Modelling future higher education-labor market imbalances: A multi-scenario approach

Economics of Education Review, Vol. 9. No. 1, pp Printed in Great Britain.

Modelling Future Higher Education-Labor Market Imbalances: A Multi-Scenario Approach

l-23.1990. X272-7757/90 $3.00 + 0.00 @ 1990 Pergamon Press plc

NICHOLAS P. GLYTSOS

Centre of Planning and Economic Research, 22 Hippokratous Street, 106 80 Athens, Greece

Abstract - In this paper a multi-scenario model is built, in the framework of neoclassical labor market conditions and the human capital theory, that generates multi-value estimates of potential graduate imbalances and gives alternative optimum solutions of full employment and non-optimum solutions with unemployment. The model in fact reflects the impact of the labor market, of individual preferences and of educational changes. Its diversity property provides the possibility of policy choices for reducing imbalances. To demonstrate how the model works in practice we constructed a set of scenarios appropriate to the Greek case, and estimated them by applying a simple time-trend projection technique. By combining the results of these scenarios, a matrix of numerical estimates of potential future graduate imbalances is obtained, and the conditions of alternative optimum solutions and the economically unwarranted social cost of non-optimum solutions are evaluated.

“An economy can ultimately absorb any number of educated people” W. Arthur Lewis, 1962

“No society can offer [good] jobs to everybody” Eli G&berg, 1979

INTRODUCTION

LASTING imbalances between the size and the occupational structure of demand for educated labor and the size and the educational composition of supply of university graduates has been a common phenomenon for developed and developing countries alike, during the 1970s and 1980s. These imbalances have been generated not only in systems with relatively open enrollment policies, but also, and perhaps more so, in systems with admission controls - numerus clausus - which are supposed to restrain the oversupply of graduates. Evidently, these controls have not been very successful as is witnessed by the more severe imbalances in several countries with such systems.’

Graduate imbalances have been generated by a set of rather common factors, with perhaps different intensities of their impact on individual countries. Rapid economic growth and the high absorption of

the public sector, on the demand side, and the baby boom of the 1950s and 196Os, on the supply side, are major causes. Also, the difficulties of finding employment after the mid-1970s enhanced, in many countries, the demand for higher education as an alternative to unemployment.* Reduced relative wages, resulting from the pressure of graduate supply on the market, the need for more teachers, and finally political pressures3 also increased graduate employment.

From a labor market perspective, graduate imbalances are the reflection of “overeducation”, variously defined4 as relative economic deterioration (decrease of relative wages)5, occupational frustration (high unemployment rates), or over- qualification (increasing shares in clerical jobs). The result is a continuous decline of the social status and the economic value of higher education.

Despite these developments, the heavy subsidization of higher education by the government, in the

[Manuscript received 20 July 1987; final revision accepted for publication 23 July 1989.1

1

2 Economics of Education Review

majority of countries, where higher education is exclusively or predominantly public, sustains a private rate of return high enough to induce an increasing demand for university education.(’

The common problem of imbalances notwith- standing, several developed countries are also very much concerned with the decline in the growth of enrollments, so much so that they consider it a serious threat to the vitality of higher education institutions.’ Developing countries are, in contrast. concerned with the rising enrollments and with the frustrated demand for higher education, which for lack of adequate educational capacity cannot be satisfied.s

Successfully devising and implementing policies, with the double purpose of reducing imbalances and of sustaining the vitality and efficiency of higher education institutions, requires a knowledge of the possible future developments in the labor market and the education system. Thence, various methods of assessing the prospects of these two sectors have been used, encompassing different degrees of emphasis on technological or economic criteria, on more or less dependence on market forces and on relative wages.Y

According to some views, graduate imbalances must be dealt with a general orientation of higher education towards the anticipated employment opportunities and with controls of admissions”’ (Atkinson, 1983, p. 60; Sanyal, 1983, p. 357; Sanyal et al., 1981, p. 148). Many others express, however, considerable scepticism on planning higher education on the basis of estimated future needs of the economy. Apart from any philosophical questions that are related to the broader scope of higher education, the objections are methodological and practical. On methodology, manpower forecasting is primarily criticized for its fixed coefficient approach, which does not allow for substitution between factor inputs, and also for ignoring the role of relative wages in labor demand. The method is also blamed for assuming unlimited financial resources for education and an unconstrained labor supply of educated workers. The practical objections refer to data availability and the difficulty of forecasting the future labor needs of the economy in medium and long-term time horizons.”

In the face of these opposing views, the reality is, at least in the majority of the OECD countries, that individual demand for higher education is the main guide to educational planning (OECD, 1983, p. 121).

THEORETICAL SETTING

The aim of this paper is to develop a methodology that, in the framework of human capital theory and the neoclassical tradition of the role of wages. generates alternative solutions of future balances in the job market for university educated workers. Along these lines, alternative sets of conditions of optima equilibria will be identified and the oppor- tunity cost to society of graduate imbalances will be evaluated. The scope of this effort is to capture the diversified influences on job creation and graduate production of various macroeconomic structural changes and of changes in the labor market and the higher education system.

Human capital theory teaches that individual demand for more education is a function of the rate of return on the investment on that education. This rate is estimated by comparing the expected lifetime additional earnings with the corresponding private cost. But since students are rather “present- oriented” (Ehrenberg and Smith, 1988. p. 298), they attach considerable weight on current earnings. in estimating the prospects of their future employment. It is also true, that due to the rather long training period for producing higher education graduates, graduate supply is inelastic to wages in the short-run.

Under these conditions, and assuming for ex- positional purposes zero elasticity supply (a vertical supply curve Si) and a normal demand curve D,, neoclassical full employment is at point 2 of Fig. la, at the fixed supply S, and w,, real wages. An increase in demand to D, would generate a shortage and raise wages to wi. Supply will respond in time to higher wages and the supply curve will shift, say to Sz, increasing employment by SiSz and reducing wages from wi to w2. The wage-elastic long-run supply curve would then be formed by connecting points such as Z and Z1 (not drawn here).

If instead of demand it is the supply that is autonomously increasing (due for instance to an extensive new program of scholarships), wages, to the extent that they are downward flexible, will be reduced toward wj, absorbing the whole or part of the new supply. In case some of the new graduates go without jobs, they will exercise some political pressure, as noted in the Introduction. So far as they are successful, they will shift the demand curve, say to Dz, and this will mean that supply creates its own demand (see Note 4).

Modefling Future Higher Education-Labour Market Imbalances

LH- -----_.

Is-01 +=

tw

-0e

IS-01 tOO

Figure 1. Private earnings. social costs and optimum equilibrium in the graduate job market.

As it turns out, one way or another, there is. in a dynamic situation. a correspondence between suc-

D = fd (w-J?+) (1)

cessive shifts of supply (vertical) and demand curves. But, in reality. high surpluses do not seem to

S = S,, (V,,) = fixed at a given V. (2)

be able to reduce substantially average graduate wages. In fact, the evidence presented in the Variables R and V are vectors identifying particular Introduction shows that disequilibria with surpluses “structures” that determine respectively the position is indeed the rule in many countries. of demand and supply curves. From (1) and (2) the

In terms of some elementary algebra. the demand off-equilibrium graduate imbalances are given by and supply curves in Fig. la can be respectively expressed as I(&, - D)l = f, (w+J-) (3)


where ]I means absolute value of the difference. There is a surplus or a shortage according to

(SO - D) >< 0. Equation (3) is derived graphically in Fig. lb from Fig. la, for the two pairs of demand and supply curves.

An imbalance entails economically unwarranted social costs. In the case of surpluses, these costs take the form of undue spending on higher education and in case of shortages the form of output foregone due to educated labor bottlenecks. The social cost (SC) will therefore be a positive function (in absolute terms) of the supply-demand gap, i.e.

WI = fC (PO’ - 01) (4)

which in an assumed linear form is depicted in Fig. lc. Substituting from (3) into (4), social cost is expressed as a function of wages and the vector R, for fixed quantities of supply. That is,

PC] = fc & (w+Kh So WI). (5)

Since So is a fixed quantity, Equation (4) can also be written as

PCI = fSC P-9 So). (6)

A corollary from Equation (6), which is a key factor for our subsequent discussion on multi-scenarios, is that the social cost for producing a given number of graduates varies according to the level of demand. And conversely, the social cost for satisfying a given demand varies according to the response of supply (exact-, over-, or under-response).

Returning to our diagramatic exposition, Equation (6) is graphically derived in Fig. Id from Fig. la, lb and lc, for corresponding pairs of demand and supply curves. This represents a scheme of social cost schedules, reflecting different conditions of graduate supply and demand. The horizontal line L is the loci of optimal full employment equilibria with zero unwarranted social costs. Above-line points on the social cost schedules, such as A corresponding to wages w4 in Fig. la, represent a surplus of A’[, and below-line points, such as lo_ corresponding to wages w5 in Fig. la, represent a shortage of kF’. In both cases unwarranted social costs are naturally present.

Can a sustained optimum equilibrium be reached? Recalling that students’ choices on higher education are heavily based on current information, which is

necessarily related to a certain phase of the business cycle, students make, in effect, long-term individual choices on enrollments - which have overall structural implications - on the basis of cyclical observations. This alone creates some a priori unstable conditions for equilibrium.

Of particular importance in this context is also the role of governments in the education-labor market system, mentioned above. As “political stabilizers”, governments have some economic destabilizing effects. They push enrollments beyond the market warranted level, through their excessive subsidization of higher education, on grounds of social equity, mobility and democracy (Levin, 1987, p. 153). But, on the other hand, governments’ contribution to this “market failure” may be matched, to a certain extent, by some “non-market success”, stemming from the same attribute of governments as political stabilizers.

To borrow Hirschman’s (1970) terminology, this matching may be achieved with an “effective voice”, that through political pressures raises graduate employment so that prospective graduates “post- pone exit” (in the form of reduced enrollments). Since unemployed graduates are not normally organized, their voice can be transmitted through “proxy advocates”, such as the press. Even a silent expression of discontent that the mere presence of the unemployed graduates communicates may set in motion the government equilibrating machine which is “[no] less automatic” (Hirschman, 1970, p. 18) than labor market forces, though the “process is not necessarily conscious” (Hirschman, p. 19, quoted from Arrow, 1963).

Graduates are actually hired at wages that are determined by institutional, administrative, social or political considerations (Levin, 1987, p. 160; Blank, 1985, p. 212) and these wages are usually higher than their equilibrium level, or, in any case, higher than in the private sector (see for instance, Blank, 1985, p. 213). This enhances the risk aversion of graduates and leads to queuing for the relatively secure public jobs (Blank, 1985, p. 215). Overall, however, this countenveight stabilizing effort on the part of the government is not adequate to eliminate imbalances. As a result, a surplus of graduates is generated (Hinchliffe, 1987, p. 144; Levin, 1987, p. 153), which may lead to “extended disequilibria over long periods of time” (Freeman, 1976, p. 52).

Excess supply can also be affected by some individual attitudes and actions reflected in the labor

Modeling Future Higher Education-Labour Market Imbalances 5

market, such as frictional unemployment (workers between jobs), occasional entrance and exit of married women or migrants and so on.

Are sustainable surplus solutions in any sense efficient? From a broader perspective, and under the above economic and non-economic conditions, these surpluses seem to identify with the kind of slack that in Hirschman’s sense is “continuously being generated as a result of some sort of entropy characteristic of human surplus-producing societies” - “[tlhere is a slacker born every minute” (1970, p. 15, italics his). If, in our case, the exogeneity of public expenditure on higher education generates some social externalities (see, for instance Mc- Mahon, 1987) and if one accepts Hirschman’s claim that a slack “fulfills some important if unintended or latent function” (ibid), then one can rationalize some “overall efficiency” stemming from economic surpluses of university educated people.

Finally, in a purely market sense, and in view of the long-lag response of higher education to market signals, a stockpile of unemployed graduates plays a short-term equilibrating role to cyclical changes of demand that “facilitates efficient job matches” (Krueger, 1988, p. 578).

THE MULTI-SCENARIO MODEL

In predicting the future,. all sorts of uncertainties and misinformations are present as to whether past experiences will be repeated, and if so as to whether they will generate similar past developments. Indivi- dual decisions and policy orientations on enrollments and on graduate employment will be accord- ingly affected and may therefore be misled. For these reasons, it would appear of little use to estimate single demand and supply functions and simulate unique forecasts of imbalances, that would be very precarious and unreliable for shaping policies. In line with our previous analysis, an effort is made to develop a methodology that overcomes some of the major weaknesses of traditional manpower forecasting and offers the flexibility that is needed for estimating reasonable ranges of values leading to sound policy choices.

With a carefully conceived set of scenarios, based on reasonable assumptions related as closely as possible to particular experiences and expectations of a country, the sensitivity of graduate labor demand and supply against various relevant factors can be evaluated. The different scenarios may refer

to different variables, different parameters or different values of the variables. The alternative values given by the combinations of complementary scenarios constitute the individual components (a series of numerical estimates) of graduate demand and supply structures, R and V respectively, that determine the positions of the two curves.

Vector R may incorporate macroeconomic variables, e.g. output and productivity, labor market variables, e.g. occupational composition, institutional factors, and non-economic variables, e.g. political pressures. Vector V may include variables that affect enrollments, e.g. individual demand for higher education, constraints on admissions and the subsidization of higher education. It also includes demographic factors, social status attitudes and migration. Finally, wages, w, as noted above, play their role by shifting the vertical supply curves as in Fig. la, which in our empirical context means that it affects enrollments. All three variables determine the social cost as in Equation (5) and Fig. Id. By combining the numerical results of the two sides, alternative estimates of future imbalances are calculated. The story behind each scenario combination is recited, so that the different conditions under which optimum and non-optimum solutions can be reached are spelled out and the social cost of each of these solutions is indicated.

In general, II scenarios of graduate demand and k scenarios of graduate supply can be constructed. The analytical content of each of these scenarios may be of different degrees of sophistication, ranging from elaborate forecasts to simple time projections.

The simulated estimates of D and So, for all n and k scenarios, for the target period, can be used to calculate all permutations of complementary scenarios separately for graduate demand and graduate supply. This will give N combinations of demand scenarios (MJ and K combinations of supply scenarios (Es) (serial combinations) and an N by K matrix of imbalances (cross-combinations), that will give the empirical social cost schedules of Fig. Id.

Arranging these combinations in increasing order of demand and of supply and using as an index of social cost the ratio of the estimated values of supply and demand, C, = &lb,, i.e. SC, = C,, where d=l,2,. ..,Nands=1,2 ,..., K,Tablelcan be obtained and through it an empirical cluster of the social cost-demand schedules can be con-


Tabk I. Ratios of new job-seekers over job openings according to different scenario combinations

Serial combinations of demand scenarios (increasing order of

potential job openings)

Serial combinations of supply scenarios

(increasing order of potential new job-seekers)

El E2 . . . . . . (Es) . . . . . . Ek

Total number of predicted job openings

during the projection

period

Total number of predicted new job- seekers during the projection period

c C ,2.................. ,k a G, . . . . . . . . . . . . . . . c,, a

i (c*=+) : Id,)

c, . . . . . . . . . . . c,, &

s, . . ($) . . . . s,

strutted, each curve representing a column of Table 1 against the fixed quantity of supply at the foot of the column (using the ratio approximation of social cost the optimum full employment line will be at point 1, not zero).

The schedule of column Et, for instance, relates the social cost indexes (Ctt, C,, , . . . C,,) with the variable quantities of demand (d,, dz, . . ., DN) of the last column of Table 1, each Dd generated by a serial-combination of demand scenarios (M,, Mz, . . ., MN) and positioned by the fixed quantity of supply 3,. We are not constructing corresponding social cost-supply schedules because they are analytically redundant, since an empirical social cost map, as in Fig. Id, encompasses all estimated values of demand along the social cost curves and all estimated values of supply between curves.

The policy corollary from Fig. 1 is that any measures that are geared to reducing existing graduate imbalances and consequently the unwarranted social cost that they entail, should induce movements either along particular social cost- demand schedules toward the zero line. at constant supply, or shift such curves, which allows for changes in supply. This suggests, as will be discussed

shortly, that different menus of policy choices can be initiated for reducing imbalances.

The possibilities offered by this flexible” scheme may satisfy the propositions made for overcoming some of the weaknesses of manpower forecasting. That is, manpower forecasting “must move . . . to predictions of both shifts in demand and changes in the quantity demanded” (Ahamad and Blaug, 1973, p. 323), must “aspire to making optimizing forecasts” (ibid) and must allow “output [to] be con- strained by a shortage of manpower” (Wery, 1978, p. 335).

THE ESTIMATING TECHNIQUE

Following the theoretical foundation and its working device - the multi-scenario model - this section deals with the methodology of estimation and the next section with the empirical application of the estimating technique.

This technique is simple but comprehensive, comprising the major factors that determine the above R and V demand and supply structures. Having as its basic premise, time-trend extrapolations, this technique can have moderate demands

Modelling Future Higher Education- Labour Market Imbalances 7

on data and has the advantage of avoiding the introduction of behavioral variables whose values must be predicted - a formidable and error prone task-for simulating future job openings and future graduate flows. The only independent variable that is used is time which is the only variable that can be predicted with absolute precision.

Concerning the assumptions that quantify the various scenarios, given the nature of the issues discussed, they express alternative developments of past experiences of a long-term period, of more recent times, or the current period. As for the data basis of these scenarios, given the theoretical background of the multi-scenario model and considering that we are dealing with structural issues, it is necessary to work in the empirical estimation on

some disaggregated level of analysis. Consequently, our scenarios are related to sectoral data.

Graduate Employment Predictions In this framework, total employment by industry

that constitutes the “reference” aggregate for pro- jecting graduate absorption, is predicted, in the spirit of production functions, on the basis of output and productivity projections. Equations (7) and (8) in Table 2 give the time-trend parameter estimates of these two variables and Equation (12) gives the predicted value of total employment in the target year. Next, the time-trend parameters of the occupational shares are estimated with Equation (9). the corresponding parameters of graduate shares by occupation and industry with Equation (10) and the

Table 2. Empirical predictive set of trend-equations

Equation number Employment predictions

Equation number Supply predictions

(i) Trend relurionships

{ii{ y, = YiU) Y,4Li = f,(r)

(9) (&J&) = CO)

I:$ tGjdLji) = g,tr)

CGfjJGji) = dr)

(12)

(13)

(14)

(15)

(16)

(17)

(ii) Predicted values

(iii) Trend relarionships

(18) (19) (19a)

(20) (iv) Predicred values

r%c = ~%J~w~,,) (&,-a,)

(21)

Y = real output L = employment

List of variables

Y/L = real productivity G = number of graduates E = number of enrollments i = industry (= 1.2,. . . A) r = years of observations (=1,2, . . .n) j = occupation (=1,2, . . . p) f = educational field (for labor demand) JI = educational field department (for labor supply) T = number of years in the target period T = the target year to = a year of observation .$ = number of years of study in the JI department.

Equations (7)-( 11) can be estimated either with the general exponential form X, = ue*’ or with the general linear form XC = a + br, where X, is any of the variables in the above equations. The estimated values of the parameters will be 6, and 6,. By setting r = r” + 5, where r, is the number assigned to the last year of observations and 7 is the number of years in the target period, the projected values of X, for the target year T are estimated.


educational field mix of graduates with Equation (11). Based on these estimated parameters, the corresponding overall employment as well as the number of graduates by occupation, industry and educational field are predicted, for the target year, with Equations (13)-(17).

Whenever a series of data for extrapolations, particularly for Equations (9)-(11) is not available and if there is enough practical justification, the base-year relevant ratios are assigned, so that h(t) = ~1, gj(t) = cz and gAt> = ~3.

The projections with (9) and (11) constitute, in effect, a top-down estimate for individual fields of study. If relevant information is available one could use the whole process directly for field projections.

Trend Equations (7) and (8) reflect, respectively, the effects on employment of growth and of industrial structural changes; trend Equation (9) manifests the evolution of the occupational structures of industries that may come from technological changes and factor substitution; and Equations (10) and (11) capture, respectively, changes in the overall qualifications of the workers by occupation and industry and changes in the educational field mix of graduates.

Deducting from these findings the corresponding base year values gives the potential net increase in graduate employment during the projection period. The old jobs that will be vacated due to retirement and deaths should also be estimated by any appropriate method that might be fit, so that the total number of graduate job openings is obtained.

Graduate Flow Predictions The prediction of the annual graduate flows from

the higher education system, for each and all years of the target period, is done with Equations (20) and (21). This is preceded by an estimate of a survival ratio of enrollments by educational department with Equation (18), allowing appropriate time lags (mostly 4 years) between enrolling and graduating classes. The enrollments that are not already in the pipeline - for predicting graduates of the later years - are projected with Equation (19), on the basis of past values.

If a series of data on enrollments and graduates is not available, the base-year values can be used in (18) and (19). Also, an alternative way of prediction is to extrapolate directly the historical trend of graduate flows with (19a).

Concluding Remarks To calculate from here the new graduates that will

be in the job market, the gross flow must be adjusted for net migration, including those who return from studies in foreign universities, and also for the working students with solid jobs who after their graduation may not appear in the job market. Furthermore, account must be taken for any foreign students who upon their graduation will leave the country. It must be pointed out that the prediction of some of these items is not often easy and the available data may be sluggish and unreliable. For these reasons no standard method of prediction can be suggested and one has to improvise according to the particular case at hand (see below the case of Greece).

EMPIRICAL APPLICATION OF THE MODEL

We experiment now with Greek data, specifying a set of scenarios, in the lines described, in order to attach alternative values to the variables contained in the set of the predictive equations just discussed. The purpose of this application is to show the step by step process of working with the multi-scenario model, to indicate any practical problems that may emerge and to pinpoint its demands on statistical data.

For the better understanding of this particular empirical application, let us first prepare the territory by briefly presenting the Greek case.

Highlights of the Greek situation In a nutshell, higher education in Greece is

exclusively public, and tuition, fees and textbooks are free for the students. A numerus clausus applies with the double purpose of restraining oversupply of graduates in the labor market and of preventing the overflow of higher education institutions, in the face of a very high demand for university education (1 out of 6 applicants is admitted). The public sector is the major employer of graduates.” The unemployment of graduates is substantial (4.2% of males and 10.0% of females in 1986) and the queuing for public sector jobs is long. Both the education system and the labor market are, for institutional and other reasons, considerably inflexible (see Glytsos, 1988).

The predominance of the public sector in the graduate job market shapes both individual prefer- ences14 and, to a considerable extent, higher education institutions behavior. Due to the lack of

Modelling Future Higher Education-Labour Market imbalances 9

selectivity in the public sector hirings and the uniformity of remuneration, students are motivated to select easier studies.15 This can be economically justified in terms of the rate of return on higher education. With given wages, the only way to increase the private rate of return, or to prevent it from falling, is to keep the private cost low. By choosing easier to enter and to study fields, the chances of failing are reduced and the possibility of repeating the effort in subsequent years and spending additional amounts of money for preparing for the very competitive entrance examinations - a usual and expensive phenomenon in Greece - are minimized.

As it turns out, therefore, the students seem to make economically rational choices, but the government as a major employer, reflecting society’s choices, does not. This generates the phenomenon of the fallacy of composition: relatively speaking, what is good for individual students is not necessarily good for the society as a whole, of which they are an important part.

The same kind of argument, in reverse, can also explain why the majority (about 75%) of Greek students in foreign universities study, as we will see shortly, mostly what can be considered more difficult fields, namely medicine and engineering. That is, with a given relatively high private cost for their education compared with the cost of their colleagues in Greece, these fields give them a better chance either to obtain jobs with negotiable wages in the private sector, or more probably to engage in self- employment activities, raising in either case the possibility of securing a satisfactory rate of return on their expenditure on higher education.

Queuing for a public sector job, often over a relatively long period of time, is worth paying the waiting cost, given the lifetime earnings, including all the fringe benefits and the security of employment that is guaranteed by constitutional decree. Although this waiting raises the apparent unemployment of graduates (see, for instance, Krueger and Summers, 1987, p. 44), some of them hold tempor- ary jobs, often in the concealed economy.

Under all of these conditions and in the spirit of the definitions of overeducation cited above, Greece seems to experience it on all counts described.

The heavy subsidization of higher education by the government, noted above, contributed among other factors - of which the social status of the university degree is very important, not only for the

students but for their parents too (see Nassiakou, 1981; Tsoucalas, 1981, p. 119) -in the demand for higher education, which resulted in enrollments growing faster than GNP, giving an elasticity of demand (actually of enrollments) for higher education equal to 1.43.

Scenarios for Greece The basis of specified scenarios. For designing a

set of scenarios for Greece, one must capture the potential consequences that the continuation, but also the non-continuation, of the long economic recession which started in 1973, may have on graduate employment and on enrollments in higher education. At the same time, one should not ignore some main structural changes, such as the increasing share of “information” occupations of any kind, that despite the recession take their own course.

Furthermore, phenomena, such as the saturation of the public sector job market and the increasing pressure of graduate supply on the private sector job market, in conjunction with the increasing substitution of graduates for other workers that it entails, should also be accounted for.

The supply side set of scenarios, on the other hand, must cover the alternatives of long-term past increases in enrollments, the stabilization of enrollments at current levels, and the potential changes in the drop-outs ratios. Potential developments in the Greeks that study abroad - currently a very considerable figure - and in the international mobility of graduates, in view and of the intra-EC free movement of labor, must also be represented in our supply scenarios.

The horizon of our projections (made by the formula X, = a + bt) extends to the year 1995, the base year is 1984 and the longest period of data used is 1961-1984.

Empirical scenarios of employment. Turning to specifics, two scenarios of output and and productivity changes are adopted: (i) long-term (1961- 1984) trends will continue (scenario A). These extrapolations are in fact based on the experience of a booming economy of the period 1961-1972 and of a recessionary economy of the period 1973-1984; (ii) the trend of the recessionary period alone (1973-1984) will continue (scenario B). Aggregate employment projections by industry are then made with Equations (7), (8) and (12) for the one-digit standard classification sectors (8 sectors).


These kinds of projections seem to be legitimate for Greece, since past experience shows that employment in most of the major non-agricultural sectors follows a trend, at least since 1974, from which annual data are available. Despite their short- term fluctuations, output and productivity changes also follow trends.

Occupational share projections are sketched by the following four scenarios: (i) a long-term base (1974-1986) occupational composition (scenario I); (ii) a medium-term base (1981-1986) occupational composition (scenario II); (iii) an “information”‘6 (and non-information) occupations projection of total employment, based on 1971 and 1981 data (keeping constant the 1981 shares of different occupational groups in total information employment) (scenario III); and (iv) projections of shares of occupational groups in total information employment (scenario IV). For these predictions Equations (9) and (13) are used.

Graduate shares, by major occupation, are projected with Equations (10) and (14), under the assumption that these shares will not decrease during the projection period. This assumption can be justified, either on the grounds of an expected rising demand for qualified workers, following technology advances, education needs for more teachers and the more sophistication of the service sector, or on the grounds of a potential higher absorption of graduates, resulting from the pressure of their supply on relative wages.

Thus, graduate shares in major occupational groups are projected in cases where there is an increasing trend, whereas the base year shares are retained in groups that have not demonstrated any long-term trend. Particularly, for clerical workers, for which the potential oversupply of graduates will find major outlets, two alternative assumptions are made: (i) the very rapidly rising long-term (1961- 1981) share of clerical workers will continue (scenario a), and (ii) the much slower moving medium- term (1971-1981) share will continue (scenario b).

Since no series of the working graduates composition by educational field is available, we adopt for the projection period the 1984 base year composition and make the projections with Equation (15). Save for engineering, the field shares of the 1984 stock are in fact very close to the corresponding average shares of graduate flows which in the last 20 or so years have not manifested any noticeable trend.

Finally, retirements during the projection period were estimated by assuming that all graduates who in 1995 will be over 65 years of age will by then be out of the labor force. The deaths of working graduates are predicted by applying extrapolated age-specific death rates and the age distribution of graduates by educational field.

Concluding, the breakdown of projected employment by f, j and i can theoretically be as disaggregated as desirable. But practically, and aside from data limitations, more aggregate groupings have a better chance of avoiding the possibility of grave errors and provide results that are more manageable and less risky in their policy implications. For testing the predictive capability of our approach as we have applied it to Greece, we conducted two different tests: first, we reestimated Equation (9) with the 1974-1980 data and predicted the 1981-1986 figures, which were then compared with actual data (Test I); second, we compared similar available estimates made in the past (in 1975 for 1980) with ex-post actual data (Test II). In both cases the results are remarkably good (see Appen- dix).

Empirical scenarios of supply. In constructing the supply scenarios, account has been taken of the fact that in the past the percentage of candidates admitted to higher education institutions was relatively constant with a slight tendency to rise (about 18-20%). The evidence of some slow long-term tendency of a reduced youth population growth and some slackening in students’ intensive preference for university education - due mainly to the alternative possibilities that the new third-level Technological Educational Institution (TEI) offer - are also accounted for. But despite all these, the gap between the demand for higher education and the places at the university is still wide enough to allow these factors to have any dampening effect on enrollments. And, on the other hand, the establishment of new regional universities, which is in progress, will enlarge educational capacity. As a result, if anything, enrollments will be expected to rise. On the basis of these views, three scenarios of future enrollments are made: (i) constant future enrollments (scenario Ga); (ii) extrapolation by (19a) of historical trends (1961-1984) of graduates (scenario Gb); and (iii) separate projections of historical trends (1961-1984) of enrollments with (19) and of first year enrollments/graduates ratios with (18) to

Model& Future Higher Education-Labour Market Imbalances 11

obtain corresponding figures of graduates with (20) and (21) (scenario Gc).

Considerable difftculties present the prediction of emigration (brain drain) and of graduates coming from foreign universities, which are to some extent related issues. The evaluation of the future course of these variables, a grave problem by itself, is actually magnified by the lack of regular primary data.

For predicting emigration, we make the reasonable assumption that those graduates leaving Greece for working abroad, but also those that after the completion of their studies in foreign universities stay to work outside Greece, will not return at least during our target period. Under this assumption, the flow of degree holders into Greece will belong to the currently graduating classes abroad, which are predicted below.

On the basis of this assumption, we adopt two extreme scenarios, one repeating the high emigration proportions of 1961-1966 (scenario EGr), the other repeating the low emigration proportions of 1972-1976 (scenario EGz) and, a zero emigration assumption (scenario EGs). The so estimated graduates will be distributed by educational field according to the corresponding composition of the two periods, computing social sciences graduates - for which data are not available - as a residual.

Formally:

G,b(T,-, = Gd(T,-T) . EGn+-,) EG+m(, - t) G&- t) EGn(,o-,

(23)

where Gd(r,,-D, G+,,-,) = correspondingly, tOta

number of graduates during the projection period To - T and during the past period to - t (i.e. 1961-1966 and 1972-1976); EC,,,+,), EC++,_,) = correspondingly the total number and the + field graduate emigrants during to - t, and m = 1,2.

As for returning graduates from studies abroad, we assume that the total number of incoming graduates during the projection period will be of the same order of magnitude as those that have applied to the Greek authorities for accreditation of their degrees - required by relevant professional associ- ations or for taking employment in the public sector - during the period 1979-1984 for which some data have been collected. ”

Concluding remarks. Some qualifications concem- ing the size of our graduate supply estimates are in order in this context. If a number of graduates will

not appear in the job market, because they already have a good job, our predicted job-seekers may be overstated. But, on the other hand, predicted job- seekers are understated, considering the presently unemployed 24,000 graduates. This understatement is further enhanced by the expected TEI diploma receivers who will compete with the university graduates for some of the available jobs (in the last few years TEIs are enrolling about 25,000 new students annually).

All in all, we construct for Greece 8 single scenarios and 16 scenario combinations for labor demand and 6 single scenarios and 9 scenario combinations for labor supply, ending up with a 16 x 9 matrix of 144 cross-combinations. Below we present a taxonomic summary chart, including the projection bases of these scenarios, an identification of major variables and forces each combination may represent, and the analytical justification of their specification (Table 3).

A fair warning is that this table needs careful reading if any useful conclusions are to be drawn from it. This is only a rough plan, indicative of the potential impact on graduate demand and supply of some complexes of overlapping factors. No quantitative single-factor distinctions of effects can be made. The plan can only be used as an interpreter of Table 4 which presents the numerical estimates of the various scenarios. This way some overall quantitative effects can be attributed to certain nexus of forces by comparing different scenario combinations. The numerical estimates give the empirical content of Table 1 and the social cost-demand curves of Fig. Id, drawn in Fig. 2.

DISCUSSION OF RESULTS

In this section, we interpret our results against the theoretical setting of the paper and the multi- scenario approach, and we explore the significance of these results for educational planning and labor market policies for graduate imbalances.

Figure 2, which is, as noted, a numerical expression of Fig. Id, depicts the two extreme scenarios, and one in-between “more realistic” scenario of Table 4 to be discussed below, for showing respectively the boundaries and the more probable outcome of the various scenario combinations. Brief- ly, the absorption (or social cost-demand) curves AL and AU correspond to the lowest and the highest (upper) supply scenarios, and the curve AR to the

Tab

le 3

. A

tax

onom

ic

sche

me

for

rein

terp

retin

g an

d as

sess

ing

scen

ario

re

sults

Scen

ario

s of

dem

and

Proj

ectio

n ba

se

Var

iabl

es

appr

oxim

ated

an

d th

e A

naly

tical

ju

stif

icat

ion

for

the

Exp

ecte

d im

pact

on

na

ture

of

the

ir

pote

ntia

l ch

ange

s sp

ecif

ied

scen

ario

s gr

adua

te

empl

oym

ent

Indu

stry

sce

nari

os

A

Out

put

and

prod

uctiv

ity

1961

-198

4

B

Out

put

and

prod

uctiv

ity

1974

-198

4

Occ

upat

ions

sc

enar

ios

I O

ccup

atio

nal

com

posi

tion

1974

-198

6

II III

IV

Occ

upat

iona

l co

mpo

sitio

n 19

81-1

986

Shar

e of

inf

orm

atio

n oc

cupa

tions

in

tot

al

empl

oym

ent

1971

-198

1

Shar

e of

occ

upat

iona

l gr

oups

in

tota

l em

ploy

men

t in

in

form

atio

n oc

cupa

tions

19

71-1

981

Gra

duat

e sk

ill

cont

ent

scen

ario

s a

Shar

e of

gra

duat

es

in

cler

ical

occ

upat

ions

19

61-1

981

Gro

wth

an

d in

dust

rial

st

ruct

ure

Proj

ectio

n pe

riod

po

tent

ially

ch

ange

s of

a p

erio

d in

clud

ing

a in

clud

es

a pe

riod

of

rece

ssio

n an

d bo

omin

g an

d a

rece

ssio

nary

pa

rt

a pe

riod

of

gro

wth

Gro

wth

an

d in

dust

rial

st

ruct

ure

chan

ges

of a

rec

essi

onar

y pe

riod

Po

tent

ially

lo

w g

row

th d

urin

g th

e pr

ojec

tion

peri

od

Com

para

tivel

y ra

pid

incr

ease

in

th

e sh

ares

of

occu

patio

ns

with

a

high

con

tent

of

gra

duat

es;

rest

ruct

urin

g of

edu

catio

nal

skill

co

nten

t in

fav

or o

f tr

aditi

onal

oc

cupa

tions

Com

para

tivel

y sl

ow i

ncre

ase

in

the

shar

es o

f oc

cupa

tions

w

ith a

hi

gh c

onte

nt

of g

radu

ates

; re

lativ

ely

stab

le s

kill

cont

ent

of

occu

patio

ns

Com

para

tivel

y ra

pid

incr

ease

in

th

e sh

ares

of

grad

uate

s in

tot

al

empl

oym

ent

Mor

e ra

pid

incr

ease

th

an i

n sc

enar

io

III

in t

he s

hare

s of

gr

adua

tes

in a

ll oc

cupa

tiona

l gr

oups

Rel

ativ

ely

mor

e ra

pid

abso

rptio

n of

gra

duat

es

in c

leri

cal

jobs

, th

an

in t

he r

ecen

t pa

st

Publ

ic s

ecto

r w

ill h

ave

to c

ontin

ue

to c

ushi

on

surp

lus

grad

uate

s

Publ

ic s

ecto

r w

ill n

ot b

e ab

le t

o co

ntin

ue

acco

mm

odat

ing

larg

e nu

mbe

rs

of g

radu

ates

, du

e to

bu

dget

co

nstr

aint

s

Intr

oduc

tion

of n

ew t

echn

olog

ies

rela

tivel

y fa

ster

in

the

priv

ate

sect

or;

shif

t of

em

ploy

men

t fr

om

publ

ic t

o pr

ivat

e se

ctor

Mor

e ra

pid

intr

oduc

tion

of n

ew

tech

nolo

gies

th

an i

n sc

enar

io

III

and

rela

tivel

y hi

gher

shi

ft o

f em

ploy

men

t fr

om p

ublic

to

priv

ate

sect

or

Falli

ng r

elat

ive

wag

es i

ncre

ase

empl

oym

ent

of g

radu

ates

in

the

pr

ivat

e se

ctor

; su

bstit

utio

n of

gr

adua

tes

for

non-

univ

ersi

ty

educ

ated

w

orke

rs

A f

aste

r th

an i

n th

e re

cent

pa

st i

ncre

ase

Sim

ilar

to t

he r

ecen

t pa

st i

ncre

ase,

sl

ight

ly

high

er

than

sce

nari

o A

Rel

ativ

ely

high

in

crea

se,

high

er t

han

in

scen

ario

s II

and

III

Rel

ativ

ely

low

inc

reas

e,

low

er t

han

in s

cena

rios

I

and

IV.

Les

s ra

pid

incr

ease

th

an

in s

cena

rios

I

and

IV

Hig

her

incr

ease

th

an i

n sc

enar

ios

I, I

I an

d II

I

Mor

e ra

pid

incr

ease

th

an i

n sc

enar

io

b

_ -

.”

_.

,_

.,

- _

-

- -

7

. .

. __

_d_,

,--_

,m

*rn

--

- .,

.”

^.

‘>

- L>

--

- _

_

b Sh

are

of g

radu

ates

in

In

crea

se

of g

radu

ates

in

cle

rica

l A

mod

erat

e ef

fect

of

falli

ng

Mod

erat

e in

crea

se

cler

ical

occ

upat

ions

w

orke

rs

sim

ilar

to t

he i

ncre

ase

in

rela

tive

wag

es o

n th

e em

ploy

men

t 19

71-1

981

the

mor

e re

cent

pas

t of

gra

duat

es;

mod

erat

e su

bstit

utio

n of

gra

duat

es

for

non-

un

iver

sity

ed

ucat

ed

wor

kers

Scen

ario

s of

sup

ply

Proj

ectio

n ba

se

Var

iabl

es

appr

oxim

ated

an

d th

e A

naly

tical

ju

stif

icat

ion

for

the

Exp

ecte

d im

pact

on

na

ture

of

the

ir

pote

ntia

l ch

ange

s sp

ecif

ied

scen

ario

s gr

adua

te

supp

ly

En

rollm

ent

scen

ario

s

Ga

Con

stan

t re

cent

ye

ar

enro

llmen

ts

Gb

Gc EG

Lon

g-te

rm

1961

-198

4 gr

adua

tes

Lon

g-te

rm

1961

-198

4 en

rollm

ents

an

d en

rollm

ent/g

radu

ate

ratio

Gra

du

ate

emig

rati

on

sc

enar

ios

EG

I R

epea

ted

1961

-196

6 pr

opor

tion

of

emig

ratio

n

Rep

eate

d 19

72-1

976

prop

ortio

n of

em

igra

tion

E’S

Z

ero

emig

ratio

n

Dec

eler

atin

g gr

owth

of

stud

ent

age

popu

latio

n;

redu

ced

pref

eren

ce

for

univ

ersi

ty

stud

ies;

re

duce

d nu

mbe

r of

app

lican

ts,

but

still

ina

dequ

ate

educ

atio

nal

faci

litie

s

Incr

ease

d fl

ows;

sus

tain

ed

pref

eren

ces

for

univ

ersi

ty

stud

ies;

br

oad

diff

eren

ces

betw

een

appl

ican

ts

and

enro

llmen

ts

Incr

easi

ng

enro

llmen

ts;

impr

oved

qu

ality

of

enro

llmen

ts

Hig

h gr

adua

te

emig

ratio

n

Low

gra

duat

e em

igra

tion

Zer

o gr

adua

te

emig

ratio

n

Falli

ng r

elat

ive

wag

es;

dim

inis

hing

so

cial

sta

tus;

dif

ficu

lties

of

em

ploy

men

t; al

tern

ativ

e th

ird

leve

l no

n-un

iver

sity

st

udie

s

Sust

aine

d so

cial

and

pol

itica

l pr

essu

res

for

high

er e

duca

tion

for

mor

e pe

ople

; m

ore

regi

onal

un

iver

sitie

s;

wea

k im

pact

of

mar

ket

forc

es

Sam

e as

abo

ve f

or G

,; a

bette

r se

lf-s

elec

tion

of c

andi

date

s fo

r un

iver

sity

st

udie

s (h

ighe

r ra

tios)

Falli

ng r

elat

ive

wag

es a

pus

h fa

ctor

; de

crea

se

in g

over

nmen

t hi

ring

s; h

ighe

r de

man

d fr

om E

C

Falli

ng r

elat

ive

wag

es a

n in

cent

ive

for

hiri

ng g

radu

ates

; sl

ight

de

crea

se

in g

over

nmen

t em

ploy

men

t; m

oder

ate

dem

and

from

EC

Mor

e fa

vora

ble

cond

ition

s of

em

ploy

men

t in

pri

vate

and

pub

lic

sect

or;

sust

aine

d un

empl

oym

ent

and

queu

ing

Flow

s as

in

rece

nt

past

Incr

easi

ng

flow

s

Incr

easi

ng

flow

s

Falli

ng s

uppl

y

Stab

le o

r ri

sing

sup

ply

depe

ndin

g on

G

scen

ario

s

Stab

le o

r ri

sing

sup

ply

depe

ndin

g on

G

scen

ario

s

Tab

le 4

. R

atio

of

new

gra

duat

e jo

b-se

eker

s ov

er

job

open

ings

du

ring

th

e pe

riod

19

85-1

995

(all

fiel

ds o

f st

udy)

Scen

ario

s of

: Sc

enar

ios

of:

(i)

outp

ut-p

rodu

ctiv

ity

(a)

enro

llmen

ts

Abs

olut

e (i

i) o

ccup

atio

nal

com

posi

tion

(b)

emig

ratio

n (i

ii)

abso

rptio

n in

cle

rica

l jo

bs

num

ber

Ga-

EG

2 G

a E

dj In

crea

sing

or

der

of p

oten

tial

job-

seek

ers)

(i

ncre

asin

g or

der

of p

oten

tial

Ga-

EG

I of

job

G

c-E

Gl

Gb-

EG

l G

c-E

G2

Gc-

EG

3 G

b-E

G2

Gb-

EG

3 op

enin

gs

job

open

ings

) (1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) 19

85-1

995

1 A

-III

-b

1.69

1.

86

1.93

h

1.98

2

2.13

2.

21

2.29

2.

35

2.42

12

1,cO

O

s B

-III

-b

1.56

1.

72

1.78

1.

83

1.97

2.

04

3 2.

11

A-

11-b

1.

46

1.61

2.

17

2.24

13

1,08

0 1.

66

1.71

1.

84

8 1.

91

1.98

2.

03

2.09

14

0,00

0 3

4 B

- II

-b

1.36

1.

50

1.55

1.

60

1.72

5

1.78

1.

85

1.89

1.

95

150,

000

5’

A-I

II-a

1.

32

1.46

1.

51

1.56

1.

67

6 1.

73

1.80

A

- II

-a

1.24

1.

84

1.36

1.

90

154.

000

1.41

1.

45

1.56

1.

62

1.68

1.

72

1.78

16

5,00

0 9

7 B

-III

-a

1.23

1.

35

1.40

1.

44

1.55

1.

61

1.67

1.

71

8 1.

76

166,

000

2 B

- 11

-a

1.16

1.

28

1.32

1.

36

1.46

9

1.52

1.

57

1.61

1.

66

176,

000

fi

A-

I-b

0.90

0.

99

1.03

1.

06

1.14

10

1.

18

1.22

1.

25

1.29

22

7,00

0 g.

B

- I-

b 0.

85

0.94

0.

97

1.00

1.

08

11

1.11

1.

15

A-

I-a

0.79

1.

18

0.87

1.

22

240,

000

s 0.

90

0.93

12

1.

00

1.03

1.

07

A-I

V-b

0.

78

1.10

0.

86

I.13

25

9,00

0 0.

89

0.92

0.

99

&

1.02

1.

06

1.09

I.

12

260,

000

ti

13

B-

I-a

0.75

0.

82

0.85

0.

88

0.94

0.

98

1.01

1.

04

1.07

27

3,00

0 :.

14

B-I

V-b

0.

74

0.81

0.

84

0.87

0.

93

0.97

1.

01

1.04

1.

07

274,

008

a

15

A-I

V-a

0.

71

0.77

0.

81

0.84

0.

90

0.93

0.

97

16

B-I

V-a

0.

68

0.75

0.

99

1.02

0.

77

0.80

28

6,00

0 0.

86

0.89

0.

92

0.94

0.

97

301,

000

Abs

olut

e nu

mbe

rs

of j

ob-s

eeke

rs

1985

- 19

95

204,

000

225,

000

233,

000

240,

000

258,

000

267,

000

277,

000

284,

000

293,

000

Model&q Future Higher Education-Labour Market Imbalances

i i I I I 1 I

I I II II I I I II II I 2 3 4 5 67, 6 9 IO II l2J 1314, I5

15

(121) (131) 040)E4x15‘% (l65) i

(1761 E-27) (240)

(166)

(25&C273~4;66) (301)

Demand scencmos (jobopenings !n thousands)

Figure 2. Boundaries of demand scenarios delimited by the two extreme supply scenarios.

realistic supply scenario. Horizontal line aa’ represents the perfect match combinations. At the cross with the absorption curves, this line gives the alternative equilibrium points PI, Pz and P3.

Zones Zr, Zi and Zy - the major areas above the horizontal line aa’ and, alternatively, under the three absorption curves - contain all the scenario combinations (on both sides) that generate “surpluses” (pessimistic combinations). Zones Z,, Z$ and Z$ - below the horizontal line aa’ and, alternatively, over the three absorption curves - include the overabsorbing scenario combinations that may create shortages (over-optimistic, or un- realistic combinations). Zones Z,, Z; and Z;, that is, some small areas in the neighbor- hood of the alternative equilibrium points - to avoid giving a unique exact solution - include full (or almost fun) employment scenarios (matching combinations). Finally, zone Z4 - the area <+X,

delimited by the more realistic demand scenarios 5 and 7 and the more realistic absorption curve AR, represents the more probable outcome that we have guessed, resulting in a potential graduate surplus of 35 to 46%. This finding is consistent with our theoretical hypothesis that disequilibrium with overall graduate surpluses and economically unwarranted social costs is rather the rule in modern economies.

These results, referred back to Fig. 1 and Equations (3)-(6) can give some insights, or even some estimates, of the role of wages in reducing imbalances and indicate levels of wages that are consistent with empirical surplus solutions. As an example, take point A of the realistic surplus solution of Fig. 2, which represents the 5th serial combination of demand scenarios A-III-a in Table 4, generating 154,000 new job openings. From Table 3 read the conditions behind this combination and


the (Ga - EGz) supply combination that positions the absorption curve at 225,000 new job-seekers.

Translating now (not done here) the 1.46 ratio of social cost index of the A solution into monetary units, e.g. by calculating higher education expenditure per graduate, one can have: (i) an empirical wage-social cost relationship from (5); (ii) the unwarranted social cost of alternative non-optimum solutions from Table 4; and (iii) the set of conditions that generate these solutions from Table 3. Com- bining properly this information, various feasible solutions can be assessed, which can guide respec- tive policies in terms of wage changes and other labor market and education adjustments. For instance, the surplus and social cost of the A solution can be reduced more through a rise in the absorption of graduates, by moving along the realistic absorption curve towards line 1 and much less (or not at all) through a reduction of graduate supply, by turning to a lower supply scenario combination entailing a downward shift of the absorption schedule (notice that in this exercise no supply combination gives less than 204,000 new graduate job-seekers).

The extent that these actions can be pursued will depend on the degree that relative wages can be reduced and the degree that the substitution of graduates for non-graduates is feasible. All this must be considered in conjunction with any other changes that are required for shifting the absorption curve, as they are highlighted in the taxonomic Table 3.

Regarding the sensitivity of our results to alternative scenarios, it appears that, on the demand side, the numerical estimates are mostly diversified by the different assumptions on occupational structure changes. On the supply side, a relatively greater sensitivity is produced by the different assumptions on enrollments than on migration, though the latter’s impact is not negligible.

In contrast, the estimates do not seem to be significantly affected by the different output and productivity scenarios. The reason perhaps is that these two variables change analogously, one affect- ing the other, leaving relatively stable the growth rate of aggregate employment.

From a policy point of view, two general ways of using these results for educational planning can be suggested. The first is to ask the more general question of what are the consequences for policies of any of the scenario combinations, given the conditions of their materialization, and the more

specific question of which of all scenario combinations are more realistic or more feasible? The second way is to reverse somehow the argument and ask how this scheme can accommodate specific, perhaps desirable, perhaps forced upon, occur- rences?

Concerning the first approach, our estimates show that the perfect (almost) matching between job openings and job-seekers could be attained, for instance, with constant enrollments (scenario Ga) and with any of the emigration flows experienced in the past (scenarios EGi and EGz), as long as this is accompanied with rapid increases in the shares of occupations mostly employing graduates (scenario I). The matching can be completed even with a rather moderate absorption of graduates in clerical occupations (scenario b). Furthermore, if instead of the conditions in scenario I, the past change in the shares of information occupations continues, other things remaining equal, even shortages of graduates may occur.

Going to the upper extreme on the supply side (job-seekers), it is evident that continuation of long- term trends in enrollments and in graduate flows (scenarios Gb and Gc), accompanied with very low or zero emigration (scenarios EGz and EGJ, i.e. combinations 7-9, will leave surpluses under almost any of the assumptions about job absorption poten- tials.

Between these extremes, perhaps a more realistic development is represented by scenario III, i.e. the anticipation of an increase in the share of information occupations in total employment, at a constant composition of occupational groups in the employment of information occupations. Realistic seems also to be the assumption that the share of graduates in clerical occupations will be rapidly rising (scenario a). This can be justified by the possible further reduction in relative wages of graduates resulting from their oversupply, and the lowering of their job aspirations.

Thus, on the demand side, more realistic scenario combinations are the A-III-a and the B-III-a. On the supply side, the constant enrollments scenario Ga is the most probable one-not excluding though, the possibility of some moderate increase in enrollments. Borrowing some of the information of the chart in Table 3, this can be justified by the arguments given above concerning youth population growth, and students’ preferences that are related to some extent to their disappointment with the labor

Modelling Future Higher Education-Labour Market Imbalances 17

market situation. Also, the recent creation of the new non-university higher education technological institutions may offer students a satisfactory alternative to university education and perhaps better prospects for work. All these factors may somehow relieve the pressure of students on university openings. On the other hand, due to the deterioration in the quality of higher education, the establishment of the new regional universities would not necessarily entail an increase in enrollments but rather a decongestion of the big city universities.

Coming to brain drain, it does not seem possible that the exodus of the early 1960s will be repeated. However, some emigration, mostly in the form of non-returning graduates after the completion of their studies, will very likely take place, attracted not only by the potential relative availability of jobs but also by the much higher wages compared with those in Greece. Emigration might also be moder- ately encouraged by the free movement of labor in the EC countries.

Under these conditions, scenario EG2 is the more probable, which together with ‘the above enrollments scenario gives an overall realistic combination on the supply side depicted by Ga - EG2. This in conjunction with the “realistic” scenarios on the demand side entails a possible “surplus” of graduates of between 35 and 46% of the economy’s potential absorbing capacity. With the given number of job-seekers, a perfect match would require at least the 1974-1986 rapid increase in the share of professional and administrative workers (scenario A-I-b), or, more generally, the rapid increase in the occupational composition suggested by the intra- marginal scenario I, or scenario IV, concerning the absorption in information occupations.

Reversing now the argument, in line with our second suggestion of using these results, suppose, for example, that the public sector will not be able to cushion in the future graduate flows to the extent done in the past. Under what conditions in the whole set up of scenario combinations can a con- sequent increase of graduate unemployment be avoided? To begin with, a possible reduction of public employment may be reflected, as a push factor, in the high emigration scenario EG,, imply- ing simultaneously the inability of the private sector to absorb more graduates (scenario b). This, in turn, suggests that many of those studying abroad may decide to work outside Greece. But, alternatively, the private sector may be able to accommodate

relatively more graduates, perhaps induced by their falling relative wages. This is reflected in the higher absorption scenario a of clerical workers, which is consistent with the low emigration scenario EG2.

Sometimes such comparisons make sense only if they are used in conjunction with some additional knowledge. For instance, past experience for Greece shows that employment in “information” occupations has increased much faster in the secondary sector - primarily a private sector - than in the tertiary sector - in which government services are included (see Note 16). This information com- bined with the insights given by scenarios III and IV, reflecting an intensive introduction of new technologies, may lead to some inferences concerning substitution of private for public sector employment. In fact, this kind of substitution would be the result of the changing occupational structure and not the result of substituting university educated workers for others with less education, due to declining relative wages.

For assessing the sensitivity of estimates of individual educational fields, we select the above two extremes, an intermediate and the realistic scenario combinations (Table 5).

The disaggregated results by field show that, generally, the natural sciences group is likely to be the more problematic in matching jobs and flows. Medicine, social sciences and humanities appear to have to face a similar order of difficulties, with some differences among scenarios, whereas engineering seems to be in a slightly better situation. Concerning particular scenarios, it is evident that in the higher supply scenario, balance can hardly be reached with almost any of the demand scenarios, for any one of the five educational fields.

The so called realistic scenario combinations giving, on the other hand, the overall “surph~s” of 46% ranges by field, between 37% for medicine and 59% for natural sciences.

CONCLUSIONS _4ND POLICY IMPLICATIONS

Based on neoclassical labor market conditions and the human capital theory, and assuming a short- term inelastic supply of university educated workers, we build an empirical model for estimating future graduate imbalances, on an idea of a multi- scenario approach. This model generates empirical multi-value estimates of potential future absorption and of potential future production of graduates and


Table 5. Ratio of new graduate job-seekers over job openings for selected scenarios, during the period 1985-1995 (by educational field)

Demand scenarios Lower

Ga-EG 1

Supply scenarios Higher Intermediate

Gb-EG3 Cc-EG 1 “Realistic”

Ga-EG2

Absolute number of job openings

1985-1995

Lower Higher Intermediate Realistic






Natural sciences A-III-b B-IV-a A- I-b A-III-a

Engineering 29,800 37,100 A-III-b 1.66 2.07 I3 -IV-a 0.67 0.83 A- I-b 0.88 1.10 A-III-a - -

Medicine A-III-b B-IV-a A- I-b A-III-a

Social sciences A-III-b B-IV-a A- I-b A-III-a

A-III-b B-IV-a A- I-b A-III-a

All fields (Absolute number of job-seekers)

A-III-b B-IV-a A- I-b A-III-a

36,000 64,100 1.84 3.28 0.74 1.31 0.98 1.74 - -

45,800 39,600 - 2.34 - 19,600 0.94 - 48,800 1.24 - 36,800 - 1.59 24.900

28,500 34,700 - 1.58 - 18,000 0.64 - 44,800 0.84 - 33,800 - 1.51 22.900

26,000 39,700 1.49 2.27 0.60 0.91 0.80 1.22 - -

30,500 30,500 - 1.74 - 17,500 0.70 - 43,600 0.94 - 32,900 - 1.37 22.300

84,600 110,800 1.77 2.33 0.71 0.93 0.95 1.24 - -

104,200 86,400 - 2.19 - 47.700 0.88 - 118,600 1.17 - 89,408 - 1.42 60.800

27,600 41,300 1.52 2.27 0.61 0.91 0.81 1.21 - -

31,000 33,800 - 1.70 - 18,200 0.68 - 45,200 0.91 - 34,100 - 1.46 23,100

204,000 293.000 240,000 22s. 000 -

1.69 2.42 0.68 0.97 0.90 1.29 - -

1.98 0.80 1.06 -

- - -

1.46

121,000 301,000 227,000 154,000

provides a series of alternative optimum solutions of labor demand-supply map. Unlike that model, full employment and many more non-optimum which by calling for policies to achieve single-value solutions with unemployment. Our social cost inter- predictions “made [policy] choice unnecessary . . . pretation of graduate surpluses suggests an econ- [which is] one of the secrets for its abiding appeal” omically optimum allocation of resources to higher (Ahamad and Blaug, 1973, p. 323), our approach education, with zero unwarranted social cost, at the has no built-in policy biases and offers ample full employment level of graduates. possibilities for policy choices.

In contrast to the traditional manpower require- Our model can, in fact, be a considerable in- ments forecasting, our model does not ignore the formative source for complementing other policies role of relative wages and of factor substitutions and (e.g. incentives to change student’s preferences for allows for labor bottlenecks in production. The higher education, labor market policies for employ- impact of feedback effects is also considered in the ing more graduates etc.) with more specific objec- construction of the various scenarios and of the tives, through its assessment of the consequences of


such policies. More specifically, the model can be is the need that the model is used in conjunction instructive in picking affordable, feasible, or with other policy tools. This is necessary for pin- optimum solutions on either the demand or the pointing the feasible and consistent with other supply side and, conversely, it can contribute to the objectives policies for reducing imbalances in either scanning of the more effective channels of policy direction. In this respect, the model has more the action for attaining desirable policy objectives (e.g. properties of an exploratory sensitivity testing shifting graduate employment from the public to the device rather than the properties of a reliable rigid private sector). instrument for targeting or planning.

The width of the empirical boundaries and the size of the matrix of scenario combinations generally depend upon the degree of flexibility of the education-labor market system and on the nature of the various linkages in that system. In principle, therefore, countries with stable structures and flexible systems may be expected to have narrower boundaries and sparse maps, which makes for easier policy choices, whereas countries with unstable structures and rigid systems may have broader boundaries and dense maps, which renders more difficult the focus of policies for reducing imbalances.

The broader is the range of scenarios, the greater

Acknowledgements - Part of the work was done during the author’s Sabbatical Leave, as a Visiting Fellow at the Woodrow Wilson School of Public and International Affairs, Princeton University, for the academic year 1987-1988. The Leave was partly financed by a Study Mission Grant from the TECO Programme of the OECD, which is gratefully acknowledged.-The author wishes to thank P. Paoakonstantinou of KEPE for her valuable collaboration’ in the quantitative analysis and Maria Anastassiou for her assistance during the revision of the paper. He also would like to thank the anonymous referees of this journal for useful comments on a previous draft, and the Associate Editor, Stephen A. Hoenack, whose suggestions gave the paper a wider perspective. Of course, the author is solely responsible for the final product.

NOTES

1. Of the OECD countries, Denmark, Finland, Greece, Sweden and Turkey have numerus clausus control of admissions (OECD, 1983, p. 26).

2. Among others, see Wu, (1982) p, 112; OECD, (1983), p. 17; Sanyal, (1983). pp. 209, and 215; Sethi, (1983) p. 36.

3. See Freeman (1976) Ginzberg (1979) and Dore and Oxenham (1984). 4. For alternative definitions of “overeducation” and their weaknesses see Rumberger (1981) pp. 4-18, and Tsang and Levin (1985), p. 93.

5. Relative wages of graduates are continuously decreasing in many countries (see for instance, Teichler er al. (1980). p. 26; Rumberger (1984), pp. 443-444; Teichler and Sanyal (1982) p. 162; Kanellopoulos (1986), pp. 124-125; Freeman (1976), p. 198, and Psacharopoulos (1977), p. 80.

6. For the adequacy of the private rate of return to sustain demand for higher education, see, for various countries, Freeman (1976). p. 2; Teichler er al. (1980). p. 33; Sethi (1983), pp. 200-201 and Sanyal er al. (1981), pp. 245.275-276. In many countries - developed and developing - the social rate of return was found to be lower than the private rate of return (Psacharopoulos, 1973, p. 59).

7. For various countries, see OECD (1983), p. 14; Debeauvais (1984), p. 26; Rumberger (1984). p. 449; Baldridge et al. (1982), p. 5.

8. The ratio of entrants to higher education over applicants, was, in mid 1960s 0.23 for Brazil, 0.42 for Chile, 0.34 for India, 0.15 for Iran, 0.40 for Pakistan, 0.43 for Senegal, 0.33 for Vietnam and 0.25 for Greece (1975). And the ratios are dropping over time (Psacharopoulos, 1977, p. 74).

9. See, for instance, Bowles (1969), p. 177, AHAMAD and BLAUG (Editors) (1973), pp. 8-25 and Teichler and Sanyal (1982), p. 152. Some very simple extrapolations of enrollments have been made, among others, by the United States National Center for Education Statistics in 1976, using constant rates of enrollments. Also the Carnegie Foundation in 1975 and the Carnegie Council in 1980, have forecasted enrollments by using enrollment trends and external factors (see Baldridge er al. (1981), P. 9).

10. For earlier efforts to adjust enrollments to job opportunities in several countries see Debeauvais (1980), p. 26.

11. For opposing views on planning education on the basis of market requirements see, for instance, AHAMAD and BLAUG (Editors), (1973). pp. 50-54. For the criticism of manpower forecasting see, among others, Ahamad and Blaug, (1973), pp. 15-16; Blitzer (1975). p. 180.

12. Projections on the basis of alternative assumptions for testing the sensitivity of core variables has


been attempted (see Hollister (1966), p. 15) but only in very few cases (see Ahamad and Blaug (1973). p. 314).

13. According to our estimates about 75% of the economically active graduates, including teachers, are employed by the public sector at large. Tsoucalas (1983), p. 47 also makes a similar educated guess.

14. A recent study, based on a survey of third year lyceum students found that about 73% of them would choose disciplines suitable for public sector jobs. The study, in fact, argues that the drive for government employment is so strong that “it seems to function less as one alternative out of many and more as the norm against which other options are weighed” (Karmas er al., 1986, p. 81).

15. Four out of the five surveyed students expressed a preference for relatively easier to enter educational fields (Karmas et al., 1986, pp. 76-80).

16. In a nutshell, information occupations are those occupations that have the primary purpose of producing, processing and distributing information. Such occupations include scientific and technical workers that are primarily engaged in research, information gatherers, market search and coordination specialists and consultative services. They also include administrative and managerial occupations, process control and supervisory occupations and some clerical occupations. Finally, they include educators, communication workers and workers in information infrastructure. (For detailed description see OECD, 1981, p. 24). In Greece, the employment in information occupations, which include the majority of graduates, has increased between 1971 and 1981 at an average annual rate of 4.7% (7.1% in the secondary sector and 4.1% in the tertiary sector), compared with a corresponding increase of 2.6% in overall non-agricultural employment (1.7% in secondary and 3.3% in the tertiary sector).

17. The number was roughly 3500 annually of which 50% (1750 persons) were engineers and 25% (825 persons) were physicians (see Souris, 1985, p. 16). These figures compare with 1300 engineers and 1180 physicians that graduated from Greek universities in the same period.

This is, however, an underestimate of the actual size, because many others, especially in social sciences, that may be self-employed or work in the private sector, may not need an official accreditation of their degrees. To avoid such biases, the total figure of graduates is distributed by educational field according to the recent field composition of Greek students in foreign countries.

REFERENCES

AHA~~AD, B. and BLAUG, M. (Editors) (1973) The Practice of Manpower Forecasting: A Collection of Case Studies. Amsterdam: Elsevier.

ARROW, K. (1963) Uncertainty and the welfare economics of medical care. Am. econ. Rev. 53,941-973. ATKINSON, G.B.J. (1983) The Economics of Education. London: Hodder & Stoughton. BALDRIDGE, J.V., KEMERER, F.R. and GREEN, K.C. (1982) Enrolbnents in the Eighties: Factors,

Actors, and Impacts. Washington, DC: American Association for Higher Education. BLANK, R.M. (1985) An analysis of workers’ choice between employment in the public and private

sectors. Ind. Lab. Rel. Rev. 38, 211-224. BLITZER, C.R. (1975) Employment and human capital formation. In Economy-Wide Models and

Developmenf Planning (Edited by BLITZER, C.R., CLARK, P.B. and TAYLOR. L.). London: Oxford University Press.

BOWLES, S. (1969) Planning Educational Systems for Economic Growth. Cambridge Massachusetts: Harvard University Press.

DEBEAUVAIS, M. (1980) The role of international organizations in the evolution of applied comparative education. In Diversity and Unity in Education (Edited by HOLMES, B.), pp. 18-30. London: George Allen & Unwin.

DORE, R. and OXENHAM, J. (1984) Educational reform and employment - an overview. In Education Versus Qualifications (Edited by OXENHAM, J.), pp. 3-40. London: George Allen & Unwin.

EHRENBERG, R.G. and SMITH, R.S. (1988) Modern Labor Economics: Theory and Public Policy. Third Edition. Glenview, Illinois: Scott Foresman.

FREEMAN, R.B. (1976) The Overeducated American. New York: Academic Press. GINZBERG, E. (1979) Good Jobs, Bad Jobs, No Jobs. Cambridge, Massachusetts Harvard University

Press. G~nsos, N.P. (1977) Manpower Requirements and the Educational System in Greece. Report for the

Project of Post-secondary Education in Greece (Directed by G. Psacharopoulos and A. Kazamias, Athens (in Greek), in “collaboration” with R. Fakiolas).

GLYXOS, N.P. (1988) Rigidities and adaptability of the Greek labor market. In Essays in Memory of George Halkiopoulos, pp. 139-186. Athens: The George Halkiopoulos Foundation (in Greek).

HINCHLIFFE, K. (1987) Education and the labour market. In Economics of Educarion: Research and Studies (Edited by PSACHAROPOULOS, G.), pp. 141-146. Oxford: Pergamon Press.


HIRSCHMAN, A.O. (1970) Exif, Voice, and Loyalty. Cambridge, Massachusetts, Harvard University Press.

HOLLISTER, R.G. (1966) Phutning Education for Manpower Needs: A Technical Evnluation of the First Stage of the Mediterranean Regional Project. Paris: OECD.

KANELLOWULOS, K.N. (1986) Income and Poverty in Greece: Determining Factors. Centre of Planning and Economic Research. Studies Series, No. 22, Athens (in Greek).

KARMAS, C., KOSTAKIS, A. and DRAGONAS, T. (1986) Confronting an Uncertain Future: Occupational Demand and Work Values of Educated Greek Youth. Athens: Centre of Planning and Economic Research, Draft Copy.

KRUEGER, A.B. and SUMMERS, L.H. (1987) Reflections on the inter-industry wage structure. In Unemployment and the Structure of Labor Markets (Edited by LANG, K. and LEONARD, J.), pp. 17-47. Oxford: Basil Blackwell.

KRUEGER, A.B. (1988) The determinants of queues for federal jobs. Ind. Lab. Rel. Rev, 41, 567-581. LEVIN, H.M. (1987) Work and education. In Economics of Education: Research and Studies (Edited by

PSACHAROWULOS, G.). pp. 146-157. Oxford: Pergamon Press. LEWIS, W.A. (1962) Education and economic development. Int. Sot. Sci. I. 14, 685-699. MCMAHON, W:W. (1987) Externalities in education. In Economics of Education: Research and Studies

(Edited bv PSACHAROF~ULOS, G.), PD. 133-137. Oxford: Pergamon Press. MINISTRY 0; NATtONAL ECONOMY. liia&aI Accounts Statistics (Ivarious issues). NASSIAKOU, M. (1981) The tendency toward learning in the Greek countryside. /. Hellenic Diasporu 8,

63-74. NATIONAL STATISTICAL SERVICE OF GREECE, Cen.ru.r of Population, 1961, 1971, 1981. NATIONAL STATISTICAL SERVICE OF GREECE, Labour Force Survey, (1981-1986). NATIONAL STATISTICAL SERVICE OF GREECE, Educational Stutistics, (1960-1982). N.4noNAL SrAnsncAL SERVICE OF GREECE, Vital Statistics, (various issues). N.-vnoNAL SrAnsncAL SERVICE OF GREECE, Stutistictd Yearbook, (1981-1983). OECD (1981) Information Activities, Electronics and Telecommunications Technologies: Impact on

Employment, Growth and Trade, Vol. I. ICCP, No. 6, Paris. OECD (1983) Policies for Higher Education in the 1980s. Paris, OECD. PSACHAROPOULOS, G. (1973) Returns to Education: An International Comparison. Amsterdam:

Elsevier. PSACHAROWULOS, G. (1977) The perverse effects of public subsidization of education. Comparative

Education Review 21, 69-90. PSACHAROPOULOS, G. (Editor) (1987) Economics of Education: Research ond Studies. Oxford:

Pergamon Press. RUMBERGER, R.W. (1981) Overeducation in the U.S. Labor Market. New York: Praeger. RUMBERGER, R.W. (1984) The job market for college graduates 1960-1990. /. Higher Educ. 55,

433-454. SANYAL, B.C., PERFECTO, W.S. and ARCELO, A.A. (1981) Higher Education and the Labour Market in

the Philippines. Paris/New Delhi: UNESCO. SANYAL, B.C. (1983) University Education and Graduate Employment in Sri-Lanka. Colombo:

UNESCO/Marga Institute. SETHI, J.D. (1983) The Crisis and Collapse of Higher Education in India. New Delhi: Vikas Publishing

House PVT Ltd. SOURIS, P. (1985) Reinsertion of Scientific Manpower in Greek Labour Market after Completion of

Post-graduate Studies Abroad. Preliminary Text, mimeo (in Greek). TEICHLER, U., HARTUNG, D. and NUTHMANN, R. (1980) Higher Education and the Needs of Society.

Windsor: NFER Publishing Company Ltd. TEICHLER, U. and SANYAL, B.C. (1982) Higher Education and the Labour Market in the Federal

Republic of Germany. Paris: The‘UNESCG Press. TSANG. M.C. and LEVIN. H.M. f1985) The economics of overeducation. Econ. Educ. Rev. 4.93-104 TSOU&AS, C. (1981) Some aspects of “over-education” . m modem Greece. J. Hellenic Diaspora 8,

109-160. TSOUCAJAS, C. (1983) Social implications of the public sector as employer in the post-war Greece.

Greek Review of Social Research 50 (in Greek). WERY, R. (1978) Manpower forecasting and the labour market. Int. Labour Rev. 117, 331-342. Wu, CHUNG-LIH (1982) An Econometric Analysis of the Demand for Higher Education in the U.S.:

I947-1978. Nankang: The Institute of Economics, Academia Sinica, Monograph Series, No. 25.


APPENDIX

Educational Fields The academic departments included in each educational field are as follows:

Nurural Sciences. Biology, geology, physics, chemistry, pharmacy, agronomy, forestry, mathematics.

Engineering. Civil, mechanical, electrical, chemical, shipbuilding, mining and metallurgy, topography, architecture.

Medicine, Denttitry, Veterinary. Medical school, dentistry, veterinary practice.

Low and Social Sciences. Law, economics, business, political sciences.

Humanities. Theology, history, archeology, Greek studies, English, French, German literature, Greek literature, fine arts, linguistics.

Testing the predictive capability of the employment projection technique According to Test I, as described in the text, there are very slight differences in either direction

between actual and predicted values in the various occupational groups, ranging, on the average, for the period, between 1 and 1.8%, giving, on balance, a roughly accurate overall prediction.

Regarding Test II, contrary to evidence elsewhere that shows “the glaring inaccuracy of employment projections” (Wery, 1978, p. 333) of occupational structures, in the case of Greece the projected values show a remarkable proximity with the actual values.

Below are presented only the average figures of Test I for the period 1981-1986, but the predictions are good even for each single year with a highest gap - of all occupational groups and of all years - of about 4% for clerical workers in 1986.’

(values in thousands)

Tests

Professional and administrative Rest of

workers Clerical workers occupations All occupations Actual Predicted Actual Predicted Actual Predicted Actual Predicted

Test I Average total employment 1981-1986 344.1 337.5 301.3 305.9 1619.0 1620.8 2264.4 2264.2

Average occupational shares 1981-1986 15.2 14.9 13.3 13.5 71.5 71.6 100.0 100.0

Test II Occupational shares (1981) 11.8 11.3 8.9 8.8 79.3 79.9 100.0 100.0

Sources: NSSG. Labour Force Surveys 1981-1986. Glytsos, 1977, pp. 31-32 (for projections for 1981).

l Graduates are not used in this test because relevant data are not available for urban and semi-urban areas for which occupational employment of row 1 is referred to.

Data and Sources Economic data. Output: Output by one-digit industry, at constant 1970 prices, is obtained from Greece’s National Accounts Statistics. Employment: For the largest part of the period, i.e. 1961-1981, employment data for the total of

Greece are only available from the 1961, 1971 and 1981 population censuses. Only for a few industries, such as manufacturing and trade, some additional, not very comparable, data are available for some

Modelling Future Higher Education- Labour Market Imbalances

intervening years from the Cenrur of Industry. We therefore constructed a series of data for the period l%l-1981, by interpolation from the data of the population census, checked whenever possible with the industry census data. The interpolation was based on the actual annual change of output and an average elasticity of employment with respect to output, calculated from the census data and the corresponding years’ output data. Then for the period 1982-1984 we used the primary employment data provided since 1981 by the Labour Force Surveys.

In the alternative productivity projection of the period 1974-1984, we also experimented, in addition to the constructed series, with the employment data of the urban and semi-urban (communities with over 2000 population) regions of the country, given in the Employment Surveys (after 1981 Labour Force Surveys). These data were first adjusted on the basis of the relation of employment of the urban and semi-urban regions to total employment by industry, obtained from the population census of 1971 and 1981. For most of the industries, the results were similar with those obtained from the constructed data series.

Occupational structures: Due to the lack of a long series of employment by occupation and industry, only aggregate estimates of the occupational composition of total employment were projected on the basis of the 1974-1986 figures of the Employmenr Surveys of the urban and semi-urban regions. Only the projections of the period 1981-1986 refer to the total of Greece on the basis of data from the Labour Force Surveys, that started in 1981. Finally the data for the “information” occupations were assembled from the 1971 and 1981 Population Censmes.

Economically active graduates: The basic year 1984 stock of economically active graduates by educational field is not available. An estimate was made by combining the data of the 1981 Population Census and the estimate made for 1984 of total graduates (active and inactive) by field, as described below. This was done by taking the field composition of each age bracket of the estimated total number of graduates and apply it to the corresponding 1981 number of total active graduates by age group, given in the census.

Migration: Migration data are taken from the Sratisticul Yearbook and refer to the figures collected up to 1977. According to these precarious migration data, collected by the Statistical Service of Greece at the country’s gates to the world, the “Professional” graduates that emigrated (including those that left for postgraduate studies) ranged between 6 and 10% of the graduating classes, during the period 1958-1977. Particularly, in the peak emigration period of 1961-1966, the equivalent of 14.6% of the graduating classes, on the average emigrated (18.0% of the natural sciences field, 31.5% of engineers, 23.2% of physicians and 24.2% of humanities majors). Then, at the later low migration period of 1972-1976 only an average of 4.1% of the graduating classes emigrated (2.7% of natural sciences, 10.0% of engineers, 6.8% of physicians and 9.8% of teachers). To the extent that “non-professional” graduates emigrated, these figures give an underestimate of the true graduate emigration.

Corresponding return data, referring only to the period 1968-1977, show that the equivalent of 5-9% of the graduating classes (including those that returned from their undergraduate or graduate studies) returned home.

Demographic data. Volume of population with university Degrees: The total figure of population with university Degrees

(stock of graduates) is not readily available for the base year 1984. It is, however, given by field of study in the 1981 Populafion Census. An estimate for 1984 was obtained by adding to the 1981 stock the flows of new graduates of 1982, 1983, and 1984 and subtracting a corresponding figure for deaths.

Deaths of graduates: The death rates by wide age groups- determined by the age break-down of the graduate stock - i.e. below 44,44-64 and 65 and over, were estimated as an average of the years 1980, 1981 and 1982. The number of deaths was taken from Vital Statisrics and the mid-year population from the Statirtical Yearbook.

Educational data. Graduate flows from Greek universities: First year enrollments and graduate flows by academic

department and school, for the period 1961-1984, are taken from the Education Statistics. Graduate flows from foreign universities: No primary data series of graduates from foreign

universities are available. Their flow was estimated on the basis of applications of graduates to the government accrediting body DIKATSA during the period 1979-1984. These data were collected by P. Souris and were taken from Souris (1985). Data of Greek Students in foreign universities are taken from the UNESCO, Stutistical Yearbook. Some unpublished figures are also taken from the Bank of Greece registration of student foreign exchange.

Note: All the above Greek statistical publications are of the National Statistical Service of Greece, except of the National Accounts Statistics, which is of the Ministry of the National Economy.

23

Modelling future higher education-labor market imbalances: A multi-scenario approach

Documents