Reflections on the present state of the brain drain and a suggested remedy

Evidence of the Magnitude of the Brain Drain During the 1960s when public concern with the brain drain was at its peak, the most readily available, seemingly reliable and comprehensive

Reflections on the Present State of the Brain Drain and a Suggested Remedy

H E R B E R T G G R U B E L

PUBLIC attention to the problems raised by the international migration of highly skilled persons, which came to be called the "brain drain ", has fluc- tuated During the 1960s newspapers and other publications often carried stories about the brain drain Governments and international organisations, such as the United Nations, were induced to gather information on the magnitude of the problem and to propose policies for dealing with it However, after the economic recession of 1971 in the United States, the accelerating inflation in Western countries and the increase in the price

of oil, the brain drain appeared to have ceased to exist as an issue of broad public concern Then, in 1974 two major official documents concerning the brain drain were published) suggesting that there remain grounds for continued public concern In February 1975, a conference was held in Bellagio, Italy, under the chairmanship of Professor Jagdish Bhagwati of the Massachusetts Institute of Technology on " T h e Brain Drain and Income Taxation ", to discuss the merits of imposing an income tax sur- charge on highly skilled persons from poor countries resident in industrial countries The revenue from this tax surcharge was to be passed on either

to an international agency, such as the United Nations, or to the govern- ments of the countries of origin of the highly skilled migrants)

The renewed interest in the brain drain expressed in these recent pub- lications and conference has prompted me to reflect on the issues in the light of some empirical observations and theories which were developed in the 1960s, but which have not become widely known

Evidence of the Magnitude of the Brain Drain

During the 1960s when public concern with the brain drain was at its peak, the most readily available, seemingly reliable and comprehensive statistics on the flow of highly skilled persons to the United States were those published by the United States Immigration and Naturalization

1 United Nations Conference on Trade and Development, " The Reverse Transfer of Technology: Economic Effects of the Outflow of Trained Personnel from Developing Countries " (Geneva: U N C T A D Trade and Development Board) GE74-45088, 15 July,

1974 (mimeographed); U.S House of Representatives, Committee on Foreign Affairs, Brain Drain: A Study of the Persistent Issue o] International Scientific Mobility (Washington: U.S Government Printing Office, 1974)

2 The papers presented at the Bellagio Conference have been published in a special issue

of World Development, Vol III, 10 (November 1975), pp 677-763; and in Bhagwati,

J N and Partington, M (ed.), Taxing the Brain: A Proposal (Amsterdam: North Holland Publishing Company, 1976), and The Brain Drain and Taxation : Theory and Empirical Evidence (Amsterdam: North Holland Publishing Company, 1976) The proposal for the brain drain tax was first made by Bhagwati, Jagdish and Dellalfar, William, " The Brain Drain and

Income Taxation ", World Development, I, 1 and 2 (February 1973), pp 91-98

Service ~; these were used in the reports published by the United States National Science Foundation? They continue to be used uncritically by many analysts and they play a key role in the publications of the United Nations Conference oll Trade and Development of 1974 5 and the United States House of Representatives ~ Yet, these statistics are very misleading; they overstate to a significant degree the gains of the United States and the losses of the rest of the world in migrants human capital because they

do not include information about re-migration or the place where the skilled person received his education The reasons for this omission are simple The United States immigration authorities do not keep a record of highly skilled persons leaving the United States and they do not distinguish in their records between persons who were educated in the United States or abroad Similarly, foreign governments do not record the return of native-born persons who in most countries can re-enter simply by showing their pass- ports

In order to obtain reliable statistics on re-migration from the United States, investigators in some Western European countries had to engage

in expensive and painstaking search of other records, for example, records

of church registration in Sweden, military records in Switzerland, and special computation by the United Kingdom Ministry of Technology, in order to obtain information about returned emigrants, z

On the basis of these investigations, re-migration into certain countries during the period from 1958 to 1969 accounts for between 40 and 90 per cent of recorded gross migrations, s In the case of Dutch engineers, in a two-year period more returned to Holland from the United States than left for the United States Probably immigration into the United States is still overestimated because during the period flows were increasing rapidly and immigrants tended to stay a number of years in the United States

To count only inflows and outflows over a given period neglects the pos- sible re-migration of the larger numbers in subsequent years Even if net immigration were zero, rising gross flows and a sojourn of several years in the United States would show a net gain for the United States within any given period of time Many highly skilled Swedish migrants are very mobile internationally, adhering to a pattern which Dr, Grran Friborg has called a ~ yo-yo" movement, by which he means that some individuals enter and depart several times within a given period These in the past

3 For an excellent collection of information on sources of data for the estimation of brain

drain flows see Friborg, G~Sran (ed.), Brain Drain Statistics: Empirical Evidence and Guide- lines, Report 6 (Stockholm: Research Council, Committee on Research E c o n o m i c s - - F E K ,

1975); and Organisation for Economic Cooperation and Development, The Utilization of Highly Qualified Personnel, Venice Conference, 25-27 October, 1971 (Paris: OECD, 1973)

4 National Science Foundation, Scientists, Engineers, and Physicians from Abroad: Trends through Fiscal Years (Washington: National Science Foundation, various years)

5 Ibid

6 Ibid

r For more details on these studies see Friborg, G., op cit., and The Utilization of Highly Qualified Personnel

The Present State of the Brain Drain 211 have sometimes been counted each time as immigrants, but never as emigrants from the United States There is every reason to believe that for many of the poorer countries also net flows are much smaller than gross flows ~

FIGURE 1

Flows of Foreign High-Level Scientific and Technical Personnel

to and Jrom North America (Selected countries, years and occupations)

F R O M AND TO T H E UNITED KINGDOM

1961-69 (All QSE) ENGINEERS

#

All graduatt

FROM AND TO S

(Swedish and fore

r Z E R L A N D

0 t o 5 5 )

5 ~

ENGINEERS

25

435

~TURAL SCIENTISTS

FISTS

z t z 323 ENGINEERS 1963-64

F R O M AND TO THE NETHERLANDS

(Dutch and fereigners)

SOURCE: Organisation for Economic Cooperation and Development, The Utilization of Highly Qualified Personnel, Venice Conference, 25-27 October, 1971 (Paris: OECD, 1973),

p 34

The magnitude of losses of human capital incurred through the emigra- tion of highly skilled persons to the United States is over-stated by United States statistics of immigration by the further fact that they do not include the country where the immigrant was educated The immigration to the

9 Friborg, G6ran, A First Preliminary Report concerning an Investigation by the Com-

mittee regarding the Migration of Scientists to and Jrom Sweden (Stockholm: Research Council, Committee on Research Economics, 1968), Report 20

United States of a person wholly educated abroad and of a person whose education was undertaken in and perhaps even paid for by the United States should be treated differently Despite the deficiency of American statistics on this matter, there are some indirect clues to the relative magnitude of the foreign and American education of immigrants to the United States In 1970, of the 13,337 scientists and engineers who received immigrant visas, 5,470, or 41 per cent., were in the United States at the time they had their visas changed from visitors', students' and other classes

of visa 1~ The overwhelming proportion of these changes in type of visa were for students who had received a significant and the most expensive part of their education in the United States, Furthermore, some of the remaining 59 per cent received their advanced training in the United States but applied for immigrant status while abroad because of the particular requirements of the type of visa under which they had entered the United States initially as students Others received their higher educa- tion in another industrial country such as Canada, Great Britain or Western Germany

With the notable exception of scientists from the United Kingdom and Switzerland, between 30 and 85 per cent of foreign-born holders of doctorates in the United States received at least their final degree in the United States In the case of the poorer countries in the list, for example, Greece and Turkey, the proportion of doctorates acquired in the United States is above three quarters (Table I),

I do not deny that there might be a "brain drain problem" deserving the attention of economists and politicians It is quite possible that the migration of highly educated persons should be controlled, even though more accurate analysis of the statistical data shows that the amount of permanent emigration by such highly educated persons is smaller than is often said on the basis of uncorrected statistics However, any specific policies for controlling the brain drain involve some social costs These costs should be compared rationally with the gains derived from reductions

of the brain drain through these policies Certain costs considered to be wo,rth incurring at a certain level of the brain drain may no,t be worth incurring below that level

Measuring the Costs in Welfare of the Brain Drain

On the basis of careful statistical techniques and using data on the composition by age of immigrants to the United States and on earnings of professional persons in the United States, an estimate of the value of the personal incomes earned by the highly educated immigrants from poor countries to the United States in the year 1970 came to $3"6 billion in one year alone, n Another estimate of the "replacement cost" of the "human

l o T h e U t i l i z a t i o n o J H i g h l y Q u a l i f i e d P e r s o n n e l

The Present State of the Brain Drain 213

Country

TABLE I

Place of Doctoral Study of Foreign-Born Scientists with Doctorates

in the United States in 1966 (%)

Area of Award

capital stock" acquired by the United States through the immigration of highly educated persons from poor countries during the 1960s came to more than $2 billion I~

It is however, wrong to put these figures forward as equivalent to gains

in welfare for the United States, as these estimates do It is especially mis- leading to compare these statistics with the figures of American aid to poor countries in the past and to argue that the United States has " s a v e d " an equivalent amount of investment in human capital In the calculus of wel- fare, it is important to consider the most fundamental fact that nearly all

of the earnings of the "brain drain" immigrants in the United States were received by them as individuals, to be consumed or disposed of as they saw fit Similarly, the value of the human capital embodied in the highly educated immigrants yielded a return which accrued to its holders As a

12 Reubens, Edwin J., " Professional Migration from the Less Developed Countries: Recent Trends, Relative Magnitudes, Economic Significance ", paper presented at the the Bellagio Conference (mimeographed)

result, residents of the United States received no direct benefits from the earnings of the highly skilled immigrants from poor countries

While it is true that the work of highly educated persons may result

in externalities, that is, benefits accruing to other residents of the United States for which they are not compensated, the importance of these extern- alities is often exaggerated For example, in the case of externalities associated with pure and therefore non-patentable knowledge produced by scientists, the benefits know no borders and are available to the migrants' countries of origin and of residence Highly educated persons also tend to pay proportionately more taxes than they claim in government services, thus contributing to the subsidy of recipients of low incomes in their new country of residence, while making such subsidies more difficult in their country of emigration While the migration of highly skilled persons tends

to create benefits of this sort in the countries receiving them and costs in the countries which they have left, the magnitude of these benefits is surely equal only to a small fraction of the migrants' incomes or the value of their education

Although the preceding argumeaats about the nature of the effects

of the brain drain on wdfare and about their measurements have been discussed widely and well in the literature produced during the 1960s, ~3 the studies produced by the United Nations Conference on Trade and Develop- ment as well as by Dr Edwin Reubens, which were published in the middle

of the 1970s, are evidence of the fact that many persons still misinterpret existing statistics and overestimate by a very large margin the cost of the brain drain Such overestimates can lead to the insrtutio,n of policies which will attempt to stop the brain drain or alleviate its effects and which will involve costs exceeding their benefits

Geometrical Analysis of Effects on Welfare o] Migration of Human Capital i~

An analysis of the effects on welfare of migration, or the flow of human capital, shows that, under a reasonable set of assumptions and using some rough empirical estimates, it is possible that in the case of some countries with large populations the emigration of highly educated persons may raise

rather than lower the incomes per caput of the remaining population

Assuming the absence of externalities, my analysis is conducted without relying on the theory of the marginal productivity of wages

For this purpose, consider Figure 2 in which Quadrant I shows the production function between labour and physical capital, represented in the conventional manner by isoquants Y0, Y1 and Y~ The production

13 T h e U n i t e d States G o v e r n m e n t p u b l i c a t i o n B r a i n D r a i n : A S t u d y o [ t h e P e r s i s t e n t

I s s u e o f I n t e r n a t i o n a l S c i e n t i f i c M o b i l i t y , c o n t a i n s a n u p - t o - d a t e set o f r e f e r e n c e s t o this l i t e r a t u r e

~a T h e a n a l y s i s in this s e c t i o n d r a w s o n a n u n p u b l i s h e d m a n u s c r i p t w h i c h I h a v e w r i t t e n

The Present State of the Brain Drain 215 function is assumed to be linearly homogeneous The innovation of the present approach lies in its view of the determination of output on the assumption that labour can perform its tasks only by cooperation with a certain number of engineers, and that the education of labourers into engineers requires the use of human capital which, according to the Fisherian view of capital, is part of the country's capital stock Quadrant

II of Figure 2 measures the number of engineers on vertical axis and shows the number of engineers required per unit of labour as the slope of the line

OF The total labour force is assumed to be equal to O-TLo, which is efficiently allocated by having OLo labourers and OE0 engineers In Quad- rant III the horizontal axis measures units of human capital and the line

OG expresses the number of units of human capital required to make an

FIGURE 2

The Efficient Allocation of Human and Physical Capital

Physical Capital

Human

Capital

TK,,

0

c /

Lli L0i

i /

/ :

TLI

9 y~

Y; Y+

TLo

Labour

III

TLI

TLo [ Engineers

engineer out of a labourer The OEo engineers require OC0 units of human capital With total capital given as O-TKo, the amount of physical capital left ovei" to equip labour is OK0, giving us the combination of inputs: A, in Quadrant I, which must lie on an isoquant such as YoYo Now consider the effect of the emigration of TL1-TL0 units of labour from this country In the long-run equilibrium the new amount of labour

is OL1, of engineers is OE1 and since the capital stock has remained un- changed and fewer engineers and therefore fewer units of human capital are required, the physical capital per worker is raised to OK1

The new point in the combination of factors is B which must be on an isoquant such as Y~Y~ below YoYo because the reduction in the factor input of labour combined with unchanged capital necessarily lowers output under the assumption of a linearly homogeneous production function Now turn to the analysis of the effect of the emigration of TL1-TLo engineers The main difference between this and the preceding case is that emigration of engineers removes human capital and therefore reduces the country's total capital stock This latter fact is shown as the inward shift of the total capital line to TK1-TK, Under the assumption of a fixed engineer-labour coefficient, the emigration of the engineers in the new equilibrium leads to further output point D in Quandrant I, which must lie on an isoquant such as Y2Y2, which must be below Y~Y1 because at point D the country has as much labour but less physical capital than at point B Under the assumption of a linear homogeneous production function, output at D must be below that at B

This model of the effects of migration on total output shows the nature

of the brain drain facing all countries, whether they have a market econ- omy or a planrted economy, in the sense that it discloses the role of skilled manpower in the determination of the productivity of labour and capital

It shows the way in which social capital can be distributed between either human or physical capital and it demonstrates that the total output of a country falls more when a skilled person leaves than when an unskilled one does so The model has a number of applications, such as the demonstra- tion of the cases of short-run adjustment arising from the emigration of skilled and unskilled labour, and it could readily be expanded to permit variable ratios ,of engineers to workers and human capital for different requirements of human capital by engineers

The line OM0 in Quadrant I of Figure 2 indicates the country's ratio of total capital to total work-force in the initial situation of equilibrium since the point M0 is the co-ordinate of O-TL0 and O-TK0, the total work-force and capital stocks, respectively Figure 3 shows that generally output per worker is an increasing function PoPo of the ratio of capital to workers in

a country This functional relationship does not need to be modified in the present analyses of human capital and two types of labour as long as the specified criteria of efficiency are met It may be assumed that initially therefore the country's output per worker is (Y/W)0 at the capital-worker

The Present State of the Brain Drain 217 ratio (K/W)o equal to O-TK0/O-TL in Figure 2 Now when TL1-TL0 labourers leave, in the new equilibrium the capital-labour ratio rises to OM1 Using this fact in Figure 3 it follows clearly that the emigration of labour raises income per caput, even though it lowers aggregate income Since it is accepted generally that welfare is an increasing function of income per caput, emigrating workers taking along no capital raise the welfare of the people remaining behind as well as their own and for this reason there is normally in most countries no opposition to general emigration

Consider now the case where engineers emigrate and take along some

of the country's capital stock This form of migration may either raise or lower income per caput of those remaining behind, depending on the relationship between the human capital per engineer and the country's ratio

of capital to labour Figure 2 shows a situation where the units of capital required to train an engineer are fewer than the country's average stock

of capital per worker This property holds geometrically whenever the angle labelled a and expressing the average capital-labour ratio is smaller than the angle h, the human capital per engineer, assuming equality of units of measurement of both forms of capital and labour and engineers For this reason the ratio OM2 in Quadrant I is between OMo and OM1 Another way of describing this property is that in the case shown in Figure

2 the proportionate decrease in labour (TL1-TLo)/O-TLo is necessarily greater than the proportionate decrease in capital (TK1-TK0)/O-TKo, so that OM2 must be steeper than OM0

The level of education at which the emigration of educated persons raises rather than lowers the average worker's endowment with capital is

an important question for analysis I have made an attempt to solve this question for India? 5 In about 1963 in India the average value of physical capital per worker was $1,197 and of human capital $66, giving a total average capital stock of $1,263 The cost of education was $28 per year for primary and $189 per year for secondary education, counting both the direct costs and earning foregone in about 1960-66 Under the: assumption that each Indian student spends six years at primary and six more at secondary school, the human capital embodied in a high school graduate is $1,302 These estimates are subject to a large margin of error and not too much faith should be placed in them Nevertheless, these results suggest that during the 1960s the emigration of an Indian student who had completed his secondary school education left unchanged the income per caput of those who remained behind

There are two reasons for believing that actually his emigration may raise Indian income per caput For one thing, our calculations dealt with the average costs of education and marginal cost is likely to be much lower

5 The following estimates draw on basic data found in Psacharopoulos, George, with

K Hinchliffe, Returns to Education: An International Comparison (Amsterdam: Elsevier

Scientific Publishing Company, 1973)

(Y/W)~

(Y~W)~

(Y/W)~

FIGURE 3

Capital-Labour Ratios and Output per C a p u t

Output/Worker

;., ,,.::: :~

oo~ ~j#

Po

)

Total Capital/Worker

(Y/Cap)*

(Y/Cap),

Income/Caput

FIGURE 4

Optimum Population

2