Wage policy, wage structure and employment in the public sector of Senegal

DRAFT FOR STA.,-T USE ONLY CPD8541 WAGE POLICY, WAGE STRUCTURE AND EMPLOYMENT IN THE PUBLIC SECTOR OF SENEGAL Peter C. Bloch (Consultant) CPD Discussion Paper No. 1985-41 May 1985 CPD Discussion Papers report on work in progress and are circulated for Bank staff use to stimulate discussion and comment. The views and interpretations are those of the authors.  t A I  Wage Policy, Wage Structure and Employment in the Public Sector of Senegal ABSTRACT This study was motivated by the mutual interest of CPD and the West Africa Region in the social, economic and financial importance of public sector employment and wages in the Senegalese economy. Its objective was to describe and analyze the appropriateness of Senegal's pay and personnel poli- cies in both the civil service and the public enterprises, in the light of the crisis in public finance on one hand and the crisis of educated un- employed on the other. Most of the paper is concerned with the facts and institutions of the modern-sector labor market, especially as they affect and are affected by employment and wages in the civil service and (with much greater information limitations) the public enterprises. Civil service pay and personnel poli- cies are found to be excessively personalized at the hiring stage, and there- after excessively rigid, encouraging rent-seeking while discouraging initiative. Much of the rapid increase in the wage bill in recent years can be explained by changes inthe composition of civil service employment, on such dimensions as grade level, average educational attainment and average age. The policy conclusions are not sanguine. As is true in most Third World countries, Senegal's civil service is found to be too costly in the aggregate even though individual employees are underpaid relative to their coequals in the private sector. This paradox implies that a reduction in the rate of growth of the government wage bill via real wage declines will be hard to implement, and that only serious attention to the problem of numbers of employees will have any lasting effect.  -- -- - - | | --  NOTES AND DEFINITIONS 1. Unless otherwise stated, all wages and salaries are expressed in thousands of Francs CFA per month (US$ 1 = 450 F. CFA in 1984, about 200 in 1980). This is in keeping with common Senegalese parlance. v 2. Unless otherwise stated, wages and salaries are defined as total compensation (Masse Salariale) considered by firm or government accountants as such, gross of withholding taxes. 3. Unless otherwise stated, employment figures are those relevant to the analysis under way[ e.g. when age distributions are shown, the total is the total number whose age is knownt when pay distributions are shown, the total is the total number who have been paid (not the total number on the payroll). Thus total employment may seem to vary from table to tablet this explanation should cover most such discrepancies. 4. Translations from French are unofficial. 5. The principal sources of data on employment and wage payments in the civil service and public enterprises are listed here and thereinafter referred to by their acronyms in the text, a) DTAI -- Direction de Traitement Automatique de l'Information, the Ministry of Finance computer center. b) CEP -- Centre des Etablissements Publics, the accounting agency in the Ministry of Finance responsible for overview of public enterprise operations c) SCS -- Service Central de la Solde, the payroll office of the Ministry of Finance d) BULGE -- Monthly computer report of itemized payroll. ,  I I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  Contents Paqe I. Employment and Wages in Senegal 1 A. Introduction 1 B. Wage Policy 4 C. Wage Differentials 8 I II. Public Sector Pay Policy 13 A. The Civil Service 13 B. The Parastatal Sector 19 III. Composition of the Pay Package 25 A. Initial Observations 25 B. Differences among Types of Employees -- Civil Service 29 1. Distribution by rank 29 2. Distribution by contract type 33 3. Distribution by function of ministry 35 4. Distribution by age 37 5. Types of allowances 39 6. Housing provided to civil servants 40 7. Social Payments 41 C. Public Enterprises 43 1. Distribution by age 46 2. Distribution by contract status 46 3. Allowances received by public enterprise employees 48 IV. Decomposition of Changes in the Public Sector Wage Bill 51 A. The Civil Service 52 B. Public Enterprises 60 V. Conclusions on Feasible Means of Managing the Wage Bill 63 Bibliography 67  LIST OF TABLES 1. Urban Labor Force, 1960-1980 2 2. Recent Wage Level and Structure Changes 5 3. Indices of Prices and Incomes 7 4. Employment and Wages in the Modern Sector, by Occupation and Race 9 5. Structure of Minimum Wages, Commerce Collective Agreement 10 6. Employment and Wages by Industry, Modern Non-Government Sector, 1974-83 12 7. Index of Base Pay of Selected Civil Service Occupations 15 8. Distribution of Civil Servants' Pay by Hierarchy 16 9. Career Ladders in the Civil Service 18 10. Composition of Pay Package, Civil Service and Public Enterprises 27 11. Distribution of Earnings of Civilian Civil Servants, 1971-84 30 12. Distribution of Earnings of Public Enterprise Employees, 1972-84 31 13. Mean Earnings of Civil Servants by Hierarchy 32 14. Employment and Wage by Sex and Contract Status, Civil Service 34 15. Employment and Wages by Ministry Function 36 16. Distribution of Employees by Age, Civil Service 38 17. Principal Allowances Received by Civil Servants 39 18. Public Enterprise Employment, 1971-84 44 19. Average Pay in the Public Enterprises 45 20. Distribution of Employees by Age, Public Enterprises 47 21. Distribution of Employees by Contract Status, Public Enterprises 47 22. Principal Allowances Received by Employees of Public Enterprises 49 23. Decomposition of Change in Civil Service Wages by Hierarchy 53 24. Decomposition of Change in Civil Service Wages by Contract Status 55 25. Decomposition of Change in Civil Service Wages by Ministry Function 56 26. Decomposition of Change in Civil Service Wages by Age 57 27. Summary of Civil Service Decomposition Exercises 59 28. Decomposition of Change in Public Enterprise Base Pay by Contract Status 61 29. Decomposition of Change in Public Enterprise Wage by Age 62  I. EMPLOYMENT AND WAGES IN SENEGAL A. Introduction The Senegalese economy has not been very successful at generating employment. The labor force participation rate has fallen since Independence, from 45.7% in 1960 to 43.3% in 1980 overall, and from 31.4% to 29.9% in the urban sector (Situation Economique, 1959-79). As Table 1 shows, urban unemployment has risen substantially, so that the fall in the employment rate (employed / urban population) has been even greater, from 29.3% to 25.0%. There are several explanations for this: first, the labor-absorption or -retention capacity of a rural sector increasingly vulnerable to drought has diminished, pushing people to migrate to the urban areas, and especially Dakar; second, the agricultural-surplus-based urban economy has not been growing rapidly, and even if it had the very low employment elasticity of modern-sector output would have kept increases in employment low; third, and of course related to the second, the set of institutions and policies built up over the past forty years to regulate factor markets and factor prices has had negative consequences on labor absorption, both directly through high costs of labor relative to capital, and indirectly through high overall production costs relative to the rest of the world at an exchange rate until recently artificially held above equilibrium by the tight linkage to the French Franc. The modern sector has never represented a large share of total employment. In 1960, it had 88,000 jobs, roughly equally split between private and public sectors, about one-third of the urban labor force and five per cent of the country's economically active population. By 1980, the modern sector still represented one-third of the urban labor force (which had more than doubled), and its share of the total had risen to eight per cent, as the urban sector  TABLE 1 URBAN LABOR FORCE, 1960 - 1980 percentage of number of workers labor force SECTOR 1960 1970 1976 1980 1960 1980 Agriculture & Fishing 1,507 4,911 5,713 85514 0.6 1.5 Mining 1,310 2,150 3,278 5,039 0.5 0.9 Manufacturing 11,602 15,064 20,872 23,647 4.6 4.1 Utilities 2,138 3,052 4,205 5,726 0.8 1.0 Construction 7,050 9,200 10,100 11,723 2.8 2.1 Commerce 6,200 10,001 15,396 18,599 2.5 3.3 Transport & Communic'ns. 3,917 6,100 15,207 17,003 1.6 3.0 Financial Services 2,316 2,102 3,044 4,208 0.9 0.7 Other Services _ 6,169 _ 8,031 9,165 12.705 2.4 2.2 TOTAL PRODUCTIVE SECTOR 42,209 60,611 86,980 107,164 16.7 18.8 PUBLIC SECTOR* 46,106 61,836 71,742 88,390 18.3 15.5 TOTAL MODERN SECTOR 88,315 122,447 158,722 195,554 35.0 34.3 INFORMAL SECTOR** 146,990 205,216 241,091 275,695 58.2 48.4 OPEN UNEMPLOYED 17,226 41,153 69,780 98,878 6.8 17.3 TOTAL URBAN LABOR FORCE 252,531 368,816 469,592 570,127 100. 100. * Civil Service and Non-Commercial Public Enterprises. ** Presumably calculated as a residualL source gives no methodology. Source. Direction de la Statistique, Situation Economigue 1959-1979.  3 rose from 16 to 23 per cent of the economically active population. Belying the conventional wisdom that the weight of the public sector has increased over time, public employment grew by "only" 3.3 per cent per year from 1960-1980, compared to 4.7 per cent per year for the "productive" sector, according to the official data based essentially on the 1976 population census. Thus the former's share of the urban sector fell from 18.3 to 15.5 per cent, and that of the latter rose from 16.7 to 18.8 per cent. The public sector has, however, grown more rapidly than the country can afford, given its stagnant ability to pay: in 1960 there were about 35 members of the total labor force to generate tax revenue to support each public employee, and only 28 in 1980. Since per capita GDP has been constant or falling, and since the cost per public employee has remained approximately the same, the burden of the public sector on the rest of the economy has been growing heavier. Simultaneously, however, the government has felt an obligation to absorb as much labor as possible in order to minimize the unemployment problem which has been emerging, most noticeably among secondary and university graduates. The public finance crisis on one hand, and the labor market crisis on the other, have forced Senegal into an untenable position for which nobody has thus far discovered feasible solutions. This paper attempts to address the facts and the consequences of this dilemma, focusing on employment and pay policy in the public sector. Both the administration and parapublic enterprises are scrutinized, with the goal of providing answers to the following issues: 1) The determinants of the level and the structure of pay 2) The composition of individual pay packages 3) Actual pay policy in both parts of the public sector, and consistency of the policy with appropriate fiscal and employment practices 4) Budgetary implications of alternative pay levels and structures, and the wage-level implications of alternative employment structures.  4 B. Wage Policy The Senegalese government has had enormous influence over the level and structure of wages in the modern sector since well before Independence. The French colonial administration established a minimum wage law in 1925, and instituted a set of labor laws comparable to those instituted in the metropole as a result of the progressive Front Populaire government of the 1930's. In deference to the politically strong trade union movement, the post-Independence government did not try to eliminate or reduce any of the significant advantages previously granted to workers in the private sector, such as pension, job security and vacation provisions as generous as many in the industrial countries. The government influences not only the lowest wage the private sector is allowed to pay, but the entire scale as well. The tripartite system of collective negotiations (trade union central, management and government) which supposedly decides on the amount by which all salaries will change is in fact unbalanced, and government has generally played the determining role (Bloch, 1978). Since the onset of the global inflation of the early 1970's, the government's wage policy has been constrained by the necessity to adjust producer and consumer prices of essential commodities to reflect changing realities in the world market. Wage changes have usually accompanied changes in consumer prices of rice, sugar, cooking oil and similar necessities which the government controls through sales or import monopolies. These basic goods constitute a larger proportion of the incomes of poorer families, and government has felt it necessary to cushion the blow of "price reality" (verite des prix) for those who can afford it least. Table 2 shows the history of changes in legal minimum wages and wage scales since 1974. In the 1974 price and wage restructuring, still the most comprehensive in Senegalese history, for example, the minimum wage was raised  TABLE 2 RECENT WAGE LEVEL AND STRUCTURE CHANGES (wage levels in Francs CPA per month) November 1974 January 1980 April 1983 Range % Increase Range % Increase July 1981 July 1982 Range % Increase July 1983 January 1985 Under 12,000 60 Under 20,000 25 General 5% General Under 27,000 15 Civil General 12 - 15,000 57 20 - 22,500 24 increase in increase 27 - 40,000 14 service increase 15 - 20,000 46 22 - 25,000 23 all wages of 2,000F 40 - 50,000 12 2,000 F of 1,515 20 - 25,000 34 25 - 30,000 22 (civil per month 50 - 60,000 10 per month F /Month 25 - 30,000 27 30 - 35,000 21 service 60 - 75,000 8 increase (equivalent 30 - 35,000 23 35 - 40,000 20 excluded) Over 75,000 6 to 5% of SMIG) 35 - 40,000 20 40 - 50,000 19 40 - 60,000 17 50 - 60,000 17 60 - 100,000 10 60 - 65,000 16 100 - 150,000 6 Over 65,000 15 Over 150,000 3 SOURCE: JO, various issues[ Chamber of Commerce.  6 by 60 per cent; highest wages by only 3 per cent; and rice, sugar and oil prices by 67, 67, and 60 per cent, respectively. The most recent increase, in January 1985, saw a 5% rise in the SMIG -- 1,515 francs per month -- joined to a 3,000 francs per sack (30 francs per kilogram) rise in the price of rice. Since the archetypal consumption of rice is considered to be one sack per household per month, the man in the street was heard to be saying that the net result of the wage and price changes was a fall in real family income of about 1,500 francs per month. It is evident that wage changes have come infrequently -- except in 1973-74 and 1980-83 -- and tend to be quite substantial. The resulting real wage pattern is anything but smooth, as Table 3 shows. The pattern has been for wage changes to lag behind general price trends, with both controlled prices and wages catching up discretely (the latter, increasingly unsuccessfully). The macroeconomic effects of this pattern cannot be predicted with any confidence in an economy as imperfectly understood as Senegal's, but surely cannot be salutary. Wage drift in the private sector in between adjustments is likely to be higher because of the jerky periodicity of the adjustments, and therefore the wage contribution (if any) to inflationary pressures in the economy is likely to be greater than if the pattern were smoother, more predictable. Overall, mean modern-sector wages have fallen secularly, both in real terms and relative to national per capita income. This suggests noticeable declines in real wages among higher-skill, higher-paid employees, unless wage drift has been more rapid in that category. Unfortunately, since the most 1. Salaire Minimum Interprofessionnel Garanti -- Guaranteed (interprofessional) minimum wage. 2. defined as unlegislated movement in wages actually paid relative to legislated minimum rates.  7 TABLE 3 INDICES OF PRICES AND INCOMES 1980 = 100 YEAR CPI-A CPI-E WPI REAL PCY REAL SMIG 1 2 3 4 5 1984 154.50* 150.40* - 101.10 84.50 1983 135.90* 137.80* - 108.50 93.00 1982 124.30 125.30 121.90 107.10 87.90 1981 105.90 111.30 109.30 96.30 96.80 1980 100.00 100.00 100.00 100.00 100.00 1979 92.00 88.50 83.90 106.30 86.90 1978 83.90 81.60 74.70 99.70 95.30 1977 81.10 76.00 66.90 109.30 98.60 1976 72.90 71.40 61.20 107.50 109.70 1975 72.10 67.30 59.50 105.30 110.90 1974 _ 54.70 55.40 47.00 101.00 99.50 1973 46.90 48.10 35.60 99.50 85.60 1972 42.20 45.40 33.30 110.50 89.60 1971 41.10 42.90 33.00 104.20 92.00 1970 39.50 40.40 33.00 106.70 95.70 1969 37.20 38.90 31.10 100.60 101.60 1968 35.80 37.30 30.10 110.10 99.20 1967 35.80 37.40 29.30 106.00 91.40 1966 34.9** 36.50 28.80 109.60 93.70 1965 34.2** 35.70 28.30 109.10 95.60 1. African Consumer Price Index, 1967 base year. 2. European Consumer Price Index, 1961 base year. 3. Construction Price Index, 1972 base year. 4. Real GDP per capita, deflated by column 1. 5. Real value of minimum wage, averaged over calendar year deflated by column 1. * Estimated on the basis of incomplete data from Direction de la Statistique. ** These two years' data do not exist for CPI-A. The entries are those of CPI-E, adjusted to the 1967 base year and then converted to 1980=100. SOURCE: 1 - 3) Direction de la Statistique. 4) Direction de la Statistique and CEM 1984. 5) JO and IMF, unpublished data.  8 recent reliable information on wage distribution by skill is from 1977, it is not possible to shed much light on this possibility for the modern private sector. The only quantitative indication that this phenomenon has occurred is based on Table 4: African skilled workers, foremen, technicians and managers all experienced more rapid wage increases between 1976 and 1977 -- a period of rapid price inflation and no adjustments in legal minima -- than did ordinary blue and white collar workers. This one piece of evidence is, however, insufficient to draw any general conclusions about secular trends. C. Wage Differentials The Senegalese economy is characterized by very wide wage and income differentials, both between sectors and among skill groups. The data presented in this Section suggest that the highly inegalitarian status Senegal has been awarded in most international rankings of income distribution (e.g. Adelman and Morris, 1973 and Jain, 1975), although based on little if any data, is probably deserved. Table 4, above, shows the wide occupational wage differentials that prevailed in the mid-'seventies. The average Non-African upper-level manager was paid thirty times the SMIG in 1977, and nearly fourteen times the average African semi-skilled or unskilled worker's wage. In the same Table, it is seen that the average Non-African was paid seven and one-half times as much as the average African, and that in each skill category the ratio ranged from 2.2:1 (high-level managers in 1977) to 4.2:1 (technicians and middle-level managers in 1976). Even African high-level managers earned thirteen times the SMIG, and more than six times as much as production workers, in 1977. By way of comparison, Table 5 shows the legal minima for a sampling of pay categories in the (quite representative) Commerce collective agreement: there has been a rapid compression of legal wage differentials, but we have no information about recent actual trends.  - ~ ~ ~ U TABLE 4 EMPLOYMENT AND WAGE STRUCTURE IN THE MODERN SECTOR (PRODUCTIVE) BY OCCUPATIONAL CATEGORY AND RACE E M P L O Y M E N T AVERAGE MONTHLY SALARY 1974 1975 1976 1977 1974 1975 1976 1977 African* 1,688 2,735 656 772 151.7 136.3 222.5 251.0 High Level managers Non-African 1,958 1,961 982 1,039 345.1 412.6 513.3 553.1 Total 1,638 1,811 396.8 424.3 Technicans and African (included 2,781 2,920 94.2 114.2 Middle Managers Non-African above) 907 959 398.1 402.2 Total 3,688 3,879 168.9 185.4 Skilled Workers African 57,506 60,620 12,788 12,144 33.8 42.6 65.4 74.4 and Foremen Non-African 1,052 892 611 542 166.7 184.6 216.1 257.8 Total 13,399 12,686 72.3 82.2 Semiskilled and African (included 50,460 51,348 36.6 40.1 Unskilled Workers Non-African above) 141 316 101.1 67.2 Total 50,601 51,664 36.8 40.3 African 59,194 63,355 66,685 67,184 37.2 46.6 46.4 52.0 TOTAL Non-African 3,010 2,853 2,641 2,856 282.7 341.3 383.0 390.0 Total 62,204 66,208 69,326 70,040 49.0 59.3 59.0 65.8 Percentage Increase African 7.0 5.3 0.7 25.3 0.0 12.1 Of Total from Non-African -5.2 -7.4 8.1 - 20.7 12.2 2.1 Previous Year Total 6.4 4.7 1.0 21.0 0.0 11.1 Ratio NA/A 7.6 7.3 8.3 7.5 * Africans (Senegalese and others). SOURCE: Direction de la Statistique, Banque de Donnes e'onomiques et financieres, 1980.  10 TABLE 5 STRUCTURE OF MINIMUM WAGES, COMERCE COLLECTIVE AGREEMENT: Ratio to SMIIG, 1974 - 1985 Class 4* Class 7A Class 9 Class 11* Pre Nov. 1974 1.86 3.18 4.76 8.65 11/74 - 1/80 1.55 2.38 3.48 5.73 1/80 - 7/81 1.47 2.27 3.21 5.27 7/81 - 7/82 1.47 2.27 3.21 5.07 7/82 - 4/83 1.44 2.17 3.05 4.76 4/83 - 1/85 1.43 2.08 2.81 4.57 1/85 - 1.41 2.03 2.72 4.39 SOURCE: JO various decrees; Chamber of Commerce. * Class 4: Skilled manual; Class 11: Top Category employee.  11 Intersectoral income differentials are large, too. Within the modern sector, the grain milling, paper and cardboard, and tobacco and matches industries all had-mean wages more than twice as high as the wood industry did, and half again as much as the average industrial wage, in 1983 (see Table 6). Although interindustry wage differentials are, of course, greatly influenced by skill composition of the labor force, it is certain that other, more subtle, factors enter as well: the bulk of grain milling is done by two companies which have a monopoly on wheat imports and thus a protected market position; similarly, tobacco and matches are produced by two companies, each with local monopolies; and paper and cardboard is dominated by one company which makes school notebooks for a captive market. The systematic research necessary to explain this and other aspects of the wage structure in Senegal's modern private sector has not yet been done; a recent attempt (Svejnar, 1984) made unimaginative use of the admittedly weak database. This paper's focus on the public sector precludes deeper investigation of this subject at this time. The modern sector's minimum and average incomes are high relative to those in the rest of the economy, although accurate data on the latter are even harder to obtain. The SMIG has fluctuated between two and two and one-half times per capita GDP since the nineteen-sixties, and in recent years has been approximately seven times as high as rural per capita income (over three times as high as average rural adults' incomes, assuming a 1:1 dependency ratio; for data see Plan (1982) -- VI Plan)). Thus the range of average incomes by race and occupational category, from expatriate managers to African farmers, is close to 100:1, and the range of extremes a multiple of that.  TABLE 6 EMPLOYMENT AND AVERAGE WAGES BY INDUSTRY, MODERN NON-GOVERNMENT SECTOR 1974 - 83 1 9 7 4 1 9 7 7 1 9 8 0 1 9 8 3 Average Average Average Average Monthly Monthly Monthly Monthly Employment Wage Employment Wage Employment Wage Employment Wage Fishing and Canning 3,784 57.2 3,715 48.0 6,478 48.6 3,567 103.7 Misc. Food Processing 2,312 43.5 3,866 48.9 3,649 78.9 3,236 106.8 Peanut Oil Mills 2,077 87.0 2,070 97.8 2,589 105.8 2,356 174.9 Grain and Flour 1,786 35.0 1,313 48.0 772 113.0 1,001 212.0 Sugar and Candy 2,294 47.8 2,224 43.0 4,876 57.9 3,844 114.1 Tobacco and Matches 503 62.6 546 80.7 557 87.4 464 182.5 Basic Textiles 3,387 51.4 3,968 53.9 1,949 73.0 4,020 123.3 Clothing and Shoes 1,713 44.3 2,545 57.7 983 67.1 905 155.4 Wood 698 47.0 936 47.7 328 69.7 303 84.2 Paper and Cardboard 226 58.3 275 63.6 169 177.8 216 191.0 Printing NA NA 402 83.7 302 93.3 821 125.2 Chemicals 1,528 77.6 1,695 93.2 1,413 99.1 2,420 150.5 Building Materials 715 65.4 829 73.2 588 99.7 782 171.5 Machinery 2,003 54.1 2,313 68.1 2,615 78.6 3,252 145.6 Total Manufacturing 23,026 55.2 26,716 59.2 27,268 72.4 27,197 136.0 Mining 1,779 70.1 1,773 92.6 1,526 92.7 1,971 175.0 Energy 2,465 78.6 2,702 76.5 2,764 63.0 3,315 186.2 Construction 6,634 42.4 7,156 66.5 8,787 75.9 6,303 160.6 Commerce NA NA 11,132 87.0 11,468 123.1 NA NA Services NA NA 20,561 62.9 10,869 101.6 21,115 128.7 Relevant Minimum Wage 12.6 18.6 23.2 29.3 Average SOURCE: Direction de la Statistique, various published (1974, 1977 and 1980) and unpublished (1983) documents.  13 II. PUBLIC SECTOR PAY POLICY The government-has generally changed the level and the structure of salaries in the civil service at the same time, or approximately the same time, as it has decreed changes in the minimum wage in the private sector. The result has been that the trend in real public sector pay has closely followed the trend in official minima for all categories in the private sector. For most of the recent period, this has meant that public sector salaries have been decreasing relative to private sector salaries of analogous employees, since the latter have probably been increasing relative to the minima. A. The Civil Service The civil service pay structure is rigidly defined. The broadest categories are called Hierarchies; there are five of these (A through E), differentiated by educational requirements: Hierarchy A - degree from University or 'Grande Ecole" -- Six sub-Hierarchies, with specific occupations assigned to each Hierarchy B -- Baccalaureat or equivalent final secondary-school diploma Four sub-Hierarchies Hierarchy C -- diploma from lower-secondary school Three sub-Hierarchies Hierarchy D -- diploma from primary school -- Three sub-Hierarchies Hierarchy E -- no diploma required. Single scale Within the Hierarchies there are four or five Classes; within the Classes are two to four Echelons. The steps in pay are defined by the "Grille Indiciaire," which assigns index points to each specified category of employee. From 1974 to 1980, the range of points was from 262 (for entry into Hierarchy E) to 6243  14 (for senior tenured university professors and the Chief Justice of the Supreme Court). In 1980, the structure was changed in two ways: first, the number of points for each Echelon was increased to reflect the 1974 progressive wage increase (which until then had been paid civil servants by means of a "solde d'attente,n or temporary pay), and second, the value of a point was increased by 7 per cent, from 37.25 to 39.86 francs per month. Since 1980, then, the range of points has been from 557 to 6579. To the resulting salary figure must be added 34 per cent in across-the-board "complements," so that base pay has ranged from 29,750 to 351,398 francs per month since 1980. Table 7 gives a sampling of the structure of base pay, both before and after the 1980 adjustment; it is evident that the 1980 adjustment has resulted in substantial compression of the civil service pay structure. Since 1980 there have been three wage increases. All of these have been across-the-board amounts: 2,000 francs per month for everyone in July 1982, 2,000 more in July 1983, and 1,515 more in January 1985. Thus the reduction in salary dispersion in the public sector, begun in 1974, is continuing up to the present. The resulting trend in distribution is shown in Table 8: within-Hierarchy as well as between-Hierarchy dispersion has fallen dramatically. The consequences for morale in the upper ranks are discussed below, in Section III.B. Promotion The regulations regarding career advancement are quite rigid,'and seniority alone is almost enough to raise individuals to the top category within their hierarchy within a determinable number of years. There is an automatic advancement at least every two years for civil servants (three years for the Echelons in the two highest Classes in Hierarchy A), and every two,  TABLE 7 INDEX OF BASE PAY OF SELECTED CIVIL SERVICE OCCUPATIONS PRE-1980 AND 1980 SCALES Ratio to Lowest Pay in Hierarchy E Pre - 1980 Scale 1980 Scale Pre - 1980 Scale 1980 Scale Representative Lowest Highest Lowest Highest Lowest Highest Lowest Highest Hierarchy Occupation Echelon Echelon Echelon Echelon Echelon Echelon Echelon Echelon A Research Director 1,878 3,802 2,215 4,054 7.1 14.5 4.0 7.3 A Veterinarian, Statistician 1,700 3,580 2,020 3,837 6.5 13.7 3.6 6.9 A Economist, Art Teacher 1,423 2,989 1,715 3,317 5.4 11.4 3.1 6.0 B Engineering Technician 1,283 2,806 1,568 3,124 4.9 10.7 2.8 5.6 B Nurse, Cartographer 821 1,765 1,140 2,092 3.1 6.7 2.0 3.8 C Animal Vaccinator 734 1,515 1,053 1,816 2.8 5.8 1.9 3.3 Rural Animator C Lighthouse Guard 560 1,010 894 1,331 2.1 3.9 1.6 2.4 D Postal Clerk, Guardian 399 891 684 1,204 1.5 3.4 1.2 2.2 E No Defined Occupation 262 477 558 789 1.0 1.8 1.0 1.4 SOURCE: Diarra (1982): Part 7. NOTES: To get base pay, multiply Pre-1980 Scale X 37.25 X 1.34. 1980 scale X 39.86 X 1.34 (see text). * e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~n  16 TABLE 8 DISTRIBUTION OF CIVIL SERVANTS' PAY BY HIERARCHY, 1976 - 1984 MONTHLY AVERAGE FOR 1976 Thousand F Per Month A B C D E Others Less than 25 69 61 185 509 1,471 2,673 25 - 40 11 72 851 5,119 1,806 3,121 40 - 60 19 1,307 6,236 6,813 107 2,698 60 - 80 96 2,144 2,217 292 5 905 80 - 100 377 1,811 314 26 1 516 100 - 120 216 946 25 9 0 400 120 - 150 275 629 14 1 0 134 150 + 468 106 2 0 0 183 Total Number 1,506 7,076 9,844 12,769 3,390 10,630 Average Pay 129.8 83.3 53.4 42.6 26.3 45.8 DECEMBER 1981 Thousand F Military & Per Month A B C D E Gendarmes Others Less than 25 2 5 6 24 10 2 94 25 - 40 2 3 13 31 1,347 2 1,362 40 - 60 13 849 3,773 1,881 9 2,222 60 - 80 2 2,060 4,237 2,857 176 91 1,542 80 - 100 223 3,042 3,773 1,391 41 1,099 947 100 - 125 705 1,965 3,730 647 22 1,080 425 125 - 150 253 2,705 1,750 168 9 999 265 150 + 1,777 4,130 1,672 307 30 2,841 520 Total Number 2,964 13,923 16,030 9,198 3,516 6,123 7,377 Average Pay 200.8 138.3 105.8 75.6 49.1 159.8 77.2 DECEMBER 1984 Thousand F Military & Per Month A B C D E Gendarmes Others Less than 25 2 6 9 14 6 2 14 25 - 39 1 3 3 11 236 59 40 - 59 1 7 107 2,613 2,343 1 2,737 60 - 79 1 655 3,818 3,088 168 27 1,810 80 - 99 33 3,765 3,069 1,752 83 694 1,090 100 - 124 588 3,200 5,345 1,013 33 1,057 923 125 - 149 375 3,861 2,802 353 17 1,275 244 150 + 2,641 5,969 2,225 306 18 4,341 439 Total Number 3,642 17,466 9,150 2,904 7,397 7,316 Average Pay 203.1 149.5 82 82 62.8 176.2 83.4 SOURCE: 1976: Plan (1978); 1981 and 1984: DTAI  17 three or four years for some contract employees. For promotion between Classes, an Evaluation Commission ranks candidates according to the annual evaluation given by their superiors, and each year, the eligible employees are promoted in the order in which they appear on the rankings.I Table 9 simulates the implications of this system for an individual's career expectations: finishing in the top half of candidates gives an incentive to effort, but only periodically, and not very frequently during the employee's career. An individual who succeeds in the most rapid promotion possible will reach the top pay category after between 20 and 24 years; if he or she entered at age 25, a typical time, by age 45 to 50 there would be no further expectation of wage increases, other than the general ones which, as shown above, have consistently been smaller for the higher ranks. For all but Hierarchy A, there is an additional means of advancement: competitive examinations for promotion into the next higher Hierarchy. The most common situation in which this happens is for entrance into training schools the graduates of which are automatically members of a given Hierarchy; in the 1984-85 school year only about 100 slots (of over 1,000) in professional training schools were allocated to already-employed civil servants who wish to move up (the rest were for outside candidates, usually recent general education graduates). This is probably the major incentive device available in the civil service; its limited extent, however, may serve mostly to discourage. Furthermore, there is no hope for those in Hierarchy A. 1. It is interesting that the Chapter of Law 61-33, of June 15, 1961, which deals with evaluation by superiors precedes the Chapter which deals with promotion. The former is dealt with in the latter only to the extent that the Evaluation Commission is supposed to "take account of" the grades given by the superiors. 2. Source: unpublished table furnished by Cellule de Controle de la Masse Salariale, Presidence de la Republique.  18 TABLE 9 CAREER LADDERS IN THE CIVIL SERVICE Hierarchy Hierarchy Hierarchy A B(l ) B(2), C, D, E NlEo~. or No. ot No. ot Echelons Years Echelons Years Echelons Years Bottom Class 2 4 2 4 4 8 Next Class 2 4 2 4 3 6 Next Class 2 4 2 4 3 6 Next Class 2 6 2 4 1 - Next Class 1 - 1 - - - Minimum Number of Times Evaluated 4 4 3 Minimum Years to Enter Top Class 18 16 20 Expected Years to Enter Top Class 22 20 24 Definitions: YEARS: Minimum years in Class before being eligible for promotion to next Class. EXPECTED YEARS: Assuming 50% success rate for the candidates for promotion. Hierarchies B(1) and B(2) are sub-hierarchies; B(1) is higher. Retirement Policy The official retirement age in the Civil Service is 55, just as it is under the Labor Code. It can be extended, in principle for a maximum of three years, if the person has dependent children to support. The person can take early retirement, qualifying for reduced pension benefits beginning at age 55, any time after he or she has worked for fifteen years. Few do the latter, unless they have found non-government jobs first. Table 16 (below) shows that few employees remain on the job after age 55; the 56-60 cohort is less than 20% as large as the 51-55 cohort. Data not reported here also show that nearly all of the 56-60 cohort is in fact 56, 57 or 58 years old. At the same time, there is evidence that some early retirement or other attrition takes place, because the 51-55 cohort is substantially less than one-half the size of the 41-50 cohort.  19 B. Policy in the Parastatal Sector Wa-ge policy-in the public enterprises is determined by a-multiplicity of laws and regulations, because the employees of the public enterprises have several different legal statuses, just as the enterprises themselves do. The enterprises can be: -- Etablissements publics a caractere administratif (EPA): firms with no intention of making profits, whose purpose is to be managerially independent of the government but under close accounting control of the Ministry of Finance's Centre des Etablissements Publics; -- Etablissements publics a caractere industriel ou commercial (EPIC): firms with a profit-making vocation, otherwise the same as the EPAs; -- Societes Nationales: firms, whether profit-making or not, which are granted a larger measure of accounting independence but which are 100% government-owned; -- Societes d'Economie Mixte: joint ventures, where the State owns a share (from 10 to 97 percent) of the capital, which are considered part of the parastatal sector but are supposed to operate with little or no government interference, neither in management nor in accounting. The workers in the parapublic sector can be of one of the following types: -- Fonctionnaires: civil servants, "detached" from the government to serve in the public enterprises; most of these work for the railway or the post office; -- Assimiles: contract employees whose contracts contain provisions linking them to civil-service conditions but who are not titulaires (tenured); most of these work for local government in Dakar; -- Solde Globale: contract employees whose contracts are independent of civil-service conditions;  20 EPACIOC: a relatively new status, especially created for the public enterprise sector -- both EPAs and EPICs -- which is intended gradually to be come the single status covering all workers in those two groups of firms; -- Conventions Collectives: workers covered by (private-sector) collective bargaining agreements and the Labor Code; -- Professions Agricoles: like the previous one, but with special provisions reflecting the unique character of agricultural employment. In June 1984, for example, the breakdown of workers by status in the thirty EPAs and EPICs under the accounting control of the Centre des Etablissements Publics was: Fonctionnaires -- 1,458 Assimiles -- 3,930 Solde Globale -- 2,369 EPACIOC -- 6,686 Conventions Collectives -- 4,574 Professions Agricoles -- 957 Others -- 1,594 Total - 21,568 Another 20,873 workers were employed in Societes Nationales and Societes d'Economie Mixte ; while a breakdown of their status was not available, it is likely that most of them are covered under the category of conventions collectives, thus having working conditions like those prevailing in the purely private sector. Many of the Societes d'Economie Mixte had, of course, been purely private firms until the government took a participation in the early 'seventies, such as the phosphate and salt mining companies and the groundnut oil mills. 1. University professors and teaching assistants, those on retainer or waiting to begin work, and those on paid leave. 2. There is some double counting here; the CEP has ex post accounting control over several Societes nationales, so that one cannot simply add the CEP figures to the others.  21 There are only a few areas in which parastatal wage policy must be discussed separately from that of the civil service on one hand, and the private sector on the other. First, the question of how to compensate managing directors of public enterprises to reflect their situation as both public employees and chief executives of firms, many of which are expected to compete in the marketplace or to conform to market tests of performance. Special arrangements have been made to differentiate directors of firms by the "importance" of their firm; since 1976 the regulations have distinguished four categories of firms by a complex, or at least unquantified, set of criteria, and award higher salaries to managing directors of more important EPICs. The most recent decree (not yet promulgated, but it will be shortly) takes the further, hitherto unanticipated step of providing for seniority increases for the managers, so that persons with substantial experience (5 years as head of an enterprise for the top class) will be able to be paid more than persons with none. The range of salaries in the newly proposed and existing scales is: Class Existing Scale(since 1980) Proposed Scale 1 407,000 F./mo. 390,000 - 510,000 F./mo. 2 352,000 328,000 - 448,000 3 297,000 281,000 - 401,000 4 242,000 250,000 - 370,000 Directors who are civil servants and hence paid according to the "grille indiciaire," which would predict a lower salary than is common in either the private sector or the public enterprises, are given a compensating "indemnity" to bring their pay up to the amount corresponding to their firm's status. This  22 type of correction is only done for the one highest-ranking manager in each enterprise, however, and even assistant directors who happen to be civil servants are paid only their civil service salary. Among high-skill employees of the parapublic sector, therefore, the incentive to leave the civil service is quite large, and this may help to explain the substantial decrease in the number of fonctionnaires in public enterprises in the past few years (see Table 21). A second area of wage policy in the parastatals that should be discussed is these firms' added burden of having to respond to the diverse assortment of pressures coming from so many different types of employees. Some enterprises, notably the post office and ISRA (the agricultural research institute), have employees from all of the employment categories discussed above. Such firms must deal with several trade unions, with different interests; for example, civil servants have their own trade -union. Given the contentious nature of Senegalese labor relations (both within the labor movement and between labor and management) it is likely to be more difficult to satisfy a diverse labor force, with different demands, different mouthpieces and different laws, than to satisfy a more homogeneous one. An interesting hypothesis to test in this regard is that wages might rise more quickly (or labor unrest occur more frequently) in firms with a non-homogeneous labor force. If the hypothesis were found to be correct, a policy reform recommendation would emerge, that public-enterprise employees should have a single, unified set of rules and working conditions, to minimize the possibility of spiraling sets of catch-up demands. This recommendation is probably worthwhile in any case, to reduce the wage and employment administration burden on personnel managers in the parastatal sector.  23 As stated above, the EPACIOC status is supposed eventually to cover all employees of EPAs and EPICs. The decree establishing the status dates from 1976 (no.-76-122 of Feb. 3, 1976), however, and still only about 30 per cent of the relevant labor force is subject to it. Presumably those with civil service and assimile status do not have to change during their careers -- unless they change jobs; it is not obvious from the decree why the others (especially those under the conventions collectives and the solde globale) have not yet become integrated into EPACIOC. A third element of parastatal policy of interest is contained in the decree mentioned above: the establishment of a single skill hierarchy for job classification and pay purposes. The scale has eight classes, arranged, according to the decree, by "increasing order of importance:" Class 1. Workers without professional qualification Four Categories, from unskilled laborer to semiskilled worker (both white and blue collar) Class 2. Workers with professional qualification Four Categories, from "ordinary professional worker" to "exceptionally qualified professional worker", again both white and blue collar Class 3. Foremen and Ordinary Technicians Six Categories Class 4. Supervisors and High-level Technicians Three Categories Class 5. Engineers and non-Decisionmaking Managers Three Categories Class 6. Managers participating in decisions Three Categories Class 7. Plant or Office Directors Three Categories Class 8. General Directors (Enterprise Chief Executives) Three Categories (but changed by the decrees discussed above to be four categories, differentiated by "importance" of the firm)  24 The range of pay under the EPACIOC scale was not settled until 1979, when a decree fixed the following ranges of pay to be applied in all the public enterprises: Class Range of Monthly Base Pay 1 21,000 - 30,000 Francs 2 34,000 - 54,000 3 47,000 - 90,000 4 110,000 - 140,000 5 150,000 - 170,000 6 180,000 - 200,000 7 210,000 - 230,000 8 245,000 - 278,000 (but see above) There are no echelons within this ranking which serve to show the seniority advancement as clearly as is the case in the civil service. Instead, there is a seniority premium added to the base pay after three years of service, equal to: 3 % of base pay after three years; 5 % of base pay after five years; 1 % more for each additional year of service from the 5th to the 15th year. A final element of interest in the public enterprise pay policy is the existence of bonuses supposed to give incentives to worker productivity. The prime de productivite, limited to 15 per cent of the wage bill for each of the eight Classes, is awarded on the basis of supervisors' quarterly evaluation of their subordinates. The total amount available is split in the following way: 6% of the total for those receiving grades from 51-601 12% 61-70 21% 71-80 34% 81-90 54% 91-100 An analysis of the consequences of this bonus is conducted in Section III.C. 1. the decree guards against grade inflation by limiting the number who can be given high grades: only 4% of a Class can get more than 90; 10% more than 80, 20% more than 70, 40% more than 60.  25 III. COMPOSITION OF THE PAY PACKAGE A. Initial Observations One of the most telling internal documents produced by the Central Payroll Service (Service Central de la Solde) of the Ministry of Finance is a monthly computer printout entitled "General Summary by Item of Pay," (Recapitulation Generale par Poste de Paie), familiarly if somewhat mysteriously nicknamed BULGE. The elements making up the government's wage bill are detailed in 250 line items, 75 or so of which are withholding lines of various kinds, and the rest are payments. Only six of the items are applicable to more than half of the government labor force, three on the payment side and three withholding. The rest is a crazy quilt of special deductions (retenues) and allowances. (indemnites and primes) some of which benefit one person ; some of which benefit all or most of a class of employee2; some of which benefit those who 34 do something out of the ordinary or have special needs4. This information is useful for several reasons: first, it demonstrates the complexity of wage administration and the consequent difficulty of analysis; second, it dramatizes the rigidity of the structure of payments, hence showing that there are unlikely to be easy ways of slowing the growth of the public sector wage bill; and third, it enables us to estimate the effective size of the civil service by counting the number who have actually been paid, as opposed to those carried on the civil service rolls. While it will not be 1. for example, using December 1984 BULGE data, there was one indemnite de fonction greffier (clerk of court's allowance) of 50,000 F. and one prime de bicyclette (bicycle allowance) of 300 F. 2. for example, an indemnite 25% des douanes paid to all 1,827 customs officers. 3. for example, an indemnite des heures suppl6mentaires (overtime pay), earned by 3,834, or a prime de rendement (performance bonus), earned by 266. 4. for example, a prime de tabac (tobacco allowance), of 500 F. or so, "earned" by 4,7Y4  26 possible in this report to undertake the kind of detailed analysis necessary to evaluate the merits of each indemnite, it may be possible to demonstrate the extent to which the complexity shown in the BULGE is an obstacle to reform. One of the principal reasons for the complexity of the itemization of the payroll is that it represents, geologically, the remnants of previous efforts at relaxing the rigidity of the grille indiciaire without creating such a large number of scales that it would become meaningless. Unless a new pay scale is created, for example, for the nine roving ambassadors (ambassadeurs itinerants), the only way to compensate them for their evident special needs is to give them each an indemnite (of 100,000 F. per month). It is clearly necessary to grant those in especially dangerous or dirty or otherwise unpleasant positionsi some additional pay, just as it is necessary to grant those with much responsibility (even though the latter probably work in pleasant, unrisky surroundings). Allowances have seemed a much better way than further complicating the grille indiciaire of tailoring pay to the individual circumstances of very different jobs within the government. The main problem with this is psychological: nearly everyone has some indemnites, and those who do not think they should and devote time and effort to lobbying for special treatment for their group2. Table 10 gives an indication of the composition of the wage bill by major categories of pay, in both the civil service and the public enterprises controlled by the Centre des Etablissements Publics. The difference between total compensation (Consommation Budgetaire) and base pay is quite substantial in both cases: base pay represents less than two thirds of the total, and 1. Firemen receive 10,000 F. per month as a prime de risque, and teachers receive 20% of base salary. Combat pay? 2. This was occurring within the Ministry of Plan during my visit in January 1985.  27 TABLE 10 CIVIL SERVICE AND PUBLIC ENTERPRISES, 1982 and 1984 C I V I L S E R V I C E P U B L I C E N T E R P R I S E December December Percentage December December Percentage 1982 1982 Point 1982 1984 Point Z % Change % % Change Straight Wages 62.8 61.0 - 1.8 62.2 65.2 + 3.0 Allowances 24.8 26.9 + 2.1 26.5 23.7 - 2.8 Social Payments 4.3 4.1 - .02 1.3 1.0 - 0.3 to Employees Employer's Social 8.1 8.0 - .01 10 10.1 + 0.1 Contributions TOTAL 100 100 Average Base Pay 72,481 77,865 7.4% 65,511 74,933 14.4% Ave. Total Pay 115,416 127,648 10.6% 105,323 114,928 9.1% SOURCE: BULGE  28 allowances (indemnites and primes) about a quarter. Social payments such as family allowances account for only four and one per cent of Civil Service and Public Enterprise wage bills, respectively, and the balance, 8 to 10 per cent, is composed of employer contributions to social insurance, especially retirement. One should be cautious about discerning trends from such short time series: these numbers can be used to demonstrate how sensitive one's conclusions are to the choice of indicators. As the second-to-last row in the Table shows, the average base salary of civil servants rose only about half as fast as that of public-enterprise workers between December 1982 and December 1984. But total compensation of civil servants rose more rapidly than that of public-enterprise workers. The reason was that between the two dates the share of Allowances was rising in the civil service and falling in public enterprises. If one had chosen to concentrate on base salaries, for example to zero in on the effect of internal promotion on the wage bill, one would judge that public enterprise wages had grown at an excessive rate, and civil service wages only moderately. If, on the other hand, one limited one's analysis to total compensation, one would measure the two-year wage change as approximately equal in the two sectors. Previous studies done in Senegal have concentrated on the latter measure, and have reached inadequate conclusions as 2 a result 1. measured as 39.86 F. times the number of index points, all multiplied by 1.34 (two general complements totalling 34% apply to everyone), plus one wage increase of 2,00U0F. for the 1982 increase, plus one more of 2,000 for the 1983 increase. Someone of index 1000 would therefore earn: (39.86 * 1,000)*1.34 + 4,000 = 57,142 F. per month. 2. Plan (1978); CEGOS (1982); Cellule de Controle (1982). All simply measure wages by Consommation budgetaire when analyzing changes over time. CEGOS does present a breakdown of the "masse salariale," but only for a single year (1981-82), and does not comment on it.  29 Neither the Civil Service nor public enterprises have homogeneous labor forces. Workers are paid a wide range of salaries, as the scales discussed above attest. Table 11 shows just how wide the distribution has been in the civil service, and Table 12 does the same for public enterprises. Clearly such a range must be explained by differences in the characteristics of employees. There are obvious skill, or rank, differences; there are different types of contractual arrangements; there are wide ranges of services provided by government departments and enterprises; public employees exhibit the typical ranges of age and experience. All these are likely to explain pay differences; which.ones are most important may have implications for the outcomes of any attempt at budget control through personnel actions. B. Differences among types of employees -- Civil Service 1) Distribution by rank Most employees in the government are either civil servants or assimiles assigned to Hierarchies A through E. The others are either contract workers under the solde globale or in the military, which has its own ranks. Table 13 shows the trend in mean earnings for workers in each hierarchy since 1976, and provides another indication of the rapidity of the compression in the government wage structure during the past decade. This reduction in wage differentials exerts a downward influence on the growth of public personnel expenditures, because the civil service is becoming increasingly educated, and hence of higher average rank; in 1971, only 25 per cent of government workers in Hierarchies had the baccalaureat or equivalent, and hence were in either Hierarchy A or B; in 1984, 42 per cent were. If wage increases had not been progressive since the early 1970's, the average government salary would have grown much more rapidly, as the analysis performed in Section IV will show.  30 TABLE 11 PERCENTAGE DISTRIBUTION OF MONTHLY EARNINGS OF CIVILIAN CIVIL SERVANTS* 1971 - 1984 Monthly Pay Category 12/1971 12/1976 12/1981 12/1984 Under 15,000 F/month 13.2 1.4 0.2 0.05 15 - 24,999 25.7 9.0 0.1 0.05 25 - 39,999 31.0 31.8 5.2 0.5 40 - 59,999 17.0 29.7 16.5 13.5 60 79,999 7.3 15.1 20.5 16.5 80 - 99,999 2.8 6.1 17.7 16.9 1