(i) Kenynesian Theory
How relevant is Keynesian theory (and its modifications) for Botswana? It is understood that the incomes and employment policy has been rectified so as to give employers much freedom in the determination of their employees' wages and salaries. This positive course of action will apparently significantly affect the distribution of labour between sectors and industries and the overall future level of employment in the country. During the period under consideration, however, wages in Botswana were not determined by the interplay of supply and demand in the labour market, but rather by the committee responsible for the implementation of the incomes and employment policy on the basis of, among other things, uniformity with similar jobs in the public service, productivity, working conditions and so on. Furthermore, the incomes and employment policy limited the role of the trade union movement as a collective bargainer over wage levels. More important is the fact that Botswana's economy is characterised by the existence of the Dutch Disease on account of the predominance of the mining sector in the economy. The Dutch Disease operates as follows: part of the mining (diamond) revenue is spent on non-traded goods (e.g. construction and services) which leads to a real appreciation ( a rise in the relative prices of non-traded goods in terms of the traded goods) (Wijnbergen, 1984). This appreciation draws resources out of the non-mining traded sectors (e.g. the movement of labour from agriculture to construction, government and general services sectors) into the non-traded sectors (Corden and Neary, 1982). It seems unlikely to us that in this case, the government, through fiscal policies (using the income tax and government expenditure) can completely influence aggregate demand and unemployment, among other things. Moreover, with three quarters of the population residing in the rural areas, outside the cash economy and leading a subsistence type of life, it is doubtful whether aggregate demand plays a decisive role in determining the level of real output.
As far as unemployment is concerned, in Botswana's context structural unemployment is illustrated by the absolute decline of the agricultural sector and a growth in the mining and public sectors, which might have resulted in some workers losing their jobs in the formal sector (Tumkaya, 1983). It has, however, been noted that the level of education has been the main determining factor in that those with no education and training, on giving up agriculture, have not been able to find employment in the growth sectors. On the other hand, those with education and training and who have been willing to take up new jobs have been more successful in adjusting to changing economic conditions. Consequently, unemployment among the uneducated persons who have given up traditional farming is not absolutely due to the failure of the economy to generate enough jobs. Account has to be taken of the mismatch between their skills and the new job requirements (Mhozya, 1987).
Unemployment and more particularly under-employment in Botswana are therefore mostly among the rural population involved in traditional agriculture (tradeable goods). In this situation it is highly doubtful whether Keynesian conditions fully apply to this small open economy. For if it did, devaluing the currency and expanding the money supply (all other things being equal) would stimulate employment. Employment would be increased in the nontradeable sectors, for example, by expanding nominal demand, leading to lower real wages through exchange rate depreciation.
(ii) The Rational Expectations Hypothesis
As for the notion of rational expectations in relation to the Botswana's situation, the question is whether economic developments and performance are systematic. Botswana maintains a stable government with stable economic policies and philosophy. In spite of this situation there are several factors which diminish the relevance (with respect to the theory under consideration) of the relationship between private agents' reactions and systematic government course of action.
It has been observed that, in a broad sense, rational expectations theory can be regarded as operating within the same framework as Keynesian theory, whereby the maintained criterion is that of an optimizing government that seeks to adopt any policy it considers best. Both theories make use of the concept of authority (government or central bank). Under both theories the system operates as follows: the government acts on the private sector and the private sector reacts through the price and wages. The entire concept of rational expectations, it is argued, deals with one-way causation, and indeed whether the government can `trick' the private sector into adopting a behaviour that it would not entertain if it had full information and utilized all that information. Like the Keynesian theory (or relative modifications) rational expectations theory does not undertake to evaluate the motives of government, or the influence of the private agents on government outside the working of the market (e.g. through lobbying, etc.) (Corden, 1987).
From the salaries and wages point of view the newly rectified version of the incomes and employment policy will eliminate the current reactions and interactions between the private agents and the government. According to the revised policy, private agents will be free in determining the levels of wages of their employees. There are, however, still many fields in which the reactions of private agents affect government decisions in Botswana. To mention a few, the customs revenue represents one of the major sources of government revenue. The amount accruing to Botswana is a result of a series of negotiations between all the member states of the Southern African Customs Union. Mineral revenue, which constitutes the major component of government revenue, is determined through long-term contracts with the mining companies in the country. In summary, a significant portion of the government revenue upon which most of the government overall expenditure is based is not free from the influence of the private agents. There is a two-way causation in the relation between government and private agents.
Furthermore, about eighty percent of Botswana's imports originate from the Republic of South Africa(RSA) and consequently a larger proportion of Botswana's inflation is imported from that country. Moreover, the value of the Pula is pegged to a basket of the currencies of Botswana's main trading partners, including among others, the RSA's Rand, the USA Dollar and the Special Drawing Rights (SDR). Due to the high degree of import dependence on the relevant country, the Rand accounts for seventy-five percent of the value of the basket. This means that fiscal policy in the RSA is likely to affect Botswana's relative prices of labour, capital and goods through its effects on employment creation, on business profits and on consumer prices. Thus, the forms of support offered to RSA farmers by the RSA government will affect the prices and availability of imported goods in Botswana. Monetary policy in the RSA will affect interest rates and exchange rates which will influence imports prices in Botswana again. The RSA's wages and employment policy will influence industrial cost, migration and remittance in Botswana (Love, 1989).
The foregoing particulars confirm the fact that private agents (including external factors) do influence Botswana government actions and policies directly. The direction of causation is two-way. Government policies certainly affect the behaviour of private agents, but at the same time, private agents also affect the governments's course of action as well as policies. This implies that the macroeconomic situation of Botswana is outside the scope of rational expectations theory.
(iii) The Neo-Ricardian Theorem
Although the Botswana government has not recorded a deficit in the last ten years, money markets are not yet well developed in the country, and therefore government deficits would be usually money-financed. In the case of money-financed deficits, it is held that the neo-Ricardian effects would not be expected if the rise in money supply neither increases inflation nor causes current account deficit, but just boosts output in a Keynesian sense. If inflation is expected, the expectation that the real value of money holdings will depreciate will lead to higher savings. The balance on current account in Botswana has been positive since 1983 on account of the favourable terms of trade. Moreover, the rise in money supply does not determine or explain the behaviour of the current account deficits in the case under consideration. Of the three causes or sources of inflation in regard to Botswana's situation, facts available suggest that cost-push inflation contributes most significantly to the overall price increases in view of the fact that price developments in Botswana are, to a large extent, linked to those in the RSA. This is because the country's imports come predominantly from the latter country, where inflation has been soaring high in recent years. The amount of inflation due to increases in money supply in Botswana is relatively small. Moreover, as already stated, it is highly doubtful whether Keynesian effects are relevant in Botswana's context, although it cannot be denied that Botswana's fiscal policies can create certain expectations at home and abroad, which may have (early) effects on the exchange rate, private savings, private investment and other factors, including the external balance. It is therefore doubtful whether the neo-Recardian situation prevails in the case of Botswana.
(iv) Supply-Side Models And Sectoral Income Effects
Supply-side economics deals with the study of factors and policies influencing and affecting the real economy as distinct from the nominal economy, that is, the physical characteristics of economic agents, and their response to variations in real (relative) prices as distinct from nominal prices, (Corden, 1987). Supply-Side models give an allowance for a `rational' behaviour and appear to be consistent with the Botswana situation. These models have been used successfully in a number of developing countries, including Tanzania (Lipumba et al, 1988).
They also allow for negative supply shocks to produce both increases in unemployment and decreases in real wages while at the same time setting wages independently of nominal demand shocks but responsive to real shocks.
It has been observed that some shock, whether nominal or real in origin, temporary or not, may be favourable, causing real wages to rise and that these real wages may not fall when the relevant shocks are reversed. This position cannot be ruled out in the case of Botswana. One example is a temporary decrease of the price of diamonds in the world market.
This situation cannot be modelled under, say, the rational expectations hypothesis. For, according to the relevant theory, if it were known that the shock was anticipated to last for a long time, rational expectations hypothesis would require the wage negotiators (trade unions) to lower their demands in the light of the new information.
The Dutch Disease theory, whose main concern is in the sectoral real income effects and the overall macroeconomic consequences of a sectoral real income boom, falls under supply-side economics. The Dutch Disease theory, as discussed earlier, adequately explains Botswana's contemporary economic situation. The uniform characteristics of all Dutch Disease models is the emphasis they place on real income and output effects on the various sectors of the economy (Corden, 1987). The foregoing details indicate that supply-side modelling is the most suitable procedure in the macroeconometric model specification and estimation for Botswana.
