This Selected Economic Issues paper examines economic development in South Africa during 1995–96. The paper highlights that in 1995, the economy of South Africa grew by 3.3 percent, the third consecutive year of economic growth, and it is expected to grow between 3½ and 4 percent in 1996. Some aspects of the unemployment problem are addressed in this paper. The paper also focuses on the implications for policy of the steps taken in 1994 and 1995 to establish an outward-oriented economy, after many years of effective autarky.

Abstract

This Selected Economic Issues paper examines economic development in South Africa during 1995–96. The paper highlights that in 1995, the economy of South Africa grew by 3.3 percent, the third consecutive year of economic growth, and it is expected to grow between 3½ and 4 percent in 1996. Some aspects of the unemployment problem are addressed in this paper. The paper also focuses on the implications for policy of the steps taken in 1994 and 1995 to establish an outward-oriented economy, after many years of effective autarky.

V. Accelerating Growth and Fiscal Policy

1. Introduction

A sustained increase in South Africa’s growth to 6 percent a year, a level consistent with declining unemployment, will require an increase in the level of fixed investment from a projected 18 percent of GDP in 1996 to about 26 percent of GDP by the turn of the century. Funding this increase in investment raises four questions. First, how much of the increase in investment can be financed from abroad and, thus, how much will domestic saving need to rise? Second, to what extent might fiscal actions aimed at stimulating domestic saving be offset by declines in private sector saving? Third, given the Government’s commitments to avoid increasing the tax burden (presently at around 26 percent of GDP) and to maintain capital expenditure at 2.7 percent of GDP, can the required public saving be generated without recourse to unrealistic cuts in noninterest current expenditure? If not, what is the role of fiscal policy in the efforts to promote the growth and employment objectives?

The chapter is organized along the lines of answering these questions. The second section discusses the issue of the level of foreign saving that can be expected to be available to the South African economy over the medium term. It is argued that while no specific level can be ascertained with certainty, there are reasons to assume that, under current economic conditions, foreign saving are unlikely to exceed 4 percent of GDP. Moreover, a pre-condition for any sustained inflow of foreign saving is that the authorities’ fiscal stance is sustainable, in the sense that the debt-to-GDP ratio declines over time, i.e., there is no public-debt trap. Section 2 concludes with an analysis of this issue.

The third section analyzes the impact of efforts to raise public sector saving on private saving behavior. It is shown that in South Africa any effort to raise public sector saving leads to a decline in private saving. However, the offset may vary depending upon the policies employed to generate the increase in public saving, i.e., the offset is likely to be lower when expenditure cuts rather than tax hikes produce the increase in public saving.

Finally, the last section discusses elements of the role of fiscal policy, as part of a wider program of action aimed to raise growth and employment. It is suggested that public sector dissaving would need to be eliminated and indeed saving be generated and that a number of changes in the composition of public expenditure would be desirable.

2. Foreign saving constraint

The level of foreign saving that can be expected to finance increased investment in South Africa depends on domestic considerations and the external environment. With respect to the former, the fiscal stance assumed by the authorities will have a significant role, as it affects the country’s risk premium. While no particular level of deficit or surplus can guarantee unlimited access to foreign saving, an unsustainable fiscal position would raise the perceived risk and result in little, if any, access to foreign resources. In this regard, there is an ongoing debate in South Africa as to whether the country is in a public-debt trap (defined as an increasing ratio of government debt to GDP), and therefore whether the present stance of fiscal policy is sustainable. Critics of current policy claim that proof that South Africa is in a debt trap can be seen from the fact that the current yield-to-maturity on government debt exceeds the growth rate of nominal GDP.

Appendix I of this chapter shows that the relationship between the yield-to-maturity on public debt and nominal GDP growth is only one element of the sufficient condition for a rising debt-to-GDP ratio. Other key elements include the overall fiscal stance, particularly the primary balance. At current rates of growth of nominal GDP, a public-debt trap would be avoided if the surplus in the primary balance averages between 1.5 percent and 2 percent of GDP. The current medium-term fiscal framework would lead to a primary surplus exceeding 1.5 percent of GDP by 1997/98, and assuming that the overall deficit continues to be reduced by 0.5 percent of GDP a year thereafter, the primary balance would show surpluses on the order of 3 percent of GDP by the turn of the century (Table 14); this would be sufficient to avoid a public-debt trap.

In these circumstances, no absolute limit on the availability of foreign saving to South Africa can be determined precisely, but a reference level can be established (and will be used in section 3 for simulations regarding the extent to which fiscal policy can generate the necessary higher domestic saving).

Table 14.

South Africa: Macroeconomic Framework—Authorities’ Fiscal Scenario

article image

In the last Selected Economic Issues paper (SM/95/21), it was pointed out that during the period 1960 to the mid-1980s investment in South Africa averaged 26 percent of GDP, and external saving played an important role in maintaining that investment level in the face of changes in the level of national saving. After the imposition of financial and economic sanctions and the re-establishment of the financial rand system, there was a steady outflow of capital that constrained investment to the level of national saving.

The end of apartheid and the recent abolition of the financial rand created the opportunity for external capital flows to return to the role they played for the quarter century before the mid-1980s, when foreign saving averaged more than 5.5 percent of GDP. However, there is still a significant risk premium that results from domestic political considerations, as well as from continued public sector dissaving through fiscal year 2000/01. Moreover, in the wake of the Mexican crisis of December 1994, financial markets have been quite cautious, and it is unlikely that South Africa could run external current account deficits in excess of 4 percent of GDP, even assuming that net foreign direct investment rise to about 2 percent of GDP a year. Furthermore, it is likely that this level of foreign saving will be available only over time, as international lending markets stabilize and political and economic uncertainties in South Africa are removed.

As shown in the scenario presented in Table 15, an investment target of 26 percent of GDP by the turn of the century along with foreign saving availability that rises over time to almost 4 percent of GDP implies that domestic saving must rise from 15.5 percent of GDP at present to 22.1 percent by FY 2000/01. 1/ Then, the key question is how much of the domestic saving increase should be generated by the public sector?

Table 15.

South Africa: Fiscal Policy—Six Percent Real Growth by FY 2000/01

article image

The reduction in 1997–99 would come from a 40 percent reduction in all subsidy programs.

3. Public sector saving target over the medium term

An increase in public sector saving is offset fully when conditions for strict Ricardian equivalence prevail. However, when conditions for strict Ricardian equivalence do not prevail, it is possible that a rise in public saving is only partly offset by lower private saving. A situation of no offset is consistent with textbook Keynesian analysis.

There is evidence that the size of the offset in developing countries is not only affected by the degree to which the conditions for Ricardian equivalence apply. It is also affected by the composition of the fiscal measures with higher offsets when the fiscal balance is strengthened by tax increases than when it is strengthened by equivalent expenditure reductions. Empirical estimates of these offsets (see Appendix II) suggest that a 1 percentage point of GDP increase of government saving is offset by a decline in private saving of 0.16 percentage point of GDP, when the public sector saving increase is generated by expenditure reductions, and 0.7 percentage point of GDP, when the public sector saving increase is generated by raising taxes.

The results of the empirical work suggest that to reach an overall increase in domestic saving of 6.6 percentage points of GDP, and assuming no change in private sector saving behavior from other sources, public sector saving would need to rise to unrealistic levels. Even if the offset is as low as 0.16 percentage point of GDP per percentage increase in government saving, the necessary hike in government saving would be just under 8 percentage points of GDP, implying continuous cuts in real expenditure throughout the remainder of the decade (Table 15).

An argument can be made, however, that increased investment and growth would “crowd in” private saving. But these increases are most unlikely to be sufficient to fund the required increase in fixed investment without a strengthening of the authorities’ current fiscal program. The mid-panel of Table 16 shows that private saving would have to rise by 4 percentage points of GDP in order to fund the increase in investment under the current fiscal program.

Table 16.

South Africa: High Growth Framework—Public and Private Savings

article image

4. Fiscal policy and structural reform

As just indicated a program of structural reforms aimed at increasing employment and growth would promote private saving and thus alleviate the burden on fiscal policy for several reasons. First, evidence from the fast growing economies in Southeast Asia suggests that faster growth leads to increased private saving. Second, accelerated efforts at reform of tax and pension systems (described in the background section of this report), as well as privatization, would also spur private saving. Third, structural reforms may “crowd in” private saving by imparting greater confidence in the economy’s long-term prospects. Fourth, implementation of a structural reform program, ceteris paribus, would eventually raise the availability of foreign saving that could be channelled into the domestic economy beyond the 4 percentage points of GDP assumed above.

The lower panel of Table 16 outlines a fiscal policy stance over the medium-term consistent with our previous analysis and the implementation of a structural reform program. While it is rather difficult to quantify the stimulus to private saving of the structural reform package, the level of public sector saving that would need to be generated would be of a far lesser scale than implied in the previous analysis and as called for today in some South African circles.

In light of the discussion in section 3, public sector saving would need to rise through a reduction in expenditure to minimize the offset. Recent work by the Fund’s Fiscal Affairs Department on fiscal policy and its impact on growth concludes that a growth promoting public sector should complement private sector activity in order to foster a more intensive utilization of existing capacity and/or reallocation of existing resources and should promote increases in the stocks of physical and human capital and technological development. 1/

In South Africa, aggregate education spending is arguably already sufficient; however, the composition of that spending is inefficient: spending at the tertiary level could be reduced with a concurrent increase in the allocation to primary and secondary education, and moreover, nearly 85 percent of current education-related expenditure is in wages; this could be scaled back through improved targeting of wages and teacher expertise with the savings shifted to infrastructure improvements. In the area of healthcare, there may be a need for more resources, and this is being addressed, in part, through the RDP. There may also be further scope to shift resources away from military expenditure and toward police and judicial services.

One area of expenditure that stands out as a drain on budgetary resources and that distorts private sector decision-making is subsidies. Also, certain nonessential economic services could be reduced or eliminated. While certain consumer subsidies may be justified on social welfare grounds, business subsidies are generally unwarranted. Appendix III of this chapter contains a list of subsidies to private sector firms that could be cut or eliminated in amounts consistent with the fiscal strategy set out in the lower panel of Table 16.

In summary, the current fiscal strategy of the authorities is sufficient to avoid a public-debt trap, but insufficient to generate sufficient domestic saving to fund a rise in investment of 8 percentage points of GDP by the turn of the century. Fiscal adjustment will thus form a central element of the structural reform effort to stimulate private saving, growth, and employment.

Primary Balance and a Declining Debt-to-GDP Ratio

The relationship between the overall fiscal position and changes in the stock of debt can be stated as follows:

Dt=Gt+Dt-1,(1)
article image

More substantively, eq. 1 can be rewritten as:

Dt=-PSt+i*((Dt+Dt-1)/2)+Dt-1,(2)
article image

Generating the sufficient condition for a decline in the stock of debt as a percentage of GDP is most easily discerned by first finding the condition under which the debt-to-GDP ratio remains constant. By definition, a constant ratio of debt-to-GDP implies that:

Dt=(1+g)*Dt-1(3)
article image

Substituting eq. 3 into eq. 2 and collecting terms yields:

PSt=Dt*((i-g)/(1+g)+(i*g)/(1+g)).(4)

Rewriting eq. 4 in terms of proportions of GDP yields:

pst=dt[-]1*((i-g)/(1+g)+(i*g)/(1+g))(5)
article image

The first factor on the right-hand side of eq. 5, (i-g)/(l+g), is the real interest rate in terms of nominal GDP. Thus, to maintain a constant debt-to-GDP ratio, the primary balance as a percentage of GDP must cover the real interest bill. The second term is the standard differential growth term that in many analyses is assumed to be second-order small and, therefore, ignored. However, in the case of South Africa, this interest growth term is almost equal to the real interest factor and, thus, cannot be assumed away. Eq. 5 implies that the debt-to-GDP ratio falls (rises) when the primary surplus is larger (smaller) than the real interest bill plus the growth factor in the interest bill.

Under present circumstances and forecasts for beyond 1996/97, dt-1 = 0.56, g = 0.1124 over the medium term, and i can range from 0.128–0.14. Substituting these parameters into eq. 5 yields a range for pst of between 0.015–0.20. Therefore, a sustainable fiscal position requires a primary balance of 1.5–2 percentage points of GDP. Under the authorities’ stated objectives, the primary balance will reach 1.7 percent of GDP in 1997/98 and over 3 percent of GDP by the turn of the century.

The Behavior of Public and Private Saving in South Africa

The analysis of the relationship between private and public saving in South Africa follows a study done by Corbo and Schmidt-Hebbel. 1/ In this research, the authors sampled 13 developing countries over the period 1968–1988 to analyze the impact of public saving on private saving. 2/ The authors first sought evidence for the existence of conditions for strict Ricardian equivalence. They reported that while there was evidence that some private agents do not face liquidity constraints, fiscal policy has a role in raising overall domestic saving as certain agents are not able to smooth out their consumption paths in a perfect manner. 3/ The second stage of their analysis was to see the effect of public saving on private saving. The model they employed of the determinants of private saving is:

Pt=b0+b1Pt-1+b2PDYt+b3PSt+b4RIt+b5CPIt+b6Wt+b7FSt+et(1)
article image

The authors found that raising public sector saving raises overall domestic saving (b3>-l), while the impact of interest rates and inflation on saving rates is minimal. The stock of wealth does appear significant but with the opposite sign as expected. Foreign saving, however, do substitute strongly for domestic saving.

Finally, the authors tested the sensitivity of private saving to improvements in public saving generated by tax increases and expenditure reductions. They found that if public saving is increased 1 percentage point of GOP through a reduction in public expenditure, private saving will be reduced by 0.16 or 0.50 percentage point of GDP depending upon the model specification of the expected permanence of the increase in government saving. If public saving is increased 1 percentage point of GDP through a hike in taxes, private saving will be reduced by 0.48 or 0.65 percentage point of GDP depending upon the model specification of the expected permanence of the increase in government saving.

These results are basically at the lower end of the spectrum regarding the offset factor for developing countries. Masson, Bayoumi, and Sarniei (1995) found in a sample of 40 developing countries that a one percentage point of GDP rise in public saving is offset by a 0.66 percentage point of GDP decrease in private saving and an even higher offset of 0.94 percentage point of GDP for a subset of high-income developing countries. 1/ Savastano found in a sample of developing countries with Fund-supported programs that a one percentage point of GDP rise in public saving is offset by a 0.70 percentage point of GDP decrease in private saving. 2/

The model of saving being used builds on the work of Corbo and Schmidt-Hebbel and is as follows:

Pt=C+b1Gt+b2FSt+b3RPGt+b4TOTt+b5ACTt+b6RIt+b7Wt+b8DISCt+b9APART+et(2)
article image
article image

DISCt and APARTt included to measure the impact of risk on saving behavior: country risk in South Africa should be captured through the financial rand discount and the imposition of official sanctions. The real price of gold is included as an explanatory variable because of the potential impact on corporate saving in the mining sector. Saving rates are scaled in terms of GDP (not disposable income) since the analysis is in terms of total private saving, including corporate saving. Wealth is measured by the Johannesburg stock market index rather than a broad monetary aggregate, and the nongold terms of trade is included as an explanatory variable (again in reflection of the inclusion of corporate saving).

Preliminary work found several of the explanatory variables not significant and these were dropped from the final regressions. The results reported below are for the following restricted model:

Pt=C+b1Gt+b2FSt+b3RPGt+et

The time series for these variables cover the period 1961–1994 and were tested for stationarity and cointegration. The data were tested for stationarity using Augmented Dickey-Fuller (ADF) tests. Stationarity was accepted at the 5 percent significance level only in the case of real interest rates. 1/ However, bivariate tests for cointegration revealed that all the explanatory variables cointegrated with private saving. 2/ In this circumstance two sets of OLS regressions were run. First, given that the series cointegrated, the long-term relationship between private and public saving described in equation 3 was analyzed using a simple OLS regression model (with data corrected for first-order autocorrelation using a Cochrane-Orcutt transformation). 3/ The results are presented in Table 17.

Table 17.

South Africa: OLS Regression Results for Saving Model

(Data corrected for first-order autocorrelation)

article image

The offset in private saving from a one percentage point of GDP increase in public saving in this case is 0.60 percentage point of GDP in private saving, and public saving was statistically significant. 1/

In the second regression, as the data was nonstationary, it was first differenced to remove trends that allowed for testing of the shorter-term dynamics between public and private saving. 2/ In this case a one percentage point of GDP rise in public saving was offset by a decline in private saving by 0.69 percentage point of GDP, and public saving was again significant (Table 18). 3/ The larger short-term offset factor may be reflecting that liquidity constraints are stronger in the short run than in the long run.

Table 18.

South Africa: OLS Regression Results for Saving Model

(First-differenced data)

article image

Possible Subsidy Reduction Program

article image
1/

The 4 percent constraint is illustrative, and the results could be adjusted for the use of higher (or lower) maximum external current account deficits.

1/

EBS/95/166 and the companion paper SM/95/275 address the issues of the composition of fiscal adjustment and growth.

1/

Corbo, V. and Schmidt-Hebbel, K., “Public Policies and Saving in Developing Countries”, Journal of Developments Economics, Vol. 36, 1991, pp.89–115.

2/

The authors carried out both individual country least squares estimations and a panel estimate.

3/

This is consistent with other studies that reject the notion of the Ricardian equivalence in developing countries. See Montiel, P. and Haque, N.U. Ricardian Equivalence, Liquidity Constraints, and the Yari-Blanchard Effect: Tests for Developing Countries, IMF Working Paper, 1987.

4/

Data is scaled to income to mitigate the problem of nonstationarity of the time series.

1/

Masson, P., Bayoumi, T., and Samiei, H., “International Evidence on the Determinants of Private Saving”, IMF WP/95/51, May 1995.

2/

Savastano, M., “Private Saving in Fund Arrangements”, Schadler, et. al., IMF Conditionality Review, OP 129, September 1995.

1/

Tests at the 1 percent level would reject stationarity

2/

Tests on foreign savings were inconclusive at the 1 and 5 percent significance levels.

3/

For a full treatment of the issues of stationarity and cointegration see, C.V.J. Granger, “Some Properties of Time Series Data and Their use in Econometric Model Specification,” Journal of Econometrics. Vol. 16, No. 1, 1981, pp. 121–130, and R.F. Engle and C.W.J. Granger, “Cointegration and Error Correction: Representation, Estimation and Testing,” Econometrica. Vol. 55, No. 2, 1987, pp. 251–276.

1/

In previous work done by FAD staff in this area, it was found that most of the offset was generated in corporate saving, while household saving behavior was generally unresponsive to changes in public savings. Moreover, corporate savings were strongly correlated (negative) with real wages, the latter was not included in the model as the series does not begin until 1970.

2/

the first-differenced data were stationary under ADF tests.

3/

Vector autoregressions (VARs) were run on the restricted model (first-differenced data) with public savings, foreign savings, the real price of gold, and private savings treated as endogenous variables; however, the VARs did not improve upon the explanatory power of the simpler OLS regressions. Also, regressions were run with proxies for distinguishing between permanent and transitory income and policy shocks in line with the work of Corbo and Schmidt-Hebbel (1991) and other literature in this area; however, the results were not robust.