located near seaports or airports, and most important of all, have access to efficient air and sea transport services; * Experience of other countries with outstanding EPZ performance all had very active joint-venture projects in the EPZs between the foreign investors and the local firms, such as those seen in Singapore, South Korea and Taiwan; As a corollary, these countries had also shown high-value added to manufactured export in the EPZs (Taiwan and South Korea as high as 60%). They also had strong backward linkage capability to continuously supply demanded quantity of parts/materials with high quality and competitive prices (i.e., South Korea's 1991 experience: 90% color T.V. parts, 25% VCR parts and 60% P.C. parts.); * The experience in Taiwan, South Korea and Singapore also shows the existence of an effective private sector export promotion agency to manage export promotion activities including a year-round international trade show center. The agency also provides technical assistance to help local firms access international market information for their specific products. The trade show center acts as a shopping mall for foreign buyers coming to town. All needed information for placing an order is under "one-roof"; * In countries with successful EPZs, there are also very active trading companies acting as a conduit to bring buyers or order to town. This is of particular importance for countries with high proportion of small-to-medium scale enterprises; * The successful EPZs around the world also keep very clean and orderly, with street names and numbers clearly shown. These EPZs are frequently visited by foreign heads of state; * Experience of countries with successful EPZs had demonstrated a foreign investor-friendly environment. At the national, local and EPZ levels, all do their best, such as visa restriction removal, and local taxi-drivers given pamphlets showing where the EPZ is located. Furthermore, the EPZ office also prepares EPZ directory on those projects in the Zones, with background companies information, giving a feeling of "you will not miss it"; and * Lastly, on the experience of private sector participation in EPZ development, it could have two different approaches: (i) public owned, but private operations and management; and (ii) private owned, private operations and management. Working Papers Annex: Egpt s Free Zones 9 Whatever the approach, the government must have a clear, transparent and predictable law and regulation encouraging private sector participation and allowing them a reasonable chance for success and profit. The government at the same time takes measures to prevent land speculations. The private sector participation in infrastructure projects such as BOT is an unstoppable trend. The government budget constraints and the public sector inefficiency in infrastructure projects are reasons for the recent active BOTs around the world. Recommendationsfor Actions Immediate Actions 29. The government officials have in many occasions stressed that the main objective for creating Free Zones in Egypt is job creation to increase employment. Among the three types of projects --- industrial, storage and service in the zones, --- the industrial and service projects, particularly apparel projects, have to some extent met the objective well. And through subcontracting to local domestic firms, some backward linkage has created, though still in a limited scale, additional jobs. Along with the creation of jobs in and out of the Free Zones, come the temporary workers. In this regard, the Egyptian experience is not much different than those in other countries with EPZs. The government should, therefore, spearhead in this direction to continue to attract labor-intensive projects to come into the Free Zones. 30. On the other hand, the desirability of continuing storage and warehouse project approvals to come into the Free Zones need a critical review. For now, there are more than half of the approved projects in the Free Zones for storage uses. Some of these storage projects are parking lots for cars waiting to be sold with unknown delivery date, while these cars are sitting on the parking lots gathering dusts. These storage projects are not labor-intensive. They do not even offer opportunity for "know-how" transfer or export promotion. Although these storage projects pay rent and a 1.0% annual fee, they sometimes occupy prime space, which might, otherwise, be of interest to prospective industrial or service project investors. The mix of, say, an apparel project, which requires a clean factory and a storage parking lot project with dusty cars next door, somehow just does not seem compatible. Most of the successful EPZs in Asia keep very clean and orderly, sometimes just like a newly designed city with street name and number. They are the "must visit" places by foreigners. In Egypt, there is no data currently available to estimate if the government is subsidizing these storage projects, because of high infrastructure cost. There is no income reason yet to continue approving storage projects. Therefore, if the government believes that the job creation for the projects to be approved in the Free Zones continues to be the highest priority, then a moratorium on storage/warehouse projects must be enacted as soon as possible. 31. Many successful EPZ starts from a small and manageable program and expands it only after gaining valuable experience. Egypt should not be an exception and should re- direct her resources, energy and time to focus on some of the existing Free Zones, which 10 Working Papers Annex: Egypt's Free Zones have a reasonable chance to success. For instance, there were seven projects approved in Ismailia's Free Zones by 1995 (see Table 3.0), but none of them, including one approved in 1989, has even become operational as of June 30, 1996. For such a Free Zone, one wonders its merit. For now, the three promising Free Zones in Egypt are: Nasr City, Alexandria and Port Said. The government should concentrate its resources and efforts to upgrade these three zones. Within the zones, consideration should be given to improve greenland, park, recreation facilities, health and medical services, zone directory, security guards, telecommunication, banks, transportation services, etc. In short, these zones must be in a "show case" condition. At the same time, the administration office should be staffed with professionals to effectively manage the Zones. 32. The Free Zones should provide options for modular design and prefabricated standard factory building either for rent or for sale to project investors for immediate occupancy, equipment installation and project start-up. Timing, when it becomes critical for investment considerations, the availability of these standardized modular factories will be of great value to investors. Foreign investors newly coming into the Zone may not know well the local construction codes and practices. The availability of these standardized factories could save time and cost. The investors will save annual rent and fee paid to GAFI during construction period. The offer and availability of such factories for choice by the investors will greatly speed up the occupancy and project operations. 33. To further assist the investors during the construction period, the Free Zones should also provide temporary storage facilities for rent, if the investors do not opt for standardized factories. The temporary storage facilities will allow the early shipment of equipment and machinery on site, and be ready for immediate installation when the factory construction is completed. The provision of this service to investors will get rid of their worrying about equipment damage, loss or theft during the construction period. 34. While concentrating time and effort to attract the desired projects into the focused Free Zones, the government should also, at the same time, remove potential obstacles and impediments to would-be investors. These include the speedy approval of work permits to foreign experts for the projects in the Free Zones, the predictable rent escalation formula, and the removal of government monopoly in shipping industry. Difficulties in obtaining work permits for foreign experts such as trainers/foreman, etc., could disrupt the smooth operations of the project, and if widely known in advance by the prospective investors, will certainly discourage more advanced projects to come to Egypt, and definitely contribute a negative image of Egypt. The current rent escalation formula is unknown. The government officials, when pressed for an answer about how future rent could be, gave no answer. This kind of uncertainty will absolutely be of no help to anyone. Furthermore, inefficient port operation, monopoly and high transportation costs, which if far exceeds the advantage of low labor cost and tax benefits, will not give any substantive advantages to prospective investors to consider having a project in the Zones (Note: USAID's June 1996 report: Egypt: Options for Increasing Market Competition in Maritime Port Services, gives detailed analysis of problems/difficulties encountered and recommendations advanced). Working Papers Annex: Egypt's Free Zones I1 35. The Free Zones need to mount an aggressive promotion campaign to attract prospective investors. In today's very competitive business world, many developing countries are offering very attractive substantive benefits and very simplified one-stop, one-window EPZs to foreign investors. Egypt's Free Zones do offer comparative benefits, but have not had an effective campaign to promote them. In today's international market, "Go get them!" is the norm of getting the business. The existing promotion related materials must be professionally updated and circulated through say, internet, international newspaper, magazines, etc. The delegation to visit prospective countries/investors should be professionally organized and pre-arranged. The invitation to prospective investors to tour the Free Zones should be guided; when needed, follow- up meetings, either individually or as a group, should be conducted. GAFI has an independent budget from rents and fees it collects and bilateral donors also expressed interest to support. So, what are Egypt's responsible officials waiting for? Egypt's Free Zones are attractive in the region in many respects including political stability. What needs to be done is an effective promotion campaign organized and carried out by professionals. Mid-term Recommendations 36. In addition to the recommendations for immediate actions listed above, several mid-term recommendations are in order: 37. There are several government agencies working for the export promotion activities, but they are mostly staffed with sub-professionals, under funded and uncoordinated. These agencies include Egyptian Export Promotion Center, General Organization for International Exhibitions and Fairs, Trade-Point, and Commercial Representations. They should be consolidated to function effectively to support export promotion. This report is of the view that certain government service support is still needed. Meanwhile, an independent private sector export promotion agency should be established. There are many such agencies in Asian countries with very good track records, which could be reviewed and adopted. In today's Egypt, export promotion is almost non-existent! 38. The Trading Company Law, currently being drafted and reviewed in the People's Parliament, should be passed as soon as possible to facilitate export promotion. If possible, Egyptian trading companies should have joint venture arrangements with trading companies in other countries, because of Egypt's lack of experience in international markets. Professional technical assistance from the NIE on how to organize and conduct trading companies operations is available. If approached, some of the NIE companies may consider setting up such a professional firm in Egypt's Free Zones, along with the establishment of such training institutes for International Trade Training, Productivity Improvement Services, and Design and Packaging Services. 12 Working Papers Annex: Egypt's Free Zones 39. All concerned, private or public sectors in Egypt, should encourage, promote and seek joint venture arrangements between the foreign investors and the local firmns whenever possible for investment projects in or out of the Free Zones. This is the most cost effective way to upgrade industrial standard and capacity. 40. A privately funded international trade show center should be established. The government, through the BOT approach, could offer some incentives or concessions to encourage private sector participation as advanced by many Asian countries. The Center will provide under "one roof' a shopping mall for all buyers coming to town. 41. The private sector participation in the establishment of Free Zones should be encouraged. Relevant laws and regulations must be enacted. Measures should be taken to prevent land speculation. The experience of BOT in Asian countries could be of help, including setting up a special high level unit in government to monitor BOT projects such as those in the Office of President in the Philippines, or in the Prime Minister's Office in Vietnam. The private sector participation in infrastructure is increasing. Whether the time is ripe or not for the private sector to develop EPZ in Egypt, this report urges the government to consider setting up such a high level responsible unit to manage and monitor BOT type projects in Egypt. 42. Experience of the countries with EPZ did encounter sun-set industries in the EPZ such as textile and garment industries. Some have been able to stand further progress by effectively dealing with the transitional restructuring of the affected industries, by moving up for higher stair-step technologies. While this report urges the government to pursue further the labor-intensive projects such as the apparel projects to come into the Free Zones, the report also urges the government to be prepared sometimes in the future (5-10 years) for the needed restructuring. ANNEX B Egypt: The Long-term Demographic Trend and Its Economic Implications Introduction 1. Egypt is the most populous Arab country, with an estimated population of over 60 million in 1996 (CAPMAS, 1995). The annual growth rate was 2.5 percent during the period of 1960-65, 2.2 percent during the period of 1965-70 and 2.0 percent in recent years (United Nations, 1995a). Although there was a temporary decline in the early 1970s, its growth did not slow down until the late 1980s. The total fertility rate declined from 4.4 births per woman in 1985-1990 to 3.8 in 1990-1995. This suggests that while the replacement-level fertility, and ultimately zero population growth, will not occur within the next decade, the downward trend is not likely to be reversed. 2. Given this fertility trends and other necessary assumptions, the questions become--what would be the future size of population, its age structure and distribution? What economic consequences can be anticipated from public policies' point of view? This study aims at assessing Egypt's future population growth and age structure and analyzing its long-term implications for employment, public expenditures on education and life-cycle savings. 3. The assessment will be cast in an economic framework which views population and economic growth as inherently integrated, postulating an explicit linkage of population growth with capital formation, employment, education and savings in the development process. As outlined below, two issues are particularly relevant in the existing literature: investment in human capital, and accumulation of savings as affected by demographic changes. 4. Coale and Hoover (1958), in their study of India, first raised the concern that whereas a rapidly growing population requires substantial investment to maintain labor productivity, the share of national output devoted to savings and investment can be adversely affected by rapid population growth. In the face of low rates of domestic savings, capital infusion through bilateral aid and multilateral institutions was seen as the key to the reduction of unemployment and to the increase of labor productivity. In addition, the implication of fertility declines for savings was emphasized, concluding that national savings rates during the 1960s were higher among countries with lower population growth rates (Bilsborrow, 1979; Gupta, 1975). 5. The view that rapid population growth impedes domestic savings has been recently challenged by those who regard population as an endogenous variable in the growth model . According to this view, economic variables and demographic structures are mutually interdependent. On the one hand, the size of the population, its growth, and age distribution as 1 This is unlike the neoclassical growth economists who treat population as an exogenous variable in the growth model. 14 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications well as the quality of human resources, affected by the extent of the investment in human capital, are the demographic factors which affect economic growth; on the other hand, the level of income affects the population growth, its age structure and quality of human resources. 6. Two important factors are emphasized in the growth model, capital formation and investment in the labor force. Recent studies (Kelley, 1994; Pritchett, 1994) have pointed out that economic growth depends on the quantity and productivity of all inputs including physical capital, human capital, and technology as well as the number of workers. Poor economic outcomes may have nothing to do with the rate of increases in the labor supply. Indeed, in recent decades, differences in the growth of the potential labor supply in the low-and middle- income countries do not explain the differences in economic growth rates. The World Bank's recent findings (1995a) suggest that "during 1965-93, growth rates of the working-age population were remarkably similar across regions, with only a few tenth of a percentage point differences. But differences in GDP growth rates were huge." In East Asia output growth exceeded expansion of the working-age population by an average of about 5 percentage points a year; in Latin America the difference was less than 1.5 percentage points; and in Sub-Saharan Africa growth in the working-age population exceeded GDP growth. In emphasizing the role of investment in the labor force, Denison (1974) argued that investment in human capital was a major source of growth because it contributed in raising the output per worker. Further empirical findings suggest that both primary and secondary school enrollment ratios played significant roles in determining the growth rate of per capita income. 7. A further perspective (the life-cycle hypothesis) on the relationship between population growth and savings (hence, investment and growth) was provided by Tobin (1967), and Leff (1969). Because consumption and earning vary over the household's life cycle, there is no match between current income and desired expenditures in any particular period; households usually save during the periods of high earnings relative to desired household consumption; while during periods of relatively low earnings, households will save less. Hence, consumption varies in response to changing demographic characteristics of households. 8. In summarizing the recent empirical findings relating to the link between savings and population change, Mason (1988) further pointed out that there are two paths by which population growth affects the aggregate savings. At the household level, declining fertility and, hence, a reduced "burden" of childbearing lead to reduced consumption and higher savings. This is "the dependency effect". At the same time, the population aging that accompanies reduced fertility increases the relative number of older households, which have lower rates of savings. This is "the rate of growth effect". The rate of growth effect may be more than offset by the dependency effect so that reduced population growth leads to higher rates of savings. Thus, depending on which effect dominates in a particular process of demographic transition, aggregate savings can be either positively or negatively related to the population growth rate. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 15 The Demographic Context Past Demographic Trends 9. Egypt's population was only about 10 million in 1897, and has increased by six times since then. It grew most rapidly after 1947 (Table B.1). The annual growth rate was 2.34 percent during 1947-1960 and 2.52 percent during 1960-1966. The population growth slowed down to 1.92 percent a year during the 1966-1976 period. This decline was due, in part, to postponement of marriage, reductions in fertility, and to some changes in the age structure echoing the effects of World War II (Cochrane, 1995). The population growth started to pick up again during the 1976-1986 period, with growth rate at 2.75 percent annually. However, it has dropped to 2.0 percent per annum during 1980-1993 (World Bank, 1995a). 10. The population growth over the past decades has changed the age structure (Table B.2). The fertility decline in the early 1970s had an immediate effect on age group of 0-4. The proportion of the population under 5 declined fromn 15.9 percent in 1960 to 13.8 percent in 1976. However, it went up to 15.3 percent in 1986 as fertility rebounded in the late 1970s and early 1980s. As fertility decreased in the early 1990s, the proportion of the population under 5 declined to 12.9 percent in 1995. The proportion of the school age population of 5-14 declined from 26.8 percent in 1960 to 25.2 percent in 1995, whereas the working age group of 15-64 increased from 53.8 percent in 1960 to 57.7 percent in 1995. The proportion of the population of 65 or over remained almost constant over the period of 1960-1986 and went up about one percentage point by 1995. In general, Egypt has a relatively young population as its fertility rate did not decline significantly until late 1980s. Sources of Growth: Fertility and Mortality 11. The major source of population growth for Egypt has been one of natural increases due to changes in fertility and mortality, although in-and-out migration was also important over the last two decades. The crude death rate declined from around 17 per 1,000 in 1970 to 8 per 1,000 in 1993 (World Bank, 1995). Life expectancy at birth increased from 51.6 years for males and 53.8 years for females in 1960 to 60.5 years for males and 63.5 years for females in 1986 (Table B.3). The recent gains in life expectancy were also noticeable, with 2.4 years gain for males from 1986 to 1991 and 3 years for females during the same period. 12. The infant mortality decline was slow over the period of 1950 to 1969 (Cochrane, 1995); however, it fell more rapidly during 1970-1993, from 158 per 1,000 live births to 64 per 1,000 live births (World Bank, 1995a). The level of fertility has been a very important factor of population change in Egypt. Data from censuses and demographic surveys such as the Egypt Fertility Survey (EFS) and the Egypt Demographic and Health Survey (EDHS) have shown that the decline in fertility did not start until late 1980s. The total fertility rates (TFR) per woman dropped from 4.7 in 1988 to 3.9 in 1992 (Table B.4). 16 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications The Demographic Trnd 13. The demographic trend has been envisaged by projections compiled by the United Nations (1995a)2. The three fertility variants presented by the UN are referred to as high, medium and low. These three variants are thought to provide reasonable and plausible future trends in fertility. The low and high fertility variants are usually thought to bracket the probable range of future population change, and the medium variant projection can be thought of as "most likely". 14. Table B.4 presents the total fertility rates under the three fertility assumptions. Under the medium-variant assumption, the fertility rate would decline to the replacement level (zero population growth) by the year 2015-2020.3 Table B.5 and Figure 2.1 present the demographic trends under the three fertility assumptions. Under the medium variant assumption, total population is estimated to be 62.9 million in 1995, and would be close to 70 million in year 2000. During the years between 1995 and 2040, there would be a net increase of 1.06 million population each year. However, population growth would be slower after 2040, reaching 117.4 million by 2050. The Economic Implications of the Demographic Trend 15. The long term population change has important economic implications. An increasingly younger population would require more resources toward investing in human capital, i.e. in provision of health and education services; a population with increasing size of the working age groups would demand more job opportunities on the one hand, and on the other hand, it would also tend to save more in view of the relatively higher current income, and the need to build-up assets for retirement; and an aging population would probably save less in view of its increasingly large size of population in retirement. This section will examine these aspects of population change in Egypt. 2 Data used by the UN for estimation of the Egyptian population are from several sources. The base population was estimated by the UN, which was consistent with the results of 1986 Egyptian census, adjusted upward by 6.7 percent due to undernumeration. The total fertility rate was based on (a) estimates for the period 1984-1992 derived from maternity-history tabulations from the Demographic Health Survey in 1988-1989 and 1992 from the Pan Arab Project for Child Development (PAPCHILD) Survey in 1991; and (b) registered births by the age of the mother to 1989. Life expectancy at birth was based on estimated level of infant and child mortality for the period of 1985-1990 derived from 1991 PAPCHILD Survey and the assumption of a Coale-Demeny South Region age pattern of mortality. International migration was based on estimated international migration for the intercensal period 1976-1986 and assumed future trends since that time. The estimate was derived as the difference between the natural increase of population and total population growth during the intercensal period. 3 This assumed target period at which fertility will stabilize is determined by taking into account a wide range of socioeconomic factors, such as population policies and program, adult literacy, enrollment levels, level of GDP, and cultural and political factors. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 17 Implications for Labor Market and Employment 16. With a current unemployment rate of over 14 percent in Egypt and disguised unemployment in the public and private sectors (Cochrane, 1995), the implications of short- term and long-term population change are particularly important. 17. Implications for the Labor Market. It appears that (Table B.7) the high, medium or low fertility assumptions affect only the 0-4 age group in the short-run, and that there is no real difference in the future trends of the working age population (at age 15-64) under any of the fertility scenarios. The working age population over the years of 1995-2050 under the medium fertility assumption are projected to be 53.1 million in 2010, whereas under the low fertility assumption, it would reach 52.5 million in 2010, a minor difference of about half a million. 18. Declining fertility will not substantially reduce the size of the working age population until after 2010, and neither will it result in any aging of the working age population within the next 15 years. Overall, Table B.8 shows a fairly stable and young age composition between 1995 and 2010. The median age, an alternative and simple summary of the aging of population, would be 27 in year 2010, indicating that half of the Egyptian population would be under 27 years of age in year 2010. In fact, due to the past and future fertility trends, the working age population would increase from 36.4 million in 1995 to 41.7 million in 2000, and reach 62.6 million in 2020. On the average, there would be an increase of one million working age population each year between years 1995 and 2020. Table B.8 also shows the percentage changes in the working-age population in six broad age categories under the medium fertility assumptions. The size of 15-24 age group in the short run would increase from 18.7 percent in 1995 to 20.4 percent by the year 2000, and the size of 40-49 age group would increase from 9.9 percent to 11.1 percent. In the medium-run, the size of 25-29 and 50-59 age groups would increase by 2 percentage points and 3 percentage points respectively between year 2000 and 2010. 19. Labor Participation and Economically Active Population. The labor force participation rates for 1986 and 1992 are presented in Table B.9. It shows that the overall participation rate in 1992 was 47.2 percent for the population at age 15 and above. The participation rates grew during 1986-1992 among various age groups except for the oldest age group of 64 and over. 20. Despite the increased overall participation rate, major gender gaps still remain. While female labor force participation rates increased during the period of 1986-1992 especially among the age groups of 30-39, 40-49 and 50-59, the participation rate for all females was only 22.0 percent in 1992 compared with 70.7 percent for all males. Male age group of 40-49 and 50-59 were almost all economically active, whereas the participation rates for females were very low, at only 27.2 percent and 17.9 percent respectively. The gender gaps also appear in other age groups. 51.8 percent and 77.2 percent of males age groups of 20-24 and 25-29 were working, or looking for work in 1992, whereas the corresponding figure for both female age groups was only 29 percent. 18 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications 21. Table B.10 shows the projections of the economically active population (at age 15-64) for the period of 1995-2050. The calculations are based on the United Nations' age-sex-specific population projection (1995c) under the medium fertility assumption and the recent age-sex- specific labor force participation rates of 1992 released by the International Labor Office (1995). It is assumed that the future labor force participation rate will remain constant at the 1992 level.4 22. It is estimated (Table B. 10) that in 1995 the economically active population (at age 15 to 64) were around 18 million. In 2000, there would be a total of 20.6 million economically active population, and would reach 32 million by the year 2020. On average, 560,000 jobs need to be created each year during the period of 1995-2020 to prevent increases in unemployment. As described below, of the 560,000 jobs needed each year for the next 5 years at least, about one fourth would arise from the entry into the labor force of the 15-24 age groups. 23. Table B.10 also shows changes in various age groups of the economically active population (at age 15-64). In the short-run, the age-groups of 15-19 and 20-24 are expected to grow by 20 percent respectively by the year 2000--the large increase was in part due to the increases in the level of labor force participation for the 15-19 and 20-24 male age groups from the lower levels experienced during the period of 1986-1992. Furthermore, the age groups of 25-29 and 30-39 are expected to grow by 6 and 8 percent respectively by the year 2000. In absolute number, there would be an increase of 287,000 of the economically active population at age 15-19 between 1995 to 2000. On average, each year there would be an increase of 57,400 economically-active population in this age group. For age group of 20-24, there would be an increase of 398,000 economically active population between 1995 and 2000. Each year, there will be an increase of 79,600 economically active population at age 20-24. Thus, combining these two age groups of 15-19 and 20-24, there will be an increase of 137,000 economically active population each year between 1995 and 2000. 24. The challenge for the economy to provide jobs for roughly 137,000 economically active population each year during 1995-2000 is compounded by the demand for jobs beyond the year 2000. Table B.10 shows the employment pressure for the working age groups of 25-29 and 50- 59 would be becoming greater during the period between 2000 and 2010 as these age groups will grow significantly in size. Implications for Public Expenditure on Education 25. Does the rapid expansion of labor supply depress economic growth? The key lies in productive investment, along with more savings, and in investment in the labor force as well. Creating labor demand depends not only on how quickly factors of production are accumulated, but also on the quality of those factors, the technology embodied in them, and how efficiently they are employed. Fast-growing economies do not simply invest more but combine physical capital and the education of workers in ways that increase output per worker. A further ' Of course, more precise employment projection requires an assessment of the overall economic conditions in the future, which is beyond the scope of this study. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 19 challenge for Egypt is how to contain the growth of demand for jobs particularly among the younger age groups. 26. The Increasing Size of School Age Population. The Medium fertility population projection (Table B.11) shows that the school age population of age 6 to 23 would increase from 24.9 million in 1995 to 27.2 million in 2000 and reach its highest of 29 million in 2001, although the proportion of this school age group relative to the total population would decline from 39.5 percent in 1995 to 31.0 percent in 2020. 27. The population projections by age group further show that student population of 6-11 (primary school education) would stay almost unchanged at 9.7 million from 1995 to 2000, and it would increase to 9.9 million by the year 2010. The size of the age group 12-14 (secondary school) would increase (from 4.5 million in 1995) to 4.9 million in 2000. The age group of 15- 17 (high school), estimated at 4.0 million in 1995 would reach 4.6 million by the year 2000. It would then continue to rise throughout the projection period until the year 2030. The 18-23 age group population (university) would rise from 6.7 million in 1995 to 9.7 million in 2010, a net increase of 3 million between 1995 and 2010. 28. Thus, with respect to the short-term trend in the school age population, there are two important issues. One is that the proportion of school age population relative to the total population will decline in the next decade; however, in terms of the absolute number, there would be a net increase of 4.2 million of school age population by the year 2010. The other one is that the school age population of 6-11 (primary school student) would decline in the next five years, whereas secondary school age population or above, especially university and college population are projected to grow in size in the next decade. It is likely that the demand for the secondary and higher education will continue to increase not only because these age groups will continue to grow but also because households will demand higher education attainment for their children and the structural changes in the economy will require such higher attainment as well. 29. The Quality of Education. Quality of primary school education as reflected in lower illiteracy rate and favorable pupil-teacher ratios in Indonesia and Malaysia have contributed to rapid economic growth (World Bank, 1995a). Table B.12 shows that the primary school enrollment ratio in Egypt increased significantly from 71 percent during 1970-75 period to 97 percent during 1989-94 period, which was very close to that of Malaysia and was lower than that of Indonesia. Nevertheless, the gender difference in enrollment rate still prevails. Females generally had a lower enrollment ratio than males. The enrollment ratio was 105 percent for males during 1989-94, whereas for females it was only 89 percent. 30. The illiteracy rate of people at age 15 or over remains high in Egypt, with 55 percent in the early 1980s; and this high illiteracy rate has not dropped significantly since then. It remained at 49 percent during the period of 1989-1994. In contrast, both Indonesia and Malaysia started with lower illiteracy rates of 28 percent and 26 percent respectively in the early 1980s and their illiteracy rates have dropped significantly over the past decade. 20 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications 31. Table B.13 shows pupil-teacher ratios. During the early 1970s, each teacher had 35 pupils in primary school in Egypt, whereas in Indonesia, each primary school teacher had only 29 pupils. In the early 1980s, the pupil-teacher ratio did not improve significantly in Egypt. Each primary school teacher had 32 pupils, whereas in Indonesia and Malaysia, each teacher had 25 and 24 pupils respectively. Although Egypt has managed to reduce the pupil-teacher ratio in recent years, it still lagged behind that of Indonesia and Malaysia. 32. Table B. 14 shows the shares of educational expenditure in the total budget. Egypt has succeeded in increasing the share of the educational expenditure in its overall budget, reaching 10.3 percent in 1993. However, Egypt still had a smaller budget on education as compared with that of Malaysia. In 1993 the expenditure on education was 20.3 percent in Malaysia, but was only 10.3 percent in Egypt. Furthermore, among its expenditure on education, Egypt has spent more on university and college education rather than basic education. Non-wage expenditures per student in Egypt were 50 times larger in higher education than in primary and secondary education (World Bank, 1991). 33. Implication for Expenditure on Education. The future population trends in Egypt has two implications for expenditure on education. Given the decline of primary school age population and the expansion of high school and college age population in the future, the immediate priority should be the improvement of basic education, particularly through increasing the female enrollment rate, reducing illiteracy, and increasing the number of teachers, and improving the quality. As the number of entering students (6-11 age group) begins to level off in the coming five years, the quality improvement of basic education can be financed in part by redirecting resources previously devoted to expanding the size of basic education services. 34. The second priority should be redirecting some portion of the current subsidy provided to university education to the secondary and vocational schools as the school age population of this age group would increase in next five years. In view of the projected labor increases in the young age groups of 15-19 and 20-24 detailed in the previous section, it seems sensible to educate all youth up to secondary school levels and encourage them to pursue vocational training. This would sharply reduce the labor force participation rate in the 15-19 and 20-24 age groups, therefore, somewhat alleviating the demand for employment. The Pace of Aging 35. The decline in fertility rates will eventually result in a decline in the proportion of children under 15 and a rise in the proportion of the elderly relative to the proportion of the working-age population. In Table B.15, the trend of age composition and dependency ratio under three fertility assumptions are presented. Under the low fertility assumption, there would be 4.5 percent elderly population in 2000, which is identical to that under the medium fertility projection. Under the high fertility assumption, there would be 4.4 percent elderly population in 2000. In 2010, the proportion of elderly would be around 4.6 to 5.1 percent under the three fertility assumptions. Therefore, declining fertility under any of the three fertility assumptions would not entail substantial increases in the relative shares of the elderly population in the next decade. The increases in the absolute number of the elderly population would also be quite insignificant until 2010. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 21 36. Along with the changes in percentage of the elderly population, Table B.15 further shows the dependency ratios under the three fertility assumptions. The dependency ratio is defined as a ratio of children under 15 plus the elderly at age 65 or over to the working-age population at age 15-64. In general, a declining fertility is associated with a decreasing dependency ratio, as the proportion of children under 15 declines. It is shown that under the medium fertility assumption, the dependency ratio would drop from 73.3 percent in 1995 to 65.6 percent in 2000. It would be only 53.6 percent by 2010, and would more or less remain at the 50 percent level through 2050. Implications for Savings 37. During a period of declining fertility, the changing age structure of the population could increase the potential for saving. As explained in the first chapter, the life cycle saving model (Leff, 1969) predicts that there are two paths by which population decline affects aggregate saving. A reduced "burden" of childbearing leads to reduced consumption and higher saving. At the same time, the population aging that accompanies reduced fertility increases the relative number of older households, which leads to a lower rates of saving. This study replicates the existing literature by using recent Egyptian data to test this hypothesis. 38. The Leffs Life-cycle Saving Model. In elaborating the logic of an inverse relation between dependency ratios and savings rates, Leff (1969) emphasizes that children constitute a heavy demand for expenditure which, in the standard national income accounting framework, is put under the heading of consumption. Because they contribute to consumption, not to production, a high ratio of dependents to the working age population might be expected to impose a constraint on a society's potential for saving. 39. Leff fitted a saving function to aggregate data on gross national saving and demographic variables for 74 countries and found that dependency ratios were an important variable affecting aggregate savings ratio, both for the 74 countries considered as a whole and within the subsets of developed and developing countries. High dependency ratios were associated with low saving rates. This hypothesis of an inverse relation between dependency ratios and savings rates has been supported by other empirical studies. Gupta (1975) employed World Bank data for the 1960s and found a strong impact of dependency ratio on saving. Analyzing international cross-section data after 1960-70, Mason (1981) also found that higher fertility depressed saving. Kelley (1986) used pooled time-series data of 67 countries and found that the impact of the young dependency ratio varied systematically with region and over time. Young dependency actually had a positive impact on saving in Africa, a negligible impact in Asia, and a negative impact only in Latin America. 40. The Empirical Model. According to Leff, saving model could be expressed by rewriting the standard national income identity to include separately the consumption of the working age 22 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications population, Cwa, and the expenditure on the dependents in the population, D. With the ex post equality of savings and investment, model is as follows: Y=Cwa+D+S+X-M (1) 41. Dividing through by Y and rearranging terms, it follows that S/Y is inversely related to the share of dependency expenditure in national income. The model estimation is therefore fairly simple, and seeks to test the proposition that the aggregate saving rate is lower, ceteris paribus, to the extent that it has more dependents in its population. In Leffs model, two dependency ratios were introduced, youth dependency ratio, Dl (population younger than 15 divided by population of age 15-64), and old-age dependency ratio, D2 (population older than 64 divided by population of age 15-64). In addition, two economic variables were introduced, log of per capita income and log of rate of growth of income. The dependent variable was the log of aggregate domestic savings ratio. 42. Accordingly, we specify a saving model as function of the level of economic growth as well as the child and elderly dependency ratios, which can be written as S=a+blC+b2E+b3G+U (2) where the dependent variable, S, is the ratio of gross domestic saving relative to gross domestic product. C is the child dependency ratio, and E is the elderly dependency ratio; both are measured as their percentage relative to the working age population of age 15-64. G is the annual growth rate of gross domestic product. Unobserved heterogeneity is captured in U which is assumed to be random and not correlated with either G or C or E. According to Leffs model specification, bl and b2 will be <0. 43. Empirical Results. In Table B. 17, the long-term trends of the gross domestic saving/GDP, the child dependency ratio, and the elderly dependency ratio during period of 1960-2010 are presented. The child dependency ratio is the ratio of children under 15 to the working age population of age 15-64, and the elderly dependency ratio is the ratio of elderly 65 or over to the working age population. 44. The long-term trend in the child and the elderly dependency ratios in Egypt could be divided into four periods. The first period is from 1960 to early 1970s, characterized by its higher child dependency ratio and lower elderly dependency ratio. The higher child dependent ratio was associated with lower gross domestic saving rates. 45. The second period is from 1975 to 1986, in which the child dependency ratio declined noticeably, a decline of five percentage points on average compared with the first period. Meanwhile, the elderly dependency ratio increased only slightly, about only half a percentage point increase on average compared with the first period. During this period, Egypt had benefited from the "dependency effect" due to the decline in child dependency ratio, and did not suffered from the "growth effect" of an increase in the elderly dependency ratio. Accordingly, the interplay of these two effects should have resulted in a favorable condition for higher saving. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 23 Table B. 17 shows that the gross domestic saving rates were as high as 15 percent on average during this period. 46. The third period is from 1987 to 1995, in which the child dependency ratio declined slightly, only about two percentage points compared with the second period; whereas the elderly dependency ratio increased about one percentage point. Therefore Egypt did not benefit substantially from the "dependency effect", whereas had suffered from an increase in elderly dependency ratio. The interplay of these two effects was associated with less favorable environment for saving. Indeed, the gross domestic saving/GDP dropped to 8 percent on average during this period. 47. The fourth period is from 1996 to 2010, in which the child dependency ratio will decrease substantially from 65.7 percent in 1995 to 46.1 percent in 2010, whereas the elderly dependency ratio will remain almost the same. The reduced "burden" of childbearing will lead to reduced consumption and higher saving on the one hand; on the other hand, the low and stable elderly dependency ratio is unlikely to result in a decrease in saving during this period. The combination of these two effects will create a much more favorable environment for saving. 48. This intuitive description on the inverse relationship between the population dependency ratios and saving is empirically tested. In Table B. 18, we use a least square multiple regression to fit the data from year 1960 to year 1992. It shows the coefficients of variables, along with the Beta coefficients, the statistic which shows the relative importance of the different independent variables in accounting for the variance in the dependent variable. Student t-values are presented in parentheses below each coefficient. The findings confirm the inverse relation between dependency ratios and savings rates. The negative coefficients indicate that, controlling for the growth rate of gross domestic product, the decrease in child dependency ratio is associated with the increase in saving; and the increase in elderly dependency ratio is significantly associated with decreases in saving. The Beta coefficients indicates that the child dependency ratio contribute the most of the statistical explanation. The t-values are all significant about the 0.01 level, which indicate that the dependency variables should be considered statistically distinct variables in the explanation of aggregate savings rates. More specifically, if elderly dependency ratio goes up one percentage point, then the saving ratio will drop 5.28 percentage point. Likewise, if the child dependency ratio decrease one percentage point, the saving ratio will increase 1.29 percentage point. Using these coefficients, it is estimated that between 1995-2010, the decline of child dependency ratio (by 0.196) could boost the aggregate savings by 25 percent (0.196 x 1.29 = 0.25) after controlling the effects of GDP growth, and of changes in the elderly dependency ratio. Similarly, everything else remains equal, the increase of elderly ratio (by 0.02) could result in an increase in consumption and a decline in savings by 1 percent (0.02 x 5.28 = 1.1). Therefore, we would expect a net increase in savings rates over the period 1995-2010 by 24 percent because of the anticipated change as the age structure. 24 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Main Findings 49. Fertility and Moralitv: Egypt has been in the demographic transition from high-fertility to low-fertility. However, the more rapid fertility decline did not start until in the mid-1980s. Replacement level fertility (Zero population growth) appears likely to occur within the next 20 to 25 years. This would imply a relatively young population within the next 15 years with the medium age of population not exceeding 27. Progress has been made in reductions of crude death rate and increases in life expectancy at birth, although the infant mortality rate still remains relatively high. 50. Population Growth: Population growth in absolute numbers will continue for several decades into the next century under any of the three fertility scenarios. With respect to the short-term population growth, under any of the three fertility scenarios, there would be a net increase of one million population each year between 1995 to 2010. With respect to the long- term population growth, there would be an increase of 1.06 million population each between year of 1995 to 2040 under the medium fertility assumption. 51. Employment and Labor Market: The general trend from the three fertility scenarios is that declining fertility would not substantially reduce the size of the working age population aged 15-64 until after 2010, and that there would be at least 52.5 million working age population in year 2010 under any of these three scenarios. If the future labor force participation rate for the population at age 15 and above remain at the 1992 level of 47.2 percent, there would be a total of 20.6 million economically active population in 2000, a total of 26.6 million in 2010 and a total of 32 million in 2020. Each year, on average, 560,000 jobs need to be created during 1995-2000. Even an immediate decline in fertility will not alleviate this pressure because it takes about twenty years for reduced fertility rates to significantly slow the growth of the labor supply. 52. Moreover, the labor market will be facing an increasing and immediate pressure from the 15-24 age group population looking for jobs. Each year, 137,000 jobs need to be provided for the 15-24 age group between 1995 and 2000. This demand for jobs will be compounded by the fact that the labor force participation rate for the 15-24 age group in the past 5 years were lower, and that the increased participation rate in the next few years will entail substantially increased demand for jobs. Beginning in 2000, demand for employment would be particularly strong among working age population of 25-29, and 50-59, as these age groups would account for much large shares of the population. 53. Expenditure on Education: The future population trend has two implications for its expenditure on education. With respect to the quality of education, there is a need to improve the basic education by reducing gender gaps in enrollment, pupils-teacher ratios, and illiteracy. 54. Furthermore, the future population decline will not result in a decline in absolute number of school age population. In short-run, there would be a net increase of 2.4 million student population of age 6-23 by the year 2000. This will require more expenditure on education. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 25 55. Within the next five years, the student population of 6-11 will be leveling off, easing the previous need for expanding facilities to accommodate this age group. This development should allow reallocation of education expenditures toward girl's enrollments, improving the pupil-teacher ratio, and reducing regional disparity in education attainment. 56. Nevertheless, other school-age population of 12-14, 15-17, and 18-23 are all expected to grow in their absolute number. It is likely that the demographic pressures at the secondary and higher education levels will continue, and will be compounded by the household demand for higher education attainment for their children. In view of the projected labor supply in the age groups of 15-24, it appears desirable to delay some of their entry into the labor market. One way to do this is to expand secondary school education and encourage vocational education. Such expansion could be facilitated by reallocating some portion of the Government expenditures on university education. 57. Savings: We used multiple regression model to fit the Egyptian data over the last decades concerning the savings rates and dependency ratios. Our finding confirms that the decrease in child dependency ratio has contributed to the increase in savings rates in Egypt, whereas the increase in elderly dependency ratio has contributed to the decrease in savings. During 1995 to 2010, the child dependency ratio would decrease from 65.7 percent to 46.1 percent, and the elderly dependency ratio would remain almost the same. This change in the age structure in the coming decade will make it more favorable for increased savings as Egypt has witnessed in the past, and as the life cycle savings hypothesis anticipates. The reduced "burden" of childcare will lead to reduced consumption and higher savings; the low and stable elderly dependency ratio is unlikely to result in a very large decrease in savings during 1995- 2010. The combination of these two effects will create a more favorable environment for savings and investment in Egypt possibly increasing the aggregate savings by about 24 percent. 26 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Table B.I. Population and Annual Growth Rates in Intercensal Years in Egypt: 1897-1996 (in thousand) Census year Total population Growth rate 1897 9,669: 1907 11,190 1.46 1917 12,718 1.28 1927 14,718 1.09 1937 15,921 1.16 1947 18,967 1 .75 1960 26 ,O085 2.34 1966 :30,076 2.52 1976 36,626 1.92 1986 48, 2e54 2.75 1996 :(Estmated) 00,603 1.80* Source: CAPMAS Statistical Y'e22 ook. 199;: :World Bank: World Developrnent Report. :1995d. Note: * This figure refers to the annual growth rate of 1989 to 1994, estimated by World Bank: Social Indicators of Development, 1995. Table B.2 Percentage Distribution of Population by Age and Sex, Egypt: 1960-1995 1960 7 1986 Age Total M F Ttal MI F Total M F Total M F .0-4 15.9 16.1 15.7 183.8 13.7 13.8 15.3 15.2 15.3 12.9 12.9 12.8 514 26.8 27.7 25.9 26.2 26.8 25.5 24.7 25.1 24.2 25.2 .24.3 25.0....... 15-64 5318 53.0 54.7 56.4 56.1 57.0 56.7 56.3 57.3 57.1 95.0 57.5:: 65+ I 23.5 .2 3.7 3.6 3.4 3.7 3.3 3.4 32 4.2 3.8 4.47: Source: Calculated from 1995 CAPMAS Statistical Year Book: United Nations: The Sex and Age Distribution of the World Populations. the 1994 Revision. Table B.3 Life Expectancy at Birth by Sex, Egypt: 1960-91 ; Year ; ; ; Male Female 1960 51.6 53.8 1976 52.7 57.7 1986 60.5 63.5 1991 . 62.9 66.4 Source: CA PMAS iStistical Yearbook. 1995. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 27 Table B.4 Total Fertility Rate Per Woman under Three Fertility Assumptions, Egypt, 1995-2050 Year Medium Low High 1990-1995 3.88 3.58 4.18 1995-2000 3.44 3.09 3.79 2000-2010 2.75 2.33 3.18 2010-2020 2.10 1.60 2.60 2020-2030 2.10 1.60 2.60 2030-2040 2.10 1.60 2.60 2040-2050 2.10 1.60 2.60 Source: United Nations: World Population Pro=cts. the 1994 Revision. Table B.5 Population Projections under Three Fertility Assumptions, Egypt: 1995-2050 (in thousand) Year Medium Low High 1995 62,931 62,345 63,510 2000 69,146 67,781 70,494 2010 81,490 77h688 85,269 2020 92,015 84,196 99,927 2030 102,254 89,384 115,898 2040 110,577 91,671 131,700 2050 117,398 91,456 147,994 Source:. United Nations: World Population ProsUects, the 1994 Revision. 28 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Table B.6 Population Projections by Urban and Rural, Egypt: 1995-2025 (in thousand) Year ''Total Urban Rural 1995 62,931 28,170 34,761 100% 44.8% 55.2% t; 20tW000 0 69,l10460 32,054 37,092 1 00%0 f 46.4% 53.6% 2005 75,424 36 733 38,691 100% 48.7% 51.30/a 2010 81,490 42,173 39,317 100% 51.8% 48.2% 2015 86,674 48,038 38,635 100% 55.4% 44%6% 20200 92,015 54,209 37,806 1 00% 58.9% 41.1% : 202500 97I301: 60,519 36,782 100% 62.2% 37.8% Source United Nations: World& Ubanization Prospects: the 1994 Revision. Table B.7 Projections of Working Age Population Aged 15-64 under Three Fertility Assumptions, Egypt: 1995-2050 (in thousand) Fertift: 1995 2000 2010 2020 2030 2040: 2050 A;ssulmption 0 :f: V:: ; Medium 36,361 41,731 53,115 62,589 66,376 74,716 77,153 High 36,361 41,731 53,682 64,928 74,851 85,207 94,168 iLow 00 0 36,361 41,731 52,542 60,216 63,631 64,578 61,523 Sources;Clculated from United Nations: The Sex and Age Distribution of the World Populations, 94 Re .: Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 29 Table B.8 Projections of Working Age Population by Age Group under the Medium Fertility Assumption, Egypt: 1995-2020 (in percentage relative to total population) Age group 1995 2000 2010 2020 15-24 18.7 20.4 19.8 17.1 25-29 8.0 7.7 9,6 9.3 30-39 13.8 13.6 14.6 18.2 40-49 9.9 11.1 11.6 13.2 50-59 6.0 6.4 9.2 10.2 60-64 2.3 2.3 2.9 4.2 Median age 21.4 22.7 26.5 31.7 Population (000) 36,361 41,731 53,115 62,589 Sources: Calculated from United Nations: The Sex and Age Distribution of the World Populations. the 1994 Revision: United Nations: World Population Prospects. the 1994 Revision Table B.9 Labor Force Participation Rate by Age and Sex, Egypt: 1986-1992 (in percentage) Age Total Male Ferale Group 1986 1992 1986 1992 1986 1992 15-19 23.35 23.9 39.44 34.7 8.87 12.4 20-24 40.9 51 8 29.3 25-29 44.06 53.8 73.02 77.2 15.73 29.0 30-39 45.86 57.4 79.25 82.6 13.14 30.8 4049 51.44 64.2 89.77 99.8 11.84 27.2 50-59 59.22 58.3 100.00 99.3 8.96 17.9 60-64 43.08 45.1 81.39 83.8 5.23 9.2 64+ 52.2] 18.8 84.43 86.0 20.7 3.5 Total(15+) 47.2 70,7 22.8 Source: ILO Yearbook of Labor Statistics. 1995. 1986 figures are from Nader Fergany "A Characterization of the Employment Problem in Egypt" in Employment and Structural Adjustment Egypt in the 1990s. ed.. Handoussa and Potter. 30 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Table B.10 Projections of Economically Active Population of 15-64 under Medium-variant Assumption, Egypt: 1995-2050 (in thousand) Age Group 1995 2000 2010 2020 2030 2040 2050 15-19 1,519 1,806 1,889 1,953 1,758 1,902 1,'901 pctincrease 1:00 119 124 129 116 125 125 20-24 2,174 2,572 3,306 3,278 3,278 3,157 3,279 pmtincrease 100 118 :152 158 151 145 151 25-29 2,672 2,820 4,008 4,212 4,367 3,937 4,266 pet increase 100 106 150 158 163 147 160 30-39 4,929 5,303 6,497 8,810 9,01: 8 9,203 8,59 pctincrease 100 108 132 179 183 187 174 40-49 3,945 4,837 5,765 7,125 9,707 9,970 10,203 pgtincrease 100 123 146 181 246 253 259 50-59 2,077 2,522 4,150 4,998 6,231 8,539 8,817 potincrease 100 121 200 241 300 411 425 60-64 646 697 1,020 1,602 1,917 2,486 3,278 pet increase 100 108 158 248 297 385 507 TotaI 17,962 20,557 26,635 31,978 36,276 39,194 40,337 Sources: Calculation based on United Nations: The sex and Age Distribution of the World Populations. the 1994 Revision; 1995 OLO Yearbook of Labor Statistics. Table B.1 I Projections of School Age Populations under Medium Fertflity Assumption, Egypt: 1995-2050 (in thousand) Age Group 1995 2000 2010 2020 2030 2040 2050 6-11 9,754 9,680 9,860 9,109 9,510 9,620 9,509 15.5% : 14.0% 12.1% 9.9% 9.3%: 8.7% 8.1% 12-14 4,468 4,909 4,808 4,877 4,601 4,865 4,696 7.1% 7.1% 5.9% 5.3% 4.5% 4.4% 4,0% 15-17 3,965 4,633 4,726 4,877 4,397 4,755 4,813 6.3% 6.7% 58 5.3% 4.3% 4.3% 4.1 %/ 18-23 6,671 8,021 9,697 9,662 9,407 9,399 9,627 10.6% 1 1.6% I11.9% 10.5% 9.2%/o 8.5% / 8.2%/a Total 24,858 27,243 29,091 20,525 27,915 28,639 20,645 39.5% 39.4% 35.7% 31.0% 27.3% 25.9% 24.4% Sources: Calculated from United Nations: World Population Prospects, the 1994 Revision. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 31 Table B.12 Primary School Enrollment Ratio and Illiteracy Rate by Country and Sex: 1970-94 (in percentage) Countiy 1970-75 1980-85 1989-94 Gross Enrollment Ratio Egypt Total 71 85 97 Male 85 94 105 Female 57 76 89 Malaysia Total 91 101 93 Male 92 101 93 Female 89 100 93 Indonesia Total 86 117 114 Male 94 120 116 Female 78 114 112 Illiteracy Rate (age 15+) Egypt Total 55 49 Female 71 61 Indonesia Total 28 16 Female 37 22 Malaysia Total 26 17 Female 35 22 Source: World Bank: Social Indicators of Development. 1995. Table B.13 Pupil-teacher Ratio, Egypt, Indonesia, and Malaysia: 1970-94 Country Primary. Secondary: Pupils per Teacher Pupils per Teacher 1970-75 1980.85 1989-94 1970-75 1980-85 1989-94 Egypt 35 32 26 34 22 21 Indonesia 29 25 23 16 17 16 Malaysia na 24 20 28 22 19 Source: World Bank: Social Indicators of Development. 1995. 32 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Table B.14 Government Expenditure on Education by Country: 1980-93 (in percentage) Counr t000:J fff; ff0 1980 1993 0 :Egypt;; lf t | 00 :83 10S.3 Ma*sia 18.3 20.3 Indonesia 8.3 10.0 Source: World Bank: World Development Report, 1995. Table B.15 Age Composition and Dependent Ratio under Three Fertility Assumptions, Egypt: 1995-2050 (in Percentage) 0AgGroup 1995 :2000 2010 2020 2030 2040 2050 Medium-Variant Proections 0414 38.1 35.1 30.0 25.3 2332 21.6 20.2 15.564 57.7 60.4 65.1 68.0 67.T 67.6$0 65.8&I 65+ 4.2 4.5 4.9 6.7 9,0 10.8 14.0' Dependent Ratio 73.3 65.6 53.6 47.1 47.7 447.90 20 High-Variant Projections 0-14 38.6 36.4 32.4 28.8 27.4 26.2 25.2 15-64 5712 59.2 63.0 65.0 64.7 64.7 633.7 65+ 4.2 4.4 4.6 6.2 7.9 9.1 11.1 Dependent Ratio 74.8 68.9 58.7 53.8 54.6 54.6 57.0 Low-Variant Projections 0-14 37.4 33.9 27.2 21.1 18.5 16,5 14.7 15-64 58.4 61.6 67.7 71.5 71.2 70A4 67,3 65+ 4.2 4.5 5.1 7.4 10.3 13.1 18.0 DependentRatio 71.2 62.3 47.7 39.9 40.4 4210 48.6 Sources: Calculated from United Nations: The Sex and Age Distribution of the World Populations. the 1994 Revision. Working Papers Annex: Egypt; The Long Term Demographic Trend and Its Economic Implications 33 Table B.16 Projections of Elderly Population under Medium Fertility Assumption, Egypt: 1995-2050 (in thousand) Age Group 1995 2000 2010 2020 2030 2040 2050 60 OT over 4,091 4,633 6,193 9,754 13,498 17,471 23,714 % of total pop 6.5% 6.7% 7.6% 10,6% 13.2% 15.8% 20.2% 65 or over 2,643 3,112 3,993 6,165 9,203 11,942 16,436 % of total pop 4,2% 4.5% 4.9% 6.7% 9.0%/o 10.8% 14.0% Source: Calculated from United Nations: World Population Prospects. the 1994 Revision. Table B.17 Gross Domestic Saving/GDP, Child Dependent Ratio, and Elderly Dependent ratio, Egypt: 1960-2010 (in percentage) Year Gross Domestic Child Dependent Elderly Dependent Savin&LQDP Ratio 0-14/15-64 Ratio 65/15 -64 1960 12.0 77.1 5.7 1965 14.0 77.4 5.9 1967 11.3 1968 8.8 1969 9.7 1970 9.4 76.1 5.8 1971 8,1 1972 6.6 1973 8.0 1974 5.7 1975 12.3 72.5 6A1 1976 16.7 71.0 6.4 1977 18.5 71.6 6.0 1978 16.4 71.1 6.2 1979 14.2 70.7 6.2 1980 15.2 70.2 6.2 1981 14.1 69.4 6.2 1982 15.2 1983 17.8 1984 14.0 1985 14.5 69.8 6.8 1986 13.8 71.2 5.9 1987 6.6 1988 9.5 1989 10.5 1990 6.8 70.5 7.1 1991 10.1 1992 10.9 70.1 6.5 1993 6.0 1994 5.9 1995 65.7 7.3 2000 58.1 7.5 2005 51.0 7.4 2010 46.1 7-5 Source: Calculation based on the United Nations Sex and Age Distribution of The World Population, the 1994 Revision: United Nations Demographic Yearbook. 1993: World Bank's World Tables. various issues; World Bank's World Development Report. various issues. 34 Working Papers Annex: The Long Term Demographic Trend and Its Economic Implications Table B.18. Multiple Regression Analysis of Gross Domestic Saving/GDP on Child Dependency Ratio and Elderly Dependent Ratio Egypt: 1970-1992 Va;00000000g:f; 00fhriable ; tf Coefficient t-value Anniual Growth Rate&of GatrosDNmiifPesti ducts 0.45 4.18 Cbil d DendentR tio --29 4.01** elderlyf Deeident Ratio -5.28 -337** RSquare 83.5% W Rj ustedRSq0ar 77.3% ** Significant at P<0.0I ANNEXC Growth and Savings in Egypt: Lessons from the Empirical Growth Literature 1. Egypt's growth performance over the past decade has been disappointing. Real per capita growth of GDP averaged 0.2% per annum during the 1986-93 period. This rate of growth was a significant decline from Egypt's performance through much of the 1970s and early 1980s. Accompanying the modest growth have been low levels of investment, and even lower levels of savings. In 1994 investment was 18% of Egypt's GDP, and savings, only 6%. An important question for policy-makers, then, is how to improve growth, savings, and investment in Egypt. 2. There are many factors that influence these variables. This paper will focus on only a few, notably fiscal and trade policy. There has been a large amount of empirical investigation of growth, savings, and investment in recent years -- both outside and inside the Bank. The theme of this paper is, what can this new empirical literature tell us about Egypt's economic performance? How is Egypt regarded in this expanding literature? The policies examined here are clearly not the only ones that are important for Egypt and the recommendation are not meant to be exhaustive. Nevertheless, the empirical growth literature has found some very robust associations between macro and trade policies, on the one hand, and growth, savings, and investment, on the other. This work suggests that Egypt could gain a lot from reform in these areas. 3. The organization of the paper is as follows: The first section reviews very briefly the relationships among savings, investment, and growth from the point of view of economic theory. The second section then surveys some of the important contributions of the new empirical growth literature. Section three presents new econometric work. We take the variables that are found to be of importance to growth when considered separately and exarnine their relationship to growth in a multivariate, panel regression framework. In that section Indonesia is taken as a good comparator for Egypt, and it is shown that differences in fiscal and trade policies can account for most of the difference between the two countries' growth rates. (Indonesia has been a very good performer in recent years, with per capita real growth averaging 5.4% between 1986 and 1993.) The fourth section of the paper surveys recent studies that have examined specific interventions aimed at raising savings and/or investment. The final section summarizes the main conclusions and policy lessons. Savings, Investment and Growth 4. Savings, investment, and growth have been at the heart of development economics since the beginning of the discipline. This is not surprising because in practice the three are closely associated. Across countries and over time higher growth rates are associated with higher investment; higher investment with higher savings; and 36 Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature higher savings with higher growth rates. Furthermore, countries with higher levels of per capita income tend to save and investment a larger fraction of national income. While these empirical relationships are well established, there is little agreement among economists about how to interpret them. The facts are consistent with different theories or models, and thus it is difficult to sort out the exact inter-relationships and directions of causality. 5. This lack of clarity is not a problem for some policy issues. For example, it is clear that high inflation retards all three: savings, investment, and growth. A policy- maker does not need to know the exact mechanism through which this works to know that monetary policy that generates high inflation is bad. For other policy issues, however, the directions of causality are important. Some policies affect savings directly: for example, introduction of a mandatory, fully funded pension system. One would need to have confidence that savings causes investment and growth in order to believe that such a policy would accelerate a country's growth and development. 6. This section briefly reviews the inter-relationships among savings, investment, and growth from the point of view of economic theory. Take first the very high correlation between savings and investment. In a world of open capital markets savings and investment can diverge. Feldstein and Horioka (1980) showed that in fact there is very little deviation of savings from investment rates. Subsequent rescarch has demonstrated that this holds for both developed and developing countries. They interpreted this finding as evidence of barriers to international capital mobility, in particular that national investors were deterred by currency and political risks from placing large shares of their portfolios abroad. An alternative interpretation is that the same factors (country characteristics and policies) that determine savings also influence investment in the same direction. In this view, the factors that create a good climate for investment also create a good environment for savings. Whatever the source of this close correlation, what it means in practice is that it is difficult to sort out at the macro level the different determinants of savings and investment. A good model that explains investment across countries and over time will also be a good model that explains savings across countries and over time. 7. From the theoretical point of view, there is good reason to expect that higher savings (=investment) would lead to higher growth. Interestingly, in the neoclassical (Solow) growth model an increase in savings does not change the steady-state growth rate. But such an increase would cause the economy to growth more rapidly in the transition from one steady state to another. If transitions are gradual and can take a decade or longer, then what we observe internationally is consistent with the Solow model. In the "new" growth theories associated with Lucas and Romer there is an even clearer role for savings and investment. In these models an increase in the savings rate leads to a permanent increase in the growth rate. The basic mechanism in both these models is straight-forward: increased savings provide more resources for capital investment, which accelerates growth, at least temporarily and perhaps permanently. Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 37 8. Most economists believe that savings cause growth in this proximate sense. The problem is that there is little agreement about what causes savings and whether there is reverse causality from investment and growth, on the one hand, to savings, on the other. Angus Deaton (1995) explains how this could happen: 9. The international correlation between growth and savings rates comes from the response of growth to investment, as predicted by a variety of growth models. Savings responds passively to investment through mechanisms that are at present not well understood. A likely candidate is the saving behavior of firms or small entrepreneurs, who retain profits in order to finance investment. In any case, such saving is done, not by the mass of households, who play little part in the process of aggregate accumulation, but by a few relatively well-off people, or by firms. 10. This view suggests a virtuous circle. High investment leads to rapid growth, which increases the profits of firms and entrepreneurs. As long as growth is anticipated to continue, these agents save and invest a large fraction of their profits. In this way savings, investment, and growth are mutually reinforcing. Empirical Growth Literature 11. It is easy to see how a virtuous circle can keep going once it has started; or alternatively how a country can get stuck in a low saving, investment, growth trap. It is more difficult to see how a country can jump-start growth when it begins with low savings and investment. An empirical literature has attempted to address this question in recent years by examining what factors and policies are robustly associated with long- term growth. The basic approach in this literature is to conduct panel regressions to explain growth and investment across countries and over time on the basis of country characteristics and economic policies. If one accepts the "mutual causality" argument, then the policies that promote growth and investment will also be ones that promote savings. This is a large and growing literature; here we just single out the most significant and robust results. 12. Stanley Fischer (1993) examined the relationship between macroeconomic policies and long-term growth in a panel regression framework. The macroeconomic policies that he considered were inflation, the fiscal surplus, and the black market premium on the exchange rate. All of these variables were found to have significant relationships with growth of per capita GNP in the post-war period. Furthermore, the magnitude of estimated effects was large. An increase in inflation of 25 points (from, say, 10% p.a. to 35% p.a.) was associated with a 1 percentage point decline in per capita growth rate. For the budget surplus, an increase of 4 percentage points of GDP added an estimated 1 percentage point to the growth rate. Fischer also investigated whether the impact on growth came through investment or total factor productivity. He found that higher inflation and higher budget deficits were associated with both lower investment and lower productivity growth. Thus, macroeconomic instability -- as evidenced by high inflation, large budget deficits, and overvalued exchange rates -- was found to be detrimental both to the quantity of investment as well to the efficiency of investment. 38 Working Papers Annex: Growth and Savings in Egypt: Lessons from the Empirical Growth Literature 13. Easterly and Rebelo (1993) looked in greater detail at the effect of fiscal policies on growth and investment. For this work they put together new data on public investment that included both government investment and public enterprise investment for a large sample of countries. Both per capita growth and growth of private investment had robust associations as follows: * positive relationship with the budget surplus; . negative association with real government consumption minus education and defense expenditures; * positive association with govermment investment in transport and communication; and * negative association with total public investment. 14. These correlations together tell an interesting story. The countries that had rapid growth of private investment and per capita GDP were ones in which the government invested in infrastructure and financed it out of own revenue (that is, did not have large budget deficits). They were also countries that had less non-productive government consumption and less non-infrastructure public investment (i.e., in state enterprises). After controlling for these variables, various tax measures had no significant association with growth or investment. The latter finding is quite interesting and suggests that taxation is not an obstacle to investment if the resulting income is used for infrastructure investment or productive expenditures on education or protection of property rights. 15. One of the hypotheses that has been difficult to test is that more open developing countries grow more rapidly. The difficulty is that trade policy is hard to measure. Some studies use exports plus imports as a share of GDP as a measure of openness, but this approach is unsatisfactory as these are endogenous variables influenced by a range of factors other than trade policy. The size of a nation, for example, influences the share of trade in GDP. In a recent study Jeffrey Sachs and Andrew Warner (1995) objectively classify countries' trade regime as open or closed based on average tariffs on intermediate and capital goods, black market premium, and extent of distortion from export marketing boards. After controlling for other factors, they find that open economies grew 2.5 percentage points faster than closed economies in the post-war period. They also show that more open economies had higher investment rates and greater investment in human capital. 16. One problem with the Sachs-Warner approach is that it does not capture any gradation from more closed to more open: economies are found to be either open or closed. Dollar (1992) developed a continuous measure of trade distortion, based on each country's price level relative to the price level predicted by its factor endowments. Countries with price levels that were high relative to their endowments were measured to be more closed. This measure of trade openness was also found to be highly correlated Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 39 with per capita growth, after controlling for investment. The estimated impact of shifting from Egypt's level of openness to Indonesia's was more than 2 percentage points of growth. 17. A question of relevance to developing countries is the effect of foreign aid on investment, saving, and growth. There is a literature that addresses this question, but much of it is beset with econometric difficulties. The basic problem is that countries' growth may be affected by exogenous factors (i.e., a drought) that causes donors to provide more aid. That transitory factor could lead to a negative association between growth and aid that could be mistakenly interpreted as evidence that aid caused bad economic performance. In a recent paper Boone (1994) gets around this problem using an instrumental variable technique to identify the permanent component of foreign aid. Boone found no association between the permanent component of aid and either growth or investment. He estimated that foreign aid goes one-for-one to consumption! Unlike the other results cited here, this is primarily a cross-sectional result and may not accurately reflect what happens as aid is increased or decreased for a particular country. Nevertheless, it is a sobering result that raises questions about how effectively aid is being utilized. 18. A final strand of the empirical growth literature looks at the effect of institutional features on investment and growth. In this literature an effort is made to measure institutional quality in areas such as corruption, strength of property rights, and efficiency of the bureaucracy. Knack and Keefer (1994) found that countries with effective institutions had more rapid growth of investment and GDP. They interpreted their finding as evidence that good public institutions were an important part of a good climate for private investment. 19. In summary, the empirical growth literature finds that good public sector management is associated with faster growth of investment and GDP. Elements of good management are low inflation, low fiscal deficits, public expenditures targeted to productive areas, efficient bureaucracies, and open trade regimes. 20. An important question for all of this literature is the extent to which these associations can be interpreted as causal relationships that can be exploited by policy- makers. In other words, it is possible that something other than these policies causes growth, and that once growth is going it is easier for governments to have good public finances and open trade regimes. If that were the case, then policy-makers who deliberately instituted these policies in a low-growth enviromnent would be disappointed at the results. The empirical growth literature addresses this causality issue in a variety of ways and provides some comfort that it has found exploitable relationships; that is, policy reform should lead to better outcomes. 21. In the case of inflation, Bruno and Easterly (1995) focus on countries that have stabilized high inflation that was 40% p.a. or above. They find that growth resumes surprisingly quickly after stabilization and that stabilizers achieve a higher growth rate of 40 Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature GDP than they had before high inflation. In other words, stabilization does not just take an economy back to its pre-inflation growth path, but seems to place it on a higher growth trajectory. Bruno and Easterly interpret this finding as evidence that the inflation crisis creates a political environment in which a wide range of refonns -- not merely stabilization -- can be introduced. 22. Easterly and Rebelo also address the causality issue in their paper, focusing on public expenditure variables. They show that growth is associated with public expenditures; this relationship could be the result of growth increasing tax revenue and hence enabling more expenditure. If that were the main mechanism at work then all public expenditures would typically rise with growth. What is interesting about the Easterly and Rebelo results is that the significant relationships are between growth and only certain public expenditures -- infrastructure, education, and defense. These selective results suggest that the causality runs from productive expenditures to growth, and not from growth to all expenditures. 23. Sachs and Warner take the approach of identifying trade liberalizers and examining their growth records after liberalization. They identify 37 developing economies that have liberalized since 1984. The average growth pattern for these reformers is very interesting. Five to ten years before reform these economies were growing at an average per capita rate of 0.1% p.a. Immediately before reform growth was typically negative (averaging -0.7% p.a.). This finding is consistent with the Bruno- Easterly notion that economic crisis may lead to reform. The trade reformers in the Sachs-Warner sample grew at an average rate of 1.3% during the first two years after reform, and at 2.0% from the third year on. Thus the estimated impact of trade liberalization in the full panel -- about 2.5 percentage points of growth -- is consistent with the average time-series experience of countries that introduced trade reforms. It should be noted that Sachs and Warner do not control for a full range of other policies; they admit that their trade openness variable is probably proxying for more general macroeconomic reform that includes trade liberalization along with other policies. Egypt in a Comparative Perspective 24. Altogether the empirical growth literature provides fairly convincing evidence that economic policy reform can spur long-tern growth. The mechanism at work is probably that these policies spur investment by making it more profitable and also increase the growth impact of any given level of investment by making it more productive. Over the long term it may be that savings responds passively in the way envisioned by Deaton above. Or, some of these policies may spur savings directly. The most obvious example is the government budget surplus, which is a direct contribution to national saving. But other policies such as trade liberalization or reform of the public bureaucracy may also have a direct effect on saving by increasing the return to saving or by strengthening property rights. As noted, policy-makers do not have to worry about the exact mechanisms as long as these policies really succeed in improving investment, saving, and growth over the long term. Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 41 25. What implications does this research have for understanding Egypt's economic performance? How does Egypt compare with other large developing countries? These are the questions to which we turn in this section. The approach here is to replicate the main results of the empirical growth literature using panel data drawn from the World Bank data base. These results can then be used to analyze the key policies that have affected Egypt's performance and to identify promising areas for reform. 26. To replicate the main results of the empirical growth literature we collected data on the following variables that have been found to be of importance for growth: * GROWTH: Growth rate of real per capita GDP (World Bank data) * BUDGET SURPLUS: Budget balance relative to GDP (World Bank data) * GOVERNMENT CONSUMPTION: Government consumption relative to GDP (World Bank data) * INFLATION: Annual inflation rate (World Bank data) * BMP: Black market premium on the exchange rate (World Currency Yearbook) * OPENNESS: Dummy for open versus closed trade regime (based on Sachs and Warner classification) 27. In carrying out this kind of research, there is a tradeoff such that as the number of variables increases, the number of countries for which data can be found declines. For the variables above, we were able to collect data for 86 countries over the period 1966-93. By averaging over four-year periods, we obtained seven observations per country. There were some missing observations, but still we ended with 474 observations. The reason to average over a period rather than to take annual observations is that we are interested in medium- to long-term relationships, not year-to-year fluctuations. To further reduce the impact of the global business cycle, we have included time dummies for each period (except one). The panel approach used here means that we are not just picking up cross- country variations, but also variations over time for each country. The panel approach increases our confidence that we are looking at relationships that can be exploited by a policy-maker. 28. The base regression is reported in column (1) of Table C.l. The R-squared for this regression is 0.27, indicating that between one-quarter and one-third of the variation over time and across countries in GDP per capita growth rates can be explained by this small number of variables. The results are generally consistent with the literature discussed above. This consistency is encouraging because our regression includes fiscal, monetary, and trade variables. The studies noted above in general did not include all of these variables, leaving open the possibility of spurious correlation. The Sachs-Warner study, for example, did not control for other macroeconomic policies. The estimated impact of openness in that study was about 2.5 percentage points per year, a very large effect on growth. With other macro variables included, the coefficient declines to 2.0; 42 Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature still it is highly significant, providing additional support to the view that an open trade regime assists developing countries to grow rapidly. 29. The regression also confirms that a large fiscal deficit, high government consumption, and high inflation all have a negative relationship with growth. For each of these variables, the t-statistic is above 2.0. With these other macro variables included, the relationship between the black market premium and growth is negative, but small and statistically insignificant. All of these macro variables are inter-related and in practice correlated. To get such clear relationships is quite interesting. In the other regressions in columns (2) to (6) of Table C.1, the variables are dropped one-by-one in order to assess the robustness of the results in light of inclusion and exclusion of different variables. In general the estimated relationship are quite robust, though coefficients vary a little bit depending on the variables included. Table C.l. Growth and Policies in 86 Countries, 1966-93 Dependent variable: growth rate of real per capita GDP n = 86 countries, 7 time period (four-year averages) Included observations = 474 (i1) (2) (3) (4) (5) (6) lBudgetSurplus 06 -- 08 08 .06. .0 (2.34) (3.41) (319) ~(2.27). (.9 Government Consumption -0O6 -.09 -- -05 -.07 .02 (2.16) (3.29) (0.82) (2.38) (0.77) Inflation -.01 -.02 I-. 0 - -.02 .. -.020 (2.63) (3.42) :(2.35) (3.47) (3.111) BMP -2,flE-6 -14.E-6 -3.OE-6 -4.9E-6 (0.8) (0.67) (1.31) (2Al1) OV0M) Openness .2.2.8. 20 (7.00) (7.29) (6:69) (7:21) (7.05.) R2 .~~ ~ ~~~ ~ ~~~ ~ ~~~27 .26 .26 .26 .27 I.8 30. The black market premium has been included in quite a few growth studies, and is sometimes interpreted as a measure of trade openness and at other times as an indicator of macroeconomic policy. As noted, the variable is insignificant in the base relationship. Its negative association with growth becomes stronger when inflation [column (4)] is dropped, suggesting that it is something of a proxy for inflationary policies. The coefficient on BMP changes only modestly when the openness dummy is dropped [column (6)] confirming that it is in fact a poor proxy for trade openness. 31. The results of the base regression can be interpreted as follows: (1) a high fiscal deficit is bad for growth, even after controlling for inflation, suggesting that such deficits draw domestic savings away from more productive uses. (2) High government consumption has a negative impact on growth, even after controlling for fiscal deficit and inflation. This presumably results from high government consumption discouraging Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 43 private sector activity. (3) Trade openness as measured by Sachs and Warner has a major impact on growth, even after controlling for a wide range of macroeconomic policies. 32. Growth regressions typically include the starting level of per capita GDP as an explanatory variable. In general, after controlling for other policy variables, the coefficient on initial income level is negative (conditional convergence). That result is commonly found in cross-section regressions. In this data set, with a panel regression on four-year averages, we found a negative relationship between growth and initial income level, but it was statistically insignificant and not reported here. 33. What can these results tell us about Egypt's growth performance over the past decade? In the empirical growth literature the East Asian economies are often taken as a reference point to which other countries' policies and growth records are compared. In this paper we take Indonesia as a relevant comparator for Egypt. The two countries are similar in a number of important ways, but different in others. Table C.2 provides some basic descriptive data on the two countries. They are large countries with similar per capita GDP and social indicators. They differ in that Indonesia has been growing much more rapidly in recent years: 5.4% per capita growth in the 1986-93 period, compared to 0.2% for Egypt. Indonesia is also far more involved in international trade and investment, with exports of 22.9% of GNP and foreign investment inflow of 4.2% of GNP in 1994. Comparable figures for Egypt were 8.1% and 2.3%, respectively. Finally, Egypt receives far more foreign aid than Indonesia, 6.4% of GNP in 1994 compared to 1.0% for Indonesia. Table C.2. Egypt and Indonesia, Comparative Indicators EygKt Indonesia Population (m.) 56.8 190.4 Per Capita GNP (1994 5) 3,720 3,600 Gini Coefficient 32 32 Infant mortality (per 1,000, 1994) 52 53 Primary Shbool Enrollment 1993 Female 89 112 Male 105 11b Secondary School Enrollrment 1993 Female 69 39 Male 81 48 Investment/ODP 1994 18% 29% Saving/(DP 1994 6% 30% DFI/GNP 1994 2.3% 4.2% ODA/ONP 1994 6.4% 1.0% ExporsGNP 1994 8.1% 22.9% Source: WDR 1996. 34. The panel regressions above can help shed some light on the different growth performance of the two economies. Table C.3 reports the values of the explanatory variables for the two economies and the estimated impact that these differences have on the growth rate. According to this analysis, the key difference between the two countries is Indonesia's openness to foreign trade. The Sachs-Warner characterization of Egypt as a closed economy is based on level of tariffs for capital goods and intermediate products 44 Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature and also on the pervasiveness of non-tariff barriers, including administrative red-tape. In our base regression the estimated impact of closedness is 2.0 percentage points of growth per annum. Table C.3. Growth and Policies in Egypt and Indonesia, 1986-93 Egypt Indonesia Estimated JImact Growth of ral per capita GDP 0.2 5.4 - Budget Surplus; i;:S : : 0 : : -5.4 -A.0 03 (%GNP) Govt. Consumption 141 9.5 X03 (% OINP) nflation 16.5 7.7 0.1 BMP 11.6 10.7 t0.0 O t3pn;;; f :; E ;; ; :; ; ;E; : j: ; Ano yes 20 Totab impact of plicies i: 35. Furthermore, Egypt has a much larger government than Indonesia, despite the fact that the two countries have similar levels of income and social indicators. Government consumption in Egypt averaged 14.1% of GDP during 1986-93, nearly 50% higher than Indonesia's 9.5%. The estimated negative impact of Egypt's larger government is 0.3 percentage point of growth. Egypt's budget deficit averaged 5.4% of GDP over this period, compared to 1.0% in Indonesia. The estimated impact of Egypt's larger deficit was 0.3 percentage point of growth. The difference in inflation (16.5% in Egypt compared to 7.7% in Indonesia) also contributed 0.1 percentage point to the difference in growth. In the area of black market premium, the two countries had identical performances; furthermore, this variable was not found to be important after controlling for the other variables. 36. Altogether the differences in macro policy and trade openness account for an estimated 2.7 percentage points in growth. The actual difference in growth was 5.2 percentage points. Thus the included variables account for more than half the difference. What can we make of this exercise? Frankly, the point estimates for the impact of different policies should not be taken too literally. The coefficients vary somewhat depending on which additional variables are included in the analysis. But the general point is quite robust. Indonesia is an example of a low-income country with very good macro policies and an open trade regime in the 1986-93 period. Careful analysis suggests that Egypt could buy itself nearly an additional 3 percentage points of growth through a reform program that emulated some of the key policies from Indonesia, notably trade liberalization, a reduction of government consumption, and a reduced fiscal deficit. Furthermore, the trade and fiscal policies are probably linked in that government reform may be needed to truly bring about a liberal trade regime. 37. What are some of the reasons to be skeptical that trade and public sector reform might generate the predicted increase in growth? In general the reason to be skeptical of this kind of econometric work is that there may be omitted variables that are crucial for Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 45 obtaining particular results. The Bank's East Asian Miracle Study (1993) analyzes in considerable detail the different factors that have contributed to growth in general and to East Asia's success in particular. One of the reasons why Indonesia is such a good comparator for Egypt is that the two countries are similar in terms of some of the additional variables that one might want to consider. For example, human capital in the two countries is quite similar, and in fact Egypt has an advantage in terms of superior secondary and tertiary school enrollment rates. There is no reason to think that Indonesia's human capital provides an advantage over Egypt that would affect the supply response to reforms. 38. Another factor that has been considered in the literature is the quality of the bureaucracy in implementing policies and providing services. Related to this is the controversial issue of the degree of corruption. There have been a number of recent efforts to measure the quality of the bureaucracy and the extent of corruption, notably the cross-country rankings developed by IRIS (0-6 scale with a lower number more inefficient or corrupt). In terms of bureaucratic quality in 1990, IRIS ranked Indonesia a 1 and Egypt a 3. In terms of corruption the rankings were 0 for Indonesia and 2 for Egypt. This is not to say that bureaucratic quality and corruption are not important issues for development efficiency. However, in explaining differences between the recent growth rates of Egypt and Indonesia, these factors do not seem to be important. Thus, the Indonesian experience provides important evidence that strong macroeconomic policies and trade openness provide a good foundation for growth even if the quality of institutions is low. More recently, Vietnam provides additional evidence in this direction. Vietnam is another country that scores very low on the IRIS rankings (2 for bureaucratic quality and 3 for corruption). But it has also gotten very good results from a program of fiscal reform and opening to the international economy. Direct Interventions on Savings and Investment 39. The previous section provided evidence that Egypt could spur growth by reducing government consumption and opening the economy to foreign trade. Because growth, investment, and savings are so closely associated in the data, the growth rate could be replaced in the previous analysis with savings or investment and similar coefficients obtained. In other words, we do not know for sure the mechanisms through which macro and trade policies operate. Their effect on savings, however, is probably indirect: policy reform directly spurs investment and growth and these in turn influence savings. The empirical literature provides quite a bit of confidence that these policy reforms are an effective way of spurring savings in the long run. But the government may also want to consider other, more direct interventions that work specifically on savings or investment. There has been quite a bit of research on different instruments, and this section summarizes some of the main findings from this work. 40. The most obvious and direct instrument for the government is its own savings. It is interesting that in theory an increase in government saving does not necessarily increase national savings. The reason for this is that private agents might adjust their 46 Working Papers Annex: Growth and Savings in Egypt: Lessons from the Empirical Growth Literature own saving in response to the government's actions. In some circumstances changes in private saving will fully offset changes in public savings (Ricardian equivalence). In practice, however, Ricardian equivalence has been typically rejected. Edwards (1995) reports offset coefficients for a large number of developing and developed countries, most of which are around 0.5. This implies that increases in government savings are an effective way of raising national savings, though the increase will not be one-for-one. This work is consistent with the regression results in the last section. That government consumption and fiscal deficits are bad for growth implies that government savings is good for growth (which would not be the case if Ricardian equivalence were the rule). 41. A second instrument that the government may consider to increase savings is financial sector reform. In repressed financial systems real interest rates for savers are often low and in theory a higher interest rate may attract more saving. But a higher interest rate will increase the return on existing financial savings producing a wealth effect that could operate in any direction. In practice there is little evidence that aggregate savings responds to higher interest rates. 42. A more promising instrument concerns pension reform, in particular replacing pay-as-you-go (PAYG) schemes with fully funded pension systems. Econometric evidence for developing countries indicates that government social security benefits (typically through PAYG schemes) lower private savings [Edwards (1995)]. This finding implies that the replacement of PAYG with fully funded pensions should raise national savings. Evidence suggests that this is exactly what happened in Chile and that the development of a fully funded system there was an important factor contributing to the large rise in national savings. Similarly, there is evidence that Singapore's fully funded Central Provident Fund boosted aggregate saving. 43. Turning to investment incentives, a range of policy instruments has been used in both developed and developing countries to influence aggregate investment and its distribution across sectors, assets, and time by affecting the rate of return on capital. In theory, investment incentives can be justified by the existence of market failures that lead to socially insufficient investment-for example, the public-good nature of some investments in research and development (R&D), imperfect information among financiers about the quality of projects, imperfect competition, and scale economies. 44. Tax concessions are by far the most commonly encountered investment incentives. They take a variety of forms: preferential tax rates for specific types of investment, tax holidays, accelerated depreciation allowances, investment tax credits, and so forth. At a very general level, tax incentives may be characterized in three ways. First, they may be automatic (available to any firm or investor meeting specific standards about the type, size, and location of the investment) or discretionary (granted on a case- by-case basis). Automatic incentives are more transparent and certain than discretionary incentives and are thus likely to be more effective than the latter, which are comparatively more responsive to noneconomic factors. Second, tax incentives may be temporary or permanent; temporary incentives are more likely to affect the timing of investment than Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 47 the long-term capital stock. Third, tax incentives may be general or selective (with eligibility determined by considerations such as the type of asset or economic sector). In general, and in the absence of market failures, selective incentives are bound to distort the allocation of investment across sectors or assets. 45. The recent investment literature exarnines the impact of tax incentives in both industrial and developing countries (Feldstein, 1987; Jorgenson and Landau, 1993; Cummins, Hassett, and Hubbard, 1995). Most empirical studies of industrial countries find only small effects (Chirinko 1993). In developing countries, the effectiveness of investment incentives is severely limited by the weakness of the tax administration system and by market distortions such as credit rationing or administrative intervention in the allocation of foreign exchange. In the presence of such conditions, tax concessions may affect the allocation of rents more than they affect investment decisions (Shah, forthcoming). 46. Some conclusions may be drawn about the effectiveness of the specific types of tax incentives most commonly used in developing countries. Preferential corporate tax rates are shown to be poor instruments for promoting new investment, because the tax concession is not directly related to the volume of new investment. Its impact on fiscal revenue loss is therefore often larger than its impact on extra capital. An extreme form of such an incentive is the granting of tax holidays. These are relatively common in developing countries and pose the added problem of encouraging the shift of taxable income into the firms enjoying the holidays (Mintz, forthcoming). Investment tax credits for R&D and the acquisition of equipment embodying advanced technologies are, by contrast, analytically more defensible and empirically more effective. 47. The use of fiscal incentives also presents several practical problems that are likely to be particularly severe in developing countries. It is difficult to identify and measure the divergences between social and private rates of return on investment that justify fiscal treatment, and the administration of any fiscal incentives places significant burdens on the tax management system. Tax concessions encourage lobbying by potential beneficiaries. The loss of fiscal revenue-implicit in incentive schemes-has an obvious opportunity cost. And attempts to fine-tune the tax system to direct the allocation of resources may lead to profound systemic distortions (Bird, 1992). 48. Stability and predictability of the tax regime are thus prerequisite to the effectiveness of investment incentives. Indeed, a regime of stable corporate tax rates at international levels offers greater promise for investment than a system of large incentives with high and unstable rates (Shah, forthcoming). In sum, from the perspective of investment, tax incentives are probably less effective than the elimination of disincentives such as unstable policies, infrastructural deficiencies, and inadequate regulatory codes. 48 Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature Summary and Conclusions 49. Cross-country evidence indicates that Egypt would grow more rapidly if it reduced the size and scope of its public sector and liberalized trade. The two reforms are generally complementary, as trade liberalization requires a streamlined and less interventionist bureaucracy. It should be emphasized that the empirical growth literature has identified an important role for the government. Education expenditures and public investment in transport and communications generally have high returns. Egypt's problem is that it has too much activity in other, non-productive areas. 50. In comparing Egypt with high-growth developing countries, one of the striking differences is in the size of the government. Government consumption in Egypt was about 50% higher than in Indonesia during the 1986-93 period. Egypt did well in terms of investment in transport and communication. In the Easterly-Rebelo data set, it had one of the highest levels of such investment. But it also had a large amount of other public investment, which has been found to have a negative impact on growth in empirical studies. Thailand provides a useful comparison: it had total public investment of 7.9% of GNP in the 1980s of which 2.7% of GNP was in transport and communication. The figures for Egypt were 18.3% and 4.4%, respectively. Thus the public sector seems to be doing too much in Egypt, in terms of both consumption and investment. In Section 3 it was shown that fiscal reforms that reduced government consumption and the fiscal deficit to the levels prevailing in Indonesia would add an estimated 0.6 percentage points to long-term growth. 51. Of greater importance is the fact that objective studies continue to find Egypt to be an economy closed to foreign trade. The estimated impact of opening up the economy is 2.0 percentage points of growth. Egypt is similar to other large developing countries, such as Pakistan and India, in that it was able to grow fairly well under an inward- oriented trade regime for a decade or more. The evidence from other countries, however, is now quite persuasive that the inward-oriented strategy runs out of steam eventually. Really opening up the economy to foreign competition will be necessary if Egypt is to significantly increase its long-term growth. This liberalization will need to be coordinated with other reforms. On the macro front, for exanple, serious trade liberalization is generally accompanied by devaluation of the real exchange rate, and some devaluation of the nominal rate may be needed to bring this about smoothly. Furthermore, dismantling on non-tariff barriers to trade is likely to call for streamlining of the bureaucracy, which is consistent with the notion of reducing government consumption noted above. 52. The cross-country experience reveals the importance of fiscal and trade reform for spurring growth, savings, and investment. Without these macro reforms, other measures are not likely to have much impact. If the macro reforms are pursued, however, then other measures can support them. In particular, there is a growing amount of evidence that fully funded pension schemes can raise domestic savings. Thus pension reform can Working Papers Annex: Growth and Savings in Egypt: Lessonsfrom the Empirical Growth Literature 49 be a nice complement to other macro reforms that aim primarily to increase growth and investment. 53. In conclusion three solid and inter-related pillars on which to build a stronger growth, savings, and investment performance in Egypt would be: * trade reform that increases openness to the international economy and spurs competition; * fiscal reform that reduces unproductive government consumption and investment; and * pension reform that yields a fully funded system, bolstering savings and taking pressure off government expenditure. ANNEXJD Pension Reform, Contractual Saving, And Capital Markets 1. Promoting high saving rates is a central policy concern due to the links between saving and capital accumulation, which affects economic growth. This chapter discusses actions in four areas that could increase the rate and improve the composition of private saving: reforming the pension system, restructuring the insurance industry, developing capital markets, and improving the tax system. Private saving sources include the domestic household and business sectors and foreign saving as well -- the share of private capital in capital flows to developing countries has been increasing. 2. Section B discusses pension reform. There is increasing evidence that generous pay-as- you-go state pensions tend to depress household saving rates.14 Econometric work for this paper also shows a negative correlation between the private saving rate and government spending on social security as percent of GDP. The argument is that mandatory saving is likely to increase household saving. Although mandatory saving may simply replace voluntary saving resulting in negligible impact on capital formation, the crowding out effect may only be partial because people may not be saving as much for old age (myopia) or because constraints on consumer borrowing may limit dissaving against pension benefits in the future. Replacing pay-as-you-go systems with fully-funded systems increases government saving (or reducing government dissaving) by introducing a transparent process in government financing of its expenditures. 3. Section C looks into the insurance industry. While the impact of contractual saving institutions such as the life insurance industry on the level of saving is arguable, what is unambiguous is the shift in the composition of saving towards long term financial assets (invested by financial institutions) instead of physical assets such as real estate, precious metals, and land (invested by individuals). With well developed contractual saving institutions, there is greater availability of long term funds for borrowing by households (e.g., for housing) and private corporations (for investment projects). In the US, for example, contractual saving institutions accounted for one-fourth of outstanding mortgage-backed securities in 1993. In addition to making long term funds available, the insurance industry supports pension reform by providing specialized products and services such as annuities and life/disability insurance. 4. Section D examines the development of capital markets. In many countries, capital markets have developed in parallel with and have supported two initiatives that have an important impact on private saving: pension reform and privatization. To the extent that pension reform and privatization programs increase household and business saving, respectively, capital market development has an impact on the level of private saving. In addition, capital markets, under certain conditions, attract private foreign saving. The nature of capital flows to developing countries has been changing towards greater private sector share and increasing share of foreign portfolio investment. The experience in Asia has been that these capital flows have been utilized to augment domestic saving, rather than replace it. It should also be noted that the share of MENA in private capital flows to developing countries was only 2.1% during the period 1990- 14 Sachs, Achieving Rapid Growth: The Road Ahead for Egypt Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 51 94. Efficient capital markets would be part of the institutional framework that would encourage the flow of overseas capital. 5. Finally, Section E deals with some of the tax policies that affect saving of both the household and business sector. In an attempt to promote saving, countries have offered tax incentives, but in so doing created distortions in the system. This section looks at the tax treatment of the most important saving vehicles and the impact of inflation on effective tax rates. The Social Insurance System Background 6. The Egyptian Social Insurance System (SIS) in Egypt has an important impact on several macroeconomic and welfare issues such as saving, redistribution, capital market development, social protection and public finance. In 1994/95, contributions by employers and employees to the pension fund was 3.5% of GDP, equivalent to about one-half of gross domestic saving. Benefit payments and fiscal transfers to the pension fund were 2.5% and 1.5% of GDP, respectively. SIS inflows (including investment income) less outflows was 5.3% of GDP in 1994/95, and accumulated reserves were 33% of GDP in mid-1995 -- by way of comparison, combined market capitalization in both stock exchanges was about 13% of GDP at end- 1995. Table D.1: Social Insurance System Indicators (percent of GDP) Contributions Benefits Transfers from Annual Surplus Aeetmulkated .___ ___ ____ ___ ______Treasury Stirplus 1985/86 5.1 2.5 1.2 5.8 38.2 1986f87 4.4 2.3 1.0 5.1 36.6 1987/88 4.3 2.3 0.9 5.6 36.2 1988/89 4.0 2.2 1.0 4.8 33.9 1989/90 3.6 2.0 0.9 4.6 31.7 1990/91 3.5 2.1 1.0 4.5 31 .9 1991/92 3.1 2.1 I.1 3.6 29.0 1992/93 3.3 2.2 1.3 4.5 30.2 1993/94 3.5 2.5 1.5 5.4 32.5 1994195 3.5 2.5 1.5 5.3 33.1 Source: Mission calculation based on data from Ministry of Social Affairs 7. Because the surplus from the SIS is invested in government projects and, in the recent past, in public enterprises, the SIS effectively works as a pay-as-you-go system. The notional reserves of the SIS at the NIB represents public debt due in the future when the SIS starts incurring a deficit. One could argue that access to cheap funds increases the propensity of governments to spend and crowds out the private sector -- during the period 1970-94, public sector share of total investment in Egypt averaged about 68%. The saving effect of a mandatory scheme is therefore offset by increased government spending. On the other hand, a truly fully funded system that invests in financial instruments at market rates has the potential of developing financial markets -- even if the system invests in govemment securities, government borrowing becomes a transparent process. Funded systems that strengthen the link between benefits and contributions (such as defined contribution plans) eliminate many of the distortions of a pay-as-you-go system. 52 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets 8. The Egyptian SIS provides pension benefits and insurance against disability, death, and loss of earnings due to unemployment or illness. The SIS covers a high proportion of the work force15 -- about 83%i6 in 1994. The ratio of pensioners17 to contributors (the system dependency ratio) is high relative to other developing countries, and relative to the ratio of population over 60 to population between the ages of 20 and 59 (demographic dependency ratio). Part of the explanation lies in the large number of survivors (mainly spouse and children) -- the ratio of old age pensioners and surviving spouses to contributors gives a ratio of 19.5%, which is lower than the average for MENA. Table D.2 gives some comparative public pension scheme and demographic indicators. Table D.2: Comparative Public Pension Scheme and Demographic Indicators (based on latest years data is available) (percent) Region Contributors to Work Pensioners to Old Age Pensioners to Population Over 60 to Force Contributors Population Over 60 Population 20-59 OECD 93.9 39.2 84.1 34.0 LAC 38.3 21.0 30.8 18.0 MENA 41.3 27.5 57.5 13.5 ASIA 23.5 11.4 22.3 13.5 AFR 6.4 8.5 24.0 12.5 EGYPT (1995) 83.1 3 778 34.3'9 14.4 Source: Ministry of Social Affairs; World Bank Contributions 9. The SIS has very high contribution rates on taxable wages. Social security contributions must be paid on workers' basic and variable wages with a maximum taxable amount for each category of wages (LE 450 per month for basic and LE 500 per month for variable). The average total taxable wage in 1994/95 was LE 267 per month for civil service employees and LE 149 per month for non-civil service employees, way below the maximum taxable amounts. In 1994/95, the basic wage was 42% of total taxable wage for civil service employees and 46% for non-civil service employees. 10. For those covered under the main program (Law 79/47), total contribution rates20 on basic wages are 41% for private sector employees, 39% for public enterprise employees, and 36% for government workers. Workers contribute 14%, the Treasury 1%, and the employer pays the remainder. The contribution rates on variable wages are lower, since no contribution is assessed on job exit indemnity. Nonetheless, the general complaint is that the contribution rates are too high in relation to the benefits. Table D.3 provides the contribution rates for both basic and variable wages for different types of workers covered under Law 79/47. Is Work force includes Egyptians working abroad, which in 1995 numbered 2.7 million out of a total work force of 19.1 million. 16 Contributors to the Social Insurance System during 1994/95 to labor force in 1995. 17 Pensioners covers old age, disability, and survivors. The number of survivors receiving benefits is about three timnes the number of old age pensioners. s Includes all beneficiaries receiving a regular pension, including surviving dependents. 19 Old age pensioners do not include surviving dependents; if surviving spouses are included, the ratio increases to 73.9%. 20 Different contribution rates are assessed for different types of benefits, such as old age pensions, disability, death, unemployment, and end of service indemnity. Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 53 Table D.3: Contribution Rates for Social Insurance (percent of wages) Basi0 Wage Variable Wage (up to Lt 450i'nlonth.) L . (up to LE 500/month) Program Worker Emnployer Gov't Total Worker Emnpoyer .ov't Total Old-age, Invalidfty, and Death 10 15 1 26 10 15 1 26 Unemnployment --- 2 - 2 2 - 2 Work lhjury --- 3 --- 3 3 - 3 Sickness (medical) 1 I 4 ... 5 1 4 - 5 Job Exit lndernity 3 2 5 - - - 0 Total 14 26 I 41 11 24 1 36 Public Sector EtE lm'ses Old-age, Invalidity,-and Death 10 15 1 26 10 15 1 26 Unemployment 2 --- 2 - 2 -- 2 Work Injury - 2 -. 2 ..- 2 - 2 Sickness (medical) 1 3 --- 4 1 3 4 Job Exit Indemnity 3 2 -- 5 --- 0 Total 14 24 1 39 11 -22 1 34 Governmnent Offices Old -age, Invalidity, and Death 10 iS 1 26 10 15 1 26 Unemnployment . - -- 0 . - . -_ 0 Work Injury - . I l -_ I _ I Sickness (nedical) 1 3 --- 4 1 3 4 Job Exit Indemnity 3 2 --- 5 -- --- 0 Total 14 21 1 36 11 19 1 31 * Employers> contribution rate is reduced by half (public employers) or one-thitd (ptivate employers) if employer provides qualifying medical care. *•*Employers' contribution can be reduced by 3% if employer provides cash and medical benefits to own employees. Source: Ministry of Social Affairs Eligibility Criteria 11. The eligibility criteria for retirement is a major variable in determining the financial viability of any defined benefit plan. The normal retirement age for Egypt (age 60 for both men and women under the major programs of Law 79/47) is lower than that of OECD countries (whose average retirement age is 64.4 years for men and 62.9 years for women), but is in line with many developing countries. Given the increasing productive life spans, many countries are improving the financial viability of their pensions systems by raising the retirement age. 12. Redistribution occurs in two main areas. First, pensioners from the agricultural sector -- about 45% of total pensioners in 1995 -- are provided the equivalent of minimum wages (LE 45 per month). Given the very low flat contributions by agricultural workers and farmers (currently LE 1 per month), the benefits are higher than what the contributions would have been able to finance. Second, within Law 79/47 which covers civil service employees and the formal sector, 54 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets the pension formula sets a minimum pension for those with a certain number of years' contribution. Benefits 13. The ratio of average monthly pensions to average covered monthly wages (replacement rate) indicates that social insurance provides monthly benefits that represent a high percentage of average taxable wages, 102% for civil service employees and 147% for non-civil service employees in 1995. In 1986, the replacement rates were 89% and 103% respectively. Inflation adjustments on pensions have been greater than the growth of average wages of contributors, with the differential greater for non-civil service workers. 14. The determination of the pension amount at retirement depends critically on the treatment of wage and price inflation. The method of computing the pensionable amount for variable wages is particularly sensitive to the inflation rate, since the basis used is the average wage during the covered period adjusted by an annual inflation factor of 2%. At a 10% annual inflation rate, the pensionable variable wage for a 30 year contribution period would only be 55% of real wages. 15. In addition, there is no automatic indexation of benefit payments during retirement to protect the value of the benefits from inflation. Prior to 1987, the purchasing power of pensions declined significantly because the Government did not raise basic pensions regularly. However, over the past ten years, annual basic pension adjustments were made to prevent the decline of basic pension benefits in real terms. Over the period 1987-96, the legislature-mandated adjustments preserved the purchasing power of the basic pensions for all but three of the past 10 years. However, pensions on variable wages are not automatically indexed or adjusted by the legislature. Thus, in addition to the adverse impact of inflation on the computation of the pensionable variable wage on which the initial pension benefits are based, the resulting variable wage pensions are not protected from inflation during retirement. Assuming 30 years contribution, 10 years of retirement and 10% inflation, the pension would only be about 14% of wage at retirement adjusted for inflation. Financial Condition 16. During the period 1985-95, the SIS has been operating on an operating surplus (contributions less outlays) due to the ratio of contributors to pensioners. Because the inflation adjustments on basic pensions approved by Parliament are financed from general revenues, total receipts by the SIS come from three sources: wage taxes, investment income, and general government revenues (i.e., transfers from the MoF). Table D.4 shows the shares of the different sources of financing for the SIS. Egypt's public pension receipts to GDP is closer to the ratios of the mature economies of OECD and the transitioning economies of ECA, and is much higher than ratios of Asia and LAC. Egypt also has the biggest differential between the ratios of public pension spending to GDP and public pension receipts to GDP. Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 55 Table D.4: Public Pension Spending Indicators (latest year when data was available) (percent) Region Public Pension Public Pension $hare of Receipts from Spending to GDP Receipts to GDP Wage Taxes Investment Income General Revenues OECD 9.2 9.1 57.4 11.0 35.1 LAC 2.0 2.4 63.8 23.0 t3.0 PCA 8.0 7.9 68.6 0.2 16.3 MENA 2.8 4.4 63.1 17.5 19.4 AFR 0.5 0.7 77.8 20.3 1.5 ASIA 1.9 5.2 61.3 24.2 14.1 Bgyt (I 995) 2.5 7.7 45.3 35.6 19.1 Note: Regional numbers are simple averages of sample country data Source: World Bank 17. Operating surpluses of the insurance funds are invested in the National Investment Bank (NIB), a government owned institution which finances primarily government projects. During the period 1980-90, NIB paid interest on the funds at 5% to 6% per year, during which period the CPI was increasing by about 18% per year. Had the insurance reserves been earning a zero real return, the reserves would have been 60% more than the amount at end- 1995 -- the reduction in value of reserves over the ten year period is equivalent to more than seven years of (1995) benefit payments. This represents significant erosion of the real purchasing power of the reserves and a net subsidy to NIB. 18. Starting July 1992, NIB raised the interest rate on incremental social security funds (including reinvested reserves) to 13%, which is the current rate. This rate is higher than current bank term deposit rates (12%) but lower than interest on NIB investment certificates (17%). Average inflation during the period 1992-96 was about 10%, providing a positive real return on incremental and reinvested funds. National Investment Bank 19. NIB is a financial institution controlled by the Ministry of Planning and is responsible for evaluating and financing the Government's investment program. Funds are disbursed to various public sector entities, including central and local governments, service agencies such as hospitals and universities, and economic agencies such as the public utilities. In the past, the NIB also financed the PEs, but since the implementation of Law 203, NIB has gradually stopped lending to PEs, although it still has significant exposure to them. NIB provides long term financing to public investment projects approved by the Parliament and charges interest to recover interest and operating costs. In the 1980s, interest rates charged to government projects and PEs were negative in real terms due to cheap funds from the SIS. However, with the increase in interest rates on social insurance funds to positive real rates since 1992, the NIB has also increased interest rates on its lending, thus raising the hurdle rates for investment projects. Currently, interest rates paid by NIB for social insurance funds and postal savings funds are 13% and 13.25%, respectively. 20. The SIS is the main source of funds for NIB. As of June 1995, the amount of SIS reserves in NIB was LE 67 billion, about 33% of GDP. Social insurance funds (new and reinvested) accounted for 68% of the funds sources of NIB during the past five years. 56 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets Investment certificates (10 years with 6 month coupons) issued to the public made up another 29% of funding sources, while deposits of postal savings fund contributed 3% (Table D.5). Over the past ten years, SIS funds available for lending averaged 4.1% of GDP -- SIS was a major financier of public investment. public investment to GDP and total investment to GDP With a policy objective of increasing the share of the private sector in total investment, the role of the NIB and the utilization of SIS funds need to be reviewed. Table D.5: NIB Funding Sources (LE million) 1 9091 1 991-92 1992-93 1993-94 994-95 Total. Social lnsurance Fund ;4,308.6 5283.1 16,37.73 7,671.1A 9,77.0 7 32928,1 Postal Savins Fundd 110.40 166.6 174.0 439.1 657.5 I1,5472 lnvestitnt Certificates 717.7 256.8 2,18100 4,6879l 5,994.3 13,837.7 Bank's.Surplus: 9338 1,536.9:0 W100.0 983.8 3,475.2 7,029.7 CTedif: t 268. 26.5 : 628.3 :52.8 810.3 3,06246 Others 1038.6 1 32. 84.4 54.6 333.9 744.3 Total$; ;; 0t00 t;0 :0000 ;t 0000 X6,477.4 8,203.3, 9,555.4 14,365.3 20,4 8.2 509,1 49.6 Source: National Investment Bank 21. The impact of the utilization of SIS funds on investment and growth depends on the response of the Government to the availability of cheap and readily available sources of funds. If the government would have made the same investments but used a different financing scheme (e.g. sale of government bonds), government utilization of SIS funds would allow potential bond buyers to shift their resources to private investment. But privileged access to public pension funds is a less transparent way of financing government projects, without the disciplining effect of capital markets. The negative real interest rates during the 1980s probably encouraged more public investments than the case where market rates were charged. In many countries, publicly managed funds required to invest a major portion of its portfolio in government securities (including lending to PEs) tend to charge below market rates, imposing a hidden tax on contributing workers. On the other hand, privately managed funds tend to achieve higher real rates of return. Figure 1 shows comparative returns for selected pension funds during the 1 980s. Working Papers Annex. Pension Reform, Contractual Saving And Capital Markets 57 Figure D.1: Comparative Performance of Selected Pension Funds Peru -37A1918 Turkey -23.8 198448) -23.4 (198048) Venezuela -15.3 519809) Egypt1 111.7108149) Ecuador -10.0 (1980-8S) Keny -3.8 (1980-90) India (1W0-90) 0-3 Singapore (190-0 3.0 Mlaysia (198090) 4.8 U.S. Oasi PubDcaDy nunaged (10090) 4.8 (1980-90) _ S . 7 NethedUands (occupational) Privately nanaged U.S.(oecup . ' (198040)- - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~8.8 U.K. (occupaonal) (1981-90) i2 Chile (AFPs) (1El1.90) -40.0 0.0 -25.0 -20.0 -15.0 40.0 4.0 0.0 5. 10.0 15.0 Percertage rabt of return afer Inftison 58 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets Private Pension Plans in Egypt 22. To complement the social insurance system, voluntary pension plans have emerged in Egypt. These plans are typically set up by employers on a defined benefit basis, with contributions made by both employers and employees. Private pension plans in Egypt are govemed by Act No. 54 (1975) and Executive Regulations Decree 78 (1977), and are under the supervision of the Egyptian Insurance Supervision Authority (EISA). Efforts are under way, with Bank support in the form of an IDF grant, to update the law and introduce appropriate regulatory and supervision arrangements. 23. The number of private pension plans have been increasing significantly -- as of June 30 1995, there were 504 plans compared to 330 in 1991. The reserves held by these funds were LE 3.3 billion as of June 30, 1995, and while these reserves represent only 1.6% of GDP, they have been growing by about 30% per year over a five year period. The number of employees covered more than doubled during the period 1990-95, to almost 0.5 million. Contributions to the funds in 1994/95 were LE 600 million, less than one percent of GDP, but with an annual growth rate of more than 30% over a five year period. Table D.6 provides data on the growth private funds over a five year period. Table D.6: Growth of Private Funds 1991-95 90/:91 = 91/92 9293 93/94 94/95 Number 330 376: 408 47:1 504: Contributions (LE ~000) 256,662 369,511 380,310 ~l585,333 602,063 Reserve Fund (LE 000) 1,379,768 1,717,928 2,129,265 2,727855 3,30,9210 Coverage 203,201 216,885 292,403 408, 21 496,386 Source: Egyptian Insurance Supervision Authority annual reports 24. A majority of these funds were actually established by employees of public sector organizations, although a recent decree forbids contributions by government organizations to these funds. Less than 10% are set up by private sector companies. Nearly all of the private pension plans operate on a defined benefit principle, where the majority of the plans provide salary related benefits in return for salary related contributions. 25. About 48% of the assets of the private funds are in fixed bank deposits while another 42% are invested in government bonds. Only about 7% are invested in equities and real estate. Lack of professional investment management capacity, the dearth of financial instruments, and the risk averse nature of these funds have been cited as the main reasons for the concentration of investments in bank deposits and government paper. 26. Private funds are covered by Law 54 (1975) and regulated by Decree No 78 (1977). Private pension funds must be registered with the EISA which requires that each fund submit annually a financial statement audited by an external auditor and every five years a financial statement prepared by an actuary registered with the EISA, which may request an external actuary to review the submissions. Investment regulations prescribe minimum and maximum investments in different types of investment vehicles. 27. To deal with the increasing number and variety of private pension funds and the expected changes in investment strategies of these funds in response to developments in the capital market, the legal and regulatory framework would have to be revised and the supervision Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 59 capacity of EISA developed. In particular, the new law and regulations would have to deal with the emergence of defined contribution plans, individualized accounts, professional fund managers, fund administrators, custodians, auditors, and actuaries. Prudential guidelines on investments would have to be developed consistent with the development of capital markets and financial infrastructure. The regulations would have to deal with pricing (fee structure), portability, and marketing issues. A decision would have to be made on guarantee arrangements. Reforming the Social Insurance System -- Short to Medium Term Measures 28. There are several reform initiatives that could be undertaken in the short to medium term to improve the efficiency and solvency of the current system. These reforms focus on: (a) improving the transparency of government utilization of SIS resources to ensure a neutral impact on government spending decisions; (b) developing a portfolio and investment strategy that supports capital market development without compromising safety objectives; and (c) correcting certain design deficiencies to improve efficiency and financial sustainability of the SIS. This section focuses on these initiatives while another section gives the longer term reform recommendations for the pension system. These short to medium term reforms would serve as a foundation for the longer term reforms described later. 29. The most critical reform initiative in the short to medium term is the improvement of the management of social security funds to achieve both rate of return and security objectives. This would entail the elimination of the special access of the NIB to SIS funds and the establishment of an investment management capability within SIS. The SIS already has plans to develop portfolio management capability and has requested technical assistance from the Bank. In addition to the hiring and training of personnel, some of the tasks necessary to ensure proper management of the reserves include development of investment objectives, designing an investment strategy by targeting a certain portfolio mix, identifying investment vehicles and participating financial intermediaries, and putting in place a control system. NIB would have to compete for the SIS funds by paying market rates (i.e., rates paid by the NIB certificates). A transition plan would have to be put in place that would reduce the amount of social security funds held by NIB. A spillover benefit of this reform is the development of capital markets and increased availability of long term funds to the private sector. The SIS should pursue technical assistance in this area. 30. Another area of reform is the retirement provisions. First, the normal retirement age of 60 should be increased over time to 65. Given the present trend of improved mortality rates, the current retirement age implies an increasing dependency ratio, thus increasing costs to the system. Second, the early retirement provisions should be revised by raising the early retirement age to 60. Workers who elect to receive pensions between the ages of 60 and 64 should receive benefits that are actuarially fair relative to the full benefit available at normal retirement age. 31. A third area of reform is the treatment of inflation. In the case of basic wages, using the average of last two years' basic wage does not take inflation into account, and also introduces an incentive for manipulation (e.g., raising wages sharply right before retirement). In the case of variable wages, the current method of using variable wages during the worker's entire career adjusted at 2% per year for inflation results in a defective pensionable wage. In both cases, one should consider using a formula that uses the average career earnings adjusted for wage inflation, although this requires a better information system. Pension benefits should also be adjusted for inflation automatically using a formula that takes into account wage inflation, rather than 60 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets awaiting legislative action. The amount of wages that would be subject to the contribution rates should also be automatically adjusted. 32. Finally, the contribution rates should be reviewed. Improved returns on reserves and less generous retirement provisions should result in lower contribution rates, although the impact of revising the pensionable wage computation to better account for inflation would result in higher costs. The net impact of all of the above reforms on the contribution rates should be determined. The cost of providing benefits to low income retirees (and whose contributions do not cover their benefits) should be estimated and its financing determined. Reforming of the Social Insurance System - Longer Term Proposals 33. A country's social security system typically has three major objectives: (a) to enable the population to shift some of their income from their working years to old age (saving or wage replacement); (b) to protect those with low incomes by providing a basic income floor during old age (redistributive or poverty alleviation); and (c) to insure against certain types of risks, such as disability, longevity, and inflation (insurance). In order to achieve all three objectives, it is recommended that a combination of systems be put in place (the multipillar system) since one system cannot efficiently achieve all objectives. Multipillar Approach 34. The reform of the pension system in Egypt should move towards the establishment of three pillars to assure adequate retirement incomes: (a) a fully funded mandatory defined benefit public pillar that insures workers' earnings up to a certain level; (b) a mandatory defined contribution private pillar that insures workers' wages above a certain level; and (c) a purely voluntary scheme that could supplement the first two pillars. In addition, the development of a competitive and stable insurance industry would provide many accompanying services such as life and disability insurance and annuity products. 35. The public pillar would provide a minimum retirement income while the compulsory and voluntary private systems would enable workers to supplement the pension from the public pillar. The three schemes should be portable across employers, and vesting should be immediate for the DC private pensions while the DB public scheme would require a minimum number of years contribution. The public pillar would achieve the objective of dealing with old age poverty which has elements of income redistribution. The private pillars have the advantage of closely linking benefits to contributions, thus minimizing the problems of evasion and manipulation. At the same time, as experienced in other countries, the private schemes should improve capital accumulation and financial market development. The Public Pillar 36. The public pillar is recommended to be a fully funded DB scheme with a required contribution from employees and employers. In order to make a smooth transition to the new system, the current basic pension scheme could be modified to form the public pillar. Under this plan, the variable pension scheme would be abolished and replaced by the mandatory private pillar discussed below. Thus, the public pillar would be built on the current basic wage pension scheme, which would be modified by reviewing contribution rates, maximum taxable amounts, redistribution objectives, automatic adjustment for inflation, and the minimization of fiscal Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 61 transfers. The SIS should undertake technical assistance to review the current system and develop a short and medium term reform plan in the context of the longer term design. The contribution rate for the public scheme would depend on the average replacement ratio (average pension payments to average covered wages), the dependency ratio (the ratio of pensioners to active workers), mortality rates, disability rates, the level of desired funding, and the rate of return on accumulated reserves. An actuarial review would be necessary. 37. Redistribution is achieved by introducing a minimum basic pension to those with low incomes. The current minimum is 50% of reference wage assuming a minimum number of 20 years' contribution; the minimum pension would be correspondingly lower if the number of years of contribution is lower. It may be prudent to review whether this formula achieves the safety net objective. Mandatory Private Schemes 38. The second pillar would be a privately managed defined contribution scheme involving compulsory contributions from earnings in excess of the public pensionable wage but below some maximum. The mandatory private pillar effectively replaces the variable pension scheme of the social insurance system. The contribution rate should allow the attainment of a certain replacement target, say 70% of pensionable wages. Under a DC scheme, the determination of such a contribution rate would largely depend on the real returns on the contributions. Disability and survivors benefits could be purchased from insurance companies and would have to be financed from an additional contribution, about 3% in many countries. The contribution rate in relation to the target replacement rate should be much lower than is currently the case for variable pensions. A contribution rate of 10% (plus 3% for disability and death insurance and 1% to 2% for management) to a fund that earns 5% in real terms (i.e., 4% over the real wage growth rate) would achieve a replacement rate of 70% of wages over the retirement period of 16 years (indexed to wage inflation), assuming that the pensioner contributed 32 years. 39. The contribution of the employers to the variable pension scheme could be merged as part of the compensation of the employees, and to the extent that the required contributions by the employees are less in the new system, workers would effectively get a pay increase. Past contributions to the variable pension scheme of the SIS could be converted into a bond carrying a market interest rate which would mature at the time the worker retires. Transition issues would have to be reviewed carefully. 40. Under a private system, participating private pension funds and fund management companies would have to be licensed and regulated. Workers could choose among the licensed funds managed by professional management companies, and mandatory contributions would be automatically withheld from wages by employers and placed in individual accounts. Workers could change employers without any impact on past contributions, thus improving labor mobility. Workers should also be able to move their accounts from one fund to another with minimal cost, thus ensuring competition among the funds. Because the system is susceptible to fraud and mismanagement, appropriate regulatory and supervision systems should be in place, including a good information flow to participants. 41. As recommended above, given that that there is a growing number of private pension funds, EISA should recommend the replacement of the 1975 Law on Private Insurance Funds and adopt appropriate prudential regulations to cover the development of individual accounts 62 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets under DC schemes managed by licensed fund professionals. This would provide the foundation for the development of a mandatory private pension scheme managed by the private sector. Since the transition to a mandatory private system would probably take some time, the development and experience of current regulatory efforts and private fund management constitute a pilot from which lessons could be leamed, allowing adjustments to be made before the mandatory private system is put in place. Furthermore, capital market development would also have to advance further, creating a more liquid and deeper financial markets. Voluntary Private Funds 42. Under the third pillar, workers would be able to make voluntary contributions with a cap, in addition to the mandatory contributions. Upon retirement, workers would be able to combine his accumulated funds with his pension account, increasing the size of his pension. This would allow firms to offer pension plans in excess of the mandatory ones, as is currently the case. These funds would be subject to the same regulatory framework as those in the mandatory schemes, and may in fact be one and the same fund. Tax Treatment 43. The consumption tax principle should be fully applied to all types of contractual savings. This would imply either: (a) the full deductibility of contributions to contractual savings and the exemption from tax of investment retums, while taxing pensions as any other source of income; or (b) not allowing the deduction of contributions while granting tax exemption status to investment retums and pension benefits. The latter approach provides cash flow advantages for the budget since no tax income is lost up front, but provides weaker incentives to workers to participate in voluntary pension funds. 44. A third alternative currently used in the Czech Republic and Australia would be to offer a govemment contribution (credit transfer) to pension members instead of a tax credit or tax exemption. This altemative would be more redistributive than the other approaches since it would also benefit nontaxpayers. In addition, it would offer a strong incentive to low and middle income workers, irrespective of whether they pay income tax, to save for their retirement. The credit transfer could be limited to active workers and could be paid only to those workers who save a specified percentage of their income and who do not withdraw their balances until they retire. As the credit transfer would be added to the individual retirement saving account of each worker, this approach would generate a higher level of long term financial resources than a tax treatment based on deductibility. 45. Currently the Egyptian tax treatment provides for tax deductibility of contributions and tax exemption of pension fund investment income and pension benefits. A review of the issues of tax treatment should be made. The Insurance Industry Introduction 46. The insurance sector can play a very important part in the development of the private sector, the emergence of a private pension system, and the modemization of the securities markets. By covering certain economic and financial risks, it enables enterprises to better Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 63 manage their financial affairs and protects households from financial losses arising from accidents or injuries. In addition, the industry, especially the life insurance sector, mobilizes long term savings that can facilitate the financing of both enterprises and households with resources that have a much longer maturity than traditional loans from the banking sector. In most developed countries, the insurance industry is a significant component of the economy, with reserves to GDP of life insurance companies ranging from 13% to 34%, and premiums to GDP ranging from 3.0% to 10.1%. The Insurance Industry in Egypt 47. The insurance industry in Egypt is underdeveloped -- life insurance premiums to GDP was an insignificant 0.2% in 1995, compared to 6% for a sample of developed economies. Total assets to GDP of all insurance companies (life and non-life) in Egypt was about 4% in 1995, while life insurance assets alone was 38% of GDP for the developed economies sample. There are 10 insurance companies in Egypt of which eight transact all classes of insurance and business and two transact only non-life. All are publicly quoted joint stock companies. Two have been set up as joint ventures with foreign investors to operate exclusively in businesses in the free zones and are not allowed to offer their services to the rest of the market. 48. The industry is characterized by a high degree of concentration. The largest company controls 50% of both life and non-life business, and three companies account for 93% of life and 89% of non-life market. The three largest insurance companies and the sole re-insurance company are state owned, thus making the sector virtually under state control. The state owned insurance companies also own shares in five of the private insurance companies. In the past, direct foreign ownership was only allowed in those companies operating in the free zones -- the exception was granted under the Investment Law which allowed for the creation of companies as joint ventures to encourage companies to operate in the free zones. Currently, regulations place a 49% limit on foreign ownership of direct insurance companies and no restrictions on foreign ownership of reinsurance companies. Investments 49. Total investments of the insurance industry as of June 30, 1995 was LE 5.4 billion, representing 2.6% of GDP. Over the period 1990/91 to 1994/95, investments have been growing by 18% per year. The state owned insurance companies accounted for 91% of the investments. The reserves of non-life business accounted for 60% of total reserves in 1995, compared to 68% in 1991. 50. The insurance companies invested the bulk of their funds in bank fixed term deposits (39% of total) and government bonds (38%). Another 18% was invested in corporate paper. The rest were used to purchase real estate and provide loans to policy holders. Lack of financial instruments and conservative investment policies have led to the concentration of investments in bank deposits and government paper. Legal and Regulatory Framework 51. In 1995, a new Law on Insurance was passed and in June 1996, a new set of regulations was issued by the EISA. In addition to solvency requirements, the regulations impose certain regulations on investments of insurance companies, making a distinction between reserves from .64 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets the life insurance business and reserves from the non-life segment. The law requires a separation of the reserves between the two businesses in companies that operate in both markets. The new regulations also liberalized the pricing of most of the insurance products, replacing price control with price reporting. The only exceptions are in the fire and motor vehicle insurance lines, which will be deregulated in 1999. The deregulation of most insurance products shifts the focus of supervision to solvency monitoring. The regulations are basically in line with international (especially EU) practices and definitions. However, solvency monitoring requires good information and technical capability on the part of the supervisor. Reform Proposals 52. With the changes in the legal and regulatory framework and the ongoing institutional development of supervision capacity, the next generation of reform efforts should focus on the issues of competition and ownership. As mentioned earlier, there is a high concentration level in the industry, which is indicative of problems of competition. To encourage competition, EISA should allow the entry of new firms -- including foreign insurance companies -- so long as they meet the licensing criteria. This means that the 49% maximum ownership by foreign firms should be abolished. EISA should also ensure a level treatment of both state owned and private insurance companies. Finally, the state owned insurance companies should be included in the current privatization program. 53. On the regulatory aspects, to complement liberalization of pricing of products and commissions, EISA should focus on disclosure requirements to the public on prices and commissions. The obligatory ceding requirements and price controls on reinsurance should be eliminated. Tight regulations on employment of foreigners should also be relaxed to enable insurance companies to acquire needed expertise quickly. Finally, to improve the transparency of the sector, financial statements of insurance companies should be available to the public, and these statements should be audited by qualified and independent auditors. Accounting standards and auditing guidelines for insurance companies should be provided by EISA. Capital Markets Introduction 54. Capital markets affect economic activity through the creation of liquidity and the reduction of transaction costs. From the point of view of the investor, liquid securities markets allow the acquisition of an asset that can be sold quickly in case of a need to access the investor's saving or alter his or her portfolio. From the point of view of the corporation issuing long term securities, it has access to a larger pool of funds and has better information about the relative cost of different ways of financing an investment. Thus, a corporation has a wider range of instruments to finance investment, and savers have more alternatives than bank deposits, precious metals, or real estate. 55. Many studies conclude that the impact of stock markets on the level or rate of domestic saving is ambiguous -- savers may merely shift funds from one saving instrument (such as bank deposits) to securities. However, some studies have shown that there is a positive relationship between private saving as percent of GDP and financial sector development, which includes capital market development, although the channels through which this relationship is defined are Working Papers Annex: Pension Reform, Contractual Saving. And Capital Markets 65 numerous. For example, stock markets are seen to increase investment, thereby increasing national income, thus increasing the level of or rate of domestic saving. 56. But what is unambiguous is that the composition of saving is improved by the introduction of liquid capital markets. Savers purchase long term securities which offer higher expected returns and enable risk diversification. In the process, more financing opportunities become available to corporations to implement projects with long term pay-offs. Investors in long term projects need not relinquish their saving for long periods -- liquid capital markets allow them to divest quickly and inexpensively. 57. In several countries, the securities markets have developed in parallel with and have supported two initiatives that have an important impact on private saving: one is pension reform (such as the transformation of pay-as-you-go to fully funded systems), and the other is privatization of state owned enterprises. To the extent that pension reform and privatization programs increase household and business saving, respectively, securities market development has an impact on the level of private saving. 58. In the case of Egypt, the development of capital markets both supports and is supported by the development of contractual saving institutions. As discussed in the previous sections, the investments of private pension funds and insurance companies have been mainly in government securities and bank term deposits. Capital markets would allow greater diversification and perhaps higher yields for these investments, thereby improving the financial performance of contractual saving institutions which should result in greater benefits to savers in the form of lower contribution rates to pensions schemes and lower premiums for insurance. At the same time, the existence of pools of funds from contractual saving institutions could be tapped through capital markets to finance investments. 59. In addition, capital markets provide an enabling institution in attracting foreign saving. The existence of liquid capital markets gives a foreign investor better exit options, thus encouraging more foreign direct investment. Foreign portfolio investors would focus on actively traded stocks in the stock market. With the greater share of private funds (77% share in 1994) in the capital flows to developing countries and the shift from bank lending to foreign direct investment and foreign portfolio investment in private capital flows, the development of capital markets becomes essential in attracting private foreign capital. But the benefits of foreign saving would be maximized if it augments, rather than replaces, domestic saving. The experience of two regions which attracted the greatest share of private capital flows gives two different outcomes -- foreign saving replaced domestic saving in Latin America and the Caribbean but augmented domestic saving in East Asia. The positive experience of Asia is attributable to macroeconomic policies and institutions that encouraged domestic saving and investment. 60. Finally, the development of capital markets and the process of privatization are mutually reinforcing. Capital markets provide more options for divestiture, while privatization increases the supply of securities in the market, thus providing the securities market with more depth. The deepening of capital markets -- as reflected in increased market capitalization to GDP, greater liquidity (higher turnover to market capitalization), and less concentration of market activity on a few stocks -- would enable capital markets to absorb the expected increase in portfolio investments from both domestic and foreign sources, and mitigate against the extreme movements of asset prices. 66 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets Revitalizing Capital Markets in Egypt 61. To revitalize the capital markets, new legislation (the Capital Markets Law) was passed in 1992 and the Executive Regulations (ERs) were issued in 1993 which, among others, instituted reforms in four areas: (a) the strengthening of the regulatory capacity of the Capital Market Authority by establishing a legal framework that provides a substantial degree of flexibility to enable the CMA to respond to changing market conditions: (b) the issuance of regulations covering all participants in the market, including brokers, investment managers, investment funds, and stock exchanges; (c) the reorganization and modernization of the stock markets, including the establishment of new trading systems and linkages between the Cairo and Alexandria stock exchanges; and (d) the improvement in disclosure rules including the development of accounting and auditing systems. 62. The CMA is charged with the responsibility for regulating and supervising the capital market industry within the framework provided by the CML. The CMA has delegated many of the regulatory powers to the stock exchanges which have established rules and regulations that govern their members and operations, although the CMA has the power to augment or revise these rules and regulations. The two exchanges have now been unified as a single market through the operation of an automated trading system. Some 65 financial intermediaries have been established and licensed by the CMA over the past two years. As of June 1996, there were 152 licensed brokers and 15 licensed mutual funds. The CMA is also expected to de-list some 200 companies whose stock has not been actively traded for some time and who have not met disclosure requirements. The CMA is involved in the modernization of the stock exchanges, development of a privately owned and managed central depository system, and reform of the clearing and settlement system. 63. The current clearance and settlement system is an inefficient manual process which increases risks and costs. Increased market activity would put a strain on the existing system. A computerized book entry clearance system would be the most efficient alternative to enhance the efficiency of securities transactions, though this would require the computerization of stock exchange and brokerage operations. A new corporation owned by brokers and other private sector investors has been formed to develop the clearance and settlement system. 64. The CML brings securities custodial services under the authority of the CMA. At present, public sector banks provide custodial services for client's securities at their major branches. This service includes the timely filing of coupons for dividend and interest payments and the handling of transfer procedures. The brokers do not have the facilities or the capital to provide custodial services. A private corporation is being set up to handle custodial services, with the intention of de-materializing shares in the future. 65. One of the important means of disciplining the market and creating investor confidence is to ensure that investors harmed financially by the actions of an intermediary can pursue a claim for compensation. The CML and the ERs do not make clear the circumstances in which a civil liability arises and the general Egyptian law is not strong in this area. The Government should specify the circumstances in which a civil liability arises and also make clear the avenues of recourse open to investors. Working Papers Annex: Pension Reform, Contractual Saving. And Capital Markets 67 Market Performance 66. The stock market has witnessed significant growth since 1990. As of June 1996, market capitalization as a percent of GDP was 13.4%, which although low compared to those of several emerging markets such as Jordan and several Asian and Latin American countries, represents a doubling of the ratio since 1990. The turnover ratio was 14.0% in 1995, still relatively low by international standards; however, during the first six months of 1996, the turnover ratio has already reached 13.0%, at which rate the 1996 ratio could double that of 1995. Moreover, the CMA is expected to delist about 200 of the 674 companies (including the OTC market) due to lack of trades, thus increasing further the turnover ratio. Many companies seek listing due to the tax benefits, even though these companies do not trade actively. Table D.7 provides some rough indicators of securities market performance. 67. Two other indicators of market performance are worth noting. The first indicator is the value of new issues (stocks and bonds), which increased from LE 2.1 billion in 1993 to LE 11.2 billion in 1995 equivalent to 5.4% of GDP and 27% of average market capitalization during the year. During the first six months of 1996, the value of new issues has already reached LE 8.7 billion and included corporate bonds of LE 0.5 billion. The other indicator is the value of trades by foreign investors. From insignificant amounts in previous years, the share of foreign investors in the value of trades during the period April to July 1996 was 33.8% During the month of July 1996, the share of foreign trades was more than 50%. Table D.7: Performance Indicators for Capital Market Nontrading conpanies Value of trading to Market capitalization New issues to market New issuis to to listed companies matket capitalization to GDP capitalization GDP 1985 34.91 3.42 4.94 na na 1986 34.88 5.23 5.48 na na 1987 35.12 4.64 7.81 na na 1988 34.92 3.25 6.70 na na 1989 35.10 5.23 5.70 na na 1990 34.73 6.73 5.28 na na 1991 34.77 4.84 7.95 na na 1992 36.43 5.50 7.80 na na 1993 39.17 4.44 8.14 15.29 1.33 1994 42.86 17.66 8.27 22.45 2.49 1995 47.18 14.04 13.38 27.23 5.48 Source: Capital Market Authority Reform Proposals 68. Significant strides have been made in developing the stock market, and these are reflected in market performance, notably the recent interest of foreign portfolio investors. Additional reform efforts are suggested below. 69. The continued and accelerated privatization of PEs would assist in the development of capital markets in two ways. One, it would increase the supply of tradable shares in the exchanges -- the book value of the PEs are estimated to account for xx% of existing market capitalization. Two, privatization provides a clear and strong signal of government commitment to market reforms. 68 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets 70. The custodial, clearing and settlement systems would have to be improved. Improved efficiency in these systems would reduce not only transaction costs but also certain types of risks. The move to privatize these functions is a step in the right direction, but would need good oversight. 71. Finally, the CMA should continuing moving towards accrediting rating agencies, both local and foreign. For rating agencies to perform their functions, there should be a timely and free flow of information. The establishment of international standards for accounting and auditing and the institutional development of the accounting and auditing profession would provide the basis for good financial information. Taxation of Saving Introduction 72. Taxation of saving and investment plays an important role in determining present and future output levels and living standards. Domestic saving, together with foreign investment, finance capital formation, which in turn raises future output by increasing the physical capital with which labor and other factors have to work. The resulting factor productivity determines wage levels and living standards. 73. This section focuses on the effects of the tax system on corporate and household saving. The interest in the effects of taxation on the composition of savings is driven by the objective of channeling savings into assets yielding the highest rate of return before taxes. A "neutral" tax treatment of alternative savings instruments will normally achieve this goal. This treatment ensures that the market, rather than the tax system, determine what types of saving are undertaken. Many departures from neutrality in existing tax treatment of savings are deliberate and intended by policymakers. Even if non-neutral treatment is intended, policymakers may well be interested in taking the costs of such policy actions into account in determining the extent of favorable treatment that they are willing to provide. But some departures from tax neutrality may well be unintended. One such departure that is well documented arises from the interaction of inflation with an unindexed tax system. Even moderate inflation can produce some dramatic differences in "effective" tax rates compared to the nominal rates prevailing in a non-inflationary environment. Personal Savings Marginal Effective Tax Rate Methodology 74. The analysis will be based on calculations of marginal effective tax rates for different types of saving. These marginal effective tax rates on saving instruments provide a useful summary measure of the incentive or disincentive provided by the tax system to undertake particular forms of savings. This methodology examines the effects of the tax system on the last act of saving worth undertaking to the saver -- the acquisition of that asset whose cost is just equal to the estimated present value of its after-tax returns over its lifetime. After determining the after tax rate of return for this marginal savings instrument, calculations are made on what rate of return the asset would earn before tax is levied. The difference between the rates of Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 69 return before and after tax is called the tax wedge, and the tax wedge divided by the before tax rate of return is called the marginal effective tax rate on this savings instrument. 75. The estimates of the marginal effective tax rates for personal savings in Egypt use actual interest rates on bank deposits to indicate the minimum acceptable after tax return for other types of personal savings. To enable the findings to be compared with those for OECD countries, the methodology used was similar to that of a OECD study.2' This procedure has the advantage of tying our estimates of marginal effective tax rates to actual capital market conditions in all countries considered. Effective tax rates are calculated for the principal types of savings vehicles available to the individual in Egypt and elsewhere: bank deposits; bonds; stocks and equities; equity invested in owner-occupied housing; private pensions and life insurance. 76. All relevant taxes affecting saving are included in the calculation of the effective tax rates and the estimates are not limited to taxation of current income from assets such as bank deposits, bonds and stocks. All of these savings instruments apart from bank deposits also offer possibilities of another form of income gain or loss to their owners; that is, where capital gains or losses occur on the sale of these assets. Where applicable, capital gains taxes are accordingly entered into our calculations of effective tax rates. Some countries also tax wealth or property, based on the value of the asset stock held by the saver. Taxation of pension saving usually involves another distinctive form of tax treatment. Contributions are frequently deductible from taxable income, while pension payouts may be taxed either at the same marginal rate applying when contributions are deducted, at a lower tax rate than that relevant for contributions, or they may not even be taxed at all. Taxes on owner equity in housing also vary distinctively from taxes on other assets, with the costs of mortgage debt being deductible from taxable income in some countries, with capital gains from sale being taxed in some countries, with a few countries taxing the imputed rental value from owner-occupied housing as income, and with many countries taxing property values. Table D.8 summarizes these heterogeneous tax provisions as they apply to various types of savings instruments in Egypt and the four OECD countries in our comparisons: Finland, France, Greece, Ireland and Italy. 77. Even a cursory examination of the tax provisions in Table D.8 indicates that some of these private saving instruments, both in Egypt and elsewhere, receive very favorable tax treatment compared to that applying to other assets. Some types of savings are given favored tax treatment because of the contribution they are expected to make to widely held social goals. Thus, housing ownership is a long-standing social policy objective in most countries. Some types of tax treatment of savings specifically aim to influence the timing of consumption. Thus, a primary social goal which is advanced in Egypt and the OECD countries is that of fostering self-reliance in retirement by giving special tax treatment to pension savings. For similar reasons of social policy, preferred tax treatment is frequently given to life insurance. More recently, share ownership is receiving preferred tax treatment in an increasing number of countries as governments try to encourage more widespread ownership of private capital. 21 The 1993 OECD study calculated marginal effective tax rates based on the assumption of a fixed before -tax rate of return of 5% as the minimum desired before-tax return (or opportunity cost) for saving. See OECD, Taxation and Household Saving, (Paris: OECD, 1994, pp. 90-91. The problem with this procedure is that the absolute size of estimated effective tax rates then depends on the size of the assumed fixed before-tax rate of return that is used. 70 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets 78. Before discussing particular effective tax rates for savings instruments, it might be helpful to explain the implications of positive, zero and negative effective tax rates in our findings. Saving is taxed under an income tax and saving instruments getting income tax treatment will accordingly have a positive effective tax rate. A zero effective tax rate implies that there is no tax penalty on savings, which is consumption tax treatment. An alternative way of way of looking at such consumption taxes is that they do not affect the timing of consumption, so that, for instance, a person saving for retirement is not penalized by the decision to save now and postpone consumption until retirement. Finally, a negative effective tax rate implies that saving is being subsidized through the tax system. Here, the after tax rate of return on a saving instrument is higher than its rate of return before tax. An example (which is not hypothetical) would occur where pension contributions are deductible from taxable income, accumulate tax free in pension funds and are not taxed when distributed as pension payments. In this example, the taxpayer is in effect able to convert taxable income into nontaxable income by making pension contributions. The Effective Tax Rates on Personal Savings in Egypt and Selected OECD Countries 79. Bank Deposits. Table D.9 shows that there is no tax on saving deposited in banks under the supervision of the Central Bank of Egypt, this form of income being exempt from tax. As recently as 1993, some incremental saving was accorded even more advantageous tax treatment, since savers were able to deduct amounts spent on acquiring new bank deposits under the general income tax up to fairly generous limits.2 This treatment is not modeled in Table D.9. 80. There are less important types of bank deposit whose returns are taxable in Egypt, however. First, deposits in a few "private" banks which are not subject to the regulations imposed on public banks by the Central Bank of Egypt are taxable.23 Interest from deposits in these private banks is subject to a 32% tax rate. Interest on deposits in foreign banks is also subject to tax at the 32% tax rate. Naturally, this nominal tax rate is also the effective tax rate if there is no inflation. As Table D.9 shows, however, at the 8% annual rate of inflation in 1996, the effective tax rate on taxable savings deposits in Egypt jumps to 80%. 81. Effective tax rates on interest from bank deposits in selected OECD countries are indicated in Table D.10. In general, effective tax rates on interest from bank deposits are the highest for any type of assets examined, both in Egypt and in the OECD countries The primary reason is that nominal interest payments are taxed as income when there is inflation and no allowance is made for the fact that such nominal interest includes an amount to compensate for the fall in the value of saving with inflation. The highest effective tax rate on bank deposits in Table D.10 is for Greece, which not coincidentally is also the country with the highest annual rate of inflation. France has the next highest effective tax rate on bank deposits, primarily because a net wealth tax of 1.5% applies to bank deposits and no other country in the table has 22 Gersovitz et al. (P.3 1) estimate that the limits on such deductions were not binding until an individual's income reached 15 times the average Egyptian income level. This deduction was also available for incremental purchases of shares and bonds. 23 The banking system in Egypt is dominated by four banks which are under the supervision of the Central Bank of Egypt. Three banks established under private laws are not registered with the Central Bank, and interest on their deposits is taxable. Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 71 such a wealth tax on deposits. To ameliorate the effects of inflation on taxation of nominal interest income, some countries such as Finland apply a lower tax rate to this type of income. 82. Government Bonds. Income from government or corporate Egyptian bonds is also not taxed in Egypt, although capital gains from the sale of these bonds was treated as taxable income until earlier in 1996. Table D.9 shows that the effective tax rate for saving in bonds in Egypt has accordingly dropped to zero from previous rates which varied from 24.1% to 8.2%, depending on assumptions about inflation and the length of time for which bonds are held. As in the case of interest on bank deposits, effective tax rates on bond interest rise with inflation if nominal and not real returns are taxed. Effective tax rates fall with the length of time for which bonds are held, because the value of future capital gains becomes lower in present value terms. 83. Of the OECD countries represented in Table D.1O, Greece also does not tax any form of income from bonds. Three of the other five OECD countries have effective tax rates on savings in bonds ranging from 11% to 51%. France has the highest effective tax rate on bonds because it also imposes a wealth tax on bond values. 84. Stocks and Equities. There is no effective tax rate on personal savings held in the form of stocks and equities in Egypt, because neither dividends nor capital gains resulting from the sale of shares are taxable as personal income. Greece has a zero effective tax rate on saving in stocks for the same reason. The effective tax rates on stocks and equities for other countries in Table D.lO range from 17% to 65% and reflect taxes on dividend income and capital gains as well as a wealth tax of 1.5% which applies in France. 85. Owner-Occupied Housing. It is fair to say that the conventional wisdom sees owner equity in housing as a lightly taxed asset in many OECD countries. By way of contrast, the OECD savings study finds it "striking" that "these rates [effective tax rates for house purchase financed by equity] taking into account all relevant taxes are not generally low compared to other assets, particularly in the case of zero inflation..."24 There are differences of opinion as to how property taxes should be treated in these effective tax rate calculations, the issue being whether property taxes should be tieated as a wealth tax on this asset or as a benefit tax which simply pays for services to property such as street maintenance. A wealth tax should be included in the effective tax rate calculations for housing, whereas benefit taxes should not be. Because it is not known whether these property taxes are really wealth or benefit taxes in the various countries, separate estimates are given for effective tax rates on housing with and without property taxes in Tables 2 and 3. 86. Table D.9 shows that because of relatively minor property and sales taxes, housing is taxed more heavily than other forms of saving in Egypt, but the total tax burden is hardly substantial. The penalty on housing relative to other types of saving becomes less marked in noninflationary conditions, however. For the two OECD countries that report property taxes, Table D.10 confirms the OECD conclusion that housing is not favored as an asset when these taxes are included in calculations of effective tax rates. Taxes on imputed rental income contribute to the substantial effective tax rates on housing equity in Greece and Italy. 87. Pensions. The tax treatment of private pension saving usually differs quite substantially from that applied to saving in other types of assets. The so-called "cashflow" method of 24 OECD, Taxation and Household Saving, (Paris: OECD, 1994), p. 1 10. 72 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets consumption taxation is commonly used, whereby pension contributions are deductible from taxable income of the individual, funds accumulate tax free while in the pension plans, but pension payments from the plans are taxable. This results in pension saving not being taxed if the marginal tax rate which applies at the time contributions are deductible from taxable income is the same as the marginal tax rate on pension payouts. 88. The Egyptian tax treatment is more generous than this cashflow method of consumption taxation. The principal difference is that pensions paid from the funds are not taxed in Egypt, so that saving for private pensions has a substantial negative effective tax rate and one which increases as inflation goes down. This tax treatment subsidizes pension saving because income which would otherwise be taxable is converted into nontaxable income. The only other country which has negative effective tax rates for pension saving in Table D. 10 is Ireland, which applies similar tax provisions to those in Egypt. 89. Insurance. Life insurance can fulfill the same purpose of saving for retirement as private pension plans and there is a corresponding case for life insurance companies to have comparable tax treatment to that given to pension savings. This appears to be approximately the case in Egypt. As in the case of contributions to pension plans, life insurance premiums are deductible from taxable personal income within generous limits (up to LE 1000 or 15% of taxable income each year). There is a tax-free "inside build-up" of the value of both types of savings and no tax is levied on pensions paid or when life insurance policies are cashed out. The only significant difference between the two types of savings is that a 3% tax is levied on insurance premiums. Accordingly, it is not surprising to see in Table D.9 that saving through life insurance has a negative effective tax rate, although the implied tax subsidy for this type of saving is not as great as for pensions. There is no comparable information on tax treatment of life insurance for the OECD countries in Table D. 10.25 90. Conclusions. In Egypt, as in much of the rest of the world, there has been a recent tendency to reduce taxation of private savings. The tax treatment of personal-savings in Egypt must now be considered to be favorable to savers, and the extent of such favorable treatment has increased substantially since the Bank's first major examination of the Egyptian tax system in 1993. Saving in housing is taxed but not as heavily as in the OECD countries with which comparisons were made. No tax is now payable at all on many forms of capital income to the individual in Egypt such as interest on most types of bank deposits or bonds and dividend income. Whereas these forms of saving have zero effective tax rates, the effective tax rate on pension savings is negative and the implied tax subsidy becomes even greater as the inflation rate falls. Life insurance receives only slightly less favorable treatment than pension contributions. 91. These forms of saving need to be treated more equally, from the points of view of equity, efficiency and maintenance of government revenues. First, the present difference in tax treatment of the different types of saving decreases real income, because some types of saving will be used to a greater extent than otherwise because of their tax advantages and not because they represent a more efficient use of resources. Secondly, widespread use of some of these tax provisions poses a potential threat to government revenue goals. In this respect, the present tax 25 While our primary interest is in life insurance as a form of saving, taxes on premiums for other types of insurance might inhibit their use in Egypt. For instance, there is a 10% tax on transportation premiums and a 20% tax on property and casualty insurance Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 73 treatment of private pensions and life insurance simply amounts to the conversion of taxable into nontaxable income for individuals saving in these forms. Finally, but perhaps most importantly, the tax provisions for some types of saving leads to an equity problem in treatment of savers and non-savers. Only a few have the discretionary income to take advantage of private pension plans and life insurance in Egypt. In these circumstances, it is somewhat disturbing when savers are subsidized through the tax system to the extent indicated by the negative effective tax rates on private pension and life insurance savings. 92. The recommended policy change of taxing private pension incomes and the cash value of life insurance policies would address all three goals of increasing efficiency and equity and maintaining government revenues. Further, it would not tax sa"ing for retirement and would align Egypt with the most frequent tax treatment of private pension savings and life insurance in OECD states. 93. Until recently, the effect of tax provisions on saving were blunted by regulation or other quantitative restrictions which offset the effects of the price incentives entailed in tax treatment.26 This helps account for the fact that life insurance and private pensions have played a minor role in the structure of Egyptian household assets and in financial intermediation. The reductions in regulation of the Egyptian economy in recent years have correspondingly increased the importance of policy instruments that work through markets such as taxes, and further economic liberalization will add to their importance. These market reforms, together with the recommended tax treatment of savings, should mean that financial intermediaries such as private pension plans and insurance companies will become much more important in Egypt's financial sector. Similar tax treatment of housing, pensions and insurance in the less regulated financial markets of the OECD economies helps account for the dominance of these forms of saving instruments in personal assets there. 94. There are inefficiencies arising from the effects of inflation on effective tax rates for personal savings in Egypt and our sample of OECD economies. The effect of inflation is greatest for taxes on interest from bank deposits or bonds, because nominal interest payments are usually taxed as income. Since most national tax systems have consumption tax treatment of some assets such as pensions or housing but income tax treatment of others such as bank deposits, this increases non-neutralities in treatment of saving vehicles as indicated by the spread in their effective tax rates. As a result, this interaction of inflation and the tax system worsens the allocation of resources. 26 In their 1993 report for the World Bank on the Egyptian tax system, Gersovitz, Gordon and Slemnrod commented that regulations have restricted the net rate of return investors can earn on insurance.. .undermining the competitiveness of insurance companies. Given the rate of return in fact offered net of tax on insurance, few investors would find investing through an insurance company attractive were it not for the benefits they receive through the insurance itself. See Mark Gersovitz, Roger H. Gordon and Joel Slemrod, "A Report on the Egyptian Tax System," pp. 45-46. These authors also suggested that tax administration in Egypt led to substantial departures from the intended effects of the tax statutes. Ibid., p.i. 74 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets 95. These effects of the interaction between inflation and the tax system are evident in our calculations of effective tax rates on savings in Egypt but have become less problematical for two reasons. First, inflation has decreased in Egypt as a result of successful fiscal policy. Second, policy changes to eliminate taxation of savings have also reduced interactions between inflation and the tax system. Our suggested policy change of taxing private pension plan payouts and the cash value of life insurance policies would also prevent erosion of government revenues andfurther reduce the interactions of inflation with the tax system in Egypt. Corporate Savings 96. To examine the corporate role in saving in Egypt, the results of a 1995 analysis of effective tax rates on investment for Egypt is reviewed. The analysis shows the effect of the tax system, as it integrates taxation of corporate and personal saving, on the cost of financing to the domestic firm in Egypt. Corporate saving in the forrn of retained earnings are one source of financing to the firm, the others being the issue of new equity and borrowing including the issue of bonds. The 1995 analysis in Table D. 11 distinguishes between the corporate, personal and aggregate effects of taxes on the cost of financing to the firm. Subsequent changes in tax provisions have completed the removal of all taxes on personal income arising from corporate business. The only remaining determinant of the cost of capital to the firm is corporate taxation, and our analysis will concentrate on its effects as shown in Table D. 11. 97. The deductibility of the costs of borrowing -- and not other sources of finance -- normally biases the financing of the firm to debt financing under corporate taxation. The advantages of debt financing are compounded in Egypt because interest income is not taxable to the individual supplying such finance. An interesting feature of the Egyptian tax system is that it does allow joint stock companies listed in the stock market to deduct an imputed interest charge for paid-up equity capital -- the interest rate applied being that for government borrowing. Table D. 11 shows that this relief is not sufficient to offset the advantages of corporate borrowing to the firm. Nor does it lower costs for retained earnings as a source of financing for the firm. Finally, this relief is not available to types of organization of the firm other than listed joint stock companies. 98. The examination of effective tax rates on sources of finance for the firm in Egypt confirms the existence of some biases whose reduction or removal could improve the functioning of capital markets for business finance in Egypt. In particular, desirable changes would offset the bias against retained earnings and against forms of business organization other than listed joint stock companies, and reduce the advantages of debt finance to corporations. Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 75 Table D.8: Summary of Tax Treatment of Savings and Other Inputs for Effective Tax Rate Calculations: Egypt and Selected OECD Countries Finland France Greece Egypt treland Italy A. Personal Taxes, Annual Inflation Rates 2.1% 2.2% 14.4% 8.0% 2.3% 4.7% Bank Deposit Interest Rates - Before Taxes1 3.19% 4.5% 15.75% 12.85% 5.42% 6.45% A. I Bank Deposits Tax rate on interest from deposits 20.0% 38.1% 15.0% 0/32%2 48.0%a 30.0°h Wealth tax on deposits 0.0%f0 1.5% 0.0% 0.0% 0.0U 0.0% A.2 Government Bonds Tax rate on interest 20.0% 18.1% 0.0% 0.0% 48.00%e 12.5% Tax rate on capital gains from bonds 25.0% 18.1% 0.0% 0.0% 0.0A 0.0% Wealth tax on bonds 0.0% 1.5%/n 0.0% 0.0% 0.0%/f 0.0°fe Turnoverrate -short term bonds 100%/o/yr 100%/elyr 100%/yr 1000/odyr 10oo/Jyr 100%//yr Turnover rate - long term bonds 20%/yr 20%o/yr 20%/O/yr 20%/yr 20%/Jyr 20%/yr A.3 Equities Tax rate on dividend income 25.0°h 59.9% 0.0 04 0.0% 48.0% 51.00f Tax rate on capital gains from securities 25.0% 18.1% 0.0% 0.0% 40.0% 0.0% Wealth tax on equities 0.9% 1.5% 0.0% 0.0% 0.0% 0.0%/O Tax rate on sale of asset 0.0% 0.3% 0.0% 0.0% 0.0%O 0.14% Tax rate on purchase of asset 0.0% 0.0% 0.0% 0.0% 1t.0%/ 0.0% Turnover rate 15%1/yr 15%/eyr 15%/yr I5%/o/yr 15s/Jyr. 15%A/yr Proportion of total return earned as dividends 33.3%h 33.3% 33.3% 33.3% 33.3% 33.3% A.4 Housing Personal Interest deductibility - top rate 25.0% 25.0% 40.0% 0.0% 48.0%/o 27.0% Income tax rate on imnputed rent 0.0% 0.0% 40.0% 0.0% 0.0% 30.6% Tax rate on capital gains 0.0% 0.0% 25.0% 0.0% . 0.0°fe 0.0°h Wealth tax on housing stocks - national 0.9% 1.5% 3.0% 0.0% 1.5% 0.0%h Wealth tax on housing stocks - local n.a 0.3% 0.0% 0.0% 0.0% 0.55% Sales tax rate 4-6% 8.0% 3.5% 2.5% 0,0%/o 0.0% Property tax rate .01% 0.00/% 0.0% 1.0% 5.0% 4.0% Proportion of earnings in irnputed income 33.3%/O 33.3°h 33.3% 33.3% 33.3%/o 33.3% A.5 Pensions, Life Insurance & Other Insurance AS.I Taxation at lhefirm level Tax rate on bond interest revenues 25.0% 0.0% n.a 0.0% 0.0% n.a Tax rate on dividend revenues 25.0% 0.0% n.a 0.0% 0.0% n.a Tax rate on capital gains on equities 25.0% 00°% n.a 0.0% 0.0% n.a Wealth tax on bonds and equities 0.0%h 0.0% n.a 0.0% 0.0% n.a Tax rate on sale of bonds and equities 0.0% 0.0% n.a 0.0% 0.0% n.a Tax rate on purchase of bonds and equities 0.0%/0 0.0% n.a 0.0% 1. 0% n.a Proportion of income paid out that is tax free 0.0% 0.0% n.a 100% 25.AD/a n.a AS.l Taxation of Individuals Proportion of tax free contributions made 0.0% 0.0% n.a 100% 25.000 n.a Percentage of fund in dividends / bonds 40/60 40/60 40/60 40/60 40/60 40/60 Percentage of real return of fund from cquities 33.3% 33.3% 33.3% 33.3% 33.3% 33.3% Turnover rate for equities in the fund 15%/yr 1 5%/.yr 15%0o/yr I 50/o/yr I 0/5/yr 15%/yr Sources: Wayne Thirsk, Barents Group; OECD, Taxation and Household Savings, 1994 From the IMF, Intemational Financial Statistics, and based on the interest rate paid by banks on term deposits. 2 Interest eamed from bank deposits under govermment supervision are exempt from taxation while interest eamed from non-government supervised deposits are subject to a 32% personal income tax rate 76 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets Table D.9: Marginal Effective Tax Rates on Personal Savings in Egypt (percent) EGYPT Late 1996 tax: Late t996 tax 1993 tax provisions - provisions - provisions - a5unied, 8% 8% Inflation No Inflation Iflaio: ;Bank Depotits :: t0 : : :f: :: ::t:: f::::; : : : : : ;: : : t :0 04 ; : - "Supervised" banks 0.0 0.0 0.0 - "Unsupervised"&banks 80.2 32.0 80.2: Bonds - Shor=t Term 0.0 0.0 24.1 Bonds -LJong Term 0.0 0.0 16.7 StocksA& quities 0.0 0.0 ~ 0.0 Owe Occrk upied Housing:100% Equity> iFirancing00: $ i :0::i : :0:0 f0:0 ;t00 : aSt:0000;Q:b. With realestaltSeste taxeS0 ;0 t :000t:0;t0:0t: :40:0 (0;161.2 .5 16:.2 b. :Withot realstale takes 5.2 2.1 5.2 Pensions. -58.0 -808 .58.0 Lifeinsurance -54.5 -78.7 -54.5 Table D.10: Marginal Effective Tax Rates on Personal Savings: Egypt and OECD Countries (percent) Finland Fraiice Greece Egy? lreland; ItaIX] Ban Deposit 10.3 82.1 85:.0 0/8Q02 56.6 66 Bonds - Short Term 12.0 50.8 0.0 0.0 36.7 37,2: Bonds-;lngTern 10.5 49.9 0.0 0.0 36.7 37.2 Stocks& Equities: 35.5 64.8 0.0 0.0 I7.1 36,0 Owner Occupied Housing: 100% Equity : fOwer Ocupied Housing - With realestate: taxes nra 53.7 n.a 16.2 n.a 46,0 OwnerfOccupied Housing - No real esatea txes 11.4 41.7 105.2 5:.2 15.7 37.6 PeniPans -standasd payout" 14.5 22.6 n.a -58.0 -:81.7 : n. Table D.A 1: Effective Tax Rates on Capital for Alternate Sources of Financing (percent) corporate Personal Aggregate Effective Tax Rates Effective :Tax :Rates Bffective Tax Rates Manufacturi*g: gAll Debtf: a) interest income exempt from taxes -67.5 0.0 -67.5 All Retained Earnings 54.0 3.4 55.0 New Equity Issues a) With prepaidcapital:deductions 3.5 0.0 3,5 b) Without prepaid capital deductions 50.5 0.0 50,5 Services: All Debt a) Interest income exempt from taxes -68.6 0.0 -8.6 : All Retained Earnings 65.1 3.4 66.2 New Equity Issues a) With prepaidcapital,deductions 7.5 0.0 7.5 b) Wiftout prepaid capital deductions 59.9 0.0 59.9 Source. Own calculations 27 The personal income tax rate on pension payouts is the same as on deductions of contributions. Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 77 Action Plan for Social Insurance System Reform Area of Reform Short to Medium Term Actions Medium to Long Term Actions * Improve Returns on Surplus * Develop portfolio management capability * manage all of SIS funds * Develop plan to place funds currently with NIB under the management of SIS * Develop an investment strategy, focusing on the identification of investment objectives * Adjust Benefits for Inflation * Review formula for treatment of inflation in determining initial benefits for both basic and variable pensions * Index benefits automatically to wage inflation * Review how to adjust maximum taxable wage automatically * Reduce Contribution Rates * Review actuarial assumptions and determine whether there is scope for reducing contribution rates for both basic and variable pensions * Introduce mandatory private pillar * Develop infrastructure for mandatory * convert variable pension scheme to private pillar by strengthening legal and mandatory defined contribution plan regulatory and supervision framework for managed by the private sector voluntary private pension schemes, strengthening the insurance industry, and developing capital markets 78 Working Paper Annex: Pension Reform, Contractual Saving, And Capital Markets Action Plan for Voluntary Pension System Area of Reform Short to Medium Term Actions Medium to Long Term Actions * Develop new legal framework * Introduce new law that allows for defined benefit individual accounts and sets a flexible framework for the regulation * Strengthen regulatory framework * develop and issue new regulations, to * adjust regulations as necessary include coverage of fund managers, administrators, custodians, auditors, and actuaries * Strengthen supervision capacity * implement institutional development of EISA Working Papers Annex: Pension Reform, Contractual Saving, And Capital Markets 79 Action Plan for Insurance Industry Reform Area of Reform Short to Medium Term Actions Medium to Long Term Actions * Improve structure and competitive * privatize one or two insurance companies * privatize the other state owned insurance conditions companies * allow more entry, including foreign firms * Deregulate prices * allow competitive pricing of fire and motor vehicle TPL insurance * Improve Supervision capacity * implement recommendations of Coopers and Lybrand * Ensure Neutrality of Tax Treatment on * Eliminate stamp duty for life insurance * Review income tax rates of insurance Saving and reduce stamp duty for nonlife companies in the context of an overall tax insurance review 80 Working Paper Annex: Pension Reform, Contractual Saving. And Capital Markets Action Plan for Capital Market Development Area of Reform Short to Medium Term Actions Medium to Long Term Actions * Increase Supply of Listed Securities * Continue privatization process * Improve Financial Infrastructure * Develop Custodial Depository and improve payments and settlement services under private sector initiative * Improve Information Quality and Access * Ensure proper accounting and auditing of financial accounts * Automate information on listed companies * Develop Prudential Regulation and * Coordinate regulation supervision of Strengthen Supervision Capacity investment managers with EISA ANNEX E EGYPT: NATURAL RESOURCE DEPLETION AND SUSTAINABLE GROWTH Countries that derive a significant share of their national income from the extraction of non-renewable natural resources are faced with several issues related to the sustainable growth of their economies. How important are non-renewable resources in the macro- economy? What are the impacts on the external balance, and output growth, of the future rate of resource extraction? How much of the resource rent should be consumed now as opposed to saved for consumption in the future? And in what should the resource rent be invested to accommodate sustainable growth? Non-Renewable Resources in the Egypt Economy Figure Ia: Share of Oil and Gas Sector In GDP Non-renewable natural (at constant prces) resources have played an important 18.0% role in Egypt's economic 146.0% development, i.e. in Egypt's case 120% - predominantly oil and natural gas. 10 0% The oil and gas sector contributed 8.0% about 10 percent of Egypt's GDP 6.0% I during the 1990's (see Figure la). 4.0% The crude oil and petroleum 2 0% 0.0% product exports have been close to 75 76 77 78 79 80 81 82 83 84 85 86 87 8B 89 90 91 92 93 94 95 US $2 billion during the 1990's, Source. Egypt Ministry of Planning and accounted for roughly 50 percent of total merchandize export (see Figure lb)1. Oil production increased Figure lb: Share of Petroleum Exports almost four-fold during 1975-1984, 100Yo from slightly more than 10 million 80% tons to more than 40 million tons, 60% and has been relatively stable at 40%- about 44-46 million tons since 20% - 1988 (see Figure 2). Marketable 0% natural gas production also grew 91 92 93 94 95 during 1975-84, and continued its Source_Central_Bank_of_Egypt, increase from less than 4 million tons of oil equivalent (toe) in 1985 to more than 10 million in 1994. While a substantial share of oil is exported, natural gas in mostly consumed at home. 'Petroleum exports (crude and products) and merchandize export figures do not include export by international oil production partners. Working Paper Annex: Natural Resource Depletion and Susstainable Growth 81 The oil resources are are primarily extracted under production sharing agreements bypintrnautionshaloil cempanis Figure 2: Oil and Gas Production (mill tons) by international oil companies whose output share 50 corresponds to their cost of 45 production and return on 40 capital. Natural gas is 35 produced under similar 25 arrangements and the price of 20 gas received by the 15 atura gas international companies is tied 10 al to the international fuel oil 5 _ . | price. The resource rent - 0 _ * the difference between the 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 economic value of oil and gas, Source: Central Bank of Egypt. and cost of production and return on capital - accrues to Egypt. The issue of the resource rent has a significant macro-economic dimension. The rent is an income to Egypt. In this respect, oil and gas reserves are income generating assets as productive capital in the manufacturing, service and agricultural sector. And, as oil and gas reserves are finite, the reserves will only generate income over a confined time period. Figure 3: Estimated Oil and Gas Rent (US $ million, current) Thus, as the non- 6000 renewable reserves of oil and gas are being 5000 Oil depleted, other forms 4000 of capital will have to 3000 be created to replace 2000 income generation 1 Gas from oil and gas. 0 * * .. . , 0 a B a s e d o n 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 estimates of the cost, normal return on capital and economic value of oil and gas resources, the resource rent (see Annex) has been on the order of US $4-4.5 billion a year during 1991-952. The oil rent fluctuated between US $3-4 billion, while the gas rent increased from US $0.5-i.0 billion (see Figure 3). The fluctuations in the oil and gas rent are due to fluctuations in international crude and fuel oil 2The economic value of oil is the price of Egypt crude oil(s), while the value of gas is approximated by the price of heavy fuel oil (f.o.b), heavy fuel and gas being close substitutes. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 82 prices. Overall the rental income from oil and gas as a share of GDP (based on current prices) declined from 14 percent in 1991 to 8 percent in 1995 (Figure 4). Projected Oil and Gas Rent Projection of future oil and gas rent depends on the magnitude of new oil and gas discoveries, world prices of crude oil, production costs, and rate of oil and gas extraction. Two scenarios are developed for the next 20 years for which assumptions are presented in Table 1. The difference between the two scenarios is a slower decline in oil slower decline in oil ~Figure 4: Resource RentlGDP production and higher real oil Fi urresourcest prices in the "high" scenario. 20%(current pces) 18% Gas reserves are at least 16% 23 trillion cubic feet, or close 14% to 60 years of current 10% production levels. Reserves 8% _ : 6% are sufficient to allow an 4%- average annual increase of 8 2% percent over the next 15 years 0% and a stabilization of 91 92 93 94 95 production thereafter beyond the scenario time horizon. Thus, no distinction in gas production rate is made Figure 5: Projected Oil and Gas Rent - 20 years between the two scenarios. (Constant US $ mill - *High case scenario) Proven oil reserves 1aD'* stand at about 3.4 billion 600 barrels, or 10-11 years of 200 current production levels. 20 . . . . . . i . . . Thus, the magnitude of -400 2 3 4 S b I t 9 1U11 12131415b 171i31 ; additional discoveries has _600 bearings on production levels -1000 over the next 20 years as can Rent Is relative to 1995-level be seen in Table 1. Projections of future oil prices entail substantial uncertainty. In the "high" scenario, real oil prices increase at an annual rate of 1.5 percent, while they are constant in the "low" scenario. The oil rent is 70 percent of crude oil prices in both scenarios (same as in the 1990's). The gas rent is fuel oil price minus long-run marginal cost of production (see Annex). Working Paper Annex: Natural Resource Depletion and Sustainable Growth 83 Table 1: Scenario Assumptions (Year 1996-2015) "High" Case "Low" Case Proven Oil Reserves 3.4 bill bbls 3.4 bill bbls Additional Oil Discoveries 1.7 bill bbls 0.6 bill bbls Gas Production Increase Year 1-15 8 %/yr 8 %/yr Year 16-20 0 %/yr 0 %/yr Oil Production Increase Year 1-6 0 %/yr 0 %/yr Year 7-20 -5 %/yr -7.5 %/yr Increase in Real Crude and Fuel Oil Prices 1.5 %/yr 0 %/yr The projected oil and gas rent for the next 20 years in the "high" scenario is presented in Figure 5. It shows an increase in rent for the next 7 years, peaking at US $ 800 million higher than the rent in 1995. The projected oil rent does not fall below 1995 levels until after year 18. The projected rent in the "low" scenario is presented in Figure 6. In this case, the oil and gas rent peaks at about US $ 400 million higher than the 1995-level after 6 years. The rent falls below the 1995-level after 8 years to US $-2,600 million in year 20. The "low" scenario may be the reality of the fiture, unless substantial new Figure 6: Projected Oil and Gas Rent - 20 Years oil reserves are discovered in (Constant US $ mill - 'Low' case scenario) the next few years. The 1500 decline in oil and gas rent 1000 500 could have negative impacts 0 e on future income in Egypt. -500 2 D 4 D u I U l I 15 1O l 18 IV 2 D The impact on future income 1000 depends largely on the use of -2000 the oil and gas rent today and -2500 - is in the domain of public -3000Years policy. * Rent is relative to 1991-level. The Competing Use of the Oil and Gas Rent Three obvious uses of the Egypt oil and gas rent are: (a) public consumption expenditures that does not increase the productive capacity of the economy; (b) public investment that does increase the productive capacity, either through investment in physical income generating capital or investment in human resources; and (c) investment in income generating financial assets, i.e. an oil and gas fund. It can be argued that Egypt's policy has been and is a combination of the three. Some aspects of the possible use of the oil and gas rent are discussed below. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 84 The public sector has dominant role in allocating the oil and gas rent. Most of the rent is made available to the central government budget, while some is invested by the oil and gas authorities in for instance gas infrastructure. The main domestic player in the oil and gas sector is the Egyptian General Petroleum Corporation (EGPC). EGPC exports and sells domestically Egypt's oil production share, whether crude or refined products. EGPC also manages oil product imports and sale of natural gas. Most of central government revenues from the oil and gas sector are from taxation of and profit transfer from EGPC. Given production sharing contracts with international partners and policy of retained after-tax profit by EGPC, the main factors that affect the level of government revenues are magnitude of oil and gas production, international oil prices, and domestic petroleum product prices. Thus, keeping domestic petroleum product prices below economic value implies a loss of government revenues and is a way of distributing the resource rent among the general population. Consumption Historically, a substantial share of the rent has been distributed to the population, although unevently, in the form of indirect consumption subsidies (prices lower than opportunity cost). Not only did this benefit larger consumers more, i.e. the relatively wealthier, but it has also resulted in excessive consumption, waste, and inefficient economic allocation of investment resources. Moreover, these subsidies can be considered a tax on the future generation as that generation's benefits of oil and gas reserves are diminished. Petroleum product retail/delivery prices and gas prices have been adjusted upwards substantially since 1990, although they have in real terms deteriorated significantly since 1993. Estimated subsidies for major petroleum products and gas are presented in Table 2, and amounted to US $2,000 million in 1991 and US $1,050 million in 1995. For petroleum products, the estimates are based on petroleum product prices and comsumption in Egypt, and international spot prices plus allowances for distribution cost (see Annex). Distribution cost from refinery to retail outlet or delivery used in the estimate is one-half of costs in the UK and Germany3. For gas, the estimate is based on bulk prices and consumption of gas, and international fuel oil prices (f.o.b) as a proxy for the economic value of gas. 3 If distribution cost equivalent to the cost in the UK or Germany is used, estimnated subsidies to petroleum products would stand at US $2,200 million in 1991 and US $1,050 million in 1995. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 85 Table 2: Estimated Implicit Subsidies to Major Petroleum Products and Gas (mill US $) 1991 1995 Gasoline -50 -295 Gasoil/diesel 725 235 Kerosene 430 110 Heavy fuel oil 520 450 Natural gas 375 550 Total 2,000 1,050 Total relative to oil rent 45% 24% The only major product with an implicit tax (-) is gasoline, and the implicit tax has increased substantially as a result of domestic price increases. However, if the gasoline retail price were to reflect the cost or road infrastructure, and social cost of traffic accidents and pollution, it should be significantly higher than the current level. The gasoline price in Egypt remains substantially below the price observed in for instance Tunisia, Morocco and most eastern European countries (and all western European countries). Implicit subsidies, in total dollar tems, to diesel and kerosene declined substantially from 1991 to 1995, while subsidies to fuel oil only declined by 13 percent. In terms of combined subsidies to the four products in 1991 and 1995, about 65 percent of the reduction can be attributed to nominal retail/delivery price increases in Egypt, 27 percent to a decline in international spot prices, 11 percent to lower domestic consumption and -3 percent to the slight nominal depreciation in the foreign exchange rate. Gas subsidies increased increased significantly from 1991 to 1995 as a result of increased gas consumption although gas bulk prices increased by more than 60 percent. While the estimated subsidies are indirect or implicit in the sense that they are not budgetary, but prevail as a consequence of price regulation, the subsidies represent a large budgetary revenue loss. The subsidies are estimated at 45 percent and 24 percent of the estimated oil rent in 1991 and 1995, respectively. Investment The use of the oil and gas rent on public expenditures for current consumption does not provide a basis for sustainable economic growth if it is accompanied by under-investment in the economy's productive capacity in the non-oil and gas sectors. Under-investment could be in areas where the public sector has a role, for example in health and education that increase human capital and in basic infrastructure that supports and increases efficiency of economic activities. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 86 For illustration, a range of rates of return for alternative uses Figure 10: Annual Incremental Investment - 20 Years of government revenues/resource (Constant US $ mill to compensate for decline in rent in Io scenario) rent are presented in Table 3. 2500- Returns to energy subsidies are 20001 estimates for Egypt in 1995 based l 500- on estimated subsidies and 1000 efficiency cost of subsidies as far as 500 - subsidies leads to higher energy o - - - - consumption. Returns to education 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 are cross-country estimates, while Yews returns to infrastructure are averages for World Bank projects from 1974-92. These rates of return can only serve as indicative for Egypt. Return to education clearly depends on more than education expenditure, and can be greatly affected by education quality variables and labor market policies. Similarly, return to infrastructure would depend on factors such as the level and quality of current infrastructure. Table 3: Illustrative Rates of Return Rate of Return Energy subsidies -5.5 % to -7.1 % Education primary 15% to 24% secondary 11% to 18 % higher 10% to 12% Infrastructure transport 18% to 21% irrigation and drainage 13% to 17% Sources: Energ, subsidies - estimate for Egypt in 1995 (see Annex); Education - Psacharopoulos 1994 (the range reflects estimated averages for and differences across regions of developing countries); Infrastructure - World Bank projects (1974-82 and 1983-92) presented in World Development Report 1994 (returns to transport projects are averages for airports, highways, ports, and railroads). Higher private sector investment could also compensate for eventual decline in oil and gas rent. Incremental annual investment to compensate for the decline in oil and gas rent in the "low" scenario is presented in Figure 10, with investments starting in year 6 to meet rent decline from year 9. The estimated annual incremental investment to yield an income (incremental GDP) equivalent to the annual rent decline is based on an incremental capital- output ratio of four and a straight line depreciation rate of ten percent. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 87 Oil and Gas Fund An alternative, or combined, use of the oil and gas rent is to invest the rent, or part of it, in financial assets to compensate for the depletion of income generating oil and gas. Many oil and gas producing countries haveestablished such funds, Egypt being one of them. Egypt's fund from oil and gas revenues stands at about LE 5.1 billion (US $1.5 billion) in fiscal year 1995/96. The fund increased from LE 0.8 billion in 1989/90 to LE 3.8 billion in 1990/91 and has increased modestly each year thereafter. The fund is raised through the issuance of bonds, primarily held by EGPC. The fund serves as an oil revenue stabilization fund in the event of declining international oil prices. The purpose of the fund is also investment and development of alternative energies to reduce domestic dependence on consumption of fossil fuels. Indirectly, the fund also serves as a transfer of oil rent from the present to the future and could cushion a decline in future oil and gas rent. An issue is, however, how large does such a fund have to be if the aim is to generate an equivalent income to the decline in rent with the reduction of oil and gas production revenues, and how much should annually be placed in the fund? This depends largely on three parameters, i.e. the annual oil and gas production to the year of depletion, the future price of oil and value of gas, and the rate of return on the financial assets. A substantial degree of uncertainty is associated with the first two parameters. The "low" case scenario of new oil discoveries of 20 percent addition to current reserves, and constant real prices of crude oil and fuel oil, is used to estimate the annual contribution to the fund. The real rate of return on the financial assets of the fund is taken to be 4 percent. Annual contributions to the fund is presented in Figure 11, and amount to two-thirds of the projected annual oil and gas rent. A contribution at that level would obviously have significant negative impact on current income. Fostering private sector investment, investing in priority infrastructure, quality education and health improvements seems like healteimativesento themanticted Figure 11: Annual Contributions to A Compensation Fund viable alteatves(Constant US $ mill per year - based on ow scenario) decline in future oil and gas rent. 4000 - 3500 _______________ 3000 40- Summary and Conclusion 2500 r 1500 The share of the oil and gas '°° sector in GDP stands at about 10 0. .; percent. Crude and petroleum 1 2 3 4 5 B 7 8 9 10 11 12 13 14 15 16 17 18 19 20 percent. CueadptoumYears product exports accounts for about 50 percent of total merchandize exports. The resource rent, i.e., value of oil and gas less production cost and normal return on investment, is estimated at about US $4.4 billion, or 8 percent of GDP. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 88 The level of future oil and gas production and consumption could have significant bearings on future income and the external balance. Proven oil reserves stand at about 3.4 billion barrels, or about 10-11 years of current production level. Gas reserves stand at about 24 trillion cubic feet. While oil production may be kept at current levels for the next 5-6 years, the magnitude of the gas reserves allow for significant production increases for at least 15 years. Two scenarios of projected oil and gas rent and net energy exports are presented. The "low" scenario could become the reality - a decline in oil production of 7.5 percent a year from year 2002. In this case, the combined oil and gas rent would fall below the 1995- level in year 2004 and thereafter decline rapidly to only 40 percent of the current level in year 2015. Net export supply of energy would decline at a faster rate than the oil and gas rent after year 2002 because of increases in domestic energy consumption. By year 2008, Egypt could be a net energy importer unless substantial oil discoveries are made. Although declines in oil and gas rent and net energy exports may be avoided during the next 6 years, alternative uses of oil and gas revenues deserves serious consideration. Petroleum product and natural gas prices have been raised significantly since 1990, but have declined in real terms since 1993. Estimated implicit subsidies to petroleum products and gas stood at more than US $1 billion in 1995, or almost 25% of the resource rent. Estirnates of the efficiency cost of energy subsidies in Egypt in 1995 suggest that energy subsidies have a rate of return to the economy of -5.5% to -7.1%. In contrast, well spent investments in education and priority basic infrastructure could yield on the order of 15-20% return. Fostering private investment already forms part of Egypt's growth strategy. Incremental private sector investments, rising to US $2 billion a year (above current levels) by the year 2014, would compensate for the anticipated decline in resource rent from oil and gas. An oil and gas fund could be a combined strategy to compensate for an eventual decline in resource rent. Egypt already has such a fund, as many other countries, but the amount required to fully compensate for anticipated rent decline would swallow two-thirds of the current rent income, thus potentially affect current income substantially in a negative direction. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 89 APPENDIX Egypt Crude Oil Prices The Egyptian crude oil prices have Figure Al: Price of Dated Brent (US $/bbl) been set based on a formula with weighted 40 35 average differentials to 30 selected internationally 25 traded crudes, of which 20 Dated Brent was the 15 most important (see 10 Figure Al for historic 5 annual average price of 0 Dated Brent). The 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 formula was simplified in January 1996 with Source: EnergyPrces andTaxes OD export crude prices set relative to the Dated Brent. The average differentials of major Egyptian crudes to the Dated Brent are presented in Table Al. The differentials are reviewed monthly. Table Al: Prices of Egyptian Crudes - 1996 (average differentials relative to Dated Brent - US $/bbl) First Quarter Second Quarter July-Aug Suez Blend -1.33 -1.63 -2.45 Ras al-Bihar -0.95 -1.03 -1.60 Zeit Bay -1.00 -1.08 -1.65 Belayim -2.03 -2.36 -3.13 Ras Budran -2.18 -2.54 -3.33 Ras Gharib -3.03 -3.44 -4.18 Source: Calculations based on monthly differentials published in Arab Oil and Gas (various 1996 issues). In the pricing formula prior to 1996, the Suez Blend served as a reference crude which price was tied through price differentials to the Dated Brent and the Iranian Heavy crude by a weighting formula. The other Egyptian crudes were tied to the Suez Blend. The Suez Blend pricing formula with differentials during 1992-95 are presented in Table A2. The price differentials of the other Egyptian crudes relative to the Suez Blend during 1992-95 were in the range of -3.85 to +0.65 (US $/bbl). Working Paper Annex: Natural Resource Depletion and Sustainable Growth 90 Table A2: Price of Suez Blend 1992-95 (differentials relative to crudes in weighting formula - US $/bbl) Weight 1992 1993 1994 1995 Dated Brent 60% -2.4 to -3.2 -2.8 to -3.2 -1.2 to -1.7 -0.6 to -1.9 Iranian Heavy 20% -0.5 -0.5 -0.5 0 to -0.5 Suez Blend 20% parity parity parity parity So-urce: Arab Oil and Gas (various issues 1992-95) Production of Suez Blend crude was about 32 percent of total production in 1993 (against about 32% in 1992), Ras Gharib about 20 percent (22.5%), Ras Budran about 19 percent (12%), Ras al-Bihar about 14 percent (11.5%), and Belayim about 13.5 percent (19%). Economic Value of Natural Gas Figure A2: Heavy Fuel Oil Price (US $Sbbl) Natural gas produced in Egypt is 25 consumed domestically 25 and there is so far no 20 pipelines for export. An 15- export parity price is 10 - therefore not available. Natural gas consumed domestically is largely 76 76 77 78 79 80 81 82 83 84 86 86 87 88 89 90 91 92 93 94 96 replacing the use of Source: Energy Prices and Taxes; OECD. petroleum products in power plants and industry. The petroleum product replaced is predominantly heavy fuel oil. The price of heavy fuel can therefore serve as a reasonable proxy for the economic value of natural gas. In fact, production agreements with international oil companies in Egypt set the price of natural gas received by the oil companies relative to the international market price of heavy fuel oil. The annual average international spot price of high sulphur heavy fuel oil is presented in Figure A2. Calculation of Oil and Gas Rent The annual realized oil and gas rent is the economic value of annual oil and gas production, less costs of production and normal return on capital. The economic value of the oil is the Egyptian crude oil prices multiplied by the quantity of the respective crude production volumes. The economic value of the natural gas is the international high sulphur heavy fuel oil price multiplied by the tons of oil equivalent quantity of gas. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 91 The cost and normal return on capital in use for oil production are not readily available. However, since most of Egyptian oil production is under production sharing contracts with international oil companies, the oil companies' production shares serve as a good proxy for the cost and return on capital. The aggregate annual production share has been in the neighboorhood of 30-33 percent in the five year period from 1990-94. Thus, the oil rent is estimated at 67-70 percent of the economic value of crude production. Since the most recent years are more relevant for assessing the role of the rent in the Egypt economy today and future policy associated with the rent, less emphasis have been placed on an accurate calculation of the rent in the 1970's and 1980's. Thus, the rent has been calculated based on 70 percent of the crude oil value, although this may understate the rent in periods of high oil prices and overstate the rent in the 1970's. In terms of natural gas, British Gas calculated the long-run marginal costs for Egyptian General Petroleum Company (EGPC) in 1994. The costs are summarized in Table A3. The long-run marginal cost of existing fields have been applied to calculate the gas rent of current and historic gas production. For historic years, the cost have been deflated. Table A3: Long-Run Marginal Cost of Gas Exploration and Production (US $ per 1000 cubic meters; 1993/94) Gas fields Base case Most likely case Existing fields 11.1 11.1 Existing discoveries 38.3 37.7 Unproven reserves 40.5 40.3 ource: EGPC/British Gas. Oil and Gas Reserves and Production Proven oil reserves in Egypt stood at about 3.4 billion barrels as of January 1, 1995, equivalent to about 10- 11 years of production at current production levels of about 45 million tons. Proven reserves have been relatively constant for several years, implying that recent discoveries have replaced annual production. However, output from existing fields is projected to decline over the coming years. Thus, substantial discoveries will have to be made to sustain annual production at about 44-46 million tons. Of 23 principal oil fields in Egypt, six were discovered in the 1980's. Seven fields account for about 80 percent of total oil production. Gupco, Petrobel, and Suco are the three largest operators in Egypt, accounting for slightly more than 80 percent of total oil production. In 1993, Gupco's annual oil production stood at somewhat more than 20 million tons, Petrobel's production at a little over 12 million tons, and Suco's production at almost 4.5 million tons. Working Paper Annex: Natural Resource Depletion and Sustainable Growth 92 Oil production by Gupco and Suco has fallen since the mid-80's, while Petrobel's output has been increasing. The companies have recently intensified efforts to increase recovery rates from existing fields by installing enhanced recovery facilities (using gas lift and water injection) and by developing new extensions to their fields. These measures contribute significantly to sustain Egypt's total annual production levels. However, from new field discoveries, production tend to be relatively small. Four fields started production in 1994 with a total production capacity of 35,000 barrels per day, or about 3.9 percent of total annual production from all production fields. Natural gas reserves were estimated at about 21 trillion cubic feet in 1995, or almost 600 billion cubic meters. In contrast, annual production in 1995 was a little more than 10 million tons. Gas production has increased at an annual rate of 8-9 percent in the 1990, and Egypt's policy is to continue gas production at or about this rate of increase. If annual production is increased at an annual rate of 8 percent until production reaches about 30 billion cubic meters, and then levels off at around 33 billion cubic meters, the reserves of almost 600 billion cubic meters will last for more than 20 more years and it is unlikely that no more discoveries are made. Petroleum Product and Gas Subsidies Retail (end-user) prices in Egypt for major petroleum products and gas bulk prices are presented in Table A4. All prices were substantially increased from 1990 to 1993, but have not been adjusted since 1993. While budgetary subsidies to petroleum products are almost nil, price regulations have kept retail prices below economic value or opportunity cost for all major products except gasoline, implying that implicit subsidies prevail. Table A4: Petroleum Product End-User Prices in Egypt Unit 1990 1990 1991 1992 1992 1992 1993 May Sept May Jan June Dec July Premium gasoline LE/ltr 0.55 0.60 0.80 0.80 1.00 1.00 1.00 Regular gasoline LE/ltr 0.50 0.55 0.70 0.70 0.90 0.90 0.90 Kerosene LE/ltr 0.10 0.10 0.20 0.30 0.30 0.30 0.40 Gas oil LE/ltr 0.10 0.10 0.20 0.30 0.30 0.30 0.40 Gas oil (for power) LE/ltr 0.15 0.15 0.25 0.25 0.35 0.35 0.45 Diesel oil LEAltr 0.09 0.09 0.18 0.27 0.27 0.27 0.36 Fuel oil LE/ton 50 50 80 80 100 130 130 Gas (bulk) LE/th m' 46.7 46.7 75 75 94 122.5 122.5 ouc:Mimsterial Decrees- ___-___ ______ -___ Working Paper Annex: Natural Res,'urce Depletion and Sustainable Growth 93 Efficiency Cost of Energy Subsidies Historic energy subsidies in developing countries have been substantial for many reasons. Policy makers' arguments have included protection of the poor and infant industry. However, subsidized energy prices (direct budgetary transfers and/or price regulations) provide reduced incentive to use energy efficiently, resulting in excessive consumption compared to the level that would prevail if prices reflected opportunity cost of energy. A methodology often used to assess the annual efficiency cost of subsidies is the estimation of deadweight loss. The deadweight loss can be seen to represent the resource waste in the economy. When a subsidy is removed, users of energy will tend to reduce energy consumption as long as energy efficiency improvements and conservation cost less than the increase in the energy price. Thus, the difference between the economic value (opportunity cost) of energy saved, and the cost of saving energy, is the deadweight loss (or resource waste) in the presence of a subsidy. The annual deadweight loss (DWL) can be estimated in two steps: (1) Qe = Qs (ps / p)e where Qe is energy consumption (tons) in the absence of subsidy, Q5 is energy consumption (tons) in the presence of subsidy, ps is the subsidized energy price (LE/ton), pe is the price at economic value of energy (LE/ton), and E (< 0) is the price elasticity of energy demand. (2) DWL = (pe _ pS)(Q5 - Qe)/2 with a linear approximation of the energy demand function in the relevant price range. Based on the consumption of major petroleum products (gasoline, kerosene, gasoil/diesel, and fuel oil) and natural gas, domestic prices, and estimated economic value of the products and gas, estimates of the deadweight loss in 1995 are presented in Table A5 for various assumption of the price elasticity of demand. The deadweight loss as a percentage of estimated subsidies can be seen as the (negative) economic rate of return to providing energy subsidies. Table A5: Estimated Annual Deadweight Loss in Egypt (1995) Deadweight loss Deadweight loss (mill US $) (% of estimated subsidies) Elasticity of demand -0.3 58 5.5% -0.4 75 7.1% -0.5 92 8.7% ARAB REPUBLIC OF EGYPT IBRD 27759 280 3 20 3 -6 Medciterranean Sea wEsTBANKANDGAZ4 / 320 ~ I Marsa Matruh Alexandrioo af el d 0 SRAEL' JORDAN \ ) ~~~~~~~~~Danianhur in El Monsron \- ( Shbinel Kom Smadil(yo OSuez \ I' Giza IRO El Fayom ARAB REPUBLIC Beni Suef OF EGYPT ElMinyao SAUDI -28, ARABIA 28- 0 Assiut Hurghoda Sohag Red Qena Sea 0 SELECTED CITIES x ®I NATIONAL CAPITAL r Luxor I ~~~~~~~~~~~~~~~Matoan RIVERS CULTIVATED AREAS Kom Ombo 240 Aswan 240 I ~~~~~~~~~~~~~~~~~~~~~~swan I.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~h 0 50 100 150 MILES 0 50 100 150 KILOMETERS [ hsmap w'os produced by the Map Design Unit of rheW frId ank fiheboundaries, colors, denomu,at,ons ond an other inormrotion show,n on Ithsmap do not Imply on the port of The Wo on Group, any judgment UDAN , on the legal shtvs ofony terntory , or any endorsement or acceptance of ! such boundories 280 32o 360 APRIL 1997 0C;;f:0f j j0000 0f Ff00X0 A;SSS ffff000000 0 0000000000f000?00 0fX0 000 0: f00;;0 00 ff 0f0f: 0 000ff0 0l;Xff;;0ff S: f:0000000000 0 00 : figitliLiEEifEti> A:;:;Ff *0:L;j.D-ldlLtSetSH t.0liLiNi:0-0iLS._:f::t:000-02titifLtiViidiLit- tLLLtLLLiLi;;00000f:LLiLa D::$itSEk,000titiN >ititESEiVi; 0 : : : : ;: : : A: 00 f: 0 : ;;: : : : : S: : : 0 : :f : :: : :: ff: ff : : : : :; ff f 0 j ff 0 00 X f:S S; f 00 ff f: : : : : : : : 7 f :: : : : : f:: A: S ; : :: :f S X : f : : 7 :; : f 0 D 0 : : f : : t: f :; : f : 0 : ti 0 : f X 0 S D 0 : f : d00:: : : f : t; SS: f: : : : : :: : : : : : ::: : : : : : : : :: : S : : : :: f : :: l : L::LLEL: :: :SSLS::i:: f C: :tS t: : : :S: : :