The simplest understanding of the peace dividend is that if governments cut their military expenditure (milex), there will be greater financial resources to allocate to more socially desirable and/or more productive expenditure categories. Given that expenditure in these other categories is, unlike milex, likely to yield a positive social rate of return, economies will grow as a result. This thinking has been strongly influenced by the events in Europe between 1991 and 1993 which led to the effective ending of the Cold War and by the continuing struggle of many low income developing countries to save sufficient resources and to allocate these to best effect.

A more sophisticated version of the peace dividend embraces the concept of the conversion of an economy, or parts of it, from a military orientation to a more civilian orientation. Clearly, this approach includes the conversion experiences of countries after, say, the first and second world wars when major reallocations of human capital and physical capital - indeed entire productive systems - took place. A peace dividend can occur at any time but it is often easiest to cut milex when a major armed conflict comes to an end.

Building on the work of Chan (1995), Intriligator (1996) emphasises that such adjustments involve considerable short term costs, in the form of unemployment of resources and various conversion costs, before benefits are realised in the medium and long term from the new allocation of resources. Disarmament, then, can be usefully considered as a type of investment process with rates of return, which vary between different wars. Intriligator suggests that the rate of return to the US economy of post - World War II conversion, based on high consumer demand and national reconstruction policies, was much higher than the rates following the Korean, Vietnam and Cold Wars. Intriligator’s work forms part of a comprehensive study of the peace dividend (Gleditsch et al 1996) ¹ which focuses on the use of macroeconometric models to determine the effect of cuts in milex in individual (mostly developed) countries, for regions and for the global economy. Generally speaking, disarmament was found to be likely to result in improved economic performance at all three levels.

The focus of this article is on developing countries where almost all conflicts have seen fought since 1945. The vast majority of these conflicts have been intrastate rather than interstate and have involved a government in armed conflict with domestic opponents either wanting to take over government or to achieve full or partial political autonomy. The particular questions this article addresses are as follows:

1. Why is the peace dividend particularly important for developing countries?

2. How many developing countries have cut milex since the late 1980s, and how big has the resulting peace dividend been?

3. What appears to have happened to the peace dividend in twelve Southern African countries?

4. How can a government wishing to capture a peace dividend to enhance its peoples welfare does so most effectively?

These are several reasons why reductions in milex are particularly important for developing countries. First, their allocations of resources to milex are relatively high. Industrial countries allocate an average of 5.5 per cent of their central governments expenditures (CGE) to defence whilst the comparable developing country figure is 11.0 per cent (IMF 1996, pp. 18-19). This has important implications for allocations to other CGE categories. Military expenditure as a proportion of combined health and education expenditures is 60 per cent for developing countries compared with 33 per cent for industrial countries (UNDP 1996, Tables 19 and 39). Second, there have been many studies of the impact of milex on economic growth in developing countries (for a recent review of this research, see Dunne 1996). The general consensus is that milex reduces the level of saving and investment and thereby reduces economic growth. This makes the task of reducing absolute poverty and its various manifestations more difficult. Third, milex frequently increases a country’s foreign indebtedness and future debt servicing burdens (Harris 1996a). Fourth, milex has important negative externalities at the political and societal levels. These include the use of the military as an instrument of human rights abuse, the propensity of the military to engage in coups d’etat, the negative impact of high levels of militarisation on democracy (Bowman 1996) and the greater likelihood of armed conflict occurring between countries which engage in arms races (Wallace 1979; Sample 1997). There appears, then, to be a very strong case for reducing milex in developing countries.

Table 1 presents estimates of the size of the peace dividend occurring between the late 1980s and the early / mid 1990s. The basic data source used was SIPRI (1996), supplemented by the IMF’s Government Finance Statistics Yearbook (1996). Twenty nine countries recorded falls in their milex as measured by the difference between their average milex and in constant prices, 1987-1989 and 1993-1995 ² . The amounts, in 1990 US dollars, are listed in column 1 and the percentage falls between the two sets of years appear in column 2. The greatest milex cuts ($2362m. per annum) were made by the 14 African countries whereas the three Asia / Pacific countries made cuts totalling only $41m. per annum. The grand total of saved milex per annum was $7036m.³ and the median cut in milex was 28.2 per cent. Six countries - Ethiopia, South Africa, Argentina, Poland, Romania and Egypt - cut their milex by over $500m. per annum and together comprised 77 per cent of the total annual reduction. Columns 3 and 4 list the milex reductions, in percentage points, as a proportion of CGE and GDP. The few examples (e.g. Cameroon) where these increased despite a fall in milex are a consequence of CGE and / or GDP falling more than milex over the period.

Despite what may seem a sizeable reduction in milex by these 29 countries, most developing countries in fact increased their milex over the period and many continue to spend amounts which are out of keeping with an objective assessment of their security requirements. The $7000m. in milex cuts can be compared with an estimated developing country military expenditure in 1994 of $140,030m. (UNDP 1996, Table 19) and to the amount required per annum to provide basic education, basic health care and nutrition, low cost water supply and sanitation and basic reproductive health by the end of the decade which the UNDP (1994, p.77) puts at $30,100 to $40,000m.

There is no certain way of knowing what happened to any dollar not spent on the military. It is, however, legitimate to suggest probabilities and possibilities, insofar as the available data allows sufficient scope for investigation. This is not to suggest that the budgetary decisions discussed below are a result of a consistent and conscious decision making process. Politicians and planners might indeed be surprised to be told of systematic patterns in their activities. In addition, revenue and expenditure decisions are typically made independently⁴.

This said, there are three broad possibilities for the financial resources saved by cutting milex:

1. The peace dividend is available for allocation to other CGE categories. This assumes that CGE does not fall as milex falls.

2. The government’s budget deficit is cut i.e. CGE does fall as milex falls.

3. Taxes are reduced i.e. central government revenue (CGR) falls.

Whilst combinations of the above are possible, it is useful to assume that they represent mutually exclusive options available to a government. In Table 2, we begin with an initial milex of 15 and a budget of 25. Milex is then cut to 5. Option 1, where a milex cut means greater non-milex shows no changes in the sizes of CGE or the budget deficit. Option 2 sees the entire milex cut going to reduce the size of CGE and the budget deficit. Option 3 illustrates a fall in CGR as a result of tax cuts, with the budget deficit remaining unchanged.

What evidence, then, would indicate which of these policy options was followed? In brief, option 1 would be illustrated by a fall in milex as a proportion of CGE and commensurate increases; in other CGE categories; option 2 would have a reduced budget deficit and option 3 would have a CGR/ tax cut, illustrated by an increase private consumption. This article investigates these possibilities for the 12 Southern African Development Community (SADC) countries from the mid - 1980s to the present, insofar as the available data allows.

The logical data source is the IMFs annual Government Finance Statistics Yearbook (GFS) of which the 1996 issue was the most recent available. The GFS reports budgeted rather than actual expenditures but a larger problem concerns its coverage of the SADC member countries. Two - Angola and Mozambique - are not included at all. The most recent data available for some countries is quite dated (Tanzania 1985, Swaziland 1989, Malawi 1988 and Zimbabwe 1991). Mauritius’ data covers 1986 to 1995, but the many breaks between years at times preclude estimation of meaningful series. In addition, the GFS entries for three countries - Namibia, South Africa and Zambia - do not include military expenditure. Some of these shortcomings were dealt with by the use of SIPRI and the IMFs International Finance Statistics Yearbook. Table 3 presents constant price data for milex for the late 1980s and the early / mid 1990s. Of the eight countries for which data are available for both sets of years, six reduced their milex and there was an overall reduction in milex of 40 per cent in constant price terms, from $5227m. to $3119m. per annum. South Africa dominates regional milex, comprising 86 and 83 per cent of total milex in the two periods and it contributed 89 per cent of the overall cuts in milex⁵. Table 4 indicates that milex as a proportion of GDP fell from an average of 4.1 to 2.7 per cent over the period.

There is a widespread belief that the military has the capacity to win a disproportionate share of CGE increases and to resist CGE cuts i.e. that trade-offs occur between milex and other CGE categories. This issue has been investigated by a number of researchers (e.g. Harris et al 1988, Apostolakis 1992) most of which have been unable to find strong evidence of trade-offs. The relevance of the trade-off hypothesis to the peace dividend is obvious: if milex is cut, the government financial resources saved can be applied to other CGE categories. We are searching, then, for evidence of commensurate increases in other CGE categories when milex is cut (all expressed as a proportion of total CGE). Statistically, we calculate Pearson correlation coefficients between milex as a proportion of CGE and other major expenditure categories. These are presented, for the few countries for which GFS data were available, in Table 5. Whilst most coefficients were of the hypothesised negative sign, only those for South Africa (all categories) and Botswana’s Economic Affairs and Services were significant at the 5 per cent level or better. Botswana’s milex increased over the period, from 11.0 per cent of CGE, 1986 - 1989 to 12.6 per cent, 1990 - 1993. The evidence of trade-offs, therefore, is restricted to South Africa, where it seems to have been a powerful factor, both between CGE categories and between components of security expenditure (see footnote 5).

Conventional wisdom (e.g. Easterly et al 1994) suggests that the size of budget deficits is inversely and causally related to economic growth. This negative impact of deficits occurs through their effects in raising interest rates (if the deficit is financed by domestic borrowing), exchange rate appreciation (if foreign borrowing) or inflation (if money is created). A related reason for cutting government expenditure is a belief in the greater efficiency of the private sector in productive activity. As a result international agencies such as the IMF and World Bank push strongly for reduced government deficits⁶. To what extent, then, have milex cuts provided a means by which governments can reduce the size of their deficits? To answer this, a series of preliminary questions need to be investigated.

First, how large a proportion of a deficit does milex make up? That is, if milex was cut, is it of sufficient size to make a difference to the overall deficit? Using SIPRI milex data and CGE data, the mean proportion for the period 1989 to the mid - 1990s is GFS 54.9 per cent (median 56.2). The range was between 25.2 for Malawi and 81.6 for Namibia. Clearly, there is scope for making significant deficit reductions via milex cuts. Second, what have been the trends in the milex - deficit relationship? Table 6 shows that the proportion rose from 27.8 per cent in the second half of the 1980s to 81.4 per cent in the early - mid 1990s. Third, what happened to CGE over the period? There was a modest fall of 1.5 percentage points in the ratio of CGE to GDP (see Table 7) between 1985 -1989 and 1990 - 1997. Fourth, what happened to deficits over the period? Table 8 reveals that these fell from a median of 5.5 per cent of GDP in the late 1980s to 5.1 per cent in the early - mid 1990s.

If a cut in milex was in effect used as a means of reducing the size of the deficit, we would expect to see, for those countries which cut milex, a commensurate reduction in their budget deficit. Statistically, the Pearson correlation coefficient between milex and budget deficit, both in current prices, would be negative⁷. An examination of Table 9 provides some support for this, with Lesotho, Malawi and Tanzania having significant negative values. Of the other countries which cut milex, South Africa’s concern with reconstruction and development (as expressed in the RDP) following the end of apartheid has meant that its large milex cuts have been swamped by development expenditures. Furthermore, the country has also experienced some of the conversion costs referred to in section 1 and as a result its deficits have risen. Zambia and Zimbabwe’s positive values indicate that milex cuts were not associated with improved budgetary positions in these countries.

6.3 Tax reductions

The third possibility is that the peace dividend has been passed on to citizens via lower taxation. This would show up in higher private consumption, which would rise as a proportion of GDP. Table 10 indicates that whilst this proportion rose in five countries, was stable in two and fell in one, the average rose only marginally. It therefore does not appear that milex reductions were used to finance tax reductions, a conclusion also borne out by fairly constant tax revenue to GDP ratios.

This article has estimated the size of the peace dividend in Southern Africa made available by milex cuts since the mid - 1980s. It has investigated three possible ‘uses’ which may have been made of the peace dividend and found evidence that it allowed greater expenditure elsewhere in South Africa and had been used to reduce the size of budget deficits, in a context of falling CGE, in several other countries.

Most discussions concerning the peace dividend (e.g. UNDP 1994, ch.3) have focussed on international responses. What advice can be given to a developing country government which is considering milex cuts as a means of financing human development initiatives? First, sustainable milex cuts are obviously best made in a context of security and this can be enhanced by the establishment of civil means of resolving conflict between, and particularly within, nations. Second, the population (and even the military) can be encouraged to accept milex cuts by earmarking most or all of the savings for desired social expenditures. These both reinforce internal security and allow progress in human development, which should be, after all, the underlying objective of a nation. There is, happily, encouraging evidence from at least one country, Costa Rica (Harris 1996b), which can be used to inspire timid politicians to push for a peace dividend.