Defence, Human Capital and Economic Development in South Africa

Table 1
Economic Performance and Labour Force Attributes

Development category/ Region	GDP per capita (US $)	Primary School enrolement (%)	Secondary School enrolement (%)	Illiteracy rate (% of population aged 15+ which is illiterate)
High income	19 590	100+	91,6	0
Upper middle-income	3 450	100+	56,3	15,1
Lower middle-income	1 530	100+	58,3	22,9
Sub-Saharan Africa	338	69	18,2	51,0

Consequently, the definition of a nation's wealth has widened to accommodate not only physical capital but also human capital as an independent factor of production required to achieve high and sustainable economic growth rates. In recognition of this relationship, developing nations have, in varying degrees, attempted to stimulate the accumulation of human capital through public education expenditure as well as government spending on health and related social services. The success of government initiatives in this regard can be gauged from the fact that by the late Eighties an average of 4,0 per cent of all developing country GDP1 was allocated to public education and health expenditure. However, although this represents an improved performance on previous standards, comparable data show that the same countries allocated almost 5,0 per cent of GDP to defence.2 (World Bank, 1990: 141-143).

This is indicative of the apparent conflict within government budgets between, on the one hand education and health expenditure and, on the other, defence spending. However, the crowding-out effect3 of defence spending on public education and health is not that simplistic. In practice, there are numerous channels by which defence spending may have an impact on the stock and quality of human capital, some of which may ultimately be of a positive nature. The aim of this article is to examine the interactions between defence, human capital and growth in the context of a developing economy, with a view to quantifying the effects for the South African situation.

THEORETICAL CONSIDERATIONS AND EMPIRICAL EVIDENCE

As stated in the introduction, investment in human capital enables a country to effectively exploit the benefits of technical and technological advancement, as well as maintaining that advancement. This view, shared by most development economists, is succinctly summarised by Harbison (1973: 3) who states that "... human resources - not capital, nor income, nor material resources - constitute the ultimate basis for the wealth of nations ... Clearly, a country which is unable to develop the skills and knowledge of its people and to utilize them effectively in the national income will be unable to develop anything else."

Although opinions differ regarding the most appropriate ways of fostering and inculcating the attributes assumed to accrue from an improvement in human capital,4 it is generally accepted that at least primary and preferably also secondary school education, as well as health expenditure, are essential first-level requisites.

A priori evaluations of the ultimate net impact of the defence burden on the development of human resources are difficult as the effect is country specific. Before attempting to quantify the nature of the trade-off in South Africa, a number of channels through which an increase in defence spending may influence the growth of human resources will be examined.

Spin-off Effects

For various reasons developing economies may suffer from a shortage of labour.5 In that case the military may be an effective instrument in mobilising human resources, promoting a disciplined work-ethic and fostering basic work skills. Moreover, as postulated by Benoit (1973), technical training is a vital element of military service and thus augments the skill content of the existing labour force. In this regard Gold and Adams (1990: 283) suggest that the private sector underinvests in research and development (since the gestation period of such investment is lengthy) so that the pre-emption by the public sector in general and the military, in particular, of the resources available for research and development is justified. Military research and development is, in turn, inextricably linked to an increase in both the quantity and quality of technological expertise and skills. Thus, even though the primary purpose of the military is not skill creation, it can make a significant contribution in that regard.6

The above reasoning may, however, be questioned at a number of levels. First, if a country's armed forces are composed of a large professional core (as opposed to being a volunteer organisation), doubts may exist as to the appropriateness to the civilian economy of educational (and other) training provided by and for the military (compare, for example, Ward et al, 1991). In addition, Ball (1983) questions the value of military-induced training and education in a country where unemployment is so high that the soldier will not be able to find gainful employment at the termination of his military service. As far as research and development is concerned, it is often suggested (compare, for example, Gold and Adams, 1990) that defence spending has a negative impact on non-defence research and development, thereby draining the economy of key resources, inhibiting its innovative abilities and militating against productivity, growth and international competitiveness. This situation is exacerbated when large numbers of highly trained scientists and engineers are drawn away by the defence sector from the civilian sector, where the latter presumably has the greater productivity potential.

In sum, therefore, the defence sector may enhance the supply of skilled labour, thereby alleviating an important growth constraint. Conversely, however, it may compete for scarce human resources with the more productive civilian economy, thereby compromising the overall productivity and efficiency of the economy.

Modernisation Factors

School systems in developing countries have progressed rapidly during the post-World War II era, so that education enrolments in Africa, the Middle East and Latin America rose from approximately 163 million in 1960 to 455 million in 1985 (Todaro, 1989: 333). Although the largest part of this increase has been in primary education, the greatest proportionate increases have occurred in secondary and tertiary education. However, despite the large outlays by governments in respect of education expenditure, human capital accumulation has been inhibited by a number of factors (compare Gillis et al, 1983: 212-215; Rees, 1979: 43-44; Todaro, 1982: 249-256; Sadie, 1980: 368-372; Schutte, 1987; Hartshorne, 1985). These include the following:

Teacher shortages resulting in larger classes and the use of less highly trained teachers.



Large dropout rates.



Deficiencies of learning methods and curricula.



Maladjustment to social needs at the primary school level (related to the fact that by far the greater majority of children in developing nations live and attend schools in rural areas).



The negative influence on learning ability exerted by parents' education, family sizes, peer group interactions and inadequate early nutrition and health care.



The inability to retain functional literacy due to limited access to written material.

In addition to the above the promotion of secondary and tertiary education in developing countries has often occurred at the expense of the provision of basic and technical education. Termed the "diploma disease" by R. Dore (see Gillis et al, 1983: 214), this phenomenon is closely related to the hierarchical and rigid structure of developing society. There is a strong tendency to judge a candidate's suitability for work on the basis of the academic credentials in his possession. This implies that the private return on education is deemed to be higher and of more importance than the social return so that education merely becomes a selection process rather than an investment process (Rees, 1979: 43). The crux of the matter is enunciated in Gillis et al (1983: 214): "... these systems are not primarily in the business of conveying knowledge and skills at all, but are principally concerned with 'certification' or 'credentialing.'" Deger (1986b: 118) states rather bluntly that: "University or high school diplomas have become the new status symbols. An Arts graduate may have little productivity from a social point of view, yet his socio-economic position is always higher than that of a technician even though the latter has much more to contribute to the industrial growth of the economy."

The reason for the latter phenomenon can, to a large degree, be attributed to the fact that there has been a perpetuation of the colonial system whereby a high premium is placed on the career benefits of higher degrees. Thus the reward system in society allocates substantially superior benefits to a degree than to other qualifications, skills or aptitudes that are perceived to be inferior. The longer-term consequence, however, is mass production of graduates with little chance of using their inappropriate expertise. Thus, the social problem of hierarchy and the educational problem of valueless higher degrees are inextricably connected with each other.

Some writers (for example, Whynes, 1979) have suggested that the military has an important role to play in dismantling social rigidities by virtue of the fact that it is a progressive institution. Moreover, military regimes are sometimes perceived to promote the modernisation of society which, in turn, dispenses with its feudal and social obligatory system (Pye, 1972) and moves towards a market-oriented, capitalist system which places a much greater premium on enhancing the quantity and quality of human capital in society.

The modernisation argument was particularly popular during the Sixties - a period which coincided with the proliferation of Third World independence. Writers such as Janowitz (1977) and Bienen (1971) claimed that the army in developing countries is largely recruited from the middle classes, therefore displaying a propensity toward breaking the rigidities of social structure imposed by the traditional, colonial classes. Moreover, the new middle class played a key role in moulding military establishments and placing them in the forefront of a progressive society. Thus, whereas modern technology was welcomed within the army and its usefulness and potential appreciated, the political system displayed inertia, inefficiency, scepticism and even greed in employing such technology.

Subsequently, however, the weight of evidence has suggested that the contention that the military is truly a modern institution in developing countries, is a fallacious one. Studies conducted during the Seventies and early Eighties in respect of ex-colonial countries in West Africa (for example, Eleazu, 1973) reach the general conclusion that there is little evidence of modernisation emanating from the military. In fact, the defence sector is accused of failing to distinguish between civilian and military cultures, perpetuating peasant and community-based values, enhancing the military elite as a bureaucratic group, exacerbating class distinctions between officers and peasants and, in the event, being incapable of or unwilling to introduce or generate technical skills in society.

In short, the evidence supporting the view that the defence force is intrinsically any more modern than any other institution (such as the judiciary or the business community) is, at best, inconclusive. On the whole, however, the influence of the military on societal attitudes appears to be minimal. Consequently, it is unlikely that the military plays a major role in dismantling social rigidities, thereby promoting the accumulation of appropriate human capital in developing countries.

Revenue Constraints

Government revenue in low-income developing countries, with relatively unsophisticated financial systems, is fairly inelastic.7 Under these conditions a significant rise in military expenditure can only occur at the expense of other government outlays, not least of which investment in human capital. Similarly, a decline in the relative importance of military activity (as experienced in South Africa from 1989) may enable the allocation of more budgetary funds to non-defence pursuits. Thus, within the constraints imposed on government revenue it is possible to postulate a negative relationship between defence and human capital.

By the same token, conventional wisdom dictates that countries with high growth rates are able to generate sufficient revenue to nurture and expand human capital. If, as has been suggested earlier, defence expenditure stimulates economic growth, then it may be claimed that defence indirectly supports human capital formation via growth. However, although theoretically plausible, the direction of causation embraced in this transmission mechanism would appear to be somewhat perverse.

A major problem encountered in attempting to quantify the inter-relationships discussed thus far is identifying a suitable variable to depict human capital. Government spending on education would appear to be appropriate by virtue of the fact that it is a crucial determinant of and a sine qua non for human capital formation. By treating education spending as the dependent variable and defence spending as the independent variable, it is generally expected that an increase in the latter will occur at the expense of the former. Thus, a negative trade-off8 is postulated.

The bulk of research on the trade-off between defence and social programmes has emanated from and been applied to industrialised nations. From the observation that there has generally been a trade-off of almost equivalent amounts between military expenditure and investment in developed countries, Smith (1980: 31-32) concludes that defence spending does not significantly affect the social wage (including education and health), the burden of higher defence outlays primarily falling on investment.

Evidence about budgetary trade-offs in developing nations is more ambiguous. The normative and traditional assumption is that governments of developing nations have displayed a greater willingness to reduce the social wage for human capital so as to accommodate an increase in defence expenditure. This willingness can be attributed to the virtual absence of institutional resistance (e.g. influential trade unions or consumer groups) to decreasing the social wage, together with a reluctance to reduce investment expenditure (Deger, 1986a: 115). Pursuing this line of thought Verner (1983: 77) quotes the then secretary-general of the United Nations who stated that military expenditures undoubtedly absorb resources that are "... substantial enough to make a considerable difference both in the level of investment for civil purposes and in the volume of resources which can be devoted to improving man's lot through social and other services." (own italics). The implication is clear - increased defence spending in developing countries may have negative consequences for socio-economic development programmes such as education.

However, the trade-off between defence and education may feasibly be positive if increases in defence spending are matched by real increases in education spending. Indeed, a number of studies have indicated that the defence-education trade-off in Latin America is largely positive. Verner (1983) used a time-series design in order to determine the nature of budgetary trade-offs in eighteen individual Latin-American countries (during the 1948 to 1979 period), and relied on regression analysis to examine each separate time series, with education spending as percentage of total public expenditure as the dependent variable, defence spending as percentage of total public expenditure as the independent variable and the annual percentage change in real per capita income as a control variable. The regression coefficient for the defence spending term indicates the direction and magnitude of the trade-off between defence and education expenditures.

Although Verner's data indicated considerable variation in the nature of the defence/ education trade-offs among countries, he found a negative trade-off in only one country (El Salvador). There was a positive trade-off in ten countries and no significant trade-off in the remaining seven. These conclusions supported the findings of Hayes (1976) who, in an analysis of Brazilian outlays on defence and various socio-economic programmes between 1950 and 1967, found that military spending did not necessarily yield negative consequences for social investment, after finding that substitutions between military allocations and allocations to other sectors do occur frequently, but that the burden of these substitutions is not severe and is distributed across all categories of social spending. Similarly, Ames and Goff (1975) found that in Latin American countries, neither education nor defence gains at the expense of the other.

These conclusions give rise to a great deal of speculation and inquiry into the apparent anomaly embodied therein. Given the fact that increases in defence spending in developing countries presumably have to occur at the cost of other government financed activities, the question arises as to why education is apparently exempt from reduction. A number of explanations may be presented. These include the following:

Military budgets (which are often committed to large capital expenditures) may be relatively immutable in the face of short-term political changes.



Ruling regimes may find it necessary to vote for larger defence budgets so as to prevent the defence force from conducting coups.



Education and defence programmes may increase in tandem with one another because both are supported by relatively strong interest groups. Hence, demands by well-organised education groups for increased education spending are matched by similar demands by powerful military groupings. As a result, both education and defence budgets benefit at the expense of socio-economic activities that are bereft of organisational pressure (Verner, 1983: 87-88).



Because the military in a developing country may be a significant political actor, involved in national decision-making and subject to the vagaries of political instability, it is less likely that large shares of the defence budget will be allocated to pure defence purposes. Instead, defence allocations may be made to ends which reinforce the political position of the armed forces, suggesting that military interests and civilian development goals may not be as disparate in the developing countries as in industrialised nations. Thus, increased spending on education may be in the interests of the military (Hayes, 1976: 21-52).



A large portion of the education budget may be fixed in terms of contracted personnel payments, so that cuts cannot be effected (Verner, 1983: 88).



In many developing nations teachers constitute the largest single highly educated group and are therefore in a position to preserve their interests, even under military regimes (Verner, 1983: 88).



A modernising and professional military establishment in a developing country may be an agent and supporter of fundamental societal change and economic growth and development. Consequently, education becomes a priority budget item for military as well as civilian-reformist regimes (Duncan, 1976: 210-214; Wynia, 1978: 54-60).

One of the problems9 identified by Verner in his study is the fact that he measured military and education expenditures only in terms of public spending, implying that measurement relative to GDP may yield different results (Verner, 1983: 89).

In an attempt to quantify the relationship between military expenditure and human capital formation, Deger (1985 and 1986b) uses the ratio of public education expenditure to GDP. A set of equations is employed to find the inter-connections between the military burden, i.e. military expenditure as a proportion of GDP (m), government education expenditure as a ratio of GDP (e), the growth rate and the savings rate (i.e. total savings as a share of GDP).10 Deger used a sample set of fifty countries, each observation being a time-series average of data for 1965 to 1973. In contrast to Ames and Goff and Verner she ultimately found that the defence burden reduces education spending, and is therefore probably strongly detrimental to human capital formation.11

Clearly then, the evidence pertaining to the trade-off in developing nations between defence spending and human capital formation is ambiguous and ambivalent, covering the entire ambit of possibilities from negative to positive. This raises an interesting consideration in respect of the future relationship between defence spending and education spending in the light of the decline since the late Eighties in the relative geo-political importance of the military. This is one of the issues to be discussed in the next sub-section which focuses the attention on the impact of defence spending on the formation of human capital in South Africa.

THE TRADE-OFF BETWEEN EDUCATION AND DEFENCE IN SOUTH AFRICA

The major focus in this section lies in the empirical analysis of the complex interrelationships between defence spending and public education expenditure in South Africa. Annual data for the thirty-year period spanning 1960 to 1990 are used, and education spending is treated as the dependent variable (i.e. the variable to be explained by others) in two regression analyses, while defence spending is entered as an independent variable (i.e. an explanatory variable). The first regression was run using the defence and education expenditures terms as percentages of total government expenditure. In the second regression the two terms were expressed as a percentage of GDP.

The first regression assumed the following functional relationship:

Education expenditure as a % of total government expenditure in any given year is explained or influenced by the combined effect of:

military expenditure as a percentage of total government expenditure;



the previous year's expenditure on education as a percentage of total government expenditure; the economic growth rate; and the previous year's economic growth rate.

In the second equation education expenditure as a % of the gross domestic product (GDP) in any given year was assumed to be explained or influenced by the combined effect of:

military spending as a percentage of GDP;

the previous year's expenditure on education as a percentage of GDP;

total government expenditure as a percentage of GDP; and

the economic growth rate.

In both cases it was found that changes in the education variable were well explained by the combined effect of the independent variables. However, there is no statistical evidence supporting either a positive or negative relationship between defence spending and education spending12. This means that military expenditure in South Africa neither promoted government spending on education, nor occurred at its expense between 1960 and 1990.

It should be pointed out that the statistical analyses conducted above are concerned only with reported public expenditures on education. Inasmuch as private education and private expenditures on public education are excluded, the data employed understate the country's commitment to education. Moreover, to the extent that the military may provide valuable training and civilian skills, some share of military expenditure may be considered as public expenditure on education. With these caveats in mind the conclusion is reached that the relationship between defence and public education spending has been largely neutral. There is therefore little, if any, evidence to support the possibility that the military has improved the productivity and efficiency of the economy via a benign positive spin-off effect on education and the enhancement of human capital. By the same token, however, military expenditure has not occurred at any great expense to public education spending. In many circles, however, the continuation (and reinforcement) of the changing trend in budgetary priorities which commenced at the end of the Eighties is expected to result in a visibly larger negative trade-off between these two areas of public spending in favour of education (and other forms of social) expenditure.

In this regard a fundamental issue is germane to the discussion. There is general consensus, based on a wide range of considerations, that the level of militarisation must be reduced significantly in a post-apartheid South Africa. This entails, inter alia, eliminating the constraints and inequities prevalent in the pre-democratic paradigm. However, consensus is lacking with regard to the most effective and appropriate way of achieving broadly-spread social upliftment. Those with a free-market slant argue in favour of reduced state intervention, while others with a social democratic bent regard the state as being of paramount importance in diluting economic exploitation and poverty. It is not the purpose or function of this study to pass judgment on the merits or lack thereof of either of these two approaches. Instead, an attempt will be made to measure the extent to which a reduction in defence expenditure can contribute towards the achievement of inter-racial parity in government expenditure on education. The following assumptions and conditions are postulated:

Desegregation of educational institutions will be accompanied by an attempt to equalise government education spending per capita at White levels prevailing before 1994.



The amount needed in 1990/91 to achieve parity at White levels was R40,5 billion (computed from Lachman and Bercuson, 1992: 22) or 13,2 per cent of GDP. Actual education spending in that year was R18,9 billion (SA Reserve Bank, 1993: S-66) or 6,2 per cent of GDP. It is assumed (by disregarding any changes in enrolment rates) that the amount required to achieve White level parity has remained and will remain unchanged in 1990/91 terms. Thus, in the light of the education budget of R19,4 billion (in 1990/91 terms) for the 1993/94 fiscal year, the parity shortfall at the beginning of the 1994/95 budgetary period amounted to R21,1 billion.



Total government expenditure will remain constant at its early-Nineties levels of over 35,0 per cent of GDP.



The defence burden will fall from 2,9 per cent in the 1993/94 fiscal year to 2,0 per cent in consequent years.



All monies saved by the decline in the defence burden will be channelled in toto to education.



No other budgetary shifts will be implemented, i.e. all other forms of government spending as a percentage of total government expenditure will remain unchanged.

Three different economic growth scenarios, viz. zero per cent, two per cent, and four per cent annual average growth, are employed to quantify the effect of the "guns versus blackboards" trade-off encompassed by the above assumptions and, in particular to determine how many years it would take to meet the 1993/94 education shortfall.

The computations reveal that, even in the event of an average annual rate of economic growth of 4,0 per cent, per capita parity with respect to education at 1990/91 White levels will only be reached by the turn of the century. Even then the level of parity achieved will be no higher than that prevailing ten years earlier. Moreover, any increases in the school-going population will serve to raise the level of expenditure required and therefore delay the achievement of parity.13 In the case of the 0 per cent and 2,0 per cent growth scenarios parity will be achieved at a yet later date, with the same caveats applying.

It can be argued with some degree of conviction that the above results - which are per se somewhat uninspiring - reflect the outcome of a number of rather generous assumptions. First, international experience has shown that if the government sector's share in total economic activity is sustained at levels exceeding 25 per cent, a consistent economic growth path of 4,0 per cent or even 2,0 per cent becomes more difficult to attain. Second, it is possible, in the light of the possible absence of internal stability during at least the initial phase of the post-apartheid era, that a significant portion of any savings generated by a reduced defence burden, may be siphoned off to the police and/ or internal peace-keeping operations. Moreover, failure to achieve spending parity within a much shorter period may lead to rising social tension and dissatisfaction with political leadership and therefore require increased military intervention to counteract internal acts of aggression. The concomitant rise in the defence burden would then dissipate any prospects of a meaningful development-enhancing peace dividend. Finally, disparities in racial spending in other social areas (for example, housing, health and pensions) also exist. Any "peace dividend" emanating from a reduced defence burden is therefore likely to be requisitioned by more than one government department. Consequently, those relying solely upon defence savings as a means toward achieving spending equality across a wide range of social areas will be disappointed in the modest progress made in this regard.

It is not the purpose of this study to explore the ways in which a non-discriminatory and equitable distribution of fiscal funds for socio-economic purposes can or should be effected in a democratic South Africa. Nor has any attempt been made to pass judgment on the desired share of government expenditure in the GDP or the "correct" prioritisation of social spending demands. Nonetheless, the crude analysis above proves that any government resources released by lower defence expenditure will make only a marginal contribution to the alleviation of poverty caused by disparate social spending.

CONCLUSION

Although it is contended by some that the military may contribute towards the promotion of the modernisation of developing societies through the enhancement of the quality and quantity of human capital by, inter alia, dismantling social rigidities, there is limited conclusive evidence to support this view. In fact, a negative trade-off between defence spending and education spending (used as a proxy for human capital formation) is generally expected. A regression analysis of the relationship between military spending and government education spending in South Africa between 1960 and 1990 is essentially statistically insignificant. This suggests that it is unlikely that military activity has served to enhance the productive capability of the South African economy via some modernising effect; nor has military expenditure significantly debilitated the flow of government resources to the formation of human capital. In short, the longer term impact of military expenditure on South Africa's stock of human capital has been largely neutral.

Future cutbacks in defence spending - even when generous assumptions are applied - are unlikely to have a significant impact on the achievement of horizontal fiscal equity, particularly in the short run. In a somewhat sobering indictment of the political friction and tension that may prevail until the turn of the century, the defence force could find itself playing a major role in the development process by assisting in the maintenance of internal stability and order - a crucial condition for both domestic and foreign investment in growth and development oriented projects.

NOTES

GDP - gross domestic product - measures the total value of all goods and services produced the factors of production located in a country.



During the late Seventies the average defence burden (defence spending as a share of GDP) in developing countries amounted to 6,5 per cent.



The proposition that government spending on defence reduces the volume of resources available for spending in other areas of the economy.



For example, Harbison (1973) emphasises the importance of secondary school education, while Blaug (1973) points out that: "... primary education yields higher social rates of return ... than any other level of education." In South Africa it is imperative that non-formal literacy programmes for adults be launched on a large scale.



These include the inaccessibility to surplus labour, a lack of work ethos, and contentment with a subsistence-like existence, outside the formal labour market.



Deger (1986b: 18) states in this respect that the standard of vocational and technological education is usually very high in developing nations.



An increase in the tax base yields a less than proportionate increase in tax revenue.



The concept of trade-off is used to describe the way in which one policy area may gain at the expense of others in the allocation of scarce resources. There is a continuum of possible trade-off patterns. At one extreme, increases in defence expenditure may be entirely paid for by decreases in education expenditure, producing a negative trade-off effect. A positive trade-off would occur if defence spending increases are matched by increases in education spending.



Others include questionable comparability of the data and the inability to measure monies not spent, or programmes not initiated in education or other policy areas because the anticipated costs are estimated to be insupportable given existing and future defence spending.



The inclusion of the growth and saving rates enables the integration of any spin-off effects from other aspects of military spending.



Deger goes on to prove that a one per cent increase in military spending as a share of GDP reduces the education budget by 1,14 per cent as a share of GDP. Then, quoting 1976 figures, she calculates that a ten per cent reduction, for one year only, in the military spending ratio for all developing countries together would have released sufficient funds to have built enough new primary schools to accommodate 100 million children, and finance a large-scale programme to extend literacy to all adults in developing countries by the end of the century.



The actual regression results obtained were as follows:

First regression (t-statistics in parentheses)

EDGOVE =6,59 - 0,12MILGOV + 0.81EDGOVE(t-1) - 0,19ARGDP

(-0,86) (7,73) (-1,64)

- 0,26ARGDP(t-1)

(-2,06)

where EDGOVE = education spending as a percentage of total government spending;

MILGOV = military spending as a percentage of total

government spending; and

ARGDP = annual rate of growth in real GDP.

t-1 denotes that the variable in question is lagged by one year.

R2 = 0,81

Durbin-Watson = 1,48

The relatively high R2 indicates that the independent variables together account fairly well for the variance in the education spending variable. However, the low t-ratio for the defence spending coefficient suggests that the budgetary trade-off between defence and education expenditure is statistically insignificant.

Second regression (t-statistics in parentheses)

EDGDP = 0,13 - 0,09MILGDP + 0,78EDGDP(t-1) + 0,06GOVGDP

(-0,61) (7,68) (1,48)

-0,07ARGDP

(-1,86)

where EDGDP = education spending as a percentage of GDP;

MILGDP = military spending as a percentage of GDP;

GOVGDP = total government spending as a percentage of GDP; and

R2=0,88

Durbin-Watson = 1,40

Again, while the regression specification provides an overall good fit of the data, the t-ratio for the coefficient for defence spending does not inspire confidence.



For instance, according to Lachman and Bercuson (1992:23) there were 1,5 million children eligible for schooling who received no formal education in 1990, and who could therefore feasibly enter the education system.