The urgent need to raise living standards and improve the distribution of income and wealth in South Africa poses a major challenge to policymakers over the medium term. Unless the rate of economic growth rises to a level well above that of the last decade, there is little prospect for reducing the current very high level of underemployment in the economy or for generating the resources needed to address social needs on a sustained basis.
This chapter provides an analysis of medium-term growth policies based on some illustrative simulations carried out on a highly aggregated model of the South African economy. The conclusions can be broadly summarized as follows: (1) a return to more rapid economic growth will require a substantially higher rate of domestic savings and investment than in the past decade as well as a reversal of the declining productivity trend of the 1970s and 1980s; (2) government savings are essential to underpin higher economic growth; thus, the scope for expanding overall government spending will be limited; and (3) if employment growth is to rise enough to begin alleviating the existing, severe underemployment problem, real wage growth must be contained.
Although medium-term prospects and policy implications cannot be gauged with certainty, two clear messages emerge from the analysis. First, fiscal resources for spending on the more disadvantaged members of society will have to come mainly from a redirection of fiscal expenditures broadly within the current overall share of expenditures in GDP. Attempts to improve the distribution of income through an expansion of fiscal expenditures are likely to be self-defeating, because they would either divert resources needed for investment or would require tax levels that would introduce strong disincentives to work and save, thereby lowering economic growth potential. Second, large real wage increases will not, over the medium term, help to improve overall standards of living or the distribution of income. Instead, they will act to reduce output growth and to raise a steep barrier between the employed and the growing number of underemployed.
Analytical Framework
The model used to produce the mediumterm scenarios has two key elements: (1) a production function that specifies the relationship between output and the factor inputs, capital, and labor; and (2) a savings-investment accounting framework that focuses on the mobilization of resources by different sectors of the economy necessary to finance any level of capital accumulation.
The production function, which is described in more detail in Chapter III, is of the simple Cobb-Douglas variety that assumes that the elasticities of output with respect to capital and labor inputs are both constant and add up to unity. The elasticity with respect to labor is assumed to be about 0.65, which reflects the estimated share of labor income in the nonprimary sector of the economy. Labor input is, in turn, assumed to be composed of a skilled and an unskilled component, which are aggregated according to a constant elasticity of substitution formula.
A trend term (multifactor productivity) is also included in the production function to account for growth from sources other than increased labor and capital inputs, including from technical progress. At best, multifactor productivity growth was stagnant in the 1980s, and productivity is currently estimated to be about 5 percent below its level at the beginning of the 1970s. The poor multifactor productivity performance may reflect a combination of inefficiencies and obstacles to the functioning of free markets and restrictions on the importation of new technology from overseas.
The savings-investment identity in the model provides a financing link between the level of fixed investment needed for real growth and the mobilization of resources in the economy. Investment finance can come from three sources of savings: the domestic private sector, the Government, and the foreign sector. The first represents forgone current consumption out of disposable income by individuals and corporations in the economy. The second represents government revenue that exceeds current expenditures and transfers. The third represents net capital inflows from abroad, which, apart from statistical errors, should be equal to the deficit on the external current account.
The savings-investment identity can be viewed from two perspectives. First, if the sources of savings are given, investment is determined by total savings. Lower savings in one sector, unless offset by higher savings in another, would lead to lower overall savings and hence lower investment. In practice, there are links between the saving decisions of different sectors—for example, lower private savings may offset, in part, higher public savings through a tighter fiscal stance. However, the precise links are hard to quantify empirically. In the model, the savings-investment identity is used as an adding-up constraint to answer questions such as: “assuming that additional foreign and private savings cannot be mobilized, what will be the effects on investment and growth of lower government savings?”
The second perspective would be to begin with a given level of investment and to examine the financing implications. Thus, a particular, desired growth path will require certain levels of employment and investment. In turn, the investment needed can be financed only out of available savings. If some sources of savings are constrained to “reasonable” limits (for example, realistic or prudent current account deficits or tolerable levels of government deficits), then the remaining sources of savings have to take up the residual financing of investment.
Scenarios
The model was used to examine several scenarios for the South African economy that are presented as variants around a main, or baseline, scenario. It should be stressed that neither the main scenario nor the variants should be taken as forecasts in the sense of representing the “most likely” outcomes. Instead, the scenarios are vehicles for examining the implications of different policies within a consistent framework. It should also be noted that the scenarios have a medium-term orientation (the actual time horizon is ten years), and no attention is paid to short-run demand-multiplier or cyclical developments. Over the medium term, it is reasonable to assume that supply-side factors are more important determinants of growth than are demand conditions.
The scenarios are based on the assumption that apartheid has been fully dismantled and that structural policies are being pursued with the aim of reducing impediments to the functioning of markets. In addition, it is assumed that monetary policy will continue to be used to bring inflation under control and that South Africa will have access to both international financial markets and to improved international technology. As a consequence, multifactor productivity is assumed to rebound to an average rate of increase of about ½ of 1 percentage point a year. While the change in multifactor productivity from its historical trend is large, the assumed multifactor productivity growth is not high by international standards.
Common to all the scenarios is the further assumption that, with the removal of sanctions, South Africa would experience an average capital inflow of 1¾ percent of GDP a year over the ten-year projection period. This inflow is lower than that which South Africa experienced in the early 1980s (3 percent of GDP) owing to the uncertainty associated with political change and to the high demand for global savings. However, by the end of the projection period, the capital inflow is assumed to rise to 2 percent of GDP.
Scenario 1: Baseline
The baseline scenario is concerned with the level of domestic savings and investment that would be needed to support a rate of long-term economic growth consistent with both an improvement in living standards and a steady decline in the underemployment rate. With the population expected to grow by over 2¼ percent a year, a meaningful increase in living standards would require GDP growth of at least 3½ percent a year (Table 7). At the same time, the objective of reducing underemployment would require employment growth of at least 3 percent a year.19 On the basis of the production function referred to above and given the assumed rate of multifactor productivity growth, the a forementioned objectives for growth in output and employment could be met only through a marked improvement in investment performance. Specifically, capital accumulation would have to be strong enough to raise the investment/GDP ratio steadily, to 27 percent by the end of the period from its current level of 19 percent.
Medium-Term Baseline Scenario
(In percent, unless otherwise indicated)
Estimated value at end of period.
Includes the savings of public corporations and public business enterprises.
Balance on current transactions of the general government on a national accounts basis.
Medium-Term Baseline Scenario
(In percent, unless otherwise indicated)
| Average 1981–90 |
1990 | Average 1991–2000 |
||
|---|---|---|---|---|
| Real GDP growth | 1.4 | –0.9 | 3.5 | |
| Employment growth | 0.7 | –0.4 | 3.0 | |
| Nonwhite underemployment rate1 | 41.7 | 41.7 | 36.6 | |
| Real wage growth | 1.4 | 1.9 | 0.7 | |
| Investment/GDP | 23.0 | 19.3 | 24.8 | |
| External current account/GDP | 0.4 | 2.2 | –1.7 | |
| Private savings/GDP2 | 23.6 | 20.9 | 21.5 | |
| Government savings/GDP3 | –0.2 | 0.6 | 1.6 | |
| Government revenue/GDP | 25.5 | 28.5 | 29.6 | |
| Government expenditure/GDP | 28.1 | 29.7 | 30.3 | |
| General government balance/GDP | –2.7 | –1.2 | –0.7 | |
| Current government expenditure per capita, 1990 rand1 | 2,372 | 2,372 | 2,568 | |
| Miscellaneous assumptions | ||||
| Multifactor productivity growth | –0.4 | –0.6 | 0.5 | |
| Labor income share | 65.9 | 66.5 | 67.4 | |
| Population growth | 2.2 | 2.0 | 2.4 | |
Estimated value at end of period.
Includes the savings of public corporations and public business enterprises.
Balance on current transactions of the general government on a national accounts basis.
Medium-Term Baseline Scenario
(In percent, unless otherwise indicated)
| Average 1981–90 |
1990 | Average 1991–2000 |
||
|---|---|---|---|---|
| Real GDP growth | 1.4 | –0.9 | 3.5 | |
| Employment growth | 0.7 | –0.4 | 3.0 | |
| Nonwhite underemployment rate1 | 41.7 | 41.7 | 36.6 | |
| Real wage growth | 1.4 | 1.9 | 0.7 | |
| Investment/GDP | 23.0 | 19.3 | 24.8 | |
| External current account/GDP | 0.4 | 2.2 | –1.7 | |
| Private savings/GDP2 | 23.6 | 20.9 | 21.5 | |
| Government savings/GDP3 | –0.2 | 0.6 | 1.6 | |
| Government revenue/GDP | 25.5 | 28.5 | 29.6 | |
| Government expenditure/GDP | 28.1 | 29.7 | 30.3 | |
| General government balance/GDP | –2.7 | –1.2 | –0.7 | |
| Current government expenditure per capita, 1990 rand1 | 2,372 | 2,372 | 2,568 | |
| Miscellaneous assumptions | ||||
| Multifactor productivity growth | –0.4 | –0.6 | 0.5 | |
| Labor income share | 65.9 | 66.5 | 67.4 | |
| Population growth | 2.2 | 2.0 | 2.4 | |
Estimated value at end of period.
Includes the savings of public corporations and public business enterprises.
Balance on current transactions of the general government on a national accounts basis.
As for real wage growth, all the scenarios were based on the assumption that labor’s share of income would rise by ¼ of 1 percentage point a year, mainly to reflect the dismantling of distortionary tax practices that favored the use of capital and artificially boosted corporate earnings. In the base-line scenario, with GDP growth of 3½ percent and employment growth of 3 percent, the drift toward a rising labor share of income would allow an average real wage growth on the order of ¾ of 1 percent a year.20
A significant portion of the required increase in investment could be financed by the anticipated swing in the external balance from surplus to deficit. However, the increased access to external savings would need to be supported by a continued effort to maintain domestic savings. The baseline scenario assumes, perhaps somewhat optimistically in view of the prospective redistribution of income, that the private sector savings ratio would rise moderately from just under 21 percent of GDP at present to 21½ percent at the end of the projection period. Even with this improvement, the public sector would be required to contribute roughly1½ percent of GDP a year to national savings, which would represent a considerable improvement over the performance of the 1980s.21
The need to maintain public savings at a level high enough to support the overall domestic savings and investment effort has clear implications for the design of budget policy: the scope for increasing public spending in relation to GDP would necessitate an increase in the overall tax revenue burden. However, with the large rise in general government revenue in relation to GDP over the 1980s—it stood at 28½ percent in 1990—the room to raise revenues further without inducing severe disincentive effects might be limited.22 Accordingly, if growth prospects are not to be jeopardized, the social backlogs would need to be redressed mainly through a redirection of budget priorities and through a basic reordering of social spending between different racial groups. It should be noted, however, that since GDP is growing more rapidly than the population in the projection period, a constant ratio of public spending to GDP would mean a marked increase in government spending per person. By the end of the projection period, if the expenditure-to-GDP ratio did not change, real per capita current government expenditure would exceed its 1990 level by nearly 10 percent.
Scenario 2: Higher Fiscal Deficit
As an alternate scenario, government savings are reduced relative to the baseline case, but there is no corresponding increase in foreign or private savings. Technological assumptions and real wage growth are the same as in the baseline case. To illustrate the possible trade-offs involved, the Government is assumed to raise its current expenditures by 1 percentage point of GDP above the baseline case. With no alternative sources of saving, the investment/GDP ratio must also fall by 1 percentage point of GDP. As a result, output and employment growth are both reduced by 0.3 percentage points a year (see Table 8 and Chart 4). Lower employment growth reduces the decline in the underemployment rate, while lower GDP growth implies that the higher ratio of current expenditures to GDP actually translates into virtually no higher growth of real per capita current expenditures than in the baseline case.23 In effect, if resources are appropriated for current expenditures, there are fewer resources for capital accumulation and less growth to generate resources in the future.
Higher Fiscal Deficit
(In percent)
End of period; end-1990 estimated at 41.7 percent.
Higher Fiscal Deficit
(In percent)
| Average Growth 1991–2000 |
Difference from baseline |
|
|---|---|---|
| GDP growth | 3.2 | –0.3 |
| Employment growth | 2.7 | –0.3 |
| Nonwhite underemployment rate1 | 38.3 | 1.7 |
| Budget balance/GDP | –1.7 | –1.0 |
| Real per capita current government expenditure growth | 0.9 | 0.1 |
End of period; end-1990 estimated at 41.7 percent.
Higher Fiscal Deficit
(In percent)
| Average Growth 1991–2000 |
Difference from baseline |
|
|---|---|---|
| GDP growth | 3.2 | –0.3 |
| Employment growth | 2.7 | –0.3 |
| Nonwhite underemployment rate1 | 38.3 | 1.7 |
| Budget balance/GDP | –1.7 | –1.0 |
| Real per capita current government expenditure growth | 0.9 | 0.1 |
End of period; end-1990 estimated at 41.7 percent.
Medium-Term Scenarios
(In percent)
Source: South African Reserve Bank, Quarterly Bulletin.Scenario 3: Higher Real Wage Growth
A second alternative scenario explores the effect of imposing rapid real wage growth on the economy as the result of, for example, increased militancy of the unions. It is assumed here that real wage growth would rise by 1¼ percent a year over the projected period or by ½ of 1 percentage point faster than in the baseline case.
The main effect of imposing higher real wage growth is to reduce the demand for labor. On the assumption that the investment performance is similar to that in the first scenario, the higher real wage growth assumed here would push employment growth down to 1½ percent a year (see Table 9 and Chart 4). As a result, GDP growth would be lowered by 1 percentage point a year. Lower employment growth also implies that the recent upward trend in the nonwhite underemployment rate is not arrested; indeed, relative to the baseline scenario, some 1¼ million fewer jobs are created for nonwhites. A further consequence of lower economic growth is that growth in real government current expenditures is much slower: real current expenditures are nearly 10 percent below the baseline level by the year 2000. In summary, the effect of high real wages could be to raise the inequality of income distribution by keeping a much larger proportion of the working population underemployed and, through lower economic growth, by limiting the scope for fiscal redistributive spending.
Higher Wage Growth
(In percent)
End of period; 1990 estimated at 41.7 percent.
Higher Wage Growth
(In percent)
| Average Growth 1991–2000 |
Difference from Baseline |
|
|---|---|---|
| GDP growth | 2.5 | –1.0 |
| Employment growth | 1.6 | –1.4 |
| Nonwhite underemployment rate1 | 46.3 | 9.7 |
| Real wage growth | 1.3 | 0.5 |
| Real per capita curren t government expenditure growth | –0.3 | –1.1 |
End of period; 1990 estimated at 41.7 percent.
Higher Wage Growth
(In percent)
| Average Growth 1991–2000 |
Difference from Baseline |
|
|---|---|---|
| GDP growth | 2.5 | –1.0 |
| Employment growth | 1.6 | –1.4 |
| Nonwhite underemployment rate1 | 46.3 | 9.7 |
| Real wage growth | 1.3 | 0.5 |
| Real per capita curren t government expenditure growth | –0.3 | –1.1 |
End of period; 1990 estimated at 41.7 percent.
Such employment growth would be consistent with an increase in white employment growth of 0.9 percent a year, which would leave the white underemployment rate approximately unchanged, and an increase in nonwhite employment of 3 ½ percent a year, which would allow for a decline in the nonwhite rate of unemployment by about ½ of 1 percentage point a year.
The real wages of nonwhites might rise a little faster than this, while those of whites might rise somewhat slower, as remaining apartheid distortions are eliminated. See Chapter III for an estimate of the scope for nonwhite wages to catch up.
If public sector investment averages 2½ percent of GDP over the projected period, providing the needed public sector savings would mean a general government deficit of under 1 percent of GDP.
For a fuller discussion of the limited scope to increase tax revenues, see Chapter VI.
If real wages were to fall relative to the baseline case, part of the employment and growth effects would be offset.
