JOURNAL
OF COMPARATNE
ECONOMICS
&13-23
(1978)
Higher Education in the USSR: An Economic Comparison with the United Kingdom’ RUTH KLINOV-MALUL Department
of Economics, Received
Hebrew May
19,1977;
University Revised
of Jerusalem, August
Jerusalem,
Israel
23,1977
Klinov-Malul, Ruth-Higher Education in the USSR: An Economic Comparison with the United Kingdom. The Soviet system of higher education is more specialized than that of the United Kingdom, both in the sense that teaching is separated from research and in the sense that institutions specialize in certain fields. The paper compares, in the light of this difference, the costs of higher education and the qualities of students in the two countries. Also, the social rate of return to higher education in the USSR is examined. The main tentative findings are that the Soviet system is not cheaper; that the “quality” of students is, on the average, lower; and that the social rate of return is close to that of tangible capital. J. Comp. Econ., March 1978, 2(l), pp. 13-23. Hebrew University, Jerusalem, Israel. Journal of Economic Literature Classification numbers: 912,053.
In the last few years the phenomenon of brain drain has spread to the USSR, mainly through the emigration of Jews to the West. As a result, an economic comparison of the Soviet system of higher education with the various systems in the West has become of more immediate interest. Before dwelling on the main points of interest, a brief outline of the Soviet system seems relevant. In the USSR institutions of higher education consist of two main types: ordinary universities, in which research and teaching are conducted simultaneously; and institutions in which the role of research is minor. Students can study in either type as regular day students or as evening students, or by correspondence. Day students do not work, and they receive state support during their studies; the standard period of study is 5 years, starting after 10 years of prior education. Evening and correspondence students usually work, although by law they receive about 10 months of paid leave during their years of study; the standard duration of study is, for them, 6 years, but it often extends to 8. The dropout rate varies considerably: for day students it is 5-7% of the starting class; for correspondence students it is near 50% (Remennikov, 1969). About half of the students take day courses, 40% study by correspondence, and the remainder take evening courses (Table 1). ’ Financial support was received from the Ford Foundation through the Israel Foundations Trustees. I wish to thank E. Kleiman and R. Gronau for their helpful comments. 13 0147-5967/78/0021~13$02.00/0 Copyright 0 1978 by Academic Press, Inc. All rights of reproduction in any form reserved. Printed in Great Britain
RUTH
14
KLINOV-MALUL
TABLE
1
DISTRIBLJ~IONOF STUDENTSIN THE USSR BY TYPE OF COURSE (PERCENTAGE) Type of course
1959
1966
Day Evening Correspondence
51
46
9 40
13 41
100
100
Total
Sources: 1959, Dewitt (1955, p. 232); 1966, Noah p. 94).
(1966,
Admission generally requires a combination of a secondary-school certificate and entrance examinations, but within this framework individual institutions have their own rules. The proportion of secondary-school graduates who continue in higher-education day courses is 22%, and in all forms of higher education, including correspondence and evening courses, about 45%. This relatively small percentage of those who continue beyond secondary school reflects the fact that during the 1960s secondary schools expanded more rapidly than institutions of higher education, and there is considerable pressure to expand the latter (Katz, 1973, pp. 82-86).* .Even this cursory summary raises several interesting questions. The degree of specialization of Soviet institutions is much greater than that of Western institutions, both in the sense that research is institutionally separated from teaching and in the sense that teaching institutions tend to specialize in certain fields. Such a structure may be the result of the generally lower mobility of the Soviet labor force. Because mobility is more regulated, it seems unnecessary to hedge against unforeseeable future changes by giving students a broad education; thus it is possible to teach a narrowly defined skill. This type of structure may be cheaper in direct teaching costs, and this in itself is worth examining. This is done in Section 1. A major obstacle to concluding anything about comparative productivity from cost comparisons is the difficulty of assessing the comparative quality of students. Quality comparisons are always dimcult, but in the case of a common market one can compare different qualities of a product by comparing market prices; this is impossible here, because the bulk of goods and services produced by university graduates is not internationally tradeable. No ready answers are offered, but several possible ways of approaching the problem are discussed in Section 2. * Demand pressure is evident in the high rate of applications per place in day studies. In 1970 the overall rate was 2.6 applications per place (Katz, 1975, p. 82).
HIGHER
EDUCATION
IN THE USSR
15
Lastly, the size of the system-the number of teachers and students-is a result of a higher degree of central planning than is usual in the West. It is of interest to consider the planning criteria and to see whether the ex-post result is different from an “equilibrium” solution by Western standards. This is done in Section 3. There is no intention in this paper of proposing a definite solution. Rather, the aim is to raise a number of questions and discuss possible approaches to them. In order to do this, the United Kingdom is taken as an example of a Western system. 1. THE DIRECT
COST OF HIGHER
EDUCATION
The direct social cost of the Soviet system of education is estimated in Table 2. This includes state budgets on current account, but excludes maintenance support given to the students by either the state or their employers. As can be seen, the cost of correspondence courses per student is less than one-tenth of the cost of day courses, but the differential in the cost per graduate is much smaller as a result of the very large differences in the dropout rate, assuming that incomplete studies are a net loss to the economy. A comparison of direct costs with those in other countries necessitates adjustments. Higher-education institutions are typically multiproduct firms, producing, in addition to graduates, research and cultural-educational services for the general public. Thus, one must take into account the fact that the mix of TABLE 2 AVERAGE BUDGET’
ANNUAL
DIRECT
EXPENDITURE
BY TYPE OF INSTITUTION
PER STUDENT
AND TYPE OF COURSE?
IN THE STATE
1965
(RUBLES)
Type of institution
Day
Correspondence
Engineering institutions Universities Agricultural schools Art Education
880 807 850 1398 688
81 79 110 110 69
882
85
Average
Source: Remennikov (1969, pp. 15-16). a Excluding maintenance support given by employers to evening and correspondence students. This consists mainly of paid leave of about 10 months during the entire period of studies, i.e., about 2 months annually. The average monthly salary paid is 100 rubles. State support to students, which averages 27% of the budget, was deducted from the day-course expenditure. b There are no data for evening courses, except for art, for which the figure is 290 rubles.
RUTH KLINOV-MALUL
16
products is different both within, but especially between, countries and separate the cost of students from that of other products. There are two main methods of separating these costs, depending on one’s assumptions about the joint supply of production factors (Layard and Verry, 1975). The first method assumes that within each institution there is a joint supply, but that the proportion of inputs and outputs varies between institutions. To separate the components one estimates a cost function by a regression analysis in which the various products are independent variables and the cost is the dependent variable. The second method assumes that the supply is not joint and that the costs of each activity can therefore be allocated according to predetermined rules. For the United Kingdom, both methods of cost calculation are available (Layard and Verry, 1975). Since they lead to similar results, and since some jointness of supply seems plausible, only the results of the regression method are reported in Table 3. No separate data on teaching and research activities exist for the USSR. However, the proportion of students attending institutions with little research out of the total (including universities) is high; therefore, the cost per student in the USSR is entirely attributed to teaching. TABLE DIRECT
COST OF STUDIES,
USSR
3
AND THE UNITED
KINGDOM
(E)
United Kingdom (1968/1969) Marginal cost
A. Annual cost Engineering Agriculture and biology Averagec
USSR (1965) Average cost per student*
Average cost per student
Undergraduates
Postgraduates
680
1610
790
410
550 -
1580 -
940 570
390 370d
1930f
1900’
1480
19OOf -
2260’ 1370
1400 1330d
B. Cumulative cost (12% discount rate) Engineering 1630e Agriculture and biology 1320e AverageC
Sources: United Kingdom, Layard and Verry (1975, pp. 70-72); USSR, Table 2. (1The exchange rate used is rubles 2.16/f, which is rubles 0.90/% divided by f0.41666/$. See United Nations, Statistical Yearbook, 1970, p. 608. b Including al1 university activities. c Including all subjects and activities. d Including universities only (i.e., excluding institutions). e Cumulated over the first 3 years. f Cumulated over the next 2 years.
HIGHER
EDUCATION
IN THE USSR
17
The first conclusion from Table 3 is that annual costs are indeed lower in the USSR compared with both marginal and average costs in the United Kingdom. There could be several possible explanations. One might be a different allocation of students’ time between homework and courses. If there were less classroom work in the USSR than in the West, this could explain some of discrepancy in direct cost. As a matter of fact, however, students in the USSR have 32-40 hours of courses weekly, which is much more than students in the United Kingdom have (Dewitt, 1955, pp. 1 l&l 11). Another explanation may be a higher teaching load and larger classes in the USSR. If there is a large fixed cost in hiring faculty, then the marginal cost of teaching is lower when the teaching load is higher. We know that the typical teaching load in the USSR is 18 hours. It is true that “teaching” is defined in a very broad sense, and includes personal supervision, etc.; still, the teaching load is much higher in the USSR than in the United kingdom. This is a real source of lower cost and is indeed due to the specialized structure of higher education. Still another factor that could influence the results of Table 3 is the exchange rate used. The effective exchange rate of rubles/f is probably higher than the official one, so that differences in “true” costs are probably even higher than those shown. On the other hand we do not know enough about the relative prices of university inputs within each economy, so that it is impossible to determine which country has the comparative advantage. Lastly, lower annual costs may reflect a lower quality of students, as measured by their potential productivity in the labor market. This is discussed in the next section. Turning to cumulative direct costs (Panel B of Table 3), two points require clarification. First, the discount rate used is the internal rate of return in the USSR. This is discussed in Section 3.3 Second, the question is whether the Soviet figures should be compared with the 3-year or the 5year United Kingdom figures. It is impossible to decide this. On the one hand, entry into higher education comes after 10 years of secondary school in the USSR, and after 12 years in the United Kingdom, so that a case can be made for treating the 5 Soviet years as equivalent to 3 years of higher education in the United Kingdom. On the other hand, this assumes equal productivities at the secondary-school level in the two countries. A comparison of the 5 years in the USSR with the 3 years in the United Kingdom still shows higher United Kingdom cost in engineering and in the average cost per student, but not in the marginal cost of undergraduate students. The main results of this incomplete comparison are: 1. In terms of total teaching hours per student, the USSR is more expensive than the United Kingdom; however, there are possible economies in the size of classes and also in the teaching load per teacher. 3 A case could be made for using different discount rates for the two countries, in which case the lower one would be used for the United Kingdom.
18
RUTH
KLINOV-MALUL
2. The lower cost in the USSR may reflect its comparative advantage in higher-education inputs, especially the lower relative wage of university teachers. 3. It is very difficult to draw conclusions from the cumulative figures. Using the 5 and 3-year comparison, I think that it is fair to conclude that undergraduate teaching costs come fairly close for the two countries, and that the difference in average cost is due to the inclusion of research in university activities in the United Kingdom. If this is correct, there are no economies from the institutional separation of teaching and research. 2. QUALITY OF GRADUATES Another question is whether graduates in the two countries are comparable in terms of quality. As mentioned in the introduction, such a comparison is difficult because there is no common labor market; also, there is hardly any indirect market for services produced by graduates of the two countries because such services are usually not internationally tradeable. Absolute salary or value-added comparisons therefore have little meaning for quality comparisons. Another possibility, which exists when markets are separate, is to compare the slopes of age-earnings profiles rather than their absolute levels. It is often found that the higher the level of education, the faster earnings rise with age (Mincer, 1974, Chap. 1). Human-capital theory explains this as the result of different investment profiles in on-the-job training, when this investment takes the form of the worker’s foregoing part of his potential income. Current income Y,, of a person with IZyears of schooling can then be written as t-1
Y”t = Yno +
r
c
cj-
c,,
(1)
j=l
where Y,, is potential earnings at period 0 of a person with 12 years of schooling, with period 0 the date of leaving school; C, is current investment in on-the-job training financed by foregoing earnings, net of depreciation; j is the time (date of leaving school = 0); and r is the rate of interest. As reported in the literature, C, declines with t as the result of a decline in gross investment and a possible rise in depreciation. Gross investment declines with t largely because the payoff period is shorter, whereas depreciation may rise with t for biological reasons. The decline begins from the highest C,, which occurs usually in the last year of schooling.4 In order to use this information to 4 While the student does not work, investment in schooling is at its maximum in terms of time, and increases with foregone earnings. It therefore reaches its peak in the last year of full-time schooling. However, as the last year is an indicator of the total capital invested in schooling, whatever is concluded about Cz will be true for the cumulated investment in n years of schooling.
HIGHER
EDUCATION
IN
THE
USSR
19
compare qualities one must assume a specific function at this stage. An exponential function is suggested: c, = CZ(l - g)‘, (2) where c is the investment in the nth year of schooling and g is a constant rate of decline in investment. Substituting from (2), (1) can be rewritten as f-1 Ynr= Y,, + rCE IX (1 -gy’Cz(1 -g)‘, (3) j=l and the addition to earnings between t - 1 and t as (1 + r) - CZ(l -g)‘. YPlf- ynct- 1) = CZ(l -gr-’ (4) The main conclusion for our purpose is that the slope of salaries with respect to age rises with Cz (and therefore with n), as long as g < r, and declines with a rise in g, given r. A flatter age-income profile is therefore an indication of a lower Cz, or of a higher g, and I suggest that these are the dimensions of quality of human capital. The remaining factor that may influence the slope of Y,, with respect to t is r, which is irrelevant for quality-and therefore is conveniently ignored. For lack of data on salaries in the USSR, we use instead figures on valueadded; in the Soviet sense, value-added is similar to NNP at market prices, in that it includes wages and gross profits. The main difference from NNP is that the Soviets exclude services from value-added. (The excluded services are transport and communications, teaching, medicine, science, finance, public administration and defense, housing services, other personal services, and business services (Yanovsky, 1965, pp. 113-l 15).) Value-added figures are shown in Table 4. In Table 5 rates of change in earnings in the United Kingdom and the United States are compared with rates of change in valueadded. There are two noteworthy features in Table 4: the peak of productivity occurs at quite a young age by Western standards-at about 35 for secondaryschool graduates, and 45 for persons with higher education, compared with 45 and 50, respectively, in the United States (Hanoch, 1967). This suggests a relatively high rate of g in the USSR. Table 4 also shows a relatively low expected number of working years above the age of 50. Table 5 shows that the rate of increase in earnings is very much lower in the USSR than in the other two countries, indicating a lower quality of education. The main deficiency of such a comparison is the assumption that on-the-job investment is financed by foregone earnings. This would be the prevalent form of finance if workers could realize a return on their investment, but that is not necessarily so in the regulated Soviet labor market. Lastly, a possible approach to quality comparison is to compare salaries of graduates of the different systems in a labor market where one finds both of
RUTH KLINOV-MALUL
20
TABLE ANNUAL
VALUE-ADDED
AND LEVEL OF EDUCATION ENGINEER= (RUBLES)
OF A “REPRESENTATIVE”
MALE
Higher education
Secondary education
As
20-22 23-29 30-39 40-49 50-59 60-69
BY AGE
4
Expected number of work years (1)
Value-added Unadjusted
(2)
Adjustedb (3)
8
3300
2640
9 9 I 1
3900 3000 2000 1500
3.510 2100 1400 150
Expected number of work years (4)
Value-added Unadjusted (5)
I 9 9 I 1
5ooo 6000 6000 5000 4000
Adjustedb
(6)
5000 5400 5400 3500 400
Source: Urlanis (197 1). D Starting to work at age 23. There is no further explanation in the source, either on the statistical sources or as regards the exact meaning of “representative.” b Corrected for expected working years, by taking the proportion of expected working years, out of the age interval.
TABLE RATE
Age
OF CHANGE”
5
OF EARNINGS OF MALE GRADUATES UNITED STATES, AND VALUE-ADDED
United Kingdomb @.A)
IN THE UNITED IN THE USSR
KINGDOM
United States 13-15 years of schooling
AND THE
USSR
16 years of schooling
Higher education Unadjusted
18-24 25-29 30-34
4.4 4.9
9.3
7.5
35-39 40-44
3.2 3.1 1
3.0
3.6
5&54 45-49 55-59 60+
-0.6 1.0 I 1.3 -3.0
0.9
1.5
0.0
0.4
Adjusted
2.0
0.8
0.0
0.0
-1.8
-4.2
-2.2
-19.5
Sources: United Kingdom, U.K. (1971); United States, Miller (1966); USSR, Table 4. 4 The rates are computed between midpoints of the age groups. b England and Wales.
HIGHER
EDUCATION
IN THE USSR
21
them, i.e., in the Israeli labor market. There are, however, no data on salaries, and the only “indicator” of quality is reported employee satisfaction derived from work-which might be relevant if one dared to assume that there is a positive correlation between it and the objective quality of work. Table 6 summarizes this degree of satisfaction. It should, of course, be taken with more than a grain of salt. First, it describes the data by continent rather than by country of origin (but the majority of Europeans come from the USSR); second, it does not differentiate among educational levels (although the majority in both groups has had higher education). The data suggest that European immigrants are less satisfied. 3. RATES OF RETURN The last point to be discussed is the following. The Soviet system provides free higher education, including living expenses. This obviously creates substantial pressure by secondary-school graduates to increase admissions to higher education.’ The questions are what criteria are used in deciding on the size of the system of higher education, and whether the final result is very different from the result that would have been arrived at in a market economy. For the latter question the conventional social rate-of-return measure is used, and the product is measured by value-added. This is difficult, because, by definition, only the returns to the “productive” sector are counted. From the nature of the services sector it is clear that a large proportion of graduates are employed in it. Therefore, a valuation of the services in this sector is needed for a more correct estimate of rates of return. This is a problem in noncommunist countries as well, but there there are two mechanisms which tend to equate returns in the private and public sectors: competition in salaries, and indirect valuation of the various public services via parliamentary debates on budget
TABLE IMMIGRANTS
6
IN 1970/1971 BY SATISFACTION
COMING
ORIGIN
Degree of satisfaction
ARer 1 year
FROM
WORK
AND
CONTINENT
OF
(PERCENTAGE)
After 2 years
After 3 years
Europe
America
Europe
America
Europe
America
36.1 39.0 14.3 10.6
42.8 33.1 13.0 11.1
36.7 38.0 15.9 9.4
39.2 39.6 14.0 7.2
43.0 39.6 12.5 4.9
43.0 47.7 6.2 3.1
100.0
100.0
100.0
100.0
100.0
100.0
Definitely satisfied Fairly satisfied Not so satisfied Not at all satisfied Total Source: Israel (1976).
22
RUTH
KLINOV-MALUL
laws. In the USSR, on the other hand, there is no clear mechanism that equates the productivity of graduates in the goods- and services-producing sectors. Another difficulty in rate-of-return calculations is the very high proportion of night and correspondence students in the USSR. Neither costs nor returns can be even approximately measured in their case: costs must include some valuation of the social price of their leisure and benefits must take into account the frequent claim that their quality is lower than that of day students. Our calculation ignores both, and is certainly an upper limit of the usual rate-ofreturn measurements. Using the data of Tables 1 through 4, together with dropout rates, the rate of return is calculated as 12.8% for day and 19.9% for correspondence students. Two comparisons seem appropriate: one with the social rate of return on university education in the United Kingdom, and the other, with returns on tangible capital in the USSR. However, in both comparisons, it should be borne in mind that the return on Soviet education is probably upward biased because returns are based on the average product (i.e., value-added) rather than the marginal product, whether of labor or of capital. The rate of return for Soviet day students is somewhat higher than that for engineers and scientists in the United Kingdom: 9.7% in 1966 (adjusted for dropouts; Layard et al., 1971, p. 138). It is also somewhat higher than the rate of return (for men) on first degrees in the United Kingdom, which is 1 l%, excluding research costs and adjusted for dropouts (Ziderman, 1973, p. 95). The day-course rate of return also compares favorably with the estimates of the rate of return on tangible capital in the USSR. The rate of return on net capital is estimated as 11% (Becker, 1969, p. 52); the rate on gross capital in manufacturing is estimated as 13-16% (Weitzman, 1970, p. 689). The calculations differ in their assumptions, and the comparison should be taken with the utmost caution. Nevertheless, it does seem possible to draw one conclusion: that in spite of the relatively easy access and the low private cost of higher education, there is no overinvestment. One explanation may be that there is an effective limitation on the number of places for day students, whereas the easy access to night and correspondence studies entails quite a high private cost if leisure is counted. Two additional comments: First, there is a very small difference between returns to tangible and human capital. This may be the result of the fact that a smaller risk premium is required in investment in human capital in a planned than in an unplanned economy. The second comment concerns the rate of return as a planning tool. Clearly, Soviet planners use variants of manpower forecasting techniques rather than rate-of-return calculations.5 However, it should be mentioned that a variant of the rate-of-return approach as a planning tool is in fact advocated by Urlanis (197 1). He claims that the social benefit of 5 The fact that the rate is surprisingly coincidence.
close to an equilibrium
rate is nevertheless
an intriguing
HIGHER
EDUCATION
IN THE USSR
23
higher education is not its value-added, but rather the educated person’s contribution to the welfare of others, where “others” include the members of his family (i.e., value-added less personal consumption). This suggestion is not satisfactory because the educated person’s consumption is affected by his family’s income (e.g., through his wife’s earnings) while the family members’ consumption is affected by the head’s income. At any rate, the use of Urlanis’ figures on differential consumption by level of education hardly affects the final rate of return; thus, according to this concept, too, the size of the system of higher education is “about right.” REFERENCES Income: 19.58-64. Berkeley and Los Angeles: Univ. of
Becker, Abraham S. Soviet National California Press, 1969. Dewitt, Nicholas. Soviet Professional Manpower. Washington, D.C.: National Science Foundation, 1955. Hanoch, Giora. “An Economic Analysis of Earnings and Schooling.” J. Human Resources 2: 3 10-329, Summer 1967. Israel, Central Bureau of Statistics. Immigrants Arrived in 1970/71-1971172. Special Series No. 530. Jerusalem, 1976. Katz, Zev. Patterns of Social Mobility in the USSR. Cambridge, Mass.: Center for International Studies, M.I.T., 1973. Layard, P. R. G., Sargan, J. D., Ager, M. E., and Jones, D. J. Quali$ed Manpower and Economic Performance. London: Penguin, 197 1. Layard, P. R. G., and Verry, D. W. “Cost Functions for University Teaching and Research.” Econ. J. 85 : 55-74, March 1975. Miller, Herman, P. Income Distribution in the U.S. U.S. Department of Commerce, Bureau of the Census; a 1960 Census Monograph. Washington, D.C.: Government Printing Office, 1966. Mincer, Jacob. Schooling, Experience and Earnings. National Bureau of Economic Research: Human Behavior and Social Institutions, No. 2. New York and London: Columbia Univ. Press, 1974. Noah, Harold J. Financing Soviet Schools. New York: Teachers College, Columbia University, 1966. Remennikov, B. M. “Economic Problems of Higher Education in the USSR”, Pt. II. Soviet Education 1l:July 1969. USSR, Central Statistical Office. The National Economy of the USSR. Various issues, 1960197 1 (Russian). United Kingdom, Department of Education and Science. Suruey of Earnings of Qualified Manpower in England and Wales 196667. Statistics of Education, Special Series No. 3. London: HMSO, 1971. Urlanis, B. “Problems of Economic Demography.” Probl. of &on. 13 : 69-89, Jan. 197 1. Weitzman, Martin L. “Soviet Postwar Economic Growth and Capital-Labor Substitution.” Amer. Econ. Reu. 60: 676492, Sept. 1970. Yanovsky, H. Anatomy of SocialAccounts Systems. London: Chapman & Hall, 1965. Ziderman, Adrian, “Rates of Return on Investment in Education: Recent Results for Britain.” J. Human Resources 8 : 85-97, Winter 1973.