0 C TO B E R 2 0 0 0 :~ :NENTOA :GROUP-- 1j 21143 ,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~ ................. .... .. ger der 4ivcrsity-f ' A PROGRAOFT ETV GRU ON., INTERNATrONAL AGRICULTURAL RESEARCH (CGIAR) ~~~~Geder an-d Diversity in the,CGJA-kR: A New, Baseline- _,..REPAREDBY:JOANJSHI VICKI WILDE :- ,. -~. ADRIAN MASTERS, - Swokn p- ~ the CGIAR A New wasking pap-- V '0-' f ' gendetf' 4iversity' A - . PROGRM OF TI~cOIVE CRGUP ON , RNER'hAIIONAL AGRICULTURAL FESEARCHU (toIAR. CGIAR. The Consultative Group on Intemational Agriculturat Research technical assistance and management consulting, training, eMd (CGIAR) is an informal association of public and private sector ,information services. The CGIAR Gender & Diversity Program is members, estabished in 1971, wihose rmission is to contibute hosted at ICRAF -(Nairobi, Kenya), with the leadership 'Qf Vicki ,through research to sustainable agriculture for food security in Wilde (vwiLcgiar.org). developing countries.TheWortd Bank, theFood and griculture The Gender,and Diversity Progamn aims ta,institutionalize the Organization the United Nations Devepment Program leadership, skills, polides ard norms thst-will ensure the centers' (UNDP), and the lnited Nations Environmnent Program (UNEP) capacit9:to manage ind capture the-full benefits of staff diversity. *jointly sponsor the CGIAR. lt is miade up of 16 international t The ptogram has seven objectives:- agricultural research centers lotated in 12 developing and 3 , To diag'nose staff diversity issues in the centers and develop a devloped countries. Today, productv and naturl resources conceptual framework for addressing those issues to enhance management are the twin pillars of CGIAR research on food crops, both equity and organizational eftectiveness. forestry, livestock, irrigetion management, aquatic resources and okicy rssues, and in its services to national agricultural research * To encIurge and"support leadership on genger and diversiy systems. T.e cen.ers employ about 1000 scientists of ~ issues ambng senior leaders in the CGIAR system. systems. The~ cena,ers emptoy about 1000 scieitists of tOO 0 different nitionalities. For more information about the CGIAR, see: To strengthen knwtledge and skills for center staff to work www.cgiar.org. - with diversity effectively. - - To enhance centers' ability to attract high quality staff from GENDER AND DIVERSITY PROGRAM diverse 'tdentity groups. The CGIAR Gender and Diversity Prograrn is about 'excellence To'support changes in policies, formal systerns and work' ,through-equity. tts 'overall goat is to assist the 16 agiculture - norms and practic*s to ensure equal opportunitres for leader- research centers of the COIAR to respond to the urgent need to s-sip, career develop_ent and involvement in deCsion-making atract, value and manage gender ,nd diversity.- In 199%; the - for women and nen of diverse identity groups. 'GIAR laundched a program-on gen-der staffirng aimed at assisting- --''- . ' .CGIAR .au d a progron gender staffing aimed at assistingTo support changes in work culture and practices that create a the interrfational agricultural research centers in their efforts to . opiabe n - * . . . - h~~~~~~~~~~ospitable and supportienvr t for diveifse'staff. promote the recnJitment, productivity, advancement and retention s e f dvst. of womren scientists and professionals In 1999, this program' To support the CGIAR, at center tnd -system levels, th institu- - o*ep e to include a wide range -of diversity issues. The program - tionalize the policies, cornmitment, knowledge and skills proYides support to the rsrch centers through small grants, necessary for working with a diverse staff effectively. .' - ' . ~ , ,CIAT Centro Internacional-de Agricultura Tropkal (COLOMBIA) CIFOR Center for Intemationat Foretry Research (INDONESIA) 'CIMMYT Centro ihtemaciona de Mejoramiento de Maiz y TrngD (MXICO), - :CIP Centro Internaciorial.de la Papa (PERU) - ICARDA International Center for-Agricultural Research in .the,Dry Areas (SYRIA) 11CLARM--, Intemational (Center for Living Aquatic Resources Management (PHILIPPINES) . - , 1. ICRAF, International Cente'r for Research in Agroforestry (KENYA) * : fi , ICRISAT Int,eationil Crops Research Institute for the S4mi-Arid Tropics (INDIA) - . ' ; - ; IFPRI Intemational-Food Policy-Research Institute (US4 C. - IWMI international Irrigition and Water Management,Institlute (SRI LANKA) . 0 , - IITAt -. Intemational Institute of Tropical Agriculture (NIGERIA) , * a 2 'ILRI Intemational Livestock Research Institute (KENYA), v - \ 0 ' IPGRI International Plant Genetics Resources Institute (ITALY) IRRI international-Rice Reseatch Institute (PHIUPPINES) . , ISI4AR International Service for National Agiculturat Research (THE NETHERLANDS) * . , 0 0 WARDA West Africa kice Development Assocation (COTE D'WvOIRE) GENDER AND DIVERS1TY IN THE CGLIR: A NEW BASELINE Prepared by: Joan Joshi Vick Wilde Adrian Masters Working Paper 25 October 2000 Table of Contents Acknowledgements ii Executive Summary v I. Introduction 1 II. General Profile of Internationally-recruited Staff in mid-1999 2 III. Gender Profile and Issues 8 IV. Diversity Profile and Issues 18 V. Analysis of Compensation and Positional Equity by Gender and Diversity 24 1. Overview 24 2. Data and Methodology - as applied to five comparison groups: 25 A Women as compared to men; B. Staff members of WB Part II origin as compared to staff members of WB Part I origin; C. Staff members of WB Part II origin, who are now citizens of WB Part I countries, as compared to staff members who have both WB Part I origin and citizenship; D. Staff members of WB Part II origin who are posted to their home regions as compared to other staff members of WB Part II origin (ie. not posted to their home regions); E. Staff members in each of the three disciplinary areas (I: management and information, II: social sciences, III: natural sciences) as compared to staff members in the other disciplinary areas. 3. Compensation Equity 27 4. Positional Equity 28 VI. Follow-up 32 Appendices: 1. World Bank Part lI/Part II Countries 2. Position Groups 3. CGIAR Human Resources Survey (1999) 4. Survey data from 1991, 1994 and 1997 5. Fields in the 1999 survey 6. Details of regression analysis discussed in Section V 7. List of Acronyms i :;I Acknowledgements The Committee of Deputy Director Generals (CDDC) has made this study of gender and diversity issues in the CGIAR possible. During International Centers Week (ICW) in 1999, the CDDC granted the Gender and Diversity Program access to the salary information of every internationally-recruited staff member worldng for the international research Centers, which provided the necessary data for equity analyses. For their trust, and for their commitment to helping us establish a factual baseline for our gender and diversity work, many thanks. Extra special thanks go to the co-authors, Joan Joshi and Adrian Masters. As a virtual team, we have enjoyed the benefits and challenges of working together while on three different continents. And, with the perspectives of three different disciplines, we were an example of "functional diversity" in action. Joan, who was the lead consultant for the systemwide Compensation Study conducted in 1999, is the one who chased down the data, checking and re-checking it, and giving us the benefit of her many years with the CGIAR System in the analysis of its findings. Adrian, as the labor economist, kept us on the straight and narrow path of statistical rigor, never letting us stray from what the data could fully support. To both, heartfelt thanks. I've thoroughly enjoye4 our teamwork. This study benefited from the counsel and guidance of several CGIAR staff members, but special thanks go to Howard Elliott, Principal Research Officer of ISNAR (a labor economist by training), who took the time to review our early drafts, providing us with technical advice on the approach as well as the institutional background for interpreting the data. Members of the Gender and Diversity Advisory Board also reviewed the draft and offered guidance. Special thanks to Ragnhild SohIberg, Board Chair of ICRISAT and Vice President, External Relations and Special Projects of Norsk Hydro; and to Roselyne Lecuyer, Human Resources Manager, ICRAF. It is the personnel and human resource officers of the 16 centers who deserve the greatest thanks. We fully recognize that providing the disaggregated information for this study required hours of digging through ifies and flling in our survey forms. To you, and to all senior managers of the CGIAR, our greatest hope is that this working paper will be useful to you. Vicki Wilde Program Leader CGIAR Gender and Diversity Program iii AT EXECUTIVE SU1MMARY The Compensation Survey relevant to CGIAR internationally-recruited staf£ conducted in 1999, presented an opportunity to update three earlier surveys on representation in the CGLAR Centers by gender and diversity of national origin. It also made possible an initial look, given limitations in the available data, at the question of equity in both compensation and classification in position groups (referred to in the following as "positional equity"). The survey can thus serve as a new baseline for the CGIAR Gender and Diversity Program, which was established on July 1, 1999, to succeed the Gender Staffing Program. This report presents an analysis of the CGLAR System as a whole, providing general guidance to individual Centers and to the Gender and Diversity Program. However, since action to address questions related to gender and diversity lies with the management and boards of individual Centers, each unit of the System also will receive a Center-specific analysis of the issues covered in confidential reports to senior management. In drawing a general profile of internationally-recruited staff members (IRS) in the CGIAR System, the report looks at representation by gender and by country of origin, using the World Bank's (WB) Part L Part II designation. Part I includes generally industrialized, donor countries (predominantly northem), while Part II includes generally lesser developed countries that are the recipients of IDA loans (predominantly southern). The data show that the total complement of 966 IRS is comprised of 162 women (17 percent) and 804 men (83 percent). This is an increase in the percentage of women from the 1991 date when statistics were first collected. At that time, women represented 12 percent of total staff. The diversity of the staff has also increased, from 43 percent staff from WB Part II countries in 1991 to the current 47 percent. Results of the survey data are also presented regarding the position group2 of staff members by gender and WB Part as well as disciplinary area, level of last degree, country of last degree, years of relevant professional experience, tenure at the respective Center, age and personal status. Tlhe report scrutinizes the data, first with respect to gender, then with respect to diversity of country of origin. It considers the implications for the System's goal of representational equity and puts forward a series of questions that invite firther investigation at both Center and System level. Section V presents an analysis of compensation and positional equity by gender and diversity, with "equal-pay-for-work-of-equal-worth" as the standard of compensation equity. Using regression analysis to control for permissible factors (position group, last degree acquired, years of relevant professional experience and tenure at the respective Center), it investigates how well the compensation and position group structure of the ' See Appendix I for list of countries by World Bank designation 2See Appendix 2 for description of position groups. v The intetationalyrecrited: staff members ite CGlA Sysin the aggregate, tead to vharacteisics:il 00pedoinately male (83%); : fairly evenly divide:d jbetween hse frOm WB Part I and WB 1 P tuntre With 100 individual couniri es of otigin rereene;0 * about hlf are ot:inPGroups IV ( Seniior Scientis ssioal)ad V: : (Scientists/Professional) the levels where benchScientists tend to cluster (48%); * a lag aoiyare natural scientists (70%); * most holdPhDidegrees(78%);: * PhDs and other terminal degreestendto be froimacademic insts insm WNB Part I counties(8%) * most have ten or more years of relevant pssil experiee (61%) ; . they;are reatveyoatur,ih53% overfage 45, 92% over age 35; *0 a substantil majorit (80)are accompanid t pbstbyas4 pouse/partner,V whuile 51% have; childreni *ith them * about bhafhefbeeinat their restdieCenter fewer han thireeas(45ercent). CGAR System accords with this equity standard with respect to the following compansons: A. Women as compared to men; B. Staff members of WB Part II origin as compared to staff members of WB Part I origin; C. Staff members of WB Part I origin, who are now citzens of WB Part I countries, as compared to staff members who have both WB Part I origin and citizenship; D. Staff members of WB Part II origin who are posted to their home regions as compared to other staff members of WB Part II origin (ie. not posted to their home regions); E. Staff members in each of the three disciplinary areas (I: management and information, II: social sciences, Im: natural sciences) as compared to staff members in the other disciplinary areas. The report acknowledges that no data is available for two vaxiables that would, and should, influence compensation and advancement to higher positions: merit or performance quality and evidence of specific managerial skills (e.g. interpersonal or supervisory skills, the ability to manage financial resources) needed for positions increasing in responsibility and authority. It is important to vi note that there is no reason to believe the unmeasured variables would impact systematically across each Comparison. With these caveats in mind, when controlled for the above-listed factors, the principal findings reLative to compensation equity are: * there is no significant difference in salary between groups in Comparisons A, C, D or E; (This is particularly encouraging with respect to Comparison A; the fact that the CGIAR Centers are largely equitable in their financial treatment of women professionals makes them exceptional in the world of science.) * it appears that staff members of WB Part I origin have a highly significant salary advantage (6.5 percent) over staff coming from a WB Part II county. In considering positional equity, the most significant findings are: * women are 12 percent less likely than men to be Research Program/ Administrative Heads or Principal Scientists (Position Groups 1-11E) or higher, and 15 percent less likely than men to be Senior Scientists/ Professionals (Position Group IV) or higher; although the CGLAR women are generally younger than the men, this finding remains true when the analysis is extended to include a control for age; * staff members from WB Part II countries are 14.9 percent less likely than those from WB Part I countries to be Senior Scientists/Professionals (Position Group IV) or higher, and 6.5 percent less likely to be Scientists/Professionals (Position Group V) or higher; when considered with age (WB Part II staff tend to be older), it appears this staff group has remained in a relatively steady state; * natural scientists are 15.3 percent less likely than other staff members to be Research Program/Administrative Heads or Principal Scientists (Position Groups II-m) or higher, and 19.6 percent less likely to be Senior Scientists/ Professionals (Position Group IV) or higher. While there is good news in the findings of this survey to be celebrated, clearly results have emerged that suggest further investigation and assertive follow-up action on the part of the Gender and Diversity Program Among the actions suggested are the following: * Continue to expand and diversify recruitment strategies to "cast the net ever more widely" to attract more women natural scientists, more women social scientists and more women in the fields of management and information. * Focus especially on the identification of sources of qualified women from developing countries (WB Part II countries). * Explicitly support CGIAR women's advancement, aiming to bring more women into positions of mid-level and senior management. * But also encourage training in managerial skills for men with degrees in the natural sciences. vii * Research and communicate best practices of sister Centers and of other organizations with respect to policies, practices and workplace culures to ensure that women are retained; include in-depth studies on work/family balance. * Seek and promote ways to respond to the needs of dual-career families. * Investigate more closely the question raised about the dominance of staff members with degrees from European and North American universities; consider a conscious effort to identify high value PhD programs offered by academic institutions in the South. * Similarly, investigate more closely reasons for the disparity between WB Part I and Part II staff members in position levels above Scientists/Professionals (Position Group V), as well as apparent discrepancies in compensation for WB Part II staff members at all levels. * Work to ensure accountability of CGLAR managers on gender and diversity issues. viii L INTRODUCTION The CGIAR is sometimes perceived as an exclusive club of white males from selected United States universities. Is this true? Or does the CGIAR--as an international system covering five continents--fully tap into the global workforce to attract the world's best scientists and managers, regardless of nationality or gender, and, in turn, benefit from their diverse skills and perspectives. Are CGIAR women and men, and developing and developed country nationals, treated equitably in terms of compensation and position classification? Based on a foundation of CGIAR data, there is encouraging news in this report. The facts show that not only are the Centers highly diverse in their staffing, they are largely equitable in their financial treatment of women professionals. That makes the CGIAR exceptional in the world of science. But the facts also show that progress is often much too slow, and that persistent issues of fairness require our immediate attention. The CGIAR's Gender Staffing Program, the forerunner of the Gender and Diversity Program, conducted surveys in 1991, 1994 and 1997. Those surveys gathered data about the women and men among the internationally-recruited staff members (IRS), their nationalities, their hierarchical levels and the disciplines they represented. The System drew significant lessons from those numbers, including the need to search aggressively for new sources of female candidates. Now, a new CGIAR Human Resources Survey has been conducted and is presented in this paper. This survey was 4esigned to update the earlier figures but, additionally, to take an intensive look at the question of equity in compensation and classification in position groups (referred to below as "positional equity") from the perspectives of both gender and diversity of national origin. The Gender and Diversity Program began its on work July 1, 1999. It builds on the previous program while incorporating new objectives and strategies mandated by an Inter-Center Consultation held in The Hague in 1998. In its new incarnation, the Program will assist the 16 agricultural research centers of the CGIAR as they respond to the urgent need to attract, value and manage gender as well as all other attributes of staff diversity, and thus fully utilize the skills and resources found within a global workforce. Importantly, the Program will also be concemed with issues relevant t, the Centers' nationally-recruited staff members (NRS). The survey reported here, however, continues the original focus on those brought on board through intemational recruitment. It will serve as a new baseline for work to achieve "excellence through equity" for this group of staff members. The data on which this report is based were collected from the Centers during the course of work on the recent CGIAR Compensation Survey and are comprised of figures valid as of September 1, 1999. The report takes a broad view of the CGIAR System as a whole. This is an important lens for the Gender and Diversity Program Leader and Advisory Board, and of undoubted interest to donors. However, since action to address gender and diversity issues lies with the management and boards of indirvidual Centers, each unit of the System will also receive a Center-specific analysis of representation, compensation and positional equity using the statistical methodology of labor economics. 1 IL GENERAL PROFILE OF ENrERNATIONALLY-RECRUYIEED STAFF IN MID-i999 Intemationally-recruited staff members (IRS) totaled 9661 in mid-1999, distributed among the 16 Centers as indicated in Table 1. Table 1. IRS by gender and World Bank Part, country of origin (percent representation at respective Center) Center | Total | Women | Men |WB Part I WB part H CIAT 84 19 (23%) 65 (77%) 47(56%) 37(44%) CIFOR 35 10 (29%) 25 (71%) 23 (66%) 12 (34%) CIiMMYT 99 20 (20%) 79 (80%) 49 (49%) 50 (51%) CIP 62 12 (19%) 50 (81%) 30 (48Y%) 32 (52%) ICARDA 76 6 (8%) 70 (92%) 28 (37%) 48 (63%) ICLARM 23 3 (13%) 20 (87%) 16 (70%) 7(30%) ICRAF 50 7 (14%) 43 (86%) 32 (64%) 18 (36%) ICRISAT 47 4 (9%) 43 (91%) 22 (47%) 25 (53%) EFPRI 67 19 (28%) 48 (72%) 45 (67%) 22 (33%) UITA 100 18 (18%) 82 (82%) 53 (53%) 47(47%) LLRI 95 16 (17%) 79 (83%) 54 (57%) 41(43%) IPGRI 39 9 (23%) 30 (77%) 22 (56%) 17 (44%) IRRI 104 11 (11%) 93 (89%) 48 (46%) 56 (54%) ISNAR 37 6 (16%) 31(84%) 24 (65%) 13 (35%) IWMI 23 2 (9%) 21 (91%) 15 (65%) 8 (35%) WARDA 25 0 (0%) 25 (100%) 8 (33%) 17(68%) TOTAL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) In 1991, the CGLAR Centers employed 1,295 internationally-recruited staff members, including 153 women, 12 percent of the totaL By 1994, with a slight overall decline to 1,224, the number of women had grown to 173 or 14 percent. Another small decline to 1,190 in 1997 saw a further increase in the number of women to 188 or 16 percent. Today, as shown in Table 1, the total of 966 includes 162 women, nonetheless an increase to 17 percent. The 25 percent reduction in the total staff during the eight years in question stems principally from constraints in System fimding, but also in part from the increasing availability of well-trained, nationally-recruited scientists. Further analysis of attributes with respect to gender is in Section m. IThe actual total is 976; ten entries have been dropped from this analysis due to incomplete data. 2 Diversity over time is viewed in this report from a number of perspectives, detailed in Section IV. Worthy of note here (also in Table 1), is that current data counts 450 staff members, 47 percent of the total, as originally from World Bank Part II countries. The World Bank (WB) categorizes countries as Part I, generally industrialized countries that are donors to the International Development Association (IDA), and Part IL, generally lesser developed countries that are the recipients of IDA loans. A complete list of the WB Part I and Part II countries is in Appendix 1. Although the earlier surveys did not use this World Bank indicator, they did take into account country and region of citizenship. The 1991 survey of 1,191 internationally- recruited staff members showed that 57 percent were citizens of Europe (358), North America (258), Australia/New Zealand (41) and Japan (20), wl4le only 43 percent were from Asia/Oceania (204), Sub-Saharan Africa (159), the Latin American/Caribbean region (108) and West Asia/North Africa (43). The tables and analyses in this survey report focus on country of origin rather than citizenship, in the belief that the former is the more relevant measure of ethnic diversity. Respondents were instructed to define and report "origin" as country of birth, unless that country did not represent the staff member's true nationality, e.g. those born as refugees or as the child of expatriates working abroad. In the end, however, the difference between origin and citizenship was slight. The survey found only a 3 percent net change of citizenship from WB Part II to WB Part I countries, four women and 26 men. Overall, the CGIAR System Centers employ staff members from 100 countries of origin (19 WB Part , 81 WB Part II) who are citizens of 97 countries (18 WB Part 1, 79 WB Part 11). In compiling data for this survey, respondents were asked to place their staff members in one of seven standardized position groups, carefully defined by levels of responsibility and authority as well as basic qualifications, in order to maximize the validity of cormparison (see Appendix 2). This is at once the most important and most difficult part of any human resources survey, because personal judgement is determinate in the final analysis. However, the position classification system outlined follows a pattern that is generally, alhough not uniformly, applied in the Centers. The 1999 breakdown by position group is shown in Table 2. It is important to note that Position Group I, the Executive Staff level, does not include Directors General There is currently one woman among the 16 Directors General in the CGIAR System. There is a notable variation across Centers in the distribution of individuals into position groups. For example, IRRI and ELRI, respectively, have 72 percent and 74 percent of internationally-recruited staff members in Position Groups V, VI and VII, whereas only 15 percent of ICRISAT's staff members occupy these position groups. In addition, ICARDA has only one out of 76 staff members in Position Group I, compared to ICLARM with 4 out of 23. This may reflect differences in the nature of the work being done by the Centers and the ratio of intemational to national staff members at these Centers, or it may stem from the difficulties of fitting the staffing scales of disparate organizations into the same taxonomy. 3 Table 2. Position group of IRS by gender and World Bank Part, country of origin (percent of group) Position Group [_Total Women Men WB Part I |WB Part El I Executive Staff 61 5 (8%) 56 (92%) 37 (61%) 24 (39%) II Research/Admin. Heads 117 13 (I 1 %) 104 (89%) 61 (52%) 56 (48%) HI Principal Scientists 126 8 (6%) 118 (94%) 63 (50%) 63 (50%) rv Sr. Scientists/Profs. 238 34 (14%) 204 (86%) 140 (59%) 98 (41%) V Scientists/Professionals 225 47 (21%) 178 (79%) 124 (55%) 101 (45%) VI Assoc. ScientistsfProfs. 80 23 (29%) 57 (71%) 45 (56%) 35 (44%) VU Post-doctoral FeRows 119 32 (27%) 87 (73%)[ 46 (39%) 73 (61%) TOTAL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) Staff disciplines recorded in the survey were coded into three general categories, as shown in Table 3: I, fields relevant to management and information sciences; II, social sciences; l, natural sciences. Table 3 displays the breakdown of disciplinary area by gender and World Bank Part in 1999. A review of past surveys indicates that there has been a growth in the percentage of social scientists compared to natural scientists, the reslit of an increasing recognition of the role of policy research in development impact. In 1991, 77 percent of staff members were natural scientists, while 14 percent were social scientists. By 1999, natural scientists represented 70 percent, and social scientists represented 21 percent of the total staff complement. The share of the staff population with management or information science degrees has remained roughly constant. Table 3. Disciplinary area of IRS by gender and World Bank Part, country of origin (percent in area) Disciplinary Area Total Women Men WB Part I WB Part II I M1gmt.flform. 88 28 (32%) 60 (68%) 56 (64%) 32 (36%) II Social Sciences 204 47 (23%) 157 (77%) 120 (59%) 84 (41%) m Natural Sciences 674 87 (13%) 587 (87%) 340 (50% 334 (50%) TOTAL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) Table 4 provides a breakdown of internationally-recruited staff members according to the level of last degree received. The most notable feature is the increase in the percentage of staff members with PhD degrees, from 73 percent in 1991 to 78 percent in 1999, with a corresponding decrease in percentage of those with master's degrees. 4 Table 4. Degree level of IRS by gender and World Bank Part, country of origin (percent holding respective degree) Degree Total Women Men WB Part I WB Part II PhD or equivalent 757 103 (14%) 654 (86%) 385 (51%) 372 (49%) MlAIMS or equiv. 109 34 (31%) 75 (69%) 71(66%) 37 (34%) BAJBS or equiv. 46 11 (24%) 35 (76%) 28 (61%) 18 (39%) Other 54 14 (26%) 40 (74%) 31 (57%) 23 (43%) TOTAIL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) In Table 5, 'Years of relevant professional experience" represents the number of years since achieving the last degree, adjusted to take account of the length of the respective program This assumed that staff members had been employed in their professions without interruption since receiving their degrees.2 The numbers were derived by subtracting the reported date of last degree from 2000. To account for the time taken to achieve a PhD, four years were subtracted from the resulting number in the case of those whose last. degree was reported as "master's or equivalent' and "other". Five years were subtracted in the case of those at the bachelor's leveL3 Using this computation, 36 percent of staff members have between 10 and 19 years of experience, while 25 percent have 20 or more years, 21 percent have from 5 to 9 years, and 18 percent have fewer than five years experience. Fully 60 percent bring more than ten years of experience to their current work at the Centers. Table 5. Years of relevant professional experience, by gender and World Bank Part, country of origin (percent of experience cohort) Number J Total Women Men WB Part I WB Part H Less than 5 175 53 (30%) 122 (70%) 100 (57%) 75 (43%) 5 to 9 207 44 (21%) 163 (79%) 114 (55%) 93 (45%) 10 to 19 346 46 (13%) 300 (87%) 198 (57%) 148 (43%) 20 or more 238 19 (8%) 219 (92%) 104 (44%) 134 (56%) TOTAL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) 2This is a reasonable assumption in the case of CGIAR staff. Although women are more likely than men to take time out of their careers (typically for child rearing), a relatively small percentage of female staff in the Centers has children (see table 11). 3 For comparison, John J. Siegfied and Wendy A. Stock, "The Labor Market for New PhD Economists," Journal of Economic Perspectives, vol. 13.3 (1999), pp. 115-34, report that those receiving US economics PhDIs in 1986 had taken an average of 6.3 years frm the bachelor's level. 5 The principal message that emerges from the tenure profiles in Table 6 is one of brevity. Almost one half of the internationally-recruited staff members (45 percent) have assumed their current positions within the last three years prior to the survey date. High turnover could be a consequence of the success of the CGIAR System in recruiting first-class scientists and managers; good people move around because they can.4 When compared with years of relevant professional experience in Table 5, these profiles suggest that many new recruits are already well along in their careers.5 Tenure is another important area of variability across Centers, however. ICRAF has recruited 30 percent of its international staff members in the last four years. For IWMI, this figure is 74 percent. Table 6. Tenure at Center by gender and World Bank Part, country of origin (percent of tenure cohort) Years | Total [_Women Men WB PartI WBPartII Less than 1 111 18 (16%) 93 (84%) 55 (50%) 56 (50%) 1 to 3 325 73 (22%) 252 (78%) 192 (59%) 133 (41%) 4 to 6 166 27 (16%) 139 (84%) 95 (57%) 71 (43%) 7to9 l 119 9 (8%) 110 (92%) 58 (49%) 61 (51%) 10 or more 245 35 (14%) 210 (86%) 116 (47%) 129 (53%) TOTAL 966t 162 (17%) 804 (83%) 516 (53%) 450 (47%) Table 7, which highlights the IRS by age group, indicates that 53 percent are between the ages of 45 and 65, the retirement age mandated by many Centers, with 92 percent over age 35. These figures tally with the findings relevant to years of experience. Table 7. Age of IRS by gender and World Bank Part, country of origin (percent of age cohort) Years ] Total Women Men WB Part I WB Part II 25-34 81 27 (33%) 54 (67%) 51 (63%) 30 (37%) 35-44 373 81(22%) 292 (78%) 235 (63%) 138 (37%) 45-54 388 44 (11%) 344 (89%) 183 (47%) 205 (53%) Over 55 124 10(8%) 114 (92%) 47(38%) 77(62%) TOTAL 966 162 (17%) 804 (83%) 516 (53%) 450 (47%) 4The kind of turnover inplied in Table 6 is not unusuaL In the US, about 30 percent of workers are not in the same job one year later. This figure includes reasons for which worker andjob separate, Le. job destruction, resignation, termination, retirement and layoffI See S.J. Davis and J. Haltiwanger, "Gross job creation, gross job destruction and employment reallocation," QarterlyJoumnal of Economics, vol. 107 (1992), pp. 819-63. s The survey asked only when staff members had been employed at their current Center, although it is known that an indeterminate number have moved from Center to Center in the course of their careers. 6 A set of comprehensive tables laying out systemwide findings on a range of variable staff attributes in 1999 is in Appendix 3, while Appendix 4 includes data from the three prior surveys. The attributes selected for analysis in the sections that follow, and the grouping within attnbutes, are those that the authors consider most illustrative of the issues under consideration, recognizing that other perspectives are possible. A complete list of the fields used in the survey is in Appendix 5. The Gender and Diversity Program will provide assistance in utilizing the 1999 survey data for different analyses upon request. All of the attributes discussed above, together with those referred to in Sections III and IV, play a role in the statistical analysis of compensation and positional equity presented in Section V. The intetnationally-recruited staff members of the CGIAR System, in the aggegate, tend I to have the following characteristics: * predominately male (83%); * fairly evenly divided between those from WB Part I and WB Part 1 gountries, with 100 individual countries of origin represented; * about half are in Position Groups IV ( Senior Sientists/Prfessionals and V (Scientists/Professional the levels where bech scientists tend to cluster (48%); * a large majoriy ame natural scentists (70%); * most hold PhD degrees (78%); * PhDs and other teminal degrees tend to be from academic insttutms im WB Part Ivountries (86%); M most have ten or more years of relevant professional experience (61%); they are relatively mature, with 53% over age 45, 92% over age 35; a: subsantial majority (80%) are acompanied at post by a spouse/partner, while 51% have children with them; * about half have been at their respective Center fewer than three years (45%). 7 IL. GENDER PROFILE AND ISSUES The examination of gender in the CGIAR Centers focuses on two issues. First is the matter of gender representation, which looks at the numbers within the internationally- recruited staff complement of the System and in individual Centers. Second, there is the matter of gender equity of opportunity. As seen in Table 1, women now represent 17 percent of the IRS, an increase from 12 percent in 1991, when the focus on gender was first sharpened. This, in all likelihood, reflects the effort in the Centers to "cast the net widely" as they recuit new staff Nonetheless, the progression has been slow, as highlighted in the bars of Chart L Chart 1. Percent of women among internationally-recruited staff in the CGLAR System by survey date 1991 199 199 199 Some Centers have been more successfbl than others in recruiting women, again as seen in Table 1 and in Chart 2 below. CurrenItly, CIFOR with 29 percent female staff members, IFPRI with 28 percent, CIAT and IPGRI, each with 23 percent, are ahead of other Centers. ICRISAT and 1WMI, each with 9 percent female staff members, ICARDA with 8 percent and WARDA, without female representation at the moment, fall well below the System average. The question is, why are women so u~nderrepresented in certain Centers? Is it insufficient effort to recruit? Is the alleged Hold boys network" operative here? Or, do other factors constrain the numbers? A look at representation by national origin, for example, tells a different and more posit*ve story. Table 1 indicates that the internationally-recruited staff members are fairly evenly divided between those originally from countries in WB Part I and those 6This is a phrase adopted by the original Gender Staffing Program to refer to the need to seek femnale candidates from new and previously untapped sources. 8 from WB Part II countries. The current 47 percent WB Part II representation is an increase from 43 percent in 1991, a laudable move toward an important System objective. In this case, WARDA with 68 percent and ICARDA with 63 percent exceed the average. Chart 2. Representation of women across centers dA I -. I ,.I CIFTOR .,. I Clip ICARDA ICLARM~ ICRAP- ICRISAT . IrAO CGI AR AVERAGE fMMu 16.8% HIGRI~ rs I ICARDA t v a , I 0%/ 5% 10%/ 15%/ 20% 25% 30% It appears that achievement of diversity by national origin is more easily attained at this point than gender balance. For example, Table 8 shows that, while men are evenly divided between WB Part I and WB Part II countries, only 31 percent of women are from WB Part II countries, with 69 percent from WB Part I countries. From another perspective, women represent 22 percent of the total WB Part I complement, but only I11 percent of the WB Part II total. Chart 3 looks at the origin of women from a regional perspective. Table 8. Country of origin of IRS by World Bank Part (percent of gender) I TCountry of Origin Gender _Total WB Part I WB Part Women 162 1 1 1 (69%) 51 (31%) Men 804 405 (50%) 3 99 (50%1) TOTAL 966 516 450 (47%) 9 Chart 3. Percent of female staff by region of country of origin South Asia = I X North America _ __ _ i MlddleEast& _____________________ a_ North Africa La2n Ameenca & _ _ CGIAR AVERAGE Caribbean _6.8% Europe C___ :__ E_______ Eastern Europe& & Central Asia East Asia _________________ &Pacfc L .I Australial _ _ _ _ _ _D: i i_ _ _ _ _ _ _ _ _ _ New Zealand Africa s 71 i 0% 5% 10% 15% 20% 25% 30% 35% Part of the explanation here comes from the nature of the Centers' research work itself The under-representation of women in science is very much an intemational phenomenon. In a table of female members of national scientific academies that appeared in the joumal Nature,7 the percentage of women among these senior scientists did not exceed 14.6 percent for any country listed. The highest pqrcentage was, indeed, for Turkey, but other WB Part II countries showed far lower percentages. India, for example, reported 3.1 percent women among its senior scientists, China reported 5.1 percent, and the Third World Academy of Sciences showed 3.9 percent. It is reasonable to assume that the figures will change as younger generations move into senior positions. According to the same report, the US National Academy of Sciences has only 6.2 percent women among its members, while the US National Science Foundation reported in 1995 that, of all people with PhDs employed in the US in the life sciences, 24 percent are women; in the social sciences, the percentage is 3l percent.8 Table 9 reports the UNESCO Gender Parity Index (GPI) for various parts of the world. The GPI is a measure of female access to education: the secondary education enrolment ratio for girls divided by the enrolment ratio for boys. Except for the East Asia and Pacific region (which unlike the UNESCO figure includes Japan), there is surprisingly strong accord between high school attendance by girls and the CGIAR representation of women. Those regions of the world where girls go on to secondary school tend to have a higher proportion of women among the internationally-recruited staff members. "Gender discrimination 'undermiines science'," Nature, vol. 402, 25 November 1999. s National Science Foundation/SRS 1995 SESTAT Inegrated Data Files. 10 Table 9. Gender parity index for various parts of the world, 19929 Region Gender Parity Index Developed Countries 1.03 Sub-Saharan Africa 0.78 Arab States 0.78 Latin America/Caribbean 1.09 Eastern Asia/Oceania 0.85 Southern Asia 0.63 World Average 0.85 Some of the international organizations that strive for national diversity also have been notably successful in improving their gender balance (see Table 10). On the whole, however, the staff members of these institutions are recruited from a broader range of disciplines than is the case with the CGL4R Centers. Table 10. Gender proffile of staff in international organizations 1° (percent female staff) Organization [Management Professional UNDP* 25.2 34.5 UN Secretariat** 24.6 38.1 World Bank** 19.0 35.4 IADB* 11.9 35.5 FAO** 9.6 23.5 African Dev. Bank* 8.9 22.0 Asian Dev. Bank** 0.0 22.0 *12/98 **12/99 The figures in Table 10 prompt a second look at the gender aspects of Table 2. Table 2 shows the breakdown of CGLAR staff members by position group. "Management" in the United Nations personnel system includes staff in levels Dl and D2; "professional" includes those classified as P1 through P5.11 It is thus apparent that these organizations 9 UNESCO World Education Report 1995. 10 Organizational Gender Issues Network, ORIGIN, Member Fact Sheet, 1999. " United National human resources staff compare CGLAR Position Group I to D2, Position Groups II and Im to DI, and Position Groups IV through VII to P5 through P2. CGLAR women comprise 8.5 percent of the staff in Position Groups L]II and m, comparable to UN DI andD2 as in "Management" in Table 10. See 1999 CGIAR Compensation Survey. 11 have brought a higher percentage of women into the managerial ranks than have the CGLAR Centers. Currently, women at the Centers fill only 8 percent of Position Group L ie., Executive Staff; and 11 percent of the jobs ranked in Position Group II, ie., Research Program and Administrative Heads. In 1991, the figures were 2 percent and 6 percent, respectively. This shows an increase, but somewhat smaller than might have been expected, given the focus on this issue and the management training directed to women over these years. Chart 4 highlights this issue. Chart 4. Percent of women and men in each position group 1:Exec,tiveStaff I ; . U: ReseardAdnhinL _I Heads f m: Prindpal Sdendsts IV: Senor Sdentists: Profesdoradsi, f-8 C==L - V: Sdents . Profeonals Fellows' i. . 0 'i. '----- : .f 0/ 10% 20%/ 30% 40% 50% 60% 70% 80% 90% 100% C° Women a Men There is a harbinger of change in this matter of gender balance, because women in the CGIAR System tend to be younger and have fewer years of relevant professional experience than the men. The figures in Tables 5 and 7, viewed from another perspective, show that 60 percent of women have from zero to 9 years of experience, while 35 percent of men are in this cohort. In addition, 67 percent of women are under age 44, as compared to 43 percent of men. Chart 5 depicts the experience comparison. The figures in the survey data set display a relationship between disciplinary area and degree level by gender. Women are more highly represented than men in the fields of management and information (17 percent of women to 7 percent of men), where the 12 terminal degree is Likely to be at the master's level This at least partially explains the disparity in percentages holding the PhD: 81 percent of men and 64 percent of women (see figures in Table 4). Chart 5. Comparing years of relevant professional experience for women and men 351/6 ------ 3400/ ;_______________________ ______________ 300% 25%1 A 15%A O0/o Less than 5 years 5-9 years 10-19 years 20 years or more 0 Women] Men Women are also more likely than men to be in the social sciences (29 percent of women to 20 percent of men), and, concomitantly, less well represented in the natural sciences (54 percent of women to 73 percent of men; see results reported in Table 3). The introduction of new disciplines into Center research may have helped open the door to female candidates, but in addition, the pool of qualified women in the natural sciences is growing rapidly. According to the National Science Foundation in the United States, for example, only 4 percent of the doctoral degrees awarded by US universities in agricultural sciences went to women in the early 1970s. By the early 1990s, that number had increased to 19 percent. In the biological sciences relevant to Center research, women earned 20 percent of the doctorates in the early 1970s, 40 percent by the early 1990s; in the socio-economic disciplines, the percentages increased in the same period from 15 percent to 35 percent.12 Chart 6 and the survey results presented in Table 6 compare the tenure distributions of women and men. In 1999, the rate at which women were hired was in line with their representation in the CGLAR System as a whole. However, the fact that 45 percent of women and only 31 percent of men in the System had been hired between 1 and 3 years prior to the survey is indicative of a conscious effort to recruit women. Whether this effort is likely to have a long-term effect depends on the retention rate of women '2 D. Merrill-Sands, "1997 CGIAR Human Resources Survey". 13 versus men. If women do not remain in the System as long as men, a higher recruitment rate is required just to keep their representation constant. Chart 6: Comparing tenure: distribution of women and men across tenure ranges 450h1w Z306= 6! 25% + _ _ __ _ _ 35% 7 30%V ss than I 1 to 3 4 to 6 7 to 9 l or more Years at current Center 'O0Women OMenI Chart 7 provides a comparison of retention profiles. The profiles are calculated by looking back at the tenure histograms from previous survey summaries. For instance, in 1991, there were 377 male staff members who had been at their Centers for between 1 and 3 years. These men, therefore, had begun their employment in 1988, 1989 or 1990. The 1999 survey reports that 79 current male employees joined their Center in those years. Thus the retention rate for men since 1991 is 21 percent. The profiles indicate that, over the long-term (since 1991), women have not been leaving any faster than men. However, considering the specific staff members who had been at their Centers for between I and 3 years when the 1994 survey was conducted, there is a large and unexplained disparity in the proportion of those women and men who have been retained through 1999. If the higher rate of hiring of women evident in Chart 6 is to translate into higher overall female representation, the issue of the retention of women must be addressed. Section V will look at whether equal qualifications, of which degree level, years of relevant professional experience and tenure at the Center are the most important factors, result in compensation and positional equity for women and men. 14 Chart 7. Comparing retention profiles: percent of women and men recorded as having tenure in the range 1-3 years and who remain at their Center, by original survey date 40% 20%- 10%/ 50e/ot/-' 1991 1994 1997 O] Women U Men Table 11 compares the personal status of intemationally-recruited women and men, higlighted in Chart 8. According to survey data, women are much more likely to be single, widowed or divorced, which is to say living alone at their posting, than men. For both groups, those who have a spouse or partner have a higher than 90 percent chance of being accompanied to the post, but women have a slightly higher probability (9.4 percent) of living apart from their spouse or partner than men (7.6 percent). Table 11. Personal status of IRS (percent of number of staff) Marital Status Children Spouse/ Spousel Single/ Gender Total partner partner not divorced/ Residing at None at post in residence in residence widowed post Women 162 96 (59%) 10 (6%) 56 (35%) 59 (36%) 103 (64%) Men 804 681 (85%) 56 (7%) 67 (8%) 438 (54%) 366 (46%) TOTAL 966 777 (80%) 66 (7%_o) 123 (13%' 497 (51%) 469 (49___ ) Women are also less likely to have children in residence with them at post. Although 59 percent are married, only 36 percent are accompanied by children. This disparity could be due to the fact that women in the CGIAR System are younger than the men, but it is not so. Chart 9 investigates this by plotting parental status by age and gender. The 15 hypothesis is not supported by the data, surprisingly, since men in the System appear to have children at a younger age than do the women. 13 Chart 8. Comparing personal status of women and men 400,0/ -- 300/6 ;l ].__ 200/eo ; ' - = I 0% '1-2 Proportion of Proportion of married with Proportion with children single/divorced/widowed non-resident spouse in residence WEWomen OMen Chart 9. Percent of women and men in each age group with children in residence 70%/ 609/. 50*/.- 40"!.- 301/. r- .- 20"!.- 10%/0 0%- Age 25r-34 Age 35-44 Age 45-54 Age 55 or Over '3AithoUgh the survey only asked whether or not staff had children with them in residence at post, there is no reason to believe that women are more likely to live apart from their children than men. 16 The results reported in Table 11 and Chart 8 bring into focus one of the most difficult issues relative to the recruitment and retention of women faced by the System, one that is changing but likely to be around for another generation. The increasing number of dual- career families and the difficulty of finding suitable employment for the spouse/partner of Center staff members is a constraint to the recruitment/retention of mep, to be sure. But it is far more of a constraint to the recruitment/retention of women, for societal reasons in all parts of the world. Spouses and partniers of staff members at Centers located in major cities, as are the intemational organizations referred to above, have a better chance of finding career opportunities than those in more remote postings. In the remote postings especially, job opportunities are fewer and/or spouse employment often is not legally permissible. Some Centers have developed spouse/partner employment policies and services that address this issue, but the search for creative responses to this real problem must proceed unabated. The analysis of data on vomen's representation, leaves the Centers with several vital questions: Since gender balance among internationally-recruited staff appeaTs more difficult to achieve than balance by national origin, what new recruitment techniques can be designed to uncover the sorces of- qualified WB Part LI women? What new recrutment techniques can be designed to attract more women natural scientists? More women social scientists? More women in the fields of management and inlrmation? Is there a bias against women wlen it comes to the recruitment for, or promotion to, top management positions? What changes are needed in Center policies, practices and work culure to ensure that women are retained? What cbanges in host country relationships or in Center services are required to respond to the needs of dual-career families? Are IRS women discouraged from childbearing by some unspoken convention? If so, what can be done to change this? 17 IV. DIVERSIY PROFILE AND ISSUES The two issues examined in regard to staff diversity are the same as those examined with respect to gender: is there appropriate representation and is there equity of opportunity? As Table 1 shows, 47 percent of the intemnationally-recruited staff members are originally from countries classified by the World Bank as Part nI. This is an increase from the 43 percent reported by the 1991 survey and a positive development from the System perspective. Nonetheless, the distribution across Centers is less favorable (see also Chart 10 below). While WARDA and ICARDA have 68 percent and 63 percent WB Part II staff members, respectively, ICLARM has only 30 percent, and CIFOR and IFPRI have 33 percent each. Gender questions posed in Section Im are relevant here, however, as WARDA has no IRS women on staff and ICARDA only 8 percent. As discussed in Section m, statistics tell only part ofthe story. Certainly, the needs of any given Center at a particular point in time, and the characteristics of the pool from which candidates in a specialized field must be drawn, are the critical factors in any recruitment effort. Percentage "quotas" are not the objective. Chart 10. Representation of WB Part II staff across Centers CUT ICARDA ____ ____ _____ -CGLAR AVERAGE ICLARM . . ICRAP I______ I______ ___I___ 46.6% ICRISAT .-______ ____ ____ __ __. a~m IEPGRI W A R D A --------- --- . - ----------------_---_-, 0% 10% 20% 30% 40% 50% 60% 70% Beyond the rubric of World Bank Part, representation across world regions is of interest and is depicted in Charts 11 and 12. Chart 11 highlights the fact that 37 percent of WB Part RI staff members originally come from countries in Africa; Chart 12 portrays the dominance of Europeans among WB Part I staff 1 8 Chart 11. Percent of WB Part II staff by region of country of origin South Asia Middle East & North Africa 9% Latin America & Caribbe 18% > East Asia & Pacific EastEurope & Central 14% Asia Chart 12. Percent of WB Part I staff by region of country of origin AustralialNew Zealand I* 6% East Asia & Pacific , | l ;>X3% North America 41% 1 .- - 9. ^. :-Europe t - ^ / 50% 19 One of the major diversity issues ("is there appropriate representation?") arises in Table 12. Table 12 looks at the country of last degree and country of origin by World Bank Part. The results show that, with a roughly equivalent split between WB Part I and Part II staff members, 86 percent have their last degree from countries in WB Part L while only 14 percent have their degrees from WB Part II countries. Table 12. Country of last degree and origin by World Bank Part (percent of gender) | 0 Country of Degree Cointry of Origin Gender Total WB Part I WB Part I WB Part I WB Part H Women 162 144 (89%) 18 (11%) 11l (69%)J 51 (31%) Men 804 689 (86%) 115 (14%) 405 (50%) 399 (50%) Total 966 833 (86%) 133 (14%) 516 (53%) 450 (47%) Chart 13 looks at the same question from a regional perspective. It shows that, although 90 internationally-recruited staff members are originally from countries in South Asia, only 47 completed their last degrees in the region. Of the 173 internationally-recruited Chart 13. Regional comparison: country of last degree versus country of origin Souti Aoa NordoA_erica - Midle East &NorthAfrica Latin America J &Canbbean - - EP . - - - _ !P Europe F ##t ; p#,eA;#4#i t A#i# East Europe &Cmdrai Asia East Asia &Padfic U Australial NewZedand 1iL Afrnca | O SO 100 150 200 250 300 350 400 450 500 JReipon of origin ORC!ion of deiree 20 staff members from countries in Africa, all but 33 took their degrees elsewhere. In fact, 80 percent of current IRS have terminal degrees from either European (315 or 33 percent) or North American universities (457 or 47 percent), while only 48 percent of the staff members are originally from countries in those regions. Clearly, students from a number of countries must look outside to pursue higher studies, either because opportunities are not available in their country or region, or not available in the specialization sought. What must be considered is whether there is an unjustified predilection in the recruitment process to discount advanced study from WB Part II countries and to favor candidates who bring degrees from Europe or North America. Table 2 and Chart 14 show a breakdown of internationally-recruited staff members by position group, and both show little difference worthy of note except for Position Groups I and VIL At the Executive Stafflevel, Position Group 1, albeit a small group overall, 61 percent of the positions are held by staff members from WB Part I countries, 39 percent by staff members from WB Part II countries. Among the Post-doctoral Fellows, exactly the reverse percentages are evident. Chart 14. Percent of WB Part I and Part II staff in each position group 1: Executive Staff 1: Research/Adm.i I Eleads - . I__ _ _ _ _ _ _ _ l _ _ _ m: Prinipal Scentists IV: Semior Scientists/ __ Profesknal V Sadentast - Prolfessioi51s L . , - .. ., .. VI: Assodate ScientsL, Proressionalsi . VII: Postdoctord Fellows E. 0% 5% 10% 15% 20% 25% 30% IOWBPartI DOWBPart I A review of the results reported in Table 5 (on years of relevant professional experience) reveals close parity of those from WB Parts I and II countries, while that of the results on Table 7 (on age) indicates that WB Part II staff tend to be somewhat older than those from WB Part I countries. Only 37 percent of all staff members under age 45 are from 21 WB Part II countries, while 55 percent of staff members over age 45 come from these countries. On average, staff members from WB Part II countries are 46.9 years old, 2.8 years older than their WB Part I colleagues. 14 The gender discussion in Section HI noted that the comparative youth of female stag in age and experience, could augur a future increase in the percentage of women on the Executive Staff In this case, the parity of WB Parts I and [ staff in experience, taken with the greater age of those in WB Part LI, may suggest the need to review promotion and recruitment criteria for top management positions. Again, disciplinary area and degree show a relationship (see Tables 3 and 4). Only 7 percent of all WB Part II staff members are in the fields of management or information, with a normal terminal degree at the master's level, compared to 11 percent of the WB Part I staff members. This tallies to some degree with the percentage of WB Part nI staff members holding the PhD degree at 83 percent to 75 percent of those from WB Part I countries. The relationship in the social sciences is 19 percent of all WB Part II, 23 percent of all WB Part I staff members; in the natural sciences, the relationship is 74 percent of all WB Part II, 66 percent of all WB Part I staff members. Chart 15. Cmpanng teure: &stribution of staff from W Parts I and II across tenure ranges 40/1~ 3.'X 301a''z'1 0 0l Less than 1 year 1-3 yeas 4-6 yeas 7-9 years Over 10 yeas 4 Both these differences between means are statistically significant at the 95% confidence level. 22 Internationally-recruited staff members from WB Part II countries tend to have slightly fewer years of tenure than those fromWB Part I countries (see Table 6): 58 percent of all WB Part II staff members have been at their Centers six years or fewer, while 66 percent of WB Part I staff members are in that tenure category. On the other hand, 42 percent of WB Part II staff members have seven or more years tenure (many over ten years), while only 34 percent of those from WB Part I countries have more than seven years. Viewed from the perspective shown in the table and in Chart 15, recruitment in the year immediately prior to the survey was close to parity; in the last six years, more staff members were recruited from WB Part I than from WB Part II countries. These figures suggest that the overall improvement in diversity percentages between 1991 and 1999 (43 percent to 47 percent) results less from efforts to recruit from WB Part II countries and more from the lower departure rate of older staff members from those countries. Section V will look at whether compensation and positional equity for staff members from countries in WB Parts I and II are systematically influenced by degree level, years of relevant professional experience and tenure at the Center. The analysis of data on diversity elicits four questions: As asked in Section m, how can the Centers come closer to. a balance of both gender and diversity? What new recruitment techniques can be designed to uncover the sources of qualified WB Part II "men? Are higher degrees eamned outside Europe and North America overly discounted in the recruitment process? in the recruitment for, or promotion to, top management positions in the Centers, does national origin play an inadvertent role? 23 V. ANALYSIS OF COMPENSATION AND POSITIONAL EQUITY BY GENDER AND DIVERSITY 1. Overview This section examines the performance of the CGIAR System with regard to equity in compensation and classification within the position group framework (the latter referred to below as "positional equity"). The standard of compensation equity used is "equal-pay-for-work-of-equal-worth". This requires that compensation be based solely on non-occupation-relevant characteristics of the job (such as level of responsibility and authority) and employment-relevant characteristics of the worker (such as academic credentials). In the context of the CGLAR, this means that salary should depend solely on position group, level of last degree acquired, years of relevant professional experience and tenure (ie. years at the current Center). Specific disciplinary area should not be a factor. Thus, for example, a natural scientist and a social scientist who have otherwise identical characteristics should receive the same compensation. Similarly, gender and country of origin, as variables, should bear no correlation with compensation. On the other hand, it is important to note that the analysis that foHows had no way to control for performance quality over time. This is a highly relevant factor, in view of the practice in most CGIAR Centers to adjust salaries annually on the basis of merit. For this reason, this analysis can only serve to highlight those areas wihin the compensation structure that warrant more detailed attention. The analysis of positional equity controls for the same three worker-related characteristics (degree level, years of relevant professional experience and tenure). It assumes that persons of identical characteristics in these respects should be in the same position group. Again, a performance quality measure is absent, as is evidence in a given staff member's performance of the specific managerial skills needed for positions increasing in responsibillky and authority. Clearly, observed interpersonal or supervisory skills, as well as the ability to manage financial resources would play appropriate roles in determining the appointment of staff members to leadership positions. Thus, again, the findings on positional equity must be considered preliminary to further investigation. Regression analyses were canied out to assess both compensation and positional equity in the CGIAR System with respect to the following comparisons: A. Women as compared to men; B. Staffmembers of WB Part II origin as compared to staff members of WB Part I origin; C. Staff members of WB Part II origin, who are now citizens of WB Part I countries, as compared to staff members who have both WB Part I origin and citizenship; 24 D. Staff members of WB Part II origin who are posted to their home regions as compared to other staff members of WB Part II origin (i.e. not posted to their home regions); E. Staff members in each of the three disciplinary areas (I: management and information, II: social sciences, Im: natural sciences) as compared to staff members in the other disciplinary areas. In each Comparison A through D, the group listed first is hypothesized to be disadvantaged; thus the second is the baseline group in each case. In Comparison E, there is no reason to suppose that any group is disadvantaged. The group of staff members whose most recent degrees are in fields that fall into disciplinary area I (management and information) was arbitrarily chosen to be the baseline group. 2. Data and Methodology Individual compensation is made up of basic salary, cost of living allowance (COLA) and other sundry benefits such as subsidies for children's education. The principal difficulty in addressing the issue of equity in the CGLAR System with these income measures is that a dollar in Kenya is not the same as a dollar in India--actual cost of living matters. 15 This means that compensation measured in US dollars is not directly comparable across Centers. The analysis reported in the subsection below on compensation equity circumvents this problem by allowing for systematic pay differentials among the Centers. This stiU leaves the problem of which measure of income to use. The data on the sundry components are inconsistently reported and have been ignored.16 The remaining possibilities are the annual cash income (which is the basic salary plus COLA) and basic salary alone. The CGIAR Compensation Survey of 1999 used the former. For the objectives of this study, however, it should be clear that within any Center, the choice is arbitrary. The regression analysis used to allow for systematic pay differentials among the Centers effectively means that the whole workforce can be viewed as being at the same Center (actually CIAT was used as the base).'7 Consequently, systemwide analyses can be carried out, and the choice between measures of income is again arbitrary. Indeed, annual cash income could be used for one Center and basic salary for another; the 15 Cost-of-living indices are published by the European Centre for Worldwide Cost of Living Comparisons and the UN, but they differ significantly. Furthermore, even if such indices were consistent and reliable, it would be difficult to select which index to apply (that of the posting location or the home country) for a workforce that repatniates much of its income. Further discussion of the indices is provided in the 1999 CGIAR Compensation Survey. 16 Since the sundry component is generally needs-based, inclusion could generate misleading results if the needs differ systematically across groups for which the extent of discrimination is being assessed For example, in certain locations women may have higher security requirements than men, or those employees with families (predomnantly men and staff from WB Part II countries) may mrke greater use of educational subsidies. 17 Technically, this is carried out by introducing dummy variables for each Center into the regression analysis. 25 difference would merely be reflected in the individual Center's systematic pay differential. Since annual cash income is, in some cases, built up from several other measures, basic salary was deemed to be less likely to suffer from errors and, therefore, is used in the following analysis. Table 13 provides a summary of staff numbers in each position group as well as basic salary compensation across the CGLAR Systenm Because of the shortcomings of basic salary as a measure of compensation discussed above, little inference about the treatment of men versus women, or of staff members who originate from WB Part I countries versus those from WB Part II countries, can be drawn from this table. Some general themes do emerge, however. Table 13. CGLI4R base salary by country of origin and gender Position Aggregate Diversity: Country of Origin Gender WB Part I WB Part I Female Male Average Average Average Average Average # Base # Base # Base Base # Base Salary, $ Salary, S Salary, S Salary, $ Salary, S 1 61 98,546 37 101,697 24 93,689 5 92,570 56 99,080 IE 117 67,762 61 69,443 56 65,932 13 67,365 104 67,812 in 126 68,373 63 68,515 63 68,231 8 68,004 118 68,398 IV 238 57,274 140 58,000 98 56,237 34 55,321 204 57,600 V 225 48,341 124 49,225 101 47,256 47 48,851 178 48,206 VI 80 35,282 45 37,963 35 31,835 23 35,591 57 35,157 VIE 119 31,104 46 36,788 73 27,523 32 31,707 87 30,882 All 966 55,472 516 58,023 450 52,548 162 48,721 804 56,833 '3Position Groups are as follows: I: Executive Staff, nI: Research ProgrmlAdministrative Heads, IL Principal Scientists, IV: Senior Scientists/Professionals, V: Scientists/Professionals, VL Associate Scienfists/Professionals, VII: Post-doctoral Fellows. 26 Table 13 also highlights the apparent discrepancies that motivate the statistical analyses in the remainder of this section. The bottom row shows that on average, CGLAR staff members originally from WB Part I countries receive US$5,475 per annum more in basic salary than do the staff members originally from WB Part II countries. In addition, on average, men are US$8,112 better off than are women. The following subsections look at whether these discrepancies can be explained by other factors (such as differences in years of relevant professional experience). One example of such an explanation is already evident in Table 13; the discrepancy between male and female salaries (Comparison A) is much smaller when we look within position groups. This indicates that the difference that emerges in the global average is largely due to the fact that women are more heavily represented in the lower position groups. In other words, when a control for position group is introduced, the male/female pay differential is reduced. Of course, controlling for other factqrs (e.g. individual characteristics) could reverse this result. In this way, the statistical analyses of the following subsections use regression analysis to control for the permissible factors discussed above (degree level, years of relevant professional experience, tenure) in order to ascertain the extent to which compensation and positional attainment differ across the groups in the five comparisons considered. 19 3. Compensation Equity The salary differential reported in Table 14 measures the extent to which the baseline group receives higher (or lower) basic salary solely as a consequence of membership. The p-value represents the reliability of the measured pay advantage.20 Itisthe probability that there is no systematic pay differential. Thus a p-value of 0.05 or lower is Table 14. Measuring salary discrepancies Comparisons Salary Advantage p-value A - Women vs. men 0.7% 0.686 B - WB Part I vs. Psrt 6.5% 0.000 C - WB Part I now Part I 4.6% 0.193 Citizens vs.- Part I origin D -WB Part HI posted to home -4.2% 0.074 Regions vs. all Part II _ E - II: Social Sciences vs. -4.9% 0.098 Mgmt. and Information - II: Natural Sciences vs. 0.4% 0.879 Mlmt. and Information I_I__ 19 All regression analysis was carried out using the Microfit package. The regressions atha for the basis for the results reported are in Appendix 6. This appendix also includes the output from the general summary regression (Regression 1) that.looks at the returs to all characteristics of individuals included in the sunvey data 20 Al the regressions reported here suffered from heteroscedasticity of the error term The reported p- values are those based on standard errors adjusted by White's method. 27 usually considered statistically significant (such figures have been highlighted in bold type). A p-value in the range 0.05 to 0.1 may be considered "borderline" significant. Thus, the results shown in Table 14 suggest the following: Comparison A: Men appear to be paid marginally more than women (0.7 percent), but the difference is not statistically significant. Comparison B: There is a 6.5 percent and highly significant salary advantage to being originally from a WB Part I country, over coming from a WB Part II country. Comparison C: Those who now have WB Part I country citizenship but who were originally from WB Part II countries have a 4.6 percent pay disadvantage as compared to those who are of both WB Part I origin and citizenship. To some degree, this appears to corroborate the results of Comparison B above; however, the figure is not statistically significant. The lack of significance emerges from the lack of data; only 30 staff members (3 percent of the total complement) originally from WB Part II countries are now WB Part I citizens. Comparison D: Staff members originally from WB Part II countries and who work in their own world regions are actually 4.2 percent better paid than their counterparts who work in regions of the world that are foreign to them The figure is on the borderline of statistical significance. Comparison E: By comparison to staff members who hold degrees in disciplinary area I (management and information), staff members with disciplinary area II (social sciences) degrees are better off by 4.9 percent. Again, this figure is on the borderline of significance. There is no appreciable difference in earnings between staff members with disciplinary area I and those with disciplinary area Ell (natural sciences) degrees. Consequently, a social scientist is likely to be better paid (by about 5 percent) than a natural scientist of similar degree level, years of experience and tenure. The finding for Comparison B (staff members from WB Part II countries compared to those from WB Part I countries), by virtue of its statistical significance as well as the size of the percentage advantage apparently accruing to Part I origin, bears investigation on a case-by-case basis by individual Centers. In all of the other comparisons analyzed, no discrepancies of statistical significance were found. 4. Positional Equity This subsection looks at whether membership in the baseline group affects the probability that an individual is in a higher position.21 For each of the five comparisons considered, this question can be asked for each tier in the CGIAR positional hierarchy. However, because Position Groups II and m (Research/Administrative Heads and Principal 21 The probabilities were obtained using logit maximum likelihood estimation. The controls were the Center dummies, level of last degree, years of relevant professional experience and tenure. 28 Scientists) are of similar status, they have been combined. Shortage of data has meant that meaningful results have not been obtainable for the analysis of membership of Position Group VII (Post-doctoral Fellows) versus membership of Position Group VI (Associate Scientist/Professionals) or higher position groups. Similarly, there is insufficient data to obtain meaningful results for Position Groups I (Executive Staff) and V (Scientists/Professionals) in Comparisons C and D. The results are reported in Table 15. The probability differential is how membership in the baseline group has. affected the probabiliy of being in the threshold position group or further up in the hierarchy.22 A negative sign means that membership in the baseline group is a hindrance to membership of higher position groups. Again, statistically significat numbers are reported in bold type. Table 15. Extent of deviation from positional equity based on level of last degree, years of relevant professional experience and tenure Comparison Group Threshold Probability p-value Position23 Differential (Appendix 6 regresion #) I (2al) 1.6% 0.242 A - Women vs. men II-m1 (2af) 12.0% 0.015 IV (2aIV) 14.9% 0.008 V (2aV) 3.2% 0.097 I (2bI) 1.2% 0.169 B - WB Part H vs. Part I ]1-RI (2bM) 5.5% 0.087 IV (2bIV) 14.9% 0.000 =_________________________ V (2bV) 6.5% 0.000 C - WB Part I now Part fl-IlI (2cf) -3.9% 0.708 I citizens vs. all Part II IV (2clV) 20.6% 0.127 D - WB Part II posted to home 1-111 (2d1) -2.5% 0.652 Regions vs. all Part I IV (2dlV) -7.2% 0.368 I (2e1) II-SS -0.6% 0.597 E - II: Social Sciences vs. IH-NS 3.1% 0.012 Mgmt. and Information 11-11 (2eml) II-SS 5.7% 0.427 - m: Natural Sciences vs. fI-NS 15.3% 0.021 Mgmt. and Information IV (2elV) II-SS 11.6% 0.224 1IH-NS 19.6% 0.027 V (2eV) 11-SS 8.9% 0.015 __EI-NS 10.3% 0.003 To oversimplify, the statistical process involved comparing the relevant charcteristics of all individuals in the indicated groups (e.g. all women and all men in Comparison A) with the characteristics of the same groups of those now in the threshold position or higher. ' Again, Position Groups are as follows: I Executive Staff, II: Research Progam/Administrative Heads, Elm Principal Scientists, IV: Senior Scientists/Professionals, V: Scientists/Professionals, VI Associate Scientists/Professionals, VII: Post-doctoral Fellows. 29 The analysis leads to the following observations: Comparison A: Women are 12 percent less likely than men to be in Position Group II-m (Research/Administrative Heads and Principal Scientists) or highe,r, and 15 percent less likely than men to be in Position Group IV (Senior Scientists/Professionals) or higher. Both percentages have a statistically significant p-value. Women are 3.2 percent disadvantaged with respect to Position Group V (Scientists/Professionals) or higher, but this result is on the borderline of significance. Since the women in the CGIAR System tend to be younger than the men, the results of this regression were tested again with age as an additional variable in order to determine whether the results indicated possible discrimination by age rather than gender. There were negligible changes in the findings. Comparison B: Staffmembers from WB Part II countries are 14.9 percent less likely than those from WB Part I countries to be in Position Group IV (Senior Scientists/ Professionals) or higher, and 6.5 percent less likely to be in Position Group V (Scientists/Professionals) or higher. Again, both percentages have a statistically significant p-value. It appears that WB Pait 11 staff members are similarly disadvantaged by 5.5 percent from joining Position Groups I-mr (Research/Administrative Heads and Principal Scientists) or higher, but this result is also on the borderline of significance. When these results are taken together with: a) the finding that the percentage of WB Part II staff has increased only from 43 percent to 47 percent, with a decline in absolute numbers, and b) the fact that this group tends to be older than staff from WB Part I countries, it appears that this staff group has remained in a relatively steady state as far as position level is concerned. Comparisons C and D: None ofthe p-values in Comparisons C and D reach the level of statistical significance, nor are any on the borderline. As noted, data on these groups are inadequate for meaningful interpretation. Comparison E: Natural scientists are relatively under represented above every threshold considered. They are as much as 15.3 percent less likely to be in Position Group 11-11 (Research/Administrative Heads and Principal Scientists) or higher and 19.6 percent less likely to be in Position Group IV (Senior Scientists/Professionals) or higher compared to similarly qualified staff with degrees in management or information science. However, when the threshold to Position Group V (Scientists/Professionals) and higher is considered, natural scientists and social scientists are similarly under-represented (by about 10.3 percent and 8.9 percent, respectively) compared to their management/ information counterparts. These percentages reach the level of statistical significance. In interpreting the above results, it is important to reiterate the two worker-related characteristics that are not available in the data set, but that are significant in movement up in the positional hierarchical. They are the quality of performance in the position from which a staff member might be promoted and evidence in that performance of skills beyond science that are relevant to the management tasks of those in higher positions. There is no reason to believe that the unmeasured variables would impact systematically across each Comparison. 30 Bearing in mind restrictions in the data at hand, it is nonetheless clear that additional investigation is called for. Results of the data analysis raise the flowing questions: In the promotion from Position Group V (Scientists/Professionals). in what may be the first significant step up in the hierarchy, are women's qualifications fairly evaluated? Are staff members from WB Paut countries unduly held in Posiion Gro-up V or below, and is their compensation unfairly constaied with respect to the positions currently held? Are natural scientists given sufficient career development oppotuiies to permit them to build the skills needed for leadership positions? 31 VL FOLLOW-UP While there is good news in the findings of this survey to be celebrated, clearly results have emerged that suggest fiuther investigation and assertive follow-up action on the part of the Gender and Diversity Program. Among the actions suggeste4 are the following: * Continue to expand and diversify recruitment strategies to "cast the net ever more widely" to attract more women natural scientists, more women social scientists and more women in the fields of management and information. * Focus especially on the identification of sources of qualified women from developing countries (WB Part II countries). * Explicitly support CGIAR women's advancement, aiming to bring more women into positions of mid-level and senior management. * But also encourage training in managerial skills for men with degrees in the natural sciences. * Research and communicate best practices of sister Centers and of other organizations with respect to policies, practices and workplace cultures to ensure that women are retained; include in-depth studies on work/family balance. - Seek and promote ways to respond to the needs of dual-career families. - Investigate more closely the question raised about the dominance of staff members with degrees from European and North American universities; consider a conscious effort to identify high value PhD programs offered by academic institutions in the South. - Similarly, investigate more closely reasons for the disparity between WB Part I and Part II staff members in position levels above Position Group V (Scientists/ Professionals), as well as apparent discrepancies in compensation for WB Part II staff members at all levels. * Work to ensure accountability of CGLAR managers on gender and diversity issues. 32 APPENDICES: 1. World Bank Part 1/Part II Countries 2. Position Groups 3. CGIAR Human Resources Survey (1999) 4. Survey data from 1991, 1994 and 1997 5. Fields in the 1999 survey 6. Details of regression analysis discussed in Section V 7. List of Acronyms Appendix I WORLD BANK COUNTRY DESIGNATIONS PART 1 COUNTRIES (26) AUSIRALIA IRELAND RUSSIAN FEDERATION AUSTRIA 1TALY SOUIH AFRICA BELGIUM JAPAN SPAIN CANADA KUWAiT SWEDEN DENMARK LUXEMBOURG SWITZERLAND FINLAND NETHERLANDS UNITED ARAB EMIRATES FRANCE NEW ZEALAND UNITED KINGDOM GERMANY NORWAY UNITED STATES ICELAND PORTUGAL PART-II COUNTRWS (134) AFGHANISTAN GAMBIA, THE NICARAGUA ALBANIA GEORGIA NIGER A.LGERIA GHANA NIGERIA ANGOLA -GREECE OMAN ARGENTINA GRENADA PAKISTIAN sARM1BIA GUATEMALA PALAU AZERBAUAN GUINEA PANAMA BANGLADESH - GUINEA BISSAU PAPUA NEW GUINEA BARBADOS GUYANA PARAGUAY BELIZE HArTI PERU BENIN HONDURAS PHILIPPINES BHLrAN - HUNGARY POLAND BOLIVIA INDIA RWANDA BOSNIA AND BERZEGOVINA INDONESIA ST. KlTTS AND NEVIS BOTSWANA IRAN, ISLAMIC REPUBLIC OF ST. LUCIA BRAZIL IRAQ ST. VINCENT & THE BURENA FASO ISRAEL GRENADINES BURUNDI JORDAN SAMOA CAMBODIA KAZAKHSTAN SAO TOME AND PRINCIPE CAMEROON KENYA SAUDI ARABIA CAPE VERDE K1RIBATI SENEGAL CENTRAL ARICAN REPUBLIC KOREA, REPUBLIC OF SERRA LEONE CHAD KYRGYZ REPUBLIC SLOVAK, REPUBLIC CHILE LAO PEOPLE'S DEM REP. SLOVENIA CHINA LAT VIA SOLOMONISLANDS COLOMBIA LEBANON SOMALIA COMOROS LESOTHO SRI LANKA CONGO, DENI. REP. OF L]BERIA SUDAN CONGO REPUBLIC OF -LIBYA SWAZILAND COSTA RICA MACEDONIA FYR OF SYRIAN ARAB REPUBLIC COTE D'IVOIRE. MADAGASCAR TAJIUSTAN CROATIA MALAWI TANZANIA CYPRUS MALAYSIA TAND CZECH REPUBLIC MALDIVES TOGO DJ]BOUTI MALI TONGA DOMICA MARSHALL ISLANDS TRINIDAD AND TOBAGO DOMINICAN REPUBLIC MAURITANIA TUNISIA ECUADOR MEXICO TURIKEY EGYPT.ARABREP`UBLIC OF MICRONESI, FED. STATES OF UGANDA EL SALVADOR MOLDOVA UZBEKISTAN EQUATORIAL GUINEA MONGOLIA VANUATU ERITREA MOROCCO VIETNAM ETHIOPIk MOZAMBIQUE YEMEN, REPUBLIC FIJI MYANMAR ZAMBIA GABON NEPAL ZIMBABWE Source: Vots and Subscriptions List, Intemational Development Association (10/99) Appendix 2 POSITION GROUPS - CGIAR COMPENSATION SURVEY Definitions of each of the seven position groups follow. Using these definitions, CGIAR Centers and other comparator organizations should designate the position groups to which, in their judgement, each staff member belongs. Group I Executive staff (excluding the Director General/CEO) Representative CGIAR titles: Deputy Director General Deputy Directors General for Research, Administration, etc., and Directors of major Divisions, Departments or Programs. Functional responsibility: Manages one of the major branches of the organizational hierarchy, usually concerned with either scientific research or finance and administration. Responsible for developing institutional strategy and policy in all areas but with focus on functional assignent. Identifies and pursues funding opportunities. Prepares and manages large budgets. Routinely negotiates on behalf of the organization complex, sensitive or contentious program or rnanagement issues. Makes major decisions in collaboration with other members of the management team. Usually selected and appointed by the Director General/CEO with the concurrence of the Board of Trustees; reports to the Director General/CEO Education/Experience: PhD in relevant area for scientists, MA/MS/MBA in relevant area for administrators; usually a minimum of 20 years experience, including experience in supervising a substantial scientific or administratiye staff Organizational impact: Represents organization to outside world, interacting with Board, donors, staff of partner and client institutions and political leaders. Supervises the work of a large number, usually approximately one half of organization's staff. Thus can have an impact of major significance on organization's intemal operations and external image. Group I Research Program, Administrative Heads and Directors Representative CGIAR titles: Program Leader for Irrigated Rice, Coordinator- Regional Program, Director-Biotechnology, Program Director-Farming Systems, Finance Head, Head of Operations, Human Resources Director, Comptroller, Director of Information/Publications. Functional responsibili : Within very broad guidelines on organizational program or administrative strategy and priorities, manages a major scientific or administrative unit of the organization with responsibility for staffing, fundraising and financial management as well as leadership of the unit's research and training program or corporate administrative and technical services. Makes a very significant professional contribution to the organization's objectives, contributes to the scientific or administrative strategy and policy of the organization and provides expert advice of the highest order both within and outside the organization. Routinely negotiates on behalf of the organization complex, sensitive or contentious program or management issues. Generally reports to the Director General/CEO or an Executive Staff member. Education/Experience: PhD or equivalent in relevant area for scientists, MAIMS/MBA or equivalent in relevant area for administrators; usually a minimum of 15 years experience, including experience supervising a substantial scientific or administrative staff Organizational impact: Represents organization to outside world, interacting with Board, donors and political leaders and with staff of partner and client institutions. Interacts extensively with staff of other organizational units. Supervises the work of a substantial number of scientific and/or administrative professional and support staff The quality of the unit's performance and of the leader's professional advice can have a significant impact on the organization's internal operations and external image. Group m Principal Scientists epresentative CGLAR titles: Principal Scientist, Distinguished Scientist Functional responsibiliiv Within very broad guidelines on organizational program strategy and priorities, conceives of, designs and provides leadership to a set of highly complex scientific research and training activities and helps to secure and manage the necessary human, financial and material support. Makes a very significant professional contribution to the organization's objectives, contributes to the science strategy and policy of the organization and provides expert advice of the highest order both within and outside the organization. Routinely negotiates on behalf of the organization complex, sensitive or contentious program or management issues. Generally reports to the Director General/CEO or an Executive Staff member, occasionally to a Program Head or Director. Education/Experience: PhD or equivalent in relevant area; usually a minimum of 15 phls years post-doctoral experience leading to an internationally recognized scientific reputation in the area of the organization's research program This may be an honorific title used to recognize exemplary service or to attract an outstanding individual to the organization's staff Organizational impact: Represents organization to outside world, interacting with Board, donors and political leaders and with staff ofpartner and client institutions in collaborative research and training. May supervise the work of a substantial number of scientific professional and support staff Thus can have a significant impact on the organization's internal operations and external image. Growt IV Senior-Scientists, Support Professionals, Administrators Representative CGIAR titles: Senior Scientist, Senior Agronomist, Senior Economist, Senior Editor, Senior Infqrmation Specialist.. Functional responsibilitv: a) Scientists - Within broad guidelines on program objectives, conceives of, designs and provides leadership to a set of complex scientific research and training activities and helps to secure and manage the necessary human, financial and material support. Makes a significant contribution to the relevant strategy and policy of the organization and provides expert advice with a very high degree of reliability to organization management and other institutions. From time to time, participates in negotiating on behalf of the organization complex, sensitive or contentious program or management issues. Generally reports to a Division Head or Director. b) Professionals, Administrators - Within broad guidelines about service expectations and priorities, manages a substantial unit of the organization charged with the day-today delivery of a corporate administrative or technical service (e.g., finance, human resources, computing or information services). Responsible for the unit's staffing and financial management and for providing leadership and direction to its staff Reviews, develops and implements relevant systems, policies and procedures. Makes a significant professional contribution to the relevant strategy and policy of the organization and provides expert advice with a very high degree of reliability to organization management. Generally reports to a Division Head or Director. EducationlExperience: PhD or equivalent in relevant area for scientists; MA/MS/MBA or equivalent in relevant area for support professionals and administrators, usually a minimum of 7 to 10 years experience. Organizational impact: a) Scientists - Represents organization to outside world, interacting with Board and donors and with staff of partner and client institutions in collaborat#ve research and training. Supervises the work of a small number of scientific professional and support staff Thus can have a very important impact on organization's internal operations and external image. b) Support Professionals, Administrators - Interacts with staff of research units and, to some extent, with Board and donors. Supervises the work of a substantial number of professionals and support staff The quality of the unit's performance and of the leader's professional advice can have a very important impact on organization's internal operations and, to some extent, on the external image. Group V Scientists, Support Professionals, Administrators Representative CGLAR titles: Scientist, Agronomist, Biologist, Forester, Accountant, Computer Programmer, Editor, Translator, Librarian. Functional responsibility: a) Scientists - Within broad guidelines on program and project objectives, conceives of; designs and provides leadership to scientific research and training activities and helps to secure and manage the necessary human, financial and material support. Makes an important professional contribution to the relevant strategy and policy of the organization and provides professional advice with a high degree of reliability to organization management and other institutions. b) Support Professionals, Administrators - Within broad guidelines about service expectations, priorities and approach to activities, manages a smaller unit of the organization charged with the day-to-day delivery of a corporate administrative or technical service (e.g., finance, human resources, computing or information services). Responsible for the unit's staffing and financial management and for providing leadership and direction to its staff Reviews, develops and implements relevant systems, policies and procedures. Makes an important professional contribution to the relevant strategy and policy of the organization and provides professional advice with a high degree of reliability to organization management. Generally reports to a Division Head. EducationExperience: PhD or equivalent in relevant area for scientists; MA/MSJMBA or equivalent in relevant area for support professionals and administrators; usually a minimum of 5 to 7 years experience. Organizational impact: a) Scientists - In a limited way, represents the organization to outside world, interacting with Board and donors and with staff ofpartner and client institutions in collaborative research and training. Supervises the work of a minimal number of scientific professional and support staff Thus can have an important impact on organization's internal operations but somewhat less on the external image. b) Support Professionals, Administrators - Interacts with staff of research units and, to some extent, with Board and donors. Supervises the work of a small number of professionals and support staff The quality of the unit's performance and of the leader's professional advice can have an important impact on organization's internal operations and, to some extent, on the extemal image. Group VI Associate-Scientists, Support Professionals, Administrators Representative CGIAR titles: Associate Scientist, Associate Agronomist, Associate Economist, Associate Administrator, Associate Expert. Functional responsibility: a) Scientists - Within relatively detailed guidelines on objectives and priorities, participates in the design of and undertakes scientific research or training duties requiring highly advanced technical expertise and professional judgement. Carries responsibilities for all aspects of activities that are substantial in scope and complexity and require an in-depth understanding of relevant science and organization protocols. Generally reports to a Division Head, sometimes via a senior scientist within the respective unit. b) Support Professionals, Administrators - Within relatively detailed guidelines about service expectations, priorities and approach to activities, provides a day-to-day corporate administrative or technical service (e.g., finance, human resources, computing or information services). Responsible for financial management of the service. Assists in developing and implementing relevant systems, policies and procedures. Provides professional advice with a high degree of reliability to organization management. Generally reports to a Division Head or a Senior Professional or Administrator. Education/Exerience: PhD or equivalent in relevant area for scientists; MA/MS/MBA or equivalent in relevant area for support professionals and administrators; usually a minimum of 2 to 3 years experience. Organizational imac: a) Interacts with staff of relevant unit and with partner and client institutions in collaborative research and/or training. May supervise a limited number of scientific support staff Thus can have a limited but important impact on organization's internal operations and external image. b) Support Professionals, Administrators - May supervise a limited number of administrative support staff or work independently but interacts with staff through out the organization who use the services in question. The quality of the performance and of the professional advice given can have a limited but important impact on organization's intemal operations and, in some cases, external image. Group VII Post-Doctoral Fellows Functional responsibilit: With relatively detailed guidelines on objectives and approach, participates in the design of and undertakes scientific research and training tasks within an activity or project, working with other staff involved in the work both within the organization and in partner institutions. Provides and accepts responsibiliy for reports on all aspects of the assigned work. Reports to a Division Head, sometimes via a more senior scientist within the respective unit. Education/Experience: PhD or equivalent in relevant area; position usually taken up immediately upon completion of degree. Organizational impact: hmpact is generally limited to the projects assigned and other members of the project team, although this may include staff ofpartner institutions and thus effect the organization's external image to some extent. May supervise a limited number of scientific support staff Appendix 3 Summary tables: 1999 CGIAR Human Resources Survey CGIAR HUMAN RESOURCES SURVEY (1999): GENDER ANALYSIS SUMMARY ..~~~~~~~~~~ . If . .~l~~ %~v Total number of international staff: 162 804 966 100% 100% 100% 17% 83% Staffing by Position Group 1: Executive Staff 5 56 61 6% 3% 7% 8% 92% II: Research Program/Admin.Heads 13 104 117 12% 8% 13% 11% 89% III: Principal Scientists 8 118 126 13% 5% 15% 6% 94% IV: Sr. Scientists/Support Professionals 34 204 238 25% 21% 25% 14% 86% V: Scientists/Support Professionals 47 178 225 23% 29% 22% 21% 79% VI: Assoc. Scientists/Support Professionals 23 57 80 8% 14% 7% 29% 71% VIl: Post-Doctoral Fellows 32 87 119 12% 20% 11% 27% 73% TOTAL 1621 804 966 100% 100% 100% 17% 83% WB Part of Country of Origin __. Part I 1111 405 516 53% 69%I 50% 22% 78% Part II 51 399 450 47% 31% 50% 1 1%O 89% TOTAL 162 804 966 100% 100% 100% 17% 83% Region of Country of Origin Africa 17 156 173 18% 10% 19% 10%1 90% Australia/ New Zealand 6 26 32 3% 4% 3% 19% 81% East Asia and Pacific 16 64 80 8% 10% 8% 20% 80% Eastem Europe and Central Asia 2 7 9 1% 1%- 1% 22% 78% Europe 39 214 253 26% 24%1 27% 15% 85% Latin America and Cadbbean 9 70 79 8% 6% 9% 11% 89% Middle East and North Africa 5 34 39 4% 3% 4% 13%0 87% North America 64 147 211 22% 40% 18% 30%1 70% South Asia 4 86 90 9% 2% 11% 4%0 96% TOTAL 162 804 966 100%1 100%, 81% 17%1 83% Page 1 ........~I~Mt- Mt TTL I OA ~ T~L OA... AOTALjTOA WB Part of Country of Citizenship Part I 115 431 546 57% 71% 54% 21% 79% Part II 47 373 420 43% 29% 46% 11% 89% TOTAL 162 804 9661 100% 100% 100% 17% 83% Region of Country of Citizenship Africa 15 146 161 17% 9% 18% 9% 91% Australia/ New Zealand 6 30 36 4% 4% 4% 17% 83% East Asia and Pacific 16 58 74 8% 10% 7% 22% 78% Eastem Europe and Central Asia 2 7 9 1% 1% 1% 22%° 78% Europe 40 223 263 27% 25% 28% 15% 85% Latin America and Caribbean 7 67 74 8% 4% 8% 9% 91% Middle East and North Africa 8 31 36 4% 3% 4% 14% 86% North America 67 162 229 24% 41% 20% 29% 71% South Asia 4 80 84 9% 2% 10% 5% 95% TOTAL 162 804 966 100% 100% 100% 17% 83% WB Part of Country of Last Degree _ Part I 144 689 833 86% 89% 86% 17% 83% Part II 18 115 133 14% 11% 14% 14% 86% TOTAL 162 804 966 100% 100% 100% 17% 83% __ I I I _ Region of Country of Last Degree Africa 6 27 33 3% 4% 3% 18% 82% Australia/ New Zealand 8 33 41 4% 5% 4% 20% 80% East Asia and Pacific 7 37 44 5% 4% 5% 16%1 84% Eastem Europe and entralAsia 1 10 1 1 1% 1% 1% 9%1 91% Europe 45 270 315 33% 28% 34% 14% 86% Latin America and Caribbean 2 8 10 1% 1% 1% 20% 80% Middle East and North Africa 2 6 8 1% 1% 1% 25% 75% North America 90 367 457 47% 56% 46% 20% 80% South Asia 1 46 47 5%1 1% 6% 2% 98% TOTAL 1621 804 966 100%l 100% 100% 17%1 83% Page 2 ~MAL~ MALE TQTAL TOTAl4. JF TOTAL- MT TAL TO OT. Tenure at Center (number of years emploi fed at Center) Less than 1 year 18 93 111 11% 11% 12% 16% 84% 1-3 years 73 252 325 34% 45% 31% 22% 78% 47- years 27 139 166 17% 17% 17% 16%h 84% 7-9 years 9 110 119 12% 6% 14% 8% 92% 1TOrTmoreAyears 351 210 245 25% 22% 26% 14% 86% TOTAL 162 804 966 100% 100% 100% 17% 83% Degree Levels (highest degree received) I.. PhD or equivalent 103 654 757 78% 64% 81% 14% 86% MA/MS or equivalent 34 75 109 11% 21% 9% 31% 69% Batchelor s 11 35 46 5% 7% 4% 24% 76% Other 14 40 54 6% 9% 5% 26% 74% TOTAL 162 804 966 100% 100% 100% 17% 83% Disciplinary Area (in which highest degre received) _ _ _ _ 1: Management/information 28 60 88 9% 17% 7% 32% 68% 11: Social Sciences 47 157 2041 21% 29% 20% 23% 77% Il1: Natural Sciences 87 587 674j 70% 54%j 73%j 13% 87% TOTAL 162 8041 9661 100%| 100%| 100%| 17%. 83% Years of Relevant Professional Experience (post PhL or equlvalent) _ __ _ _ _ _ Less than 5 years <1 531 122 175 18% 33% 15% 30% 70% 5 - 9 years __ 44 163 2071 21% 27% 20% 21% 79% 10- 1 9 years il 46 300 346 36% 28% 37% 13% 87% _20 or more years __ 191 219_ 238 25% 12% 27% 8% 92% TOTAL_________________________ n 1621 8041 966t 100%1 I 0%1Z I 00%| 17%1 83% Page 3 u .. . ....~hA R _ __ __ _________~~~~~~~ TOTA A ~ 'r iTTL,~sp4 ijr Age (years) 25-34 27 54 81 8% 17% 7% 33% 67% 35-44 81 292 373 39% 50% 36% 22% 78% 45-54 44 344 388 40% 27% 43% 11% 89% Over 55 10 114 124 13% 6% 14% 8% 92% TOTAL 162 804 966 100% 100% 100% 17% 83% Marital status (number of staffl Married with spouse/partner in residence 96 681 777 80% 59% 85% 12%h 88% Married wiout spouse/partner in residence 10 56 66 7% 6% 7% 15% 85% Single/divorced/ivdowed 56 67 123 13% 35% 8% 46% 54% TOTAL 162 804 966 100% 100% 100% 17% 83% Children (number of staff) With children in residence S9 438 497 51% 36% 54%1 12% 88% No children in residence 103 366 469 49% 64% 46% 22% 78% TOTAL 1621 804 966 100% 100% 100% 17% 83% Page 4 CGIAR HUMAN RESOURCES SURVEY (1999): DIVERSITY ANALYSIS SUMMARY Total . ..~~~. ~~. ~. . .~~. Part ijN~~~ tOTAL ~~ * T0TALO 1V:T A ro6w'TOTA owTOA Total number of international staff:._ 516 450 966 100% 100% 100% 53% 47% Staffing by Position Group_ __. I: Executive Staff 37 24 61 6% 7% 5% 61% 39% II: Research Program/Admin. Heads 61 56 117 12% 12% 12% 52% 48% III: Principal Scientists 63 63 126 13% 12% 14% 50% 50% IV: Sr. Scientists/Support Professionals 140_ 98 238 25% 27% 22% 59% 41% V: Scientists/Support Professionals 124 101 225 23% 24% 22% 55% 45% VI: Assoc. ScientiststSupport Profesionals 45 35 80 8% 9% 8% 56% 44% VII: Post-Doctoral Fellows 46 73 119 12% 9% 16% 39% 61% TOTAL 516 450 966 100% 100% 100% 53% 47% WB Part of Country of Origin Part I 516 0 516 53% 100% 0% 100% 0% Part II 0 450 450 47% 0% 100°h_ 0% 100% TOTAL 516 450 966 10 0%1 IOOO 53° 47% Region of Country of Origin Africa 4 169 173 18% 1% 38% 2% 98% Australia/ New Zealand 32 0 32 3% 6% 0% 100% 0% East Asia and Pacific 17. 63 80 8% 3% 140% 21% 79% Eastern Europe and Central Asia 2 7 9 1% 0% 2%1 22% 78% Europe 250 3 253 26% 48% 1%1 99%6 1% Latin America and Caribbean 0 79 79 8% 0% 18% 0% 100% Middle East and North Africa 0 39 39 4% 0% 9% 0% 100% North America 2111 0 211 22% 41% 0%1 100%1 0% South Asia 0° 90 90 9% 0% 20%1 0%1 100%h TOTAL 516| 450 966 100% 100% - 62%1 53%1 47% Page 5 % PrlwsPr%~PrI% WB Part of Country of Citizenship Part I 515 31 546 57% 100% 7% 94% 6% Part 11 1 419 420 43% 0% 93%° 0% 100% TOTAL 516 450 966 100% 100% 100% 53% 47% Region of Country of Citizenship Africa 3 158 161 17% 1% 35% 2% 98% Australia/ New Zealand 33 3 36 4% 6% 1% 92% 8% East Asia and Pacific 17 57 74 8% 3% 13% 23%0/ 77% Eastern Europe and Central Asia 21 7 9 1 % 0% 2% 22% 78% Europe 250 13 263 27% 48% 3% 95% 5% Latin America and Canbbean 0 74 74 8% 0% 16% 0% 100%f Middle East and North Africa 0 36 36 4% 0% 8% 0% 100% North America 211 18 229 24% 41% 4% 92% 8% South Asia 0 84 84 9% 0% 19% 0% 100% TOTAL 516 450 966 100% 100% 100% 53% 47% WB Part of Country of Last Degree Part I 510 323 833 86% 99% 72%hl 61% 39% Part 11 6 127 133 14% 1% 28% 5% 95% TOTAL 516 450 966 100% 100% 100% 53%1 47% Region of Country of Last Degree Africa 3 30 33 3% 1% 7%1 9% 91% Australia/ New Zealand 30 111 411 4% 6% 2% 73% 27% East Asia and Pacific _--_- 15 29 44 5% 3% 6% 34% 66% Eastern Europe and Central Asia 2 9 11 1% 0% 2% 18% 82% Europe _ 221 94 315 33% 43% 21% 70% 30% Latin Amerca and Caribbean 0 10 10 1% 0% 2% 0% 100% Middle East and North Africa 0 8 8 1% 0% 2% 0% 100% North America - ll 2431 214| 4571 47%, 47%. 48%1 53% 47% South Asia 21 451 471 5%1 0%1 10%1 4%1 96% TOTAL ___ 5161 4501 9661 100%1 100%1 100%1 53%1 47% Page 6 Tenure at Center (number of years emplo fed at Center) Less than 1 year 55 56 111 11% 11% 12% 50% 50% 1-3 years 192 133 325 34% 37% 30% 59% 41% 4-6 years 95 71 166 17% 18% 16% 57% 43% 7-9 years 58 61 119 12% 11% 14%° 49% 51% 10 or more years 116 129 245 25% 22% 29%h 47% 53% TOTArL 516 450 966 100% 100% 100% 53% 47% TOTAL_ Degree Levels (highest degree received) 385 372 77858__4 PhD or equivalent _ 385 372 757 78% 75% 83% 51% 49% MA/MS or equivalent 72 37 109 11% 14% 8% 66% 34% Bachelors 28 18 46 5% 5% 4% 61% 39% Other 31 23 54 6% 6% 5_ % 57% 43% TOTAL 516 450 96 100% 100% 100% 53% 47% Disciplinary Area (in which highest deg e received 1: Management/information 56 32 88 9% 11% 7% 64% 36% II: Social Sciences 120 84 204 21% 23% 19% 59% 41% ill: Natural Sciences 340 334 674, 70%, 66% 74% 50% 50% TOTAL 516 450 9661 100%1__ 100%1 100%1 53%1 47% Years of Relevant Professional Experience (post PhD or equivale nt Less_than5_year's 100 75 175 18% 19% 17% 57% 43% 5 - 9 years __ 114 93 207 21% 22% 21% 55% 45% 10-19 years 1 198 148 346 36% 38% 33%1 57% 43% 20 or more years 1l 104 1341 2381 25%1 20%1 30%1 44% 56% TOTAL __ 516, 4501 9661 100%1 100%1 100%1 53%; 47% Page 7 t!b.^X ..... X . 0 <; 4 1 -.": mv . P A L 11~ 1~r 1 1~~ ITOA T. TOTA l IITTlf~ Al jIWTTI Age (years)___ 25-34 51 30 81 8% 10% 7% 63% 37% 35-44 235 138 373 39% 46% 31% 63% 37% 45-54 183 205 388 40% 35% 46% 47% 53% Over 55 47 77 124 13% 9% 17% 38% 62% TOTAL 516 450 966 100% 100% 100% 53% 47% Marital Status (number of staff) Married with spouse/partner in residence 395 382 777 80% 77% 85% 51% 49% Married w/out spouse/partner in residence 33 33 66 7% 6% 7% 50% 50% Single/divorcedlwidowed 88 35 123 13% 17% 8% 72% 28% TOTAL 516 450 9661 100% 100% 100% 53% 47% Children (number of staff) With children in residence I 244 250 494 51% 47% 56% 49% 51% No children in residence 272 200 472 49% 53% 44% 58% 42% TOTAL i 516 4501 9661 100%t 100% 100% 530% 47% Page 8 Appendix 4 Table 1. Summary Table, 1997 CGIAR Human Resources Survey Table 2. Summary Table, 1994 CGLAR Human Resources Survey Table 3. Summary Table, 1991 CGLAR Human Resources Survey Annex 14.-Table 1 1997 Human Resources Survey TABLE 1: 1997 HUMAN RESOURCES SURVEY - SUMMARY ',-,,,,,~ ' ' --_: _j Question 1. Total number of international staff 1002 188 1190 100% 100% 100% 84% 16% Question 2. Staffing by level - by recruited _ _ _ senior manageinent/admninistration 84 6 90 8% 8% 3% 93% 7% departnent hcads/researci thrust leaders 159 21 180 15% 16% 11% 88% 12% senior and/or principal scientists 379 47 426 36% 38% 25% 89% 11% junior or associate scientists 112 25 137 12% 11% 13% 82% 18% visiting scientists/research fellows 67 20 87 7% 7% 11% 77% 23% posidoctoral scientists/fellows 90 26 116 10% 9% 14% 78% 22% associate experts 52 23 75 6% 5% 12% 69% 31% other internationally recruited 60 20 80 7% 6% 11% 75% 25% administrative staff/or professional suppot staff TOTAL 1003 188 1191 100% 100% 100% 84% 16% Question 3. Age (years) 20-30 38 23 61 5% 4% 12% 62% 38% 31-40 267 89 356 31% 27% 48% 75% 25% 41-50 410 57 467 40% 42% 31% 88% 12% 51-60 228 17 245 21% 23% 9% 93% 7% 61 and above 30 0 30 3% 3% 0% 100% 0% TrOTAL- 973 186 1159 100% 100% 100% 84% 16% Question 4. Nationialilt_ Asia/Oceania 151 25 176 15% 15% 13% 86% 14% Latin Amnerica/Caribbean 78 8 86 7% 8% 4% 91% 9% Sub-Saliaran Africa 153 11 164 14% 15% 6% 93% 7% West Asia/Norlh Africa 57 7 64 5% 6% 4% 89% 11% Northt America 184 52 236 20% 18% 28% 78% 22% Europe 319 80 399 34% 32% 43% 80% 20% Australia/New Zealatnd 39 5 44 4% 4% 3/ 3% 89% 11% Japan 21 0 21 2% 2% 0% 100% 0% TOTAL 1002 188 1190 100% 100% 100% 84%1 16% AnnexJq,Table 1 - 1997 Human Resources Sutrvey Table 1 - 1 Annex 41:-Table 1 1997 Human Resources Survey QUESTION #: .: M % o:as :% __________________-__ .MA . FEMAE TOTAL TO.A. M TOTAL FT. TA TAL TOTAL Question 5. Tenure at Ceinter (number of years _ employed at Center) ._.__ Less than I 125 41 166 14% 12% 220/ 75% 25 1-3 321 76 397 33% 32% 40% 81% 19% 4-6 176 36 212 18% 18% 19% 83% 17% 7-9 145 18 163 14% 14% 10% 89% 11% More than 10 236 17 253 21% 24% 9%M 93% 7% TOTAL 1003 188 1191 100% 100% 100% 84% 16% Quecsion 6. Location/ Posting Headquarters 689 141 830 70% 69% 75% 83% 17% Outposted (regional or field position) 313 47 360 30% 31% 25% 87% 13% TOTAL 1002 188 1190 100% 100% 100% 84% 16% Question 7. Funding source Fixed term, renewable appointine'tt 777 134 911 77% 78% 72% 85% 15% Special project - non-renewable 142 24 166 14% 14% 13% 86% 14% Donor funded positons 80 29 109 9% 8% 16% 73% 27% TOTAL 999 187 1186 100% 100% 100% 84% 16% Qucstion 8. Staff on part-time contracts (<75%) 8 1 9 I . 1% 1% 89% 11% Question 9. Degree levels (htighest degree received) . PM.D. or equivalent 791 99 890 75% 77% 58% 89% 11° Msc/MA/ or equivalenlt 146 51 197 17% 14% 30% 74% 26% Olher 85 20 105 9% 8% 12% 81% 19% TOTAL 1022 170 1192 100% 100% 100% 86% 14% Question 10. Discipline (in whicIh higlhest degree __....._ received) Crop sciences 380 5 1 431 36% 38% 27% 88% 12% Animal sciences 48 4 52 4%° 5%o 2% 92% _8% Cellulat sciences (microbiology) 61 16 771c % 6% 9% 79% 21% Annex 4-Table 1 - 1997 Human Resources Survey Table I - 2 Annex L4Table 1 1997 Human Resources Survey QUESTION 1 . .: . - _ ;:: : : % o :: M as % Ps% 7w M .______-______:______:_____.__ _ :__ MALE -: FEMALE TOTAL TOTAL 7 M TOTAL F TOAL TOTAL TOTAL Forestry/agroforestiy 31 6 37 3% 3% 3% 84% 16% Olher biological sciences 67 18 85 7% 7% 10% 79% 21% Cliemistry 4 0 4 0% 0%. 0% 100% O% Physical sciences 10 0 10 1% 1%/c 0% 100% 0%/ Environmental/soil and resource inngt. sciences 88 12 100 8% 9% 6% 88% 12% Engineering 37 8 45 4% 4% 4% 82% 18% Social/economic sciences 157 38 195 16% 16% 20% 81% 19% Coinputer/infonnation sciences 29 12 41 3% 3% 6% 71% 29% Matliematics/statistics 8 2 10 1% 1% 1% 80% 20% Management/administration 46 11 57 5% 5% 6% 81% 19% Otlher (specify) 35 10 45 4% 3% 5% 78% 22% TOTAL 1001 188 1189 100% 100% 100% 84% 16% Question I 1. Slaff actively engaged in biotechnology 64 17 81 0% °6% 9% 79% 21%. research_ Question 12. Years of relevant professional experience (post Msc or equiv.) . < 5 years 91 36 127 12% 10%/c 22% 72% 28% 5 -9 years 136 50 186 18% 15% 30% 73% 27% 10-19 years 333 54 387 37% 38% 33% 86% 14% 20-30 years 266 19 285 27% 30% 12% 93% 7% > 30 years 52 5 57 5% 6% 3% 91% 9% TOTAL 878 164 1042 100% 100% 100% 84% 16% Question 13. Marital status (number of slaff) . ._._. married w/spouse in residence 824 82 906 76% 82% 44% 91% 9% married v/out spouse in residence 63 14 77 6% 6% 7% 82% 18% single/divorced/widowed 115 92 207 17% 11% 49% 56% 44% TOTAL 1002 188 1190 100% 100% 100% 84% 16% Question 14. Children (number of staff) . ._._ Willt childreni 787 73 860 72% 79% 39% 92% 8% No children 215 115 330 28%° 21% 61% 65% 35% TOTAL 1002 188 1190 100% 100% 100% 84%° 16% Annex 4-Table 1 - 1997 Human Resources Survey Table 1 - 3 Annex 4- Table 1 1997 Human Resources Survey .Q .S T I #. : EMALM -,A I :A M T -i O : T T TOTAj Part IlI. Additional Information for Analysis of . ._. Gender Staffing 18. Number of locally-recruited scientists (1997) 258 201 459 n/a n/a n/a 56% 44% 19. Number of locally-recruited senior managers/ 115 81 196 n/a n/a in/a 59% 41% adinin. (1997) 20. International consultants hired in 1996 258 59 317 n/a n/a n/a 81% 19% 21. Regional and/or national consultants hired in 1996 164 58 222 n/a n/a n/a 774% 26% 22. Spouses of internationally-recruited staff 3 17 20 n/a n/a n/a 15% 85% lired as consultants (1996) 23. Shtort-course group trainees (in headquarters 876 170 1046 n/a n/a n/a 84% 16% and regions) in 1996 _ _ 24. Ph.D. trainees in 1996 201 121 322 n/a n/a n/a 62% 38% 25. Msc trainees in 1996 128 45 173 n/a n/a n/a 74% 26% Annex 4-Table 1 - 1997 Human Resources Survey Table I - 4 Annex 4-Table 2 - 1994 Human Resources Survey T'ABLE 2: 1994 IIUMAN RESOURCES SURVEY - SUMMARY __. QUESTION # . %of M as Fas% M.%:ro*:: F 0/ row M:. :---: - .. FEAL T TOTAL O M F .TOT TOAL TOTAL Question 1. Total numiiber of international staff 1051 173 1224 100%/ 100% 100% 86% 14 %o Question 2. Staffing by level - by recruited _____ senior managemenit/administration 84 5 89 7% . 8% 3 % 94% 6% department lieads/researcli thrust leaders 148 15 163 13% 14% 9% 91% 9% senior and/or principal scientists 393 39 432 35% 37% 23% 91% 9% junior or associate scientists 134 19 153 13% 13% 11% 88% 12% visiting scientists/researcli fellows 71 17 88 7% 7% 10% 81% 19% postdoctoral scientists/fellows 103 30 133 11% 10% 17% 77% 23% associate experts 49 16 65 5% 5% 9% 75% 25% otlher internationally recruited 69 32 101 8% 7% 18% 68% 32% administrative staff/or professional support staff ___ _ _ TOTAL 1051 173 1224 100% 100% 100% 86% 14%/o Quiestion 3. Age (years) 20-30 40 26_ 66 5% 4% 15% 61% 39% 31-40 325 82 407 33% 31% 47% 80% 20% 41-50 431 55 486 40% 41% 32% 89% 11% 51-60 231 9 240 20% 22% 5% 96% 4% 61 and above 24 1 25 2% 2% 1% 96% 4% TOTAL 1051 173 1224 100% 100% 100% 86% 14% Qtiestion 4. Nationality Asia/Oceania 190 17 207 17% 18% 10% 92% 8% Latin America/Caribbeat 98 4 102 8% 9% 2% 96% 4% Sub-Saharami Africa 168 15 183 15% 16% 9% 92% 8% West Asia/Nortlm Africa 54 7 61 5%1 5% 4% 89% 11% North America 178 55 233 19% 17% 32% 76% 24% Europe 309 71 380 31% 29% 4 1% 81% 19% Australia/New Zealand 34 3 37 3% 3% 2% 92% 8% JapalF 21 1 22 2% 2% 1% 95% 5% Annex 4i-Table 2 - 1994 Human Resources Survey Table 2 - 1 Annex4- Table 2 - 1994 Human Resources Survey QUESTION # o Ms ass% M r F~%row: ALE FEMALE: TOTL TALAL M OT. F TOTAL TOTAL TOTAL TOTAL 1052 173 1225 100% 100% 10(% 86% 14% Qucslion 5. Tenure at Center (numilber of ycars employed at Center) Less than 1 142 39 181 15% 14% 23% 78% 22% 1-3 336 70 406 33% 32% 40% 83% 17% 4-6 __ 202 27 229 19% 19% 16% 88% 12% 7-9 134 23 157 13% 13% 13% 85% 15% More than 10 237 14 251 21% 230% 8% 94%/o 6% TOTAL 1051 173 1224 100% 100% 100% 86% 14% Question 6. Location/ Posting _ . .__ Headquarters 734 142 876 72% 70%o 82% 84% 16% Oulposted (regional or field position) 317 31 348 28% 30% 18% 91%° 9% TOTAL 1051 173 1224 100% 100% 100% 86% 14% Question 7. Funding source .__ In TAC approved core staff positions 667 92 759 64% 65% 55% 88% 12% Otlier staff positions 355 74 429 36% 35% 45% 83% 17% TOTAL 1022 166 1188, 100% 100% 100% 86% 14% Question 8. Staff on parthtime contracts (<75%) 12 5 17 0% 0% 0% 0% 0% Question 9. Degree levels (highest degree received) _ . Ph.D. or equivalent _ 792 95 887 72% 75% 55% 89% 11% MscIMAI or equivalent 161 52 213 17% 15% 30% 76% 24% Otiher 98 26 124 10% 9% 15% 79% 21% TOTAL 1051 173 1224 100% 100% 100% 86% 14% Question 10. Discipline (in whiich liigliest degree received) .__._.__ Crop sciences 388 43 431 35% _ 37%, 25% 90% 10% Animal sciences 60 9 69 6% 6% 5% 87% 13% Cellular sciences (microbiology) 75 19 941 11% 80 20% Annex 4-Table 2 - 1994 Human Resources Survey Table 2 - 2 Annex 4 -Table 2 - 1994 Human Resources Survey Foresty/agroforestry 37 3 40 3% 4% 2% 93% 8% Other biological sciences 94 12 106 9% 9% 7% 89% 11% Chemistry 6 1 7 1% 1% 1% 86% 14% Physical sciences '7 0 7 1% 1% 0% 100% 0% Environmental/soil and resource mngt. sciences 83 10 93 8% 8% 6% 89% 11% Engineering _ 46 2 48 4% 4% 1% 96% 4% SociaU/econoinic sciences 145 43 188 15% 14% 25% 77% 23% Computer/infornnation sciences . 29 7 36 3% 3%1 4% 81% 19% Mathematics/s1atistics 12 13 1% 1% 1% 92% 8% Management/adninistration 1 4. _5 16 61 5% 4% 9% 74% 26% Other (specif) 1 24 7 31 3% 2% 4% 77% 23% TOTAL 1051 173 1224 100% 100% 100% 86% 14% Question 11. Staff actively engaged in biotechnology 73 21 94 0% 0% 0% 78% 22% research I Question 12. Years of relevant professional experience I (post Msc or equiv.) < 5 years 166 50 216 18% 16% 29% 77% 23% 5 - 9 years 185 36 221 18% 18% 21% 84% 16% 10-19 years 362 58 420 34%. 34% 34% 86% 14% 20-30 years 284 27 311 25%1 27% 160% 91% 9% > 30 years 54 1 55 4% 5% 1% 98% 2% TOTAL 1051 172 1223 100% 100% 100% 86% 14% Question 13. Marital status (number of staff) married w/spouse in residence 857 76 933 77% 82% 44% 92% 8% married w/out spouse in residence -67_ 9 76 6% 6% 5% 88% 12% single/divorced/widowed 121 87 208 17% 12%/s 51% 58% 42% TOTAL 1045 172 1217 100% 100% 100% 86% 14% Question 14. Children (number of staff) With children 851 73 924 76% 81% 42% 92% 8% No children 194 99 293 24% 19% 58% 66% 34% TOTAL 1045 172 1217 100% 100% 100%I 86% 14% Annex 4-Table 2 - 1994 Human Resources Survey Table 2 - 3 Annex 4-Table 2 - 1994 Human Resources Survey P t KAIWi Y ,i:il ;il400g[; :i i ii PA%Jeti0 g :j .. .l| ;;|50i|Wt|M Spt Part III. Additional Information for Analysis of Gender Staffing 18. Number of locally-recruited scientists (1994) 311 139 450 n/a i/a n/a 69% 31% 19. Number of locally-recruited senior managers/ 119 28 147 n/a n/a n/a 81% 19% admin. (1994) I 20. International consultants hired in 1994 199 38 237 n/a n/a n/a 84% 16% 21. Regional and/or national consultants hired in 1994 105 32 137 n/a n/a n/a 77% 23% 22. Spouses of internationally-recruited staff 2 15 17 n/a n/a n/a 12% 88% hired as consultants 23. Shiort-course group trainees (in headquarters 1894 417 2311 n/a n/a n/a 82% 18% and regions) in 1994 ___ 24. Ph.D. trainees in 1994 212 75 287 n/a n/a n/a 74% 26% 25. Msc trainees in 1994 158 47 205 n/a n/a n/a 77% 23% Annex 4jTable 2 - 1994 Human Resources Survey Table 2 - 4 Annex qf-Table 3. 1991 Human Resources Survey TABLE 3: 1991 HUMAN RESOURCES SURVEY - SUMMARY .USTQ .. I.:z .-Ms% Fs % %o r ~.. .... MAL' "''A'"' '''-A EtTA M _'~A _~ tOA _7A TOA Question 1. Total number of international staff 1142 153 1295 100% 100% 100% 88% 12% Queslion 2. Staffing by level - by recruited I senior management/administration 86 .2 88 7% 8% 1% 98% 22%l deparlment heads/research thrust leaders 134 9 143 11% 12% 6% 94% 6% senior and/or principal scientists 519 49 568 44% 45% 32% 91% 9% unior or associate scientists 85 26 111 9% 7% 17% 77% 23% visiting scientists/research fellows 130 14 144 11% 11% 9% 90% 10% postdoctoral scientists/fellows 88 19 107 8% 8% 12% 82% 18% associate experts 18 8 26 2% 2% 5% 69% 31% other internationally recruited 82 26 108 8% 7% 17% 76% 24% administrative staff/or professional support staff . .__ TOTAL 1142 153 1295 100% 100% 100% 88% 12% Question 3. Age (years) I 20-30 49 17 66 6% 5% 12% 74% 26% 31-40 336 63 399 33% 32% 44% 84% 16% 41-50 430 48 478 40% 41% 34% 90% 10% 51.60 197 13 210 18% 19% 9% 94% 6% 61 and above 42 2 44 4% 4% 1% 95% 5% TOTAL 1054 143 1197 100% 100% 100% 88% 12% Question 4. Nationality Asia/Oceania 187 17 204 17% 18% 12% 92% 8% Latin America/Caribbean 100 8 108 9% 10% 6% 93% 7% Sub-Saharan Africa 150 9 159 13% 14% 6% 94% 6% West Asia/North Africa 40 3 43 4% 4% 2% 93% 7% North America 203 55 258 22% 19% 38% 79% 21%1 Europe 310 48 358 30% 30% 33% 87% 13% AustraliaJNew Zealand 37 4 41 3% 4% 3% 90% 10% Japan 20 0 20 2% 2% 0% 100% oc0% Annex q-Table 3 - 1991 Human Resources Survey Table 3 - 1 Annex 4-Table 3 1991 Human Resources Survey TOTAL 1047 144 1191 100% 100% 100% 88% 12% Qucstion 5. Tenure at Center (number of years I _ employed at Center) _________ ___ ___ Less than I | 133 34 167 14% 13% 24% 80% 20% 1-3 377 57 434 36%D/ 36% 40% 87% 13% 4-6 186 33 219 18% 18% 23% 85% 15% 7-9 126 5 131 11% 12% 3% 96% 4% More than 10 225 15 240 20% 21% 10% 94% 6% TOTAL 1047 144 1191 100% 100% 100% 88% 12% 730 113 843 71% 69% 78% 87% 13% OutHosted resional or field osition 321 31 352 29% 31% 22% 91% 9% TOu(oTed (regionat or fteid position) 1051 144 1195 100% 100% 100% 88% 12% TOTAL. Question 7. Funding source L in TAC approved core staff positions 771 100 871 79% 80% 76%-/ 89 11% Other staff positions 195 3 1 226 21% 20% 24% 86% 14% TOTAL - 1 966 131 1097 100%/ 100% 100% 88% 12% Question 8. Staff on part-time contracts (<75%) I11 2 13 100%/ 100% 100% 85% 15% Question 9. Degree levels (hiighest degree received) _ Ph.D. or equivalent 799 77 876 73% 76% 53% 91% ° 9% Msc/MA/ or equivalent 158 46 204 17% 15% 32% 77% 23% Olher 95 21 116 10% 9% 15% 82% 18% TOTAL 1052 144 1196 100% 100% 100% 88% 12% Question 10. Discipline (in which highest degree received) Crop sciences 366 29 395 33% 35% 20% 93% 7% Animal sciences 71 5 76 6% 71 3% 93% 7% Cclular sciences (microbiOI0oY) 75 159 941 8% 7%1 13%1 80° 20% Annex i-Table 3 - 1891 Human Resources Survey Table 3 - 2 Annexp-Table 3 1991 Human Resources Survey . ... . ...:;- .:.. .*... .* ..* . . Forestry/agroforestry 20 1 21 2% 2% 1%. 95% 5% Other biological sciences 102 . 9 111 9% 10°% 16%0 92% 8% Chtemnistry 19 0 9 1% 1%° 0% 100% 0% P=ysical sciences 10 0 10 1% 1% 0% 100% 0% EnvironmentaUsoil and resource mngt. sciences 85 3 88 7% 8% 2% 97% 3% Engineering 44 0 44 4% 4% 0% 100% 0% S&iaUeconomic sciences 131 38 169 14% 13% 27% 78% 22%!° Computer/information sciences 29 17 46 4% 3% 12% 63% 37% Matliematicstslatistlics 8 2 10 1% 1% 1% 80% 20% Managementladministration 59 6 65 5% 6% 4% 91% 9% Othier (specify) 37 14 51 4% 4% 10% 73% 27% TOTAL 1046 143 1189 100% 100% 100% 88% 12% Question 11. Staffactively engaged in biotechnology 68 24 92 0% 0% 0% 74% 26% research Question 12. Years of relevant professional experience (post Msc or equiv.) . < 5 years 72 10 82 7% 7% 9% 88% 12% 5-9 years 169 29 198 18% 1.7% 26% 85% 15%° 10-19 years 430 47 477 43% 43% 42% 90% 10% 20-30 years 276 21 297 27% 28%° 19% 93% 7% > 30 years 56 5 61 5% 6% 4% 92% 8% TOTAL 1003 112 1115 100% 100% 100%N 90% 10% Question 13. Marital status (number of staff) married w/spouse in residence 881 69 950 79% 83% 48% 93% 7% inarried w/out spouse in residence 55 8 63 5% 5% 6%. 87% 13% sin le/divorced/widowed 127 68 195 16% 12% 47% 65% 35% TOTAL 1063 145 1208 100% 100% 100% 88% 12% Question 14. Cbildren (number of staff) With cliildren 859 69 928 78% 82% 50% 93% 7°A ,No chtildren 185 70 255 22% 18% 50% 73% 27% TOTAL 1044 139 1183 100% 100% 100%° 88% 12% Annex I-Table 3 - 1991 Human Resources Survey Table 3 - 3 Annex 4-Table 3 1991 Human Resources Survey Part III. Additional Information for Analysis of ____________ Gender Staffing 18. Number of locally-recruited scientists (1991)* - 109 109 n/a nh n/a 0% 100% 19. Number of locally-recruited senior managers190 26 26 n/a n/a n/a 0% 100% admin. (1991)* 1 20. International consultants hired in 1991 41 41 n/a n/a n/a 0% 100% 21. Regional and/or national consultants bired in 1991 34 34 n/a n/a n/a 0% 100% 22, Spouses of internationally-recruited staff 17 17 n/a n/a n/a 0% 100% hired as consultants 23. Short-course group trainees (in headquarters 19 19 n/a n/a n/a 0% 100% and regions) in 1991 24. Ph.D. trainees in 1991 45 45 n/a n/a n/a 0% 100% 25. MSc trainees in 1991 48 48 n/a n/a n/a 0% 100% * Data for total number of locally-recruited scientists and managers estimated from center publications in order to draw comparisons to 1994 and 1997. Annex q.Table 3 - 1991 Human Resources Survey Table 3 - 4 Appendix 5 Fields in the 1999 survey Position group I through VII Age (in groups 25-34, 35-44, 45-54, over 55) Degree: Last degree earned (PhD or equivalent, MSc/MA or equivalent, Bachelors, Other) Country where earned Region of country where earned (using World Bank designations) World Bank Part of country where earned Years or relevant professional experience, post PhD or equivalent (less than 5 years, 5-9, 10-19, more than 20)* Disciplinary area in which highest degree earned (I=Mgmt/Information, II=Social Sciences, LI=Natural Sciences) Birthplace: Country of origin Region of country of origin World Bank Part of country of origin Citizenship: Country of citizenship Region of country of citizenship World Bank Part of country of citizenship Tenure at Center: (less than 1 year, 1-3, 4-6, 7-9, more than 10)* Personal status: Married with resident sppuse/partner in residence Married without spouse/partner in residence Single/diyorcedlwidowed Children in residence at post Basic salary * The variable for tenure was constructed by subtracting the reported date of last degree from 2000. To account for the time taken to achieve a PhD, those whose last degree was reported as Master's or "other" had 4 years subtracted. Those with only a Bachelor's degree had 5 years subtracted. For comparison, John J. Siegfied and Wendy A Stock, 'The Labor Market for New PhD Economists", Journal of Economic Perspectives, voL 13.3 (1999) pp.115-34, report that those receiving US Economics PhD's in 1986 had taken an average of 6.3 years from the Bachelor's leveL Appendix 6 Annex 1: Survey and Methodology Survey Full Details of regression analysis discussed in section V Data The data were constructed from the CGLAR 1999 survey. Throughout the analysis information on the same 966 individuals was used. The survey collected data on 976 individuals. The 10 people dropped were: CIP, 5 men and 1 woman, lack of base salary data; IVMI, 1 man, insufficient data; ICRISAT, 2 men and 1 woman, joined centre in 35th millennium. Variable names and descriptions BSALARY Base salary in US dollars. LSALARY Logarithm of Base salary CONST: An artificial variable incorporated for technical reasons CIFOR..WARDA Dummy variables which indicate menmbership of appropriate centre. (E.g. CIFOR is 1 for people who work there and 0 otherwise). Reference group, CIAT. PGI ...PGVI Dummy variables for position group. Reference group, PG VII. FEMALE Dummy variable. Reference group Males BA. .OTHER Dummy for educational attainment. Reference group, PhD's FCII, FCEII Dummies for field codgs of last degree. FCII is social sciences FCII is natural sciences. Reference group: management/information. OWBPII Dummy for individuals bom in WB part II countries. Reference group, those born in WB part I countries. MARRIED Dummy variable. Reference group, single/divorced/widowed. CHELDREN Dummy variable. Reference group, no dependent children. AGE3544 ... AGE55 Dummy variables for bracketed age groups: 35-44, 45-54, over 55. Reference group: under 35's. YRX Years of relevant experience. This variable was constructed by subtracting the reported date of last degree from 2000. To account for the time taken to achieve a PhD, those whose last degree was reported as Masters or "other" had 4 years subtracted. Those with only a Bachelor's degree had 5 years subtracted. YJC Year joined centre. OWR Works in own world region Each of the salary regressions that follow (ie. I through le) sufferered from heteroscedasticity. Also reported after each OLS output, are the standard errors adjusted by White's method. The implied p-values were those reported in the text. 1 Regressionl. Snmmary Salary Regression: Ordinary Least Squares Estimation Dependent variable is LSALARY 966 observations used for eatimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.8163 2.0428 6.7632[.000] CIFOR .21143 .036279 5.8279[.000] C1MNY2 -.0076071 .027503 -.27659[.782) CIP -.17048 .031081 -5.4849[.000] ICARDA -.27123 .029006 -9.3509[.000) ICLARM .10796 .042336 2.5500[.011] ICRAF .078846 .032858 2.3996[.017] ICRISAT -.12574 .034569 -3.6374[.000] IFPRI .34761 .031314 11.1009[.000] IITA -.063106 .028532 -2.2117[.027) ILRI .093470 .029004 3.2226[.001] IPGRI .13629 .035808 3.8061[.000l IRRI .010055 .027451 .36629[.714] ISNAR .045116 .037244 1.2114[.226] IWWI .13168 .042737 3.0813[.002] WARDA -.12464 .041696 -2.9891[.003] PGI .97307 .034378 28.3047[.000] PGII .70638 .028681 24.6286[.000] PGIII .67381 .029548 22.8042[.000] PGOV .55816 .024305 22.9649[.000] PGV .39822 .022516 17.6861[.000] PGVI .16976 .026764 6.3429[.000] FENALE -.016122 .016803 -.95947[.338] BA -.076946 .029249 -2.6307[.009] MAX -.036780 .020529 -1.7917[.074) OTHER -.026483 .029595 -.89485[.371] FCII .056505 .026790 2.1092[.035] FCIII .0062834 .024987 .25146[.802] OWBPII -.071666 .013377 -5.3572[.000° MARRIED .015762 .019989 .78853[.431) CHILDREN -.0054711 .014220 -.38475[.701] AGE3544 .063716 .024041 2.6503[.008] AGE4554 .10155 .027558 3.6849[.000] AE55 .14183 .034769 4.0791[.000] YRX .0084366 .0011355 7.4298[.000] YJC -.0018207 .0010218 -1.7819[.075] ONR -.013906 .0146.10 -.95181[.341] R-Squared .80680 R-Bar-Squared .79931 S.E. of Regression .17755 F-stat. F( 36, 929) 107.7621[.000] Mean of Dependent Variable 10.8492 S.D. of Dependent Variable .39633 Residual Sum of Squares 29.2861 Equation Log-likelihood 317.8987 Akaike Info. Criterion 280.8987 Schwarz Bayesian Criterion 190.7452 DW-statistic 1.7904 Diagnostic Tests * Test Statistics * IM Version * F Version * * * * A:Serial Correlation*CHSQ( 1)= 9.7979[.002]*F( 1, 928)= 9.5090[.002] * * * * B:Functional Form *CHSQ( 1)= Z0.3275[.000]*FC 1, 928)= 19.9477[.000) * * * * C:Norsality *CHSQ( 2)= 364.4162[.000]* Not applicable * * * * D:Heteroscedasticity*CHSQ( 1)- 24.6874[.000]*F( 1, 964)= 25.2824[.000] A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 2 Regreesionl: White's Heteroscedasticity Adjusted Standard Errors Dependent variable is LSALARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.8163 2.2181 6.2288[.0003 CIFOR .21143 .034318 6.1610[.000] CIMMYT -.0076071 .027465 -.27697[.782] CIP -.17048 .039232 -4.3453[.000] ICARDA -.27123 .027512 -9.8588[.000] IChAZM .10796 .032202 3.3526t.001] ICRAF .078846 .028120 2.8039[.005) ICRISAT -.12574 .036047 -3.4884t.001] IFPRI .34761 .033641 10.3328t.000] IITA -.063106 .029497 -2.1394[.033] ILRI .093470 .031851 2.9346t.003] IPGRI .13629 .030480 4.4715[.000] IRRI .010055 .031356 .32066[.749] ISNAR _045116 .045023 1.0021[.317] rRIW .13168 .029540 4.4579[.000] WRRDA -.12464 .043354 -2.8748[.004] PGI .97307 .038277 25.4215[.000] PGII _70638 .035424 19.9406[.000] PGIII .67381 .034310 19.6389[.000] PGIV .55816 .028979 19.2606[.000] PGV .39822 .027420 14.5231[.000] PGVI -16976 .033377 5.0861[.000] FE5M^LE -.016122 .017901 -.90063[.368) BA -.076946 .032507 -2.3670E.018) MA -.036780 .024549 -1.4982[.134] OTHER -.026483 .036139 -.73280[.464] FCII .056505 .029099 1.9418[.052) FCIII .0062834 .025908 .24253f.808) ANBPII -.071666 .014723 -4.8676[.000] MARRIED .015762 .020971 .75163r.452] CHILDREN -.0054711 .013544 -.40397t.686] AGE3544 .063716 .031170 2.0441[.041] AGE4554 .10155 .033531 3.0284[.003] AGE55 .14183 .04141a 3.4247[.001] YPX .0084366 .0011651 7.2410E.0001 YJC -.0018207 .0011104 -1.6397[.101) CWR -.013906 .015392 -.90343t.367] 3 Regression la: Ordinary Least Squares Estimation Dependent variable is LSALARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-RatiotProb] CONST 13.1879 2.0565 6.4129[.000J CIFOR .21082 .037203 5.6668t.000] CIMMYT -.022714 .028007 -.81098[.418] CIP -.18669 .031571 -5.9134[.000] ICARDA -.29007 .029383 -9.8721[.000t ICLARM .11221 .043319 2.59031.010) ICRA? .071164 .033182 2.1447[.032] ICRISAT -.14346 .035125 -4.0842t*000] IFPRI .36590 .030381 12.0437E.000] IITA -.083306 .028383 -2.9351[.003] ILRI .096423 .028025 3.4406[.001] IPGRI .112S3 .036002 3.1340t.002] IRRI -.0035772 .027615 -.12954[.897] ISNAR .079996 .037308 2.1442[.032] rWoI .13039 .043526 2.9956[.003] WARDA -.14370 .042053 -3.4172[.001] PGI 1.0423 -033620 31.0025[.000] PGII .74784 .028090 26.6226[.000] PGIII .72664 .029138 24.9381[.000] PGIV .60669 .023168 26.1870[.000] PGV .43224 .022037 19.6140[.000] PGVI .19274 .027183 7.0905[.000] FfEl'LE -.0070908 .016578 -.42772[.669] BA -.079867 .028467 -2.8056[.005J MAR -.037506 .019666 -1.9072[.057] OTHER -.OZ6929 .0274Z4 -.98194[.326] YRX .010068 .9327E-3 10.7940[.000] 10C -.00149.52 .0010301l -1.4515[.147] R-Squared .79405 R-Bar-Squared .78813 S.E. of Regression .18243 F-stat. F( 27, 938) 133.9476[.000] Mean of Dependent Variable 10.8492 S.D. of Dependent Variable .39633 Residual St= of Squares 31.2178 Equation Log-likelihood 287.0455 Akaike Info. Criterion 259.0455 Schwarz Bayesian Criterion 190.8212 DW-statistic 1.7809 Diagnostic Tests * Test Statistics * LM Version * F Version * * * * A:Serial Correlation*CHSQ( 1)= 10.8398[.001]*F( 1, 937)- 10.6337[.001] * * * * B:Functional Form *CHSQ( 1)= 13.4229[.0002*F( 1, 937)= 13.2034t.000] * * * * C:Normsality *aHSQ( 2)= 410.4361E.000]* Not applicable * * * * D:Heteroscedasticity*C2SQ( 1)- 27.7345[.000]*F( 1, 964)- 28.4952[.000J A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 4 Regression la: White's Heteroscedasticity Adjusted Standard Errors Dependent variable is LSALARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.1879 2.2192 5.9426[.000] CIFOR .21082 .036466 5.7813[.000] CZDfMYT -.022714 .028027 -.81040[.418] CIP -.18669 .040245 -4.6389[.000] ICARDA -.29007 .027985 -10.3653E.000J IC1RMR .11221 .033373 3.3623[.001] ICRAF .071164 .028551 2.4925[.013] ICRISAT -.14346 .036722 -3.9065[.000] IPPRI .36590 .032888 11.1257[.000] IITA -.083306 .029390 -2.8345[.005] ILRI .096423 .030274 3.1850t.001] IPGRI .11283 .030670 3.6789t.000] IRRI -.0035772 .033274 -.10751[.914] ISNAR .079996 .046512 1.7199t.086] IWMI .13039 .031250 4.1724[.000] IWPfA -.14370 .042998 -3.3421t.001] PGI 1.0423 .036774 28.3429t.0003 PGII .74784 .035239 21.2218[.000] PGIII .72664 .034048 21.3418(.000] PGTV 60669 .028298 21.4390[.000] PGV .43224 .027474 15.7326t-000] PGVI .19274 .035Z56 5.4668E.000] FENALE -.0070908 .017507 -.40503[.686] BA -.079867 .029710 -2.6882t.007J MA -.037506 .023275 -1.6114[.107] OTHER -.026929 .034279 -.78557t.432] YRX .010068 .0010110 9.9585[.000] YJC -.0014952 .0011138 -1.3424t.180] 5 Regressħon Ib: Ordinary Least Squares Estimation Dependent variable is LSALARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.6197 2.0266 6.7203[.000] CIFOR .20614 .036650 5.6244[.000] CINMMT -.019189 .027596 -.69535t.487] CIP -.18094 .031116 -5.8150[.000] ICARDA -.27573 .028991 -9.5107[.000] ICLARM .10597 .042676 2.4831[.013] ICRAF .069269 .032676 2.1198[.034] ICRISAT -.13490 .034619 -3.8966f.000J IFPRI .36020 .029928 12.0355[.000] IITA -.080397 .027960 -2.8754[.004J ILRI .094306 .027584 3.4189[.001] IPGRI .11679 .035472 3.2925[.001] IRRI -.0012513 .027125 -.046132[.963] ISNAR .070847 .036776 1.9264[.054] INM .12833 .042842 2.9953[.003] NARDA -.12374 .041379 -2.9903[.003] PGI 1.0162 .033419 30.4074[.0001 PGII .72726 .027888 26.0778[.000] PGIII .70543 .028904 24.4060[.000] PGIV .58435 .023166 25.2242[.000] PGV .41578 _021921 18.9672[.000] PGVI .17802 .026915 6.6140[.000] BA -.090690 .028073 -3.2305[.001] MA -.045739 .019258 -2.3750[.018] OTHER -_034628 .026952 -1.2848[.199] YX .010793 .9253E-3 11.6636[.000] Y.xC -.0016934 .0010I50 -1.6683[.096] ONBPII -.065138 .01212.7 -5.3714[.000] R-Squared .80016 R-Bar-Squared .79441 S.E. of Regression .17971 F-stat. F) 27, 938) 139.1023[.000J Mean of Dependent Variable 10.8492 S.D. of Dependent Variable .39633 Residual Sum of Squares 30.2922 Equation Log-likelihood 301.5838 Akaike Info. Criterion 273.5838 Schwarz Bayesian Criterion 205.3595 DW-statistic 1.8017 Diagnostic Tests * Test Statistics * LM Version * F Version * * * * A:Serial Correlation*CHSQ( 1)= 8.7267[.003]*F( 1, 937)= 8.5419[.004J * * * * B:Functional Form *CHSQ( 1)= 20.0713[.000]*F( 1, 937)= 19.8818[.000] * * * * C:Nor"ality *CESQ( 2)= 378.3511[.000)* Not applicable * * * * D:Heteroscedasticity*CHSQ( I)- 23.2395[.000]*F( 1, 964)= 23.7631(.000) A:Lagrange nfltiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 6 Regression Ib: W4hite's Heteroscedasticity Adjusted Standard Errors Dependent variable is LSALARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.6197 2.2196 6.1362[.0001 CIFOR .20614 .034602 5.9574[.000] CInMYT -.019189 .027728 -.69206[.489) CIP -.18094 .039104 -4.6272[.000] ICARDA -.Z7573 .027427 -10.0530[.000] ICLARM .10597 .033926 3.1235[.002] ICRAF .069269 .027941 2.4791[.013] ICRISAT -.13490 .036468 -3.6990[.000] IFPRI .36020 .032800 10.9819t.0001 IITA -.080397 .029053 -2.7672[.006) ILRI .094306 .029499 3.1969[.001] IPGRI .11679 .029816 3.9171[.000) IRRI -.0012513 .032059 -.039033[.969] ISNAR .070847 .045376 1.5613[.119] IWMI _12833 .029927 4.2880[.000] VWPRDA -.12374 .042825 -2.8894[.004] PGI 1.0162 .036560 27.7950[.000] PGII .72726 .034656 20.9849[.000) PGIII .70543 .033645 20.9668[.000] PGIV .58435 .027471 21.2714[.000] PGV .4157T .026675 15.5865[.000] PGVI .17802 .034243 5.1985[.000] BA -_090690 .029640 -3.0597[.002] MAX -.045739 .022628 -2.0213[.044] OTHER -.034628 .034284 -1.0100[.313] YRX .010793 .0010306 10.4719[.000] YJC -_0016934 .0011131 -1.5213[.129) OWBPII -.065138 .012319 -5.2878[.000] ************************************************************ ***************** 7 Regression la: Ordinary Least Squares Estimation Dependent variable is LSALARY 546 observations used for estimation from 1 to 546 Regressor Coefficient Standard Error T-RatioEProb] CONST 8.6246 2.6149 3.2983[.001] CIFOR .19921 .041365 4.8159[.000] CnMoa' -.028992 .032775 -.88457[.377J CIP -.12762 .038045 -3.3544[.001] ICARDA -.27078 .038862 -6.9679[.000] ICLARM .093251 .046952 1.9861[.048] ICRAF .067898 .037809 1.7958[.073] ICRISAT -.13757 .043421 -3.1683[.002] IFPRI .31670 .032275 9.8125[.000] IITA -.11210 .033963 -3.3007[.001] ILRI .080222 .031999 2.5070[.012] IPGRI .10861 .040178 2.7031[.007] IRRI .060774 .032282 1.8826[.060] ISNAR .015895 .041466 .38333[.702] i1Wc .11251 .045940 2.4490[.015] EZAR1A -.14666 .056684 -2.5873[.010] PGI .88646 .040026 22.1473[.000] PGII .61407 .035523 17.2862[.000J PGIII .58655 .036169 16.2172[.000] PGIV .48058 .029628 16.2203[.000] PGV .30302 .028590 10.5988[.000] PGVI .096927 .033961 2.8541[.0043 BA -.10839 .030502 -3.5535[.000] MAk -.031126 .021135 -1.4727[.141] OTHER -.025267 .033180 -.76151[.447] YRX .010965 .0011296 9.7068[.000] YJC .8638E-3 .0013105 .65912[.510] OWBPII -.045797 .031485 -1.4546[.146] R-Sguared .80037 R-Bar-Squared .78996 S.E. of Regression .16165 F-stat. F( 27, 518) 76.9183[.000] Mean of Dependent Variable 10.9095 S.D. of Dependent Variable .35271 Residual Sum of Squares 13.5353 Equation Log-likelihood 234.6280 Akaike Info. Criterion 206.6280 Schwarz Bayesian Criterion 146.3913 DW-statistic 1.9712 Diagnostic Tests * Test Statistics * LM Version * F Version * * * * A:Serial Correlation*CHSQ( 1)= .098345[.754]*F( 1, 517)= .093139[.760] * * * * B:Functional For *CHSQ( 1)= 1.8705[.171)*F( 1, 517)= 1.7772[.183] * * * * C:Normality *CHSQ( 2)- 445.7506[.000]* Not applicable * * * * D:Heteroscedasticity*CHSQc 1)= 5.9669[.015]*Ft 1, 544)- 6.0107[.015) A:Lagrange multiplier test of residual serial correlation B:Ransey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 8 Rgressħon lc: Mbite's Heteroscedasticity Adjusted Standard Errors Dependent variable is LSALARY 546 observations used for estimation from 1 to 546 Regressor Coefficient Standard Error T-Ratio [Prob] CONST 8.6246 2.8344 3.0428t.0023 CIFOR .19921 .038322 5.1983[.000] CINMYT -.028992 .032219 -.89986t.369] CIP -.12762 .049607 -2.5726[.010] ICARDA -.27078 .035178 -7.6976[.000] ICLARM .093251 .038408 2.4279[.016] ICRAF .067898 .032379 2.0970[.036J ICRISAT -.13757 .048057 -2.8627[.004] IFPRI .31670 .036948 8.5715t.000] IITA -.11210 .034350 -3.2634[.001] ILRI .080222 .033810 2.3727[.018] IPGRI .10861 .036803 2.9510[.003] IRRI .060774 .039965 1.52071.129] ISNAR .015895 .048686 .32648t 744] IWI .11251 .032860 3.4239[.001] VM,RDA -.14666 .059722 -2.4557[.014] PGI .88646 .045754 19.3743[.000] PGII .61407 .045180 13.5915[.000] PGIII .S865 .043706 13.4205tv000] PGIV .48058 .036738 13.0814[.000] PGV .30302 .036258 8.3573[.000] PGVI .096927 .045532 2.1288r.034] BA -.10839 .032732 -3.3114[.001] MP. -.031126 .023086 -1.3483[.178] OTHER -.025267 .044357 -.56964[.569] YRX .010965 .0011967 9.1626[.000] YJC .&S38E-3 .0014235 .60679[.544] ONBPII -.045797 .035174 -1.3020[.193] 9 Regression Id: Ordinary Least Squares Estimation Dependent variable is LSALARY 450 observations used for estimation from 517 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 20.3065 2.9616 6.8567[.000] CIFOR .20171 .062638 3.2203[.001] CIMMYT -.020521 .042274 -.48543[.628] CIP -.24483 .047963 -5.1045[.000] ICARDA -.27649 .042480 -6.5087[.000] ICLARM .14250 .078322 1.8195[.070] ICRAF .068980 .057132 1.2074[.228] ICRISAT -.15498 .052391 -2.9581[.003] IFPRI .47961 .051069 9.3913[.000] IITA -.055597 .046324 -1.2002[.231] ILRI .10186 .046811 2.1760[.030] IPGRI .14982 .055701 2.6898[.007] IRRI -.039608 .043837 -.90352[.367] ISNAR .17827 .062626 2.8466[.005] IWWI -11175 .075930 1.4718[.142] WARDA -.11623 .057769 -2.0120[.045] PGI 1.1380 .052631 21.6230[.000] PGII .79718 .042214 18.8841[.000] PGIII .78946 _044068 17.9147[.000] PGIV .64108 .035019 18.3067[-000] PGV .49348 .03Z089 15.3783[.000] PGVI .19157 .041240 - 4.6452[.000J BA -.0952.81 .047221 -2.0178[.044] MA -.10243 .034816 -2.9420[.003] OTEER -.038726 .04Z654 -.90793[.364] iRX .010759 .0014018 7.6751[.000] YJC -.0051211 .0014828 -3.4537[.001] OWR .042039 .021755 1.9324[.054] R-Squared .82934 R-Bar-Squared .81842 S.E. of Regression .18689 F-stat. F( 27, 422) 75.9520[.000] Mean of Dependent Variable 10.7802 S.D. of Dependent Variable .43859 Residual Sum of Squares 14.7401 Equation Log-likelihood 130.6791 Akaike Info. Criterion 102.6791 Schwarz Bayesian Criterion 45.1496 DW-statistic 2.0136 Diagnostic Tests * Test Statistics * LM Version * F Version * * * * A:Serial Correlation*COSQ( 1)- .024945[.875]*F( 1, 421)1 .023338[.879] * * * * B:Functional Form *CHSQ( 11= 2.7752E. 0961*F( 1, 421)= 2.6124[.107] * * * * C:Normality *CHSQ( 2)- 107.9760[.000]* Not applicable * * * * D:Heteroscedasticity*CHSQ( 1)= 17.8766[.000]*F( 1, 448)= 18.5334[.000] A:Lagrange mnltiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 10 Rfegression 1d: White's Heteroscedasticity Adjusted Standard Errors Dependent variable is LSALARY 450 observations used for estimation from 517 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 20.3065 3.3848 5.9994[.000] CIFOR .20171 .059405 3.3956[.001) CIMMYT -.OZ0521 .047054 -.43612[.663] CIP -.24483 .061367 -3.9895[.000) ICAFRDA -.27649 .044087 -6.2714[.000] ICLPJPM .14250 .071583 1.9907[.047) ICRAF .068980 .049208 1.4018[.162] ICRISAT -.15498 .059814 -2.5910[.010) IPPRI .47961 .049768 9.6369[.000] IITA -.055597 .050992 -1.0903[.276] ILRI .10186 .055411 1.8383t.067) IPGRI .14982 .051007 2.9373[.003) IRRI -_039608 .049582 -.79884[.425] ISNAR .17827 .082342 2.1650[.031) IWMI .11175 .062685 1.7828[.075] NAkRDA -.11623 .063800 -1.8218[.069] PGI 1.1380 .055352 20.5600[.000) PGII .79718 .051723 15.4125[.000] PGIII .78946 .049434 15.9700[.000) PGIV .64108 .039459 16.2468[.000) PGV .49348 .038420 12.8443[.000) PGVI .19157 .053766 3.5630[.000) BA -.095211 .058199 -1.6177r.106] MA -.10243 .044196 -2.3176[.021) OTHER -.0387Z6 .052Z38 -.74135r.459) YRX .010759 .0015361 7.0041r.000) YJC -.0051211 .0016965 -3.01861.003) OWR .042039 .023450 1.7928r.074] 11 Regression Il: Ordinary Least Squares Estimation Dependent variable is LSAIARY 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.9270 2.0788 6.6997[.000] CIFOR .21528 .037040 5.8120[.000] CIMMYT -.019196 .027888 -.68833[.491J CIP -.18359 .031425 -5.8420[.000] ICARDA -.28404 .029223 -9.7196[.000] ICLARM .11787 .043117 2,7336[.006] ICRAF .076945 .033084 2.3257[.020] ICRISAT -.13386 .035078 -3.8161[.000] IFPRI .34132 .031134 10.9628[.000] IITA -.072518 .028425 -2.5512[.011] ILRI .098219 .027874 3.5237[.000J IPGRI .12177 .035990 3.3835[.001] IRRI .0062537 .027533 .22713[.820] ISNAR .062301 .037518 1.6606[.097] flMI .13510 .043307 3.1196[.002J 1ARDA -.14297 .041661 -3.4318[.001] PGI 1.0316 .033770 30.5471[.000] PGII .74894 .028019 26.7299[.0001 PGIII .72107 .028998 24.8660t-000] PGIV .60565 .023054 26.2709[.000] PGV .43161 .021948 19.6646[.000] PGVI .19118 .027055 7.0663[.000] BA -.076858 .029683 -2.5893[.010] MA -.037158 .020770 -1.7891r.074] OTHER -.027511 .030076 -.91471[.361] YRX .010384 .9296E-3 11.1703[.000] YJC -.0018731 .0010398 -1.8015[.072] FCII .049332 .027186 1.8146[.070] FCIII -.00402.9S .025149 -.16024[.873] R-Squared .79634 R-Bar-Squared .79025 S.E. of Regression .18151 F-stat. F( 28, 937) 130.8474[.000] Mean of Dependent Variable 10.8492 S.D. of Dependent Variable .39633 Residual Sum of Squares 30.8718 Equation Log-likelihood 292.4294 Akaike Info. Criterion 263.4294 Schwarz Bayesian Criterion 192.7685 DW-statistic 1.7794 Diagnostic Tests * Test Statistics * LM Version * F Version * * * * A:Serial Correlation*CHSQ( 1)= 10.8362[.001]*F( 1, 936)= 10.6188[.001J * B:Functional Form *CHSQ( 1)= 14.6579[.000]*F( 1, 936)= 14.4215[.000] * * * * C!Normality *CHSQ( 2)= 395.5253[.000]* Not applicable * D:Heteroscedasticity*CHSQ( 1)= 27.0393[.000]*F( 1, 964)= 27.7604[.000] A:Lagrange nmltiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values 12 Regrezzion le: White's Heteroscedasticity Adjusted Standard Errorz Dependent variable is LSALARY 966 observations used for estimnation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 13.9270 2.2211 6.2702[.0003 CIFOR .21528 .036913 5.8320[.0001 CIMMYT -_019196 .027643 -.69443[.488] CIP -.18359 .039349 -4.6656[.000] ICARDA -.28404 .027741 -10.2390[.000] ICLARM .11787 .033759 3.4913[.001] ICRAF .076945 .028327 2.7163[.007] ICRISAT -.13386 .036141 -3.7039[.000] IFPRI .34132 .034312 9.9475(.000] IITA -.072518 .029156 -2.4872[.0133 ILRI _098219 .030131 3.2597[.001] IPGRI .12177 .030775 3.9568[.000] IRRI .0062537 .032388 .19309[.847] ISNAR .062301 .047708 1.3059[.192] 3iMI .13510 .030579 4.4181[.000] VUUMA -.14297 .042649 -3.3523[.001] PGI 1.0316 .036823 28.0141[.000] PGII .74894 .034889 21.4662[.000] PGIII _72107 .033830 21.3144[.000] PGIV .60565 .028002 21.6286[.000] PGV .43161 .027249 15.8395[.000] PGVI .19118 .034833 5.4884[.000] BA -_076858 .032289 -2.3803[.017] MA -.037158 .025574 -1.4530[.147] OTHER -.OZ7511 .036482 -.75410[.451] YRX .010384 .0010099 10.2819[.000] YJC -.0018731 .0011124 -1.6839[.093] FCII .049332 .029812 1.6548[.0981 FCIII -.0040298 .026468 -.15226[.879] 13 Regresslon 2aI: Logit Maximum Likelihood Estimation The estimation method converged after 8 iterations Dependent variable is PGI 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-RatioEProb] CONST -118.5138 49.7694 -2.3813[.017] CIFOR -.81911 .93769 -.87353[.383] CINMYT -2.5516 1.1383 -2.2416[.025] CIP -.69352 .77093 -.89959f.369] ICARDA -1.0578 .75911 -1.3935[.164] ICLARM .65015 .79348 .81937[.413] ICRAF -.33977 .77021 -.44114[.659] ICRISAT .67740 .64150 1.0560[.291] IFPRI .30559 .66332 .46069[.645] IITA -.58571 .71969 -.81383[.416] ILRI -.28550 .63791 -.44755[.655] IPGRI .011985 .73741 .016254[.987] IRRI -1.1698 .70365 -1.6625[.097] ISNAR -.91925 .87426 -1.0515[.2931 IRMI .23141 .83384 .27752[.781J WRRDA -.074248 .92089 -.080626[.936] BA -1.1403 1.0668 -1.0689[.285] NA 1.3754 .38839 3.5412[.000J OTHER .21900 .61736 .35474[.723] YRX .14548 .019658 7.4002r.000) YJC .057044 .024942 2.2871[.022] FEMALE -.62093 .52986 -1.1719r.242] Factor for the calculation of marginal effects = .026908 Maximized value of the log-likelihood function =-177.9264 Akaike Information Criterion --199.9264 Schwarz Bayesian Criterion =-253.5312 Hannan-Quinn Criterion --220.3342 Mean of PGI = .063147 Mean of fitted PGI - .011387 Goodness of fit = .93168 Pesaran-Tiwermann test statistic --556.5801[.000] Pseudo-R-Squared = .21802 14 Regression 2a}II: Logit Maximum Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIIIUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 17.7900 30.9636 .57455[.566] CIFOR .21427 .53233 .40252[.687J CIMMYT -.86894 .40077 -2.1682[.030] CIP .64128 .42444 1.5109[.131] ICARDA -.52134 .41465 -1.2573[.209] ICLARM .76438 .57324 1.3334[.183] ICRAF .51784 .43363 1.1942[.233] ICRISAT 1.9002 .47868 3.9697[.000] IFPRI .38090 .42959 .88666[.375] IITA .47053 .39175 1.2011[.230] ILRI -1.9788 .47057 -4.2052[.000] IPGRI -.39753 .50538 -.78660[.432] IRRI -2.1764 .46413 -4.6893[.000] ISNAR -.83378 .51136 -1.6305[.103] IWMI .77931 .59135 1.3179[.188] WRRDA -.40587 .63238 -.64181[.521] BA -1.5014 .55587 -2.7010[.007] NA .52782 .28373 1.8603[.063] OTSER -1.2629 .409-91 -3.0808[.002] YRX .16586 .013771 12.0445[.000] YJC -.010351 .015514 -.66720[.505] FE4IS -.67684 .27654 -2.4475[.015] Factor for the calculation of marginal effects = .17741 Maximized value of the log-likelihood function =-413.3415 Akaike Information Criterion =-435.3415 Schwarz Bayesian Criterion =-488.9463 Hannan-Quinn Criterion =-455.7492 Mean of PGIIIUP = .31470 Mean of fitted PGIIIUP = .25259 Goodness of fit = .80124 Pesaran-Tiaermann test statistic = -31.3948r.000] Pseudo-R-Squared = .31297 15 Regression 2aIV: Logit Maximum Likelihood Bstimation The estimation method converged after 6 iterations Dependent variable is PGIVUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -58.4693 32.9270 -1.7757[.076] CIFOR .86449 .49549 1.7447[.081] CI2YT .93729 .38830 2.4138[.016] CIP 2.6978 .51169 5.2722[.000] ICARDA .92544 .40690 2.2744[.023] ICLARM 1.2469 .58131 2.1449[.032] ICRAF 1.4813 .45816 3.2331[.001] ICRISAT 2.4513 .55116 4.4476[.000] IFPRI 1.2154 .40806 2.9786[.003] IITA 1.4995 .38218 3.9236[.000] ILRI -1.0687 .39877 -2.6800[.007] IPGRI 1.4997 .50265 2.9836[.003] IRRI -1.2433 .40198 -3.0930[.002] ISNAR .60350 .50693 1.1905[.234] ilea 1.8892 .62387 3.0283[.003] V4ARDA 1.0615 .55296 1.9196[.055] BA -1.9146 .46402 -4.1261[.000] MA .068707 .26922 .25520[.799] OTBER -1.0533 .36921 -2.8528[.004) YRX .19204 .014961 12.8363[.000] YJC .028045 .016487 1.7011[.089] FEMALE -.62133 .23188 -2.6795[.008] Factor for the calculation of marginal effects = .23912 Maximized value of the log-likelihood function =-435.5441 Akaike Information Criterion =-457.5441 Schwarz Bayesien Criterion --511.1489 Hannan-Quinn Criterion =-477.9518 Mean of PGIVUP - .56108 Mean of fitted PGIVUP = .58178 Goodness of fit = .79917 Pesaran-Timermann test statistic = -2.4756[.013] Pseudo-R-Squared = .34243 16 Regression 2aV: Logit Maximm Likelihood Estimation The estimation method converged after 7 iterations Dependent variable is PGVUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -69.2706 44.0490 -1.5726[.116] CIFOR .61622 .53133 1.1598[.246) CINMT 1.6042 .55272 2.9024[.004] CIP 1.9667 .61688 3.1882[.001) ICARDA 1.3500 .51781 2.6072[.009] ICLARM 1.7465 .85105 2.0522[.040J ICERAF 2.0137 .71011 2.8358[.005J ICRISAT 2.5184 .85452 2.9471[.003] IFPRI 1.2147 .45722 2.6567[.008] IITA .85376 .42231 2.0217[.043] ILRI .094289 .41471 .22736[.820J IPGRI 2.0670 .72801 2.8392[.005J IRRI -.77202 .39870 -1.9363[.053] ISNAR -1.2059 .55259 -2.1823[.0293 INMI 1.969S .84644 2.3268[.020J VPDA .34163 .59562 .57357[.566] BA -1.2031 .43566 -2.7616[.006] MA -.086253 .30954 -.27865[.781] OTHER -.89716 .45868 -1.9560[.051] YRX .23420 .021056 11.1225[.000) YJC .034072 .02204-0 1.5459[.122] FERLE -.40392 .24296 -1.6625[.097] Factor for the calculation of marginal effects = .078421 Maximized value of the log-likelihood function -320.4971 Akaike Information Criterion --342.4971 Schwarz Bayesian Criterion --396.1019 Hannan-Quinn Criterion =-362.9049 Mean of PGVUP = .79400 Mean of fitted PGVUP = .84990 Goodness of fit = .85921 Pesaran-Timmermann test statistic - 4.9315[.000] Pseudo-R-Squared = .34768 17 Regression 2bI: Logit Maximum Likelihood Estimation The estimation method converged after 8 iterations Dependent variable is PGI 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -111.0616 50.2141 -2.2118[.027] CIFOR -.90287 .93288 -.96783t.333] CIMMYT -2.4838 1.1332 -2.1919[.029] CIP -.67082 .76707 -.87452[.382] ICARDA -.96960 .76170 -1.2729[.203] ICIARM .63651 .79769 .79794[.425] ICRFA -.39174 .76939 -.50916[.611] ICRISAT .75295 .64031 1.1759[.240] IFPRI .15910 .66124 .24060[.810J IITA -.60690 .72707 -.83471[.404] ILRI -.29788 .63492 -.46917[.639] IPGRI -.10406 .74243 -.14016[.889] IRRI -1.1819 .70646 -1.6730[.0951 ISNAR -.95048 .87192 -1.0901[.276] INMI .28664 .83033 .34521[.730] WVDA .19001 .92606 .20518[.837] EA -1.2834 1.0722 -1.1970[.232] NAx 1.2826 .38193 3.3583[.001] OTHER .11160 .62226 .17934t.858] YRX .14959 .019659 7.6094[.000] YJC .053340 .025161 2.1199t.034) OWNBPII -.42605 .30985 -1.3750[.169] Factor for the calculation of marginal effects = .027134 Maximized value of the log-likelihood function =-177.7301 Akaike Information Criterion =-199.7301 Schwa.rz Bayesian Criterion --253.3349 Hannan-Quinn Criterion =-220.1378 Mean of PGI = .063147 Mean of fitted PGI = .010352 Goodness of fit = .93271 Pesaran-Timae,mann test statistic =-584.0120[.000] Pseudo-R-Squared .21888 18 Re4ression 2bIII: Logit Maximum Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIIIUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 23.0200 30.7635 .74829[.454] CIFOR .15376 .53144 .28932[.772] CIMMfYT -.86091 .39824 -2.1618[.031] CIP .63282 .42003 1.5066[.132] ICARDA -.41121 .41372 -.99392[.3211 ICLARM .79457 .56887 1.3968[.163] ICRAF .50967 .43136 1.1815[.238] ICRISAT 1.9427 .47676 4.0748[.0003 IFPRI .29271 .42953 .68147f.496J IITA .48074 .39006 1.2325[.218] I1RT -1.9698 .47032 -4.1883t.000] IPGRI -.43310 .50818 -.85225[.3943 IPRI -2.1498 .46447 -4.6285(.0003 ISNAR -.84070 .51184 -1.6425[.1013 flMI .83209 .58538 1.4215[.156] NARDA -.18030 .62947 -.28644[.775] BA -1.6466 .56291 -2.9252[.004] MA .39016 .27778 1.4045[.1601 OTHER -1.3661 .41015 -3.3306[.001J YRX .17117 .013856 12.35291.000] YJC -.012985 .015412 -.84250[.400] OhDPII -.30963 .18051 -1.7153[.087] Factor for the calculation of marginal effects 5 .17798 Maximized value of the log-likelihood function --415.0417 Akaike Infornation Criterion =-437.0417 Schwarz Bayesian Criterion --490.6465 Hannan-Quinn Criterion =-457.4495 Mean of PGIIIUP = .31470 Mean of fitted PGIIIUP = .25052 Goodness of fit = .80124 Pesaran-Tinm&rmann test statistic = -31.6199[E000] Pseudo-R-Squared = .31014 19 Regression 9bIV: Logit Maximum Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIVUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -55.9780 32.6444 -1.7148[.087] CIFOR .78708 .49482 1.5906[.112] CIMMYT .94014 .38711 2.4286[.015] CIP 2.6550 .50779 5.2284[.000] ICARDA 1.1095 .40942 2.7099[.007] ICLARM 1.2478 .57624 2.1653[.031] ICRAF 1.4646 .45922 3.1894[.001] ICRISAT 2.5362 .55430 4.5755[.000] IFPRI 1.1155 .40816 2.7331[.006] IITA 1.5228 .38227 3.9836[.000] ILRI -1.0791 .40077 -2.6927[.007] IPGRI 1.5185 .5102? 2.9759[.003] IRRI -1.1842 .40384 -2.9323[.003] ISNAR .57837 .51264 1.1282[.260] IRMI 1.8678 .61569 3.0337[.002] 1VRDA 1.3196 .55052 2.3970[.017J BA -2.0563 .47115 -4.3644[.000] MA -.097056 .26663 -.36400[.716] OTHER -1.1966 .36971 -3.2366[.001] YRX .20115 .015153 13.2746[.000] YJC .026835 .016344 1.6419[.101] OWBPII -.62424 .17519 -3.5632[.000] Factor for the calculation of marginal effects = .23914 Maximized value of the log-likelihood function =-432.7091 Akaike Information Criterion =-454.7091 Schwarz Bayesian Criterion s-508.3139 Hannan-Quinn Criterion =-475.1169 Mean of PGIVUP - .56108 Mean of fitted PGIVUP = .57453 Goodness of fit = .80228 Pesaran-Ti-mermann test statistic = -2.5450[.011] Pseudo-R-Squared = .34671 20 Regression 2bV: Logit Maximum Likelihood Estimation The estimation method converged after 7 iterations Dependent variable is PGVUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -68.5830 44.6069 -1.5375t.125] CIFOR .52219 .53262 .98041[.327] CIMMYT 1.6128 .55551 2.9034t.004] CIP 1.9781 .62235 3.1784[.002] ICARDA 1.6382 .52557 3.1170[.0023 ICLARM 1.6756 .83851 1.9982[.046] ICRAF 1.9785 .70713 2.7979[.005] ICRISAT 2.6336 .86526 3.0436[.002] IFPRI 1.1618 .46360 2.5061[.012] IITA .90942 .42592 2.1352[.033) ILRI .12785 .41820 .30571[.760] IPGRI 2.2161 .73014 3.0351(.002) IRRI -.68620 .40140 -1.7095[.088] ISNAR -1.2941 .56404 -2.2943[.0223 IWMI 1.8802 .84500 2.2251t.026) 1kPDA .58795 .59907 .98144[.327) BA -1.3784 .44121 -3.1240[.002) MP -.27563 .31145 -.88498[.3763 OTHER -1.0674 .45902 -2.3255[.020) YRX .24655 .021547 11.4424[.000) YJC .033835 .022318 1.5160(.130) OWBPII -.86700 .21051 -4.1186[.000] Factor for the calculation of marginal effects = .075533 Maximized value of the log-likelihood function =-313.1210 Akaike Information Criterion =-335.1210 Schwarz Bayesian Criterion --388.7258 Hannan-Quinn Criterion -355.5287 Mean of PGVUP = .79400 Mean of fitted PGVUP = .84576 Goodness of fit - .86542 Pesaran-TimmermAnn test statistic - 5.2461[.000J Pseudo-R-Squared = .36270 21 Regression 2cIII: Logit Maximm Likelihood Estimation The esti-mtion method converged after 6 iterations Dependent variable is PGIIIUP 546 observations used for estimation from 1 to 546 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 42.6111 47.2578 .90168[.368] CIFOR .69048 .67943 1.0163[.310] CIMMYT -1.3994 .59217 -2.3631[.018] CIP .98341 .57479 1.7109[.088J ICARDA -.18850 .62877 -.29978(.764] ICLARM 1.2131 .71888 1.6875[.092] ICRAF 1.1562 .57171 2.0224[.044] ICRISAT 2.4134 .67446 3.5782[.000] IFPRI .49683 .54167 .91722[.359J IITA 1.2634 .53607 2.3568[.019] ILRI -1.5457 .60432 -2.5578[.011] IPGRI .29169 .64626 .45135[.652] IRRI -1.3912 .61208 -2.2729[.023] ISNAR -.43455 .64225 -.67661[.499] IDMI 1.1619 .71617 1.6224[.105] WARDA .44707 1.0780 .41472[.679] BA -1.5213 .68310 -2.2270[.026] i$R .21223 .35551 .59696[.551J OTEER -1.6296 .55073 -2.9590r.0031 YRX .17685 .019220 9.2015[.000] YJC -.023032 .023691 -.97216[.331J OWBPII .21773 .58066 .37496[.708] Factor for the calculation of marginal effects 5 .17668 Maximized value of the log-likelihood function =-233.4834 Akaike Information Criterion =-255.4834 Schwarz Bayesian Criterion =-302.8122 Hannan-Quinn Criterion -273.9846 Mean of PGIIIUP = .31136 Mean of fitted PGIIIUP = .26007 Goodness of fit = .80952 Pesaran-Ti;mmezmnn test statistic = -23.0306[.000] Pseudo-R-Squared .31048 22 Regression 2cr7: Logit Maximu Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIVUP 546 observations used for estimation from 1 to 546 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -61.9158 46.7204 -1.3252[.186] CIFOR 1.0573 .61697 1.7137[.087] CIMMYT 1.2022 .51168 2.3496[.019] CIP 3.4461 .78684 4.3796[.000] ICARDA 1.4192 .61308 2.3148[.021] ICLAR.M 1.9063 .76750 2.4838[.013] ICRAF 1.4917 .58509 2.5495[.011J ICRISAT 2.2634 .70100 3.2288[.001] IFPRI .90826 .49956 1.8181[.070] IITA 2.2612 .53314 4.2412[.000) ILRI -1.0114 .52153 -1.9393[.053] IPGRI 1.4317 .63447 2.2566[.024] IRRI -.47918 .53163 -.90135[.3683 ISNAR .71579 .64049 1.1176[.264] IRMI 1.7835 .69622 2.5616[.011] VQRDA 2.0223 .84158 2.4030[.017] EA -1.6515 .56418 -2.9272[.004] MA. .064472 .32816 .19646[.844] OTBER -1.1907 .50156 -2.3740[.018] YRX .20063 .020487 9.7932t.000] YJC .029679 .023398 1.2685[.205] OWBPII -.87653 .57272 -1.5305(.127J ractor for the calculation of marginal effects = .23494 Maximized value of the log-likelihood function =-247.8721 Akaike Information Criterion --269.8721 Schwarz Bayesian Criterion =-317.2009 Hannan-Quinn Criterion =-288.3733 Mean of PGIVUP = .57326 Mean of fitted PGIVUP = .58974 Goodness of fit - .81502 Pesaran-Timmaermann test statistic = -1.16031.246] Pseudo-R-Squared = .33471 23 Regression 2dlII: Logit Maximum Likelihood Estimation The estimation method converged after 7 iterations Dependent variable is PGIIIUP 450 observations used for estimation from 517 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 14.8412 42.1759 .35189[.725] CIFOR -.38639 .89294 -.43271[.665] CINMYT -.53798 .55973 -.96113[.3371 CIP .29694 .63794 .46546[.642] ICARDA -.77172 .58048 -1.3295[.184] ICLARM .25768 .97248 .26497[.791] ICRAF -.27925 .72597 --38466[.701] ICRISAT 1.5024 .68066 2.2073[.028] IFPRI -.0057610 .71426 -.0080657[.994] IITA -.43779 .62860 -.69646[.487] ILRI -2.3219 .75410 -3.0790[.002] IPGRI -1.4100 .84068 -1.6772[.094] IRRI -3.1053 .75103 -4.1347[.000] ISNAR -1.1826 .89610 -1.3198[.188] IWMI .63764 .98682 .64615[.519] WARDA -.67729 .80420 -.84219[.400] BA -2.4770 1.0331 -2.3976[.017) NA .65599 .47491 1.3813[.168] OTHER -1.2542 .64003 -1.9597[.051] YRX .17004 .020404 8.3336[.000] YJC -.0088622 .021120 -.41962[.675] OWR .14626 .32398 .45146[.652] Factor for the calculation of marginal effects = .16992 Maximized value of the log-likelihood function =-184.4468 Akaike Information Criterion =-206.4468 Schwarz Bayesian Criterion --251.6485 Hannan-Quinn Criterion =-224.2625 Mean of PGIIIUP = .31778 Mean of fitted PGIIIUP = .28000 Goodness of fit - .81111 Pesaran-Timterm&nn test statistic - -19.0840[.000] Pseudo-R-Squared = .34438 24 Regression 2dIV: Logit Maxinn Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIVUP 450 observations used for estimation from 517 to 966 Regressor Coefficient Standard Error T-RatiofProb] CONST -42.4120 46.7289 -.90762[.365) CIFOR .35910 .85323 .42087[.674) CIMMYT .21287 .58018 .36690[.714] CIP 1.6476 .70523 2.3363[.020] ICARDA .49583 .58697 .84473[.399] ICLARM .054690 .97103 .056321[.955] ICRAF 1.2347 .81953 1.5066[.133] ICRISAT 2.7481 .94558 2.9062[.004J IFPRI 1.2956 .67018 1.9333[.054] IITA .31740 .61922 .51257[.609] ILRI -1.3752 .65736 -2.0920[.037] IPGRI 1.2185 .77353 1.5752[.1161 IRRI -2.2391 .62066 -3.6076[.000] ISNAR .50883 .85862 .59262[.554] IWMI 2.1489 1.3027 1.6496[.100] ARfDA .44491 .78004 .57037[.569] EA -3.1977 .86911 -3.6793[.000] MA -.39732 .46626 -.85214t.395] OTYHlR -1.5250 .5769S -2.6430[.009] YRX .20735 .022566 9.1888[.000] YJC .019930 .023389 .85213[.395] CWR .29602 .32844 .90127[.368] Factor for the calculation of marginal effects = .24467 Maximized value of the log-likelihood function --189.1197 Akaike Information Criterion =-211.1197 Schwarz Bayesian Criterion --256.3214 Hannan-Quinn Criterion =-228.9353 Mean of PGIVUP = .53556 Mean of fitted PGIVUP = .52667 Goodness of fit = .80889 Pesaran-Timermann test statistic - -2.8289[.005] Pseudo-R-Sguared = .39146 25 Regression 2el: Logit Maxim; Likelihood Estimation The estimation method converged after 8 iterations Dependent variable is PGI 966 observations used fox estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -70.6400 50.6268 -1.3953[.163) CIFOR -.87311 .96477 -.90499[.366] CIMMAYT -2.4970 1.1525 -2.1665[.031J CIP -.65656 .77367 -.84863[.396] ICARDA -.75604 .76528 -.98793[.323J ICLARM .90543 .78994 1.1462[.252] ICRAF -.16374 .77684 -.21078[.833] ICRISAT .98025 .65734 1.4912[.136] IFPRI -.53061 .68683 -.77255[.440] IITA -.39834 .74207 -.53679[.592] ILRI -.26662 .65581 -.40654[.684] IPGRI .27689 .76536 .36178[.718] IRRI -.96184 .72115 -1.3338[.183] ISNAR -1.5938 .91022 -1.7510[.080] IhMI .27849 .85887 .32425[.746] WARDA -.20608 .96116 -.21440[.830] BA -1.4763 1.1015 -1.3403[.180] MiA .91531 .45510 2.0112[.0451 OTHR -.45687 .72635 -.62898[.530] YRX .15674 .020609 7.6052[.000] YJC .033283 .025356 1.3126[.190] FCII .28786 .54433 .52883[.597] FCIII -1.3641 .54218 -2.5159[.012] Factor for the calculation of marginal effects = .022508 Maximized value of the log-likelihood function =-169.3822 Akaike Information Criterion --192.3822 Schwaez Bayesian Criterion =-248.4236 Hannan-Quinn Criterion =-213.7176 Mean of PGI = .063147 Mean of fitted PGI = .018634 Goodness of fit = .93271 Pesaran-Timmermann test statistic =-431.7517[.000] Pseudo-R-Squared = .25557 26 Regression 2elII: Logit Maxi-mn Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIIIUP 966 observations used for estimation from .1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST 39.5234 31.6375 1.2493[.212] CIFOR .18165 .53648 .33860[.735] CIMMYT -.83700 .39962 -2.0945[.036] CIP .62389 .42332 1.4738[.141] ICARDA -.38617 .41282 -.93544[.350] ICLAPM .83899 .56469 1.4857[.138] ICRAF .59938 .43556 1.3761[.169] ICRISAT 2.0235 .48159 4.2017[.000] IFPRI _045083 .44695 .10087[.920] IITA .56467 .39510 1.4292[.153J ILRI -1.8957 .46939 -4.0386[.000] IPGRI -.31808 .51351 -.61943[.5361 IPRI -2.0870 .46686 -4.4704[.000] ISNAR -.97735 .52724 -1.8537[.0643 IWMa .84958 .59381 1.4307[.153J vRmDA -.37006 .64205 -.57637[.565] BA -1.9077 .59706 -3.1952[.001] MA .16067 .30830 .52116[.6021 OTHER -1.7560 .46678 -3.7619[.000] IRX .16769 .013708 12.2328[.000] YJC -_020960 .015829 -1.3241[.186] FCII -.31816 .40015 -.79510[.427] FCIII -.86212 .37409 -2.3046[.021] Factor for the calculation of marginal effects = .17773 Maximized value of the log-likelihood function =-412.1751 Akaike Information Criterion =-435.1751 Schwarz Bayesian Criterion =-491.2165 Harnan-Quinn Criterion =-456.5105 Mean of PGII1UP = .31470 Mean of fitted PGIIIUP = .25569 Goodness of fit = .80228 Pesaren-Tirermnnn test statistic = -31.0213E.000] Pseudo-R-Squared = .31491 27 Regression 2eIV: Logit Maximum Likelihood Estimation The estimation method converged after 6 iterations Dependent variable is PGIVUP 966 observations used for estimation from I to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -42.9070 33.3656 -1.2860[.199] CIFOR .84468 .49728 1.6986[.090] CIMMYT .93213 .38636 2.4126[.016J CIP 2.6097 .50824 5.1349(.000] ICARDA 1.0163 .40469 2.51131.012] ICLARM 1.2656 .56836 2.2267[.026] ICRAF 1.5692 .46104 3.4035[.001] ICR5SAT 2.5759 .55526 4.6392[.000] IFPRI .99628 .42156 2.3633[.018] IITA 1.5634 .38347 4.0768[.000] ILRI -1.0031 .39696 -2.5269[.012] IPGRI 1.6254 .51022 3.1857[.001] IRRI -1.1716 .40514 -2.8917[.004] ISNAR .56763 .51427 1.1038[.270] IWMI 1.9302 .61813 3.1226[.002] WRRDA 1.2015 .56139 2.1403[.033] BA -2.1964 .48833 -4.4978[.000J MA -.25011 .28661 -.87266[.383] OTHER -1.5141 .40970 -3.6957[.000] YRX .19496 .014981 13.0131[.000] YJC .020531 .01S8&5 1.2305[.219] FCII -.48551 .39913 -1.2164[.224] FCIII -.82045 .37133 -2.2095[.027] Factor for the calculation of marginal effects = .23866 Maximized value of the log-likelihood function =-436.1273 Akaike Information Criterion -459.1273 Schwarz Bayesian Criterion --515.1687 Hannan-Quinn Criterion =-480.4627 Mean of PGIVUP = .56108 Mean of fitted PGIVUP - .56418 Goodness of fit - .79607 Pesaran-Timmermann test statistic = -2.9809[.003] Pseudo-R-Squared = .34155 28 Regression 2eV: Logit Maximu Likelihood Estimation The estimation method converged after 7 iterations Dependent variable is PGVUP 966 observations used for estimation from 1 to 966 Regressor Coefficient Standard Error T-Ratio[Prob] CONST -55.9165 44.6075 -1.2535[.210] CIFOR .63532 .54082 1.1747[.240] CIMMYT 1.5871 .55109 2.8800[.004] CIP 1.9291 .63594 3.0334[.002) ICARDA 1.3498 .51679 2.6118[.009] ICLARM 1.5800 .83490 1.89251.059] ICRAF Z.1056 .71361 2.9506[.003) ICRISAT 2.6184 .84175 3.1106[.002] IFPRI 1.08Z0 .47213 2.2917[.022] IITA .90800 .42593 2.1318[.033] ILRI .10421 .41444 .25145[.802] IPGRI 2.2708 .73804 3.0769[.002] IRRI -.71259 .40319 -1.7674[.077] ISNAR -1.2554 .56755 -2.2120[.027) rem 2.0343 .84528 2.4067[.016] iGMRA .52898 .60597 .87295[.383) EA -1.6383 .46278 -3.5401[.000] MA -.41814 .32721 -1.2779[.202] 0XER -1_5975 .49241 -3.2442[.001] YRR .24101 .021517 11.2012[.000) YJC .OZ7953 .OZZ303 1.2533[.210] ECII -1.1729 .48295 -2.4286[.015) FCIII -1-3630 .45894 -2.9700r.003] Factor for the calculation of marginal effects = .075770 Maximized value of the log-likelihood function =-317.0417 Akaike Information Criterion -340.0417 Schwarz Bayesian Criterion =-396.0831 Hannan-Qu.inn Criterion =-361.3771 Mean of PGVWP - .79400 Mean of fitted PGVUP = .84886 Goodness of fit = .86853 Pesaran-Ti-nrmaxn test statistic = 5.4274[.000] Pseudo-R-Squared = .35472 29 Appendix 7 List of Acronyms CGLAR Consultative Group on Intemational Agriculural Research CGIAR Centers CIAT Centro Internacional de Agricuwura Tropical (Colombia) CIFOR Center for International Forestry Research (Indonesia) CLMMYT Centro Internacional de Mejoramiento de Maiz y Trigo (Mexico) CIP Centro lnternacional de la Papa (Peru) ICARDA International Center for Agricultural Research in the Dry Areas (Syria) ICLARM International Center for Living Aquatic Resources Management (Malaysia) ICRAF Tnternational Center for Research in Agroforestry (Kenya) ICRISAT International Crops Research Institute for the Semi-Arid Tropics (India) IFPRI International Food Policy Research Institute (USA) ElTA Intemational Institute of Tropical Agriculture (Nigeria) llUR International Livestock Research Institute (Kenya) IPGRI International Plant Genetics Resources Institute (Italy) IRRI International Rice Research Institute (Philippines) ISNAR International Service for National Agricultural Research (The Netherlands) 1WMI International Water Management Institute (Sri Lanka) WARDA West Africa Rice Development Association (Cote d'Ivoire)