Gearing the economy for growth: Economic resilience through strategic macroeconomic interventions

Read as PDF

Copyright © 2025

Inclusive Society Institute

PO Box 12609

Mill Street

Cape Town, 8010

South Africa

235-515 NPO

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means without the permission in writing from the Inclusive Society Institute

DISCLAIMER

Views expressed in this report do not necessarily represent the views of the Inclusive Society Institute or its Board or Council members.

April 2025

Author: Prof Jan van Heerden

Editor: Daryl Swanepoel

Table of Contents

Introduction

     Models used

     The core principles of the CGE models used in the simulations

          The baseline forecast

          The policy forecasts

Macro-economic interventions

     1. Alternative debt-reduction roadmap

     2. Youth solidarity tax

     3. Limited public salary bill increases

Conclusion

Bibliography

List of Figures

Figure 1 The baseline and policy forecasts

Figure 2 Debt to GDP ratio (policy simulation)

Figure 3 Increase in public sector debt above the baseline

Figure 4 Government Deficits (policy simulation)

Figure 5 Investment Expenditure and Imports (deviations from the baseline)

Figure 6 GDP and its components (cont.) (deviations from the baseline)

Figure 7 Industry survivors (relative to the baseline)

Figure 8 Short run industry losers

Figure 9 Real GDP % deviation from the baseline

Figure 10 Change in the income tax rate on rich households

Figure 11 Annual deviation from the baseline in the main GDP expenditure components (% change)

Figure 12 Cumulative deviation from the baseline in main GDP expenditure components

Figure 13 Cumulative % change from the baseline in Industry output

Figure 14 Deviation from the baseline in household consumption by income group

Figure 15 Cumulative deviation from the baseline in the main GDP expenditure components levels (revenue not recycled)

Figure 16 Cumulative deviation from the baseline in the main GDP expenditure components (revenue recycled)

Figure 17 Cumulative deviation from the baseline in the Supply Side components of GDP, as well as real wages and the CPI.

Introduction

The report below is a summary of a more technical modelling report of three macroeconomic simulations done for the Inclusive Society Institute in 2024. The three macroeconomic interventions are listed in the textbox at below.

Macro-economic interventions

Alternative debt-reduction roadmap – increase debt to GDP to 100 percent over the next five years followed by a systematic reduction. Proviso: increased debt ringfenced purely for economic infrastructure development, NO consumption spending.

Youth solidarity tax: Increase personal tax by one percent for all taxpayers earning R600,001+ per annum. In the form of a transitional tax for five years only. Ringfenced for youth enterprise development to be administered by the Solidarity Fund. See if this can be funded by delaying inflationary bracket adjustments over the next two or three years.

Limit the public salary bill increases to inflation

Models used

The study was done using Computable General Equilibrium (CGE) models. The Department of Economics at the University of Pretoria has various CGE models, with the same core set of equations, but with different extensions, depending on the tasks at hand. We have used three different models for the simulations: (i) The national model was used for the first macroeconomic intervention where the debt-to-GDP ratio was increased temporarily; (ii) An older version of the provincial TERM model was used for the remaining two macroeconomic interventions, since the national model does not have income taxes (personal or corporate) integrated into the model, and (iii) the latest version of our provincial TERM model was used for some sectoral interventions, included in the technical report, but not reported on here.

The core principles of the CGE models used in the simulations

A CGE model is basically a system of thousands of equations describing the relationships that exist in the economy. It uses data from Statistics South Africa (StatsSA), namely the Supply and Use tables, which resembles a snapshot of the economy at a specific point in time. The Supply and Use tables show that the economy is in equilibrium, or balanced, in the sense that the supply and demand for every commodity or factor of production is equal at that point in time. So, Computable General Equilibrium (CGE) starts with a general equilibrium of all goods, services and factors of production in the economy, and then uses a computer to solve a very large system of equations that describes all the relationships between role players in the economy. We start with equilibrium and then change one or two things in the model, such as a tax rate or spending by the government, or investment by the private sector, and then let the computer calculate the values that will hold in a new equilibrium. A CGE model could be a national model that studies the economy of the country at large, or a multi-regional CGE model which treats each province of South Africa as a separate economy. The regional model not only studies the equilibria of each province, but could also calculate the effects that actions of one province would have on the national economy as well as on all other province.

The equations of the CGE model describe the relationships between all the role-players in the economy, namely the Industries, Households, Investors, Governments and Foreigner buyers, who could be foreign industries, households, investors, or governments, or merely buyers from other provinces in the case of the regional models.

Unlike the Australian CGE model, we have multiple households in our South African models, namely twelve groups distinguished by income: we have the poorest decile split into two ventiles; then the eight next deciles by income, and finally the richest decile also split into two ventiles. The model is therefore very well suited to measure changes in the income distribution or poverty levels. In the older models we had 48 household groups, by income as well as race group.

The basic production function for all industry production is the so-called Leontief function, which means that inputs into the production process are used in fixed proportions: if an industry wants to double output, it has to double all inputs, irrespective of input prices. The Leontief production function is generally accepted by economists world-wide: for example, to produce one car, the factory needs four wheels, one engine, one steering wheel, etc., and to produce ten of those cars, it will need ten times as many wheels, engines, steering wheels, and so forth. Even though we assume that the basic production function has this simple form, industries can choose where to buy their intermediate inputs from, according to CES[1] demand functions: depending on relative prices and the substitutability of the goods from different sources, industries may decide to source their inputs from the local market, or from abroad, depending on price.

Transportation cost in the model plays a very important role, and is included in the purchaser’s price of a commodity. A buyer would therefore rather source from a nearby industry than from one far away. Distances between the capitals of the nine provinces as well as to the different ports of import and export are used to build a gravity mechanism, determining where industries would source their inputs from, or which ports of import and export they will use.

The demand for factors of production also follows the CES functional form: an average level of factors grows proportionally with output, according to the Leontief function, but an industry could substitute labour for capital or vice versa. Labour demand is also modelled using the CES demand functional form: once the industry has decided about the capital-labour ratio it prefers, different occupational groups – we have 11 in the model – are chosen: different wage levels play a strong role, as well as the substitutability of one occupation for another. 

Export demand is a very simple function of domestic versus foreign prices of commodities: if South Africa’s price levels increase relative to world prices, which are exogenous in our model, then the demand for South African commodities decrease. Since South Africa is a small open economy, exports play a very significant role in the model.

Household demand follows the LES (Linear Expenditure System) format: households have a subsistence demand for all commodities, which they will buy first, without looking at the price of the goods. Then they will use the money left over in their budgets to buy “luxury” components of the same goods. For example, we have a subsistence demand for petrol to drive to work every day, but when we go on holiday, the demand for petrol becomes a luxury demand. Subsistence demand is not a function of the price of a good – it is a must have - but luxury demand is a function of the disposable incomes of households.

Unlike households, who maximise utility, given their budget constraints, and industries, who minimize costs or maximise profits, given certain quantities to be produced, we do not have a theory of government in the model. We model the government in two possible ways: (i) exogenously, with its behaviour determined by the modeler (from information published by the government itself, such as in the regular medium-term budget information sessions), or (ii) endogenously, and specifically tied to household behaviour. Governments receive direct and indirect tax income, as well as transfers from all the role players in the economy. The indirect taxes could be modelled as GST’s or a VAT system, and we use the latter. Individual households as well as companies pay direct income taxes as well, while we also have production taxes and subsidies in the model. The government could borrow and build up debt over time, which needs to be serviced on an annual basis. In our simulation we keep the annual government budgets balanced: they give grants to all households, but need to raise either indirect or direct taxes to finance the grants, or a combination of the two types of taxes.

The general method of performing dynamic simulations with a CGE model is to start with a baseline forecast of the macroeconomy, without the economic interventions that we try to implement. A second forecast – the policy simulation – is then performed and compared to the baseline. We generally only report the deviations from the baseline that happen as a result of the policy simulations. So, to some extent we always perform two forecasts: one with the current state of affairs, and a second with some external intervention into the economy. We are then interested in the difference between the two forecasts.

The baseline forecast

The task at hand is to compare the results of three simulations, suggested by the Inclusive Society Institute, to a “baseline” forecast. Any economic forecast is difficult, so we try to keep the baseline forecast as simple as possible. We would typically incorporate macroeconomic forecast data from the National Treasury of South Africa and the International Monetary Fund, amongst others. Specifically, we adopt forecasts for the GDP expenditure components, employment and population growth. We do a “vanilla” forecast, which means that we assume the economy would run on a smooth and moderate path, rather than jumping up and down in a zig-zag fashion. For this study we adopted the strategy of gradually bringing the economy to GDP real growth of around 2% in 2026, and keep it at that level throughout the forecast period.

The policy forecasts

The baseline graphs of all the variables are usually quite smooth, especially of the future forecast values of the variables. We then apply “shocks” to the model – suggested by the Inclusive Society Institute -  by changing one or more of the exogenous variables in the model, which will then alter the smooth path of all the model variables to higher or lower levels than in the baseline. In Figure 1 below the blue line depicts the baseline path of some variable from 2023 to 2039. In 2026 some policy measure is implemented which alters the path of the variable to the orange line. We are typically interested in the size of the deviation from the baseline, and not necessarily the absolute values of the baseline and policy variable values.

Figure 1 The baseline and policy forecasts

Macro-economic interventions

1. Alternative debt-reduction roadmap

The first macroeconomic intervention to be modelled was to “increase debt-to-GDP to 100 percent over the next five years followed by a systematic reduction”, with the proviso that “increased debt (would be) ringfenced purely for economic infrastructure development (and) NO consumption spending”.

It might sound like simple tasks to “increase debt-to-GDP ratio” and “build economic infrastructure” but unfortunately neither task is straight forward.

Debt is built up over time when government expenditure exceeds government income from year to year, so, to increase the debt-to-GDP ratio, the government needs to spend more, or receive less  income, given the size of the GDP. To build infrastructure (and not spend on consumption goods) is tricky in our models. The model database has as its core the National Supply and Use Tables, published occasionally[2] by Statistics South Africa. UN-DESA (2018) describes the Supply and Use Tables (SUTs) as follows:

The SUTs describe the whole economy by industry (for example, the motor vehicle industry) and by product (for example, sports goods). The tables show links between components of GVA, industry inputs and outputs, and product supply and use. The SUTs link different institutional sectors of the economy (for example, non-financial corporations) together with details of imports and exports of goods and services, final consumption expenditure of government, household and non-profit institutions serving households (referred to as NPISHs), and capital formation.

The Supply and Use Tables form part of the system of National Accounts, measuring all aspects of the GDP of a country. However, infrastructure per se is not part of the measurement of GDP!

We therefore cannot simply increase government expenditure on “infrastructure” to increase the government deficit in a particular year, and hence the debt over time. The industries building the infrastructure do have their activities recorded in the SUTs, such as all the commodities that they buy or sell, and the values thereof, but the value of old or new infrastructure does not form part of GDP. So, how do we respond to the goal of simulating the government increasing their debt and building infrastructure? (And answer questions such as” how much growth will this bring?” and “how many jobs would be created?”).

Simulation Design

The industry most closely linked to the building of infrastructure is the Construction Industry, so one might let the government buy more Construction Goods and Services, increasing the deficit in that way. However, even if the government would buy all the available Construction Goods and Services, it would not increase the deficit enough to reach the 100% goal of the debt-to-GDP ratio. Moreover, the government does not actually buy any Construction Goods and Services in the model: of their total expenditure, less than 1% goes to Construction Goods and Services. So, even if they would start buying all available Construction Goods and Services, it would not lead to more or better infrastructure, since they are not directly involved in infrastructure development. Private firms are hired to build infrastructure, and they do buy significant amounts of Construction Goods and Services, when they increase their investment spending.

Even though the government does not buy much Construction Goods or Services, or spend much directly on infrastructure, we could let them subsidize everyone else who does! So, in our simulation we let the government borrow funds (increase their debt) and use the funds to subsidize all investment expenditure. All Investment expenditure in the model takes place according to a fixed recipe: one unit of investment has specific inputs into its production process, and Construction Goods and Services form 64,9% of all domestic inputs into the production of Investment Goods. This means that any R100 spent on Investment in the model would automatically buy R64,90 worth of Construction Goods or Services, and the rest on other inputs such as transport equipment, real estate, business services, etc., which are converted into what we know as infrastructure.

In our view all capital goods in the economy, whether privately or publicly owned, form part of the total economic infrastructure. Capital goods are produced through investment activities and Construction Goods and services form more than half of all investment. So, a subsidy on investment would be the best avenue in our model to improve the general infrastructure in the model.

Simulation Results

Figure 2 shows the results for the debt-to-GDP ratio in our simulation. We could not reach a higher level than 84% of GDP without crashing the model, using a subsidy on all investment activities. CGE models are designed for small shocks to their variables: we typically start with the economy in equilibrium and then apply small changes to one or more variables, and let the computer find a new equilibrium accordingly. We then compare the two equilibria and make conclusions about the differences.  Trying to achieve a debt-to-GDP ratio of 100% is an abnormal type of simulation to try with any CGE model.

Figure 2 Debt to GDP ratio (policy simulation)

To achieve the debt-to-GDP ratios in Figure 2, we needed to apply a subsidy on all investment done by all industries in the model of 30% in 2024, an additional 20% in 2025 and an additional 10% in 2026. To get the debt-to-GDP ratio back to its starting value, we then had to remove the subsidies again from 2027 to 2030. Figure 3 shows the level of extra debt to be generated above the baseline to achieve the said goals, peaking at R639bn in 2027.

Figure 3 Increase in public sector debt above the baseline

The government’s debt increases from one year to the next if they spend more than their income, that is, when they incur a budget deficit in a specific year. If they continuously incur deficits, the debt will grow larger and larger.

Figure 4 shows how the government deficit has to increase significantly at the start of the simulation, to achieve the debt-to-GDP ratios in Figure 2, and subsequently decrease again to get back to the baseline levels of debt-to-GDP.

Figure 4 Government Deficits (policy simulation)

The GDP components

The blue line in Figure 5 shows the deviation from the baseline in total Investment expenditure, which is the sum of all investment expenditure by all industries. In the beginning of the graph investment expenditure shoots up to 20% above the baseline, because all investment expenditure is subsidized by the government. It becomes profitable to invest and all role-players therefore respond to the government’s subsidy, and buy investment goods. The graph looks much like the one in Figure 4, because the deficit there was exactly caused by the introduction of a subsidy on investment expenditure in the first place, and then the reverse policy later on. The orange line in Figure 5 shows the deviation in Import expenditures from the baseline. Buying investment goods typically entails a high proportion of Imports (24% of all new investment expenditure is on imports), so we expect that the Imports graph should look similar to the Investment graph.

Figure 5 Investment Expenditure and Imports (deviations from the baseline)

Figure 6 shows the trends in GDP itself, as well as the remaining GDP components, Household Consumption, Government Expenditure and Exports. All the graphs show the deviation from the baseline, in percentage points. For example, the blue line shows that real GDP moves above the baseline and increases to 2,2 percentage points above the baseline in 2027. This is a wonderful result! The government borrows money and subsidize investment expenditure, and then the economy grows 2,2 percent higher than without the government’s intervention. It then decreases and lies below the baseline from 2029 onwards. The baseline portrays what would have happened without the policy intervention. So, the policy intervention of increasing government debt in the short run, is good for real GDP in the short run. However, when we apply the contractionary counter policy in the latter part of the simulation, real GDP would be worse than what it would have been for the last decade of the simulation.

Figure 6 GDP and its components (cont.) (deviations from the baseline)

The orange line shows that Household Consumption starts off very poorly as a result of the policy simulation, and then grows significantly above the baseline for five years, before following real GDP into the negative zone. The first two periods are very significant and poses a dilemma for the policy makers. However, GDP is reacting well to the policy intervention in the short run: it is the massive investment increases that cause the positive results in GDP, while households fare not good at all. The increased demand for investment goods as a result of the subsidy, pushes up the demand for construction goods, which forms half of all the components of investment. Construction goods also forms significant parts of many goods that households buy, such as real estate, transport equipment, all transport services, business and other services. The said goods’ prices increase relative to other goods, with the result that the basket of goods that the typical household buys, becomes more expensive than without the policy intervention. Households move above the baseline when we start to reverse the investment subsidy policy intervention in 2026. However, in the long run household consumption follows the rest of the economy into the red zone.

Exports perform better than in the baseline until 2030, and then also falls below it. The export equation in the model is very simple: export demand increases if South African prices increase at a slower rate than foreign prices (or decrease faster). With investment subsidies all industries benefit, and except for the prices of construction goods and a few other ingredients of a unit of investment, most prices fall, relative to the baseline. South African industries thus become more competitive in the world and therefore export more.

Industry winners and losers

Of the 56 industries in our National Model, only 6 end up above the baseline at the end of the simulation period, as shown in Figure 7. It is clear that the forms of their graphs over time are very similar to the graph of total investment in Figure 5 above, since all of them are closely related to the production of investment goods, or the construction industry itself.

Figure 7 Industry survivors (relative to the baseline)

A second series of industry graphs simulates that of Household Consumption in Figure 6, and reading through the industry names reminds one of the goods that we as households buy on a regular basis: Bakery, Dairy and Grain products, Education and Health services and Water.

Figure 8 Short run industry losers

Discussion

Increasing the debt-to-GDP ratio temporarily to allow for infrastructure development shows positive results for all industries directly related to the production of investment goods, such as the Construction industry and other hard-core Manufacturing industries. Exports also perform well in the short to medium term because the industry subsidies make South African firms more competitive than before due to lower costs of production. Unfortunately, Households are paying the price for this policy intervention in the short run because many of the goods in their shopping baskets have become relatively more expensive. In the long run all the GDP components fall below the baseline, while only a handful of the 56 industries produce more per annum than what they would have produced in the baseline.

Productivity improvement

Fortunately, the story does not end here. Investment expenditure forms part of GDP, alongside household expenditure, government expenditure, exports and imports. It is a straight forward result that a subsidy on investment expenditure would lead to increased investment and increased GDP, if the decrease in household consumption does not outweigh the increases in investment.

However, why does GDP dip below the baseline towards the end of the simulation period and not benefit from the better infrastructure? Because we have not simulated the effects of better infrastructure yet! Until now we have only subsidized investment expenditure for three years, and then reversed the subsidy subsequently.

Let’s assume that the new infrastructure would improve total factor productivity in all industries, from 2027 onwards, when the subsidy on investment is completed. Figure 9 shows the percentage deviation from the baseline in real GDP as a result of the initial (three year) subsidy on investment and the subsequent removal of the subsidy. There are two graphs, namely, the original graph of GDP from Figure 6 above, and the resulting GDP (orange line) when we let total factor productivity in all industries improve by a mere 0,3% per annum.

The conclusion could be made that it could be a good policy measure to borrow funds for infrastructure development, because (i) in the short to medium term the investment itself would lead to significant growth in GDP, and (ii) in the longer run the productivity in the economy would improve enough to secure a permanently higher level of growth over time.

Figure 9 Real GDP % deviation from the baseline

2. Youth solidarity tax

The task at hand in this Section was defined as “Youth solidarity tax: Increase personal income tax by one percent for all taxpayers earning R600,001+ per annum. In the form of a transitional tax for five years only. Ringfenced for youth enterprise development to be administered by the Solidarity Fund. See if this can be funded by delaying inflationary bracket adjustments over the next two or three years.”

Simulation Design

“Youth enterprise development” is quite difficult to define in terms of our CGE model. The labour force in the model does not have an age dimension, and we are convinced that the youth is working in all of the industries in the model. What the model does have, is a labour force divided into ten different occupational groups. So, in this simulation we are modelling a one per cent increase in the income tax rate on all “rich” households, and recycling the tax revenue in the form of wage subsidies to three of the ten occupational groups. The three chosen groups are elementary workers, operators and skilled agricultural workers. In reality, wage subsidies are paid to industries for extra workers hired, but in the model, they are paid to all the workers in the three said occupational groups. Since the cost of labour comes down, industries would certainly hire more labour in each of the three categories, including new young workers.

Figure 10 Change in the income tax rate on rich households

In this simulation we define “rich” households as everybody in the three top deciles according to labour income received[3].  Figure 10 shows the change in the income tax rate for the top three deciles of households, above the baseline.

Simulation Results

The simulation results look very good, and could serve as part of a policy recommendation to the government.

The GDP Components

All the GDP components benefit from this policy intervention. Government consumption (yellow line) is not allowed to deviate from the baseline because all the new income tax revenue is recycled into wage subsidies, while all other government expenditures remain the same. The graph of Real GDP is hidden behind the Household consumption graph. Exports is the biggest winner, and it stems from the decreased labour cost to firms as result of the wage subsidies to some occupational groups. The CGE model is a neo-classical model, which means that all the savings on labour are given through to the consumers of the goods produced. South African firms become more competitive in the world because they are able to charge lower prices for their goods as a result of the wage subsidies. Lower labour cost implies that more labour will be hired, which increases the purchasing power of households, leading to increased household consumption. Lower cost of production also leads to increased investment, since investment goods are produced using capital and (cheaper) labour, just like all other goods.

Figure 11 Annual deviation from the baseline in the main GDP expenditure components (% change)

Figure 11 shows the annual differences between the policy simulation results and the baseline during the five years of higher direct taxes, recycled as wage subsidies. After five years all the GDP expenditure items move back to the baseline. That means that cumulatively over time the five years would result in permanent increases in all the variables above the baseline. This is shown in Figure 12: the expenditure components increase to levels above the baseline during the five years of the policy intervention, and then remain there. Take investment as an example: in Figure 11 we see that investment grows 0,1% faster than in the baseline, for five years. Then it grows at baseline levels again. Five periods of 0,1% growth each adds up to about 0,5% cumulative growth, which is the level portrayed in Figure 12.

Figure 12 Cumulative deviation from the baseline in main GDP expenditure components

Industry Winners and Losers

Industries are very dependent on the GDP expenditure components that they are connected with. For example, the Food industry is strongly related to Household expenditure, while the Mining industries are related to Export demand. If all the GDP expenditure components have positive trends as in Figure 12, we could expect all industries to also have similar positive outcomes. This is indeed confirmed by all 30 industry graphs in Figure 13. It is clear that some industries gain much more than others, but in general, this policy intervention does not show any real losers.

Figure 13 Cumulative % change from the baseline in Industry output.

Discussion

This policy intervention looks very promising, with mostly positive results on the macroeconomic as well as industry levels. The increased income taxes would certainly harm some industries that supply consumer goods to the rich group of households who pay extra taxes in the simulation, but their harm should be compared to the benefits of other groups, to determine whether the policy should be implemented or not.

Are these results too good to be true? What is the catch here? If this is such a good policy measure, why does the government not just implement it permanently? Wouldn’t there be a significant cost to the economy if the rich has to pay more income tax?

Figure 14 shows the changes in household consumption for the twelve income groups in the model. The richest 3 deciles (four groups) who pay higher percentages of income tax are depicted by the bottom four lines in the graph. One of the groups (D8) increase their consumption, while V19 stays more or less on the baseline, with D9 and V20 dropping below the baseline in the simulation. The moral of the story is that higher income tax rates do not necessarily mean that household consumption of the taxpayers would decrease. It could, but it does not always happen. In the simulation the wage subsidy causes the cost of production and price levels to decrease, which could lead to an increase in consumption. The poorer groups – V1 to D7 – do not pay more tax, and all consume more as a result of (i) lower price levels of consumption goods and services, and (ii) higher total wage income for three of the occupational groups, due to wage subsidies.

Figure 14 Deviation from the baseline in household consumption by income group

3. Limited public salary bill increases

The requested policy intervention was to “limit the public salary bill increases to inflation”. We briefly report the results here, because most of the explanations of what would happen has been done in Section 3 above. Limiting the public salary bill is contractionary government policy, if they do not spend the savings on other priorities. We could expect similar results to those reported above.

Simulation Design

In the model we have three government “industries”, namely General government, Education and Health and Social Services. It is necessary to think about the government like that because they are employers of factors of production, such as labour and capital, and they produce something, namely “Government Services”. So, in the policy simulation we fix all real wages of the three government industries for seven years, starting in 2024, which means that nominal wage increases are limited to the rate of inflation. Since other industries are not limited in the same fashion, government wages seem to be lower than other wages and the model would let the government employ more labour. We therefore also need to fix the number of workers in government to avoid anomalous results.

Simulation Results

Case 1: Government revenue not recycled

All the macroeconomic results shown in Figure 15  could have been anticipated from what we learned in Section 3 above. Contractionary government policies lead to decreased total demand in the economy, resulting in decreasing price indices. South African goods seem to be cheaper than those of the rest of the world, so that exports increase. Household consumption would be less than in the baseline, because all the workers in the public sector receive lower wages than in the baseline, and therefore have less buying power. In Figure 15 below only exports lie above the baseline, while the other GDP expenditure components look very similar to those in Figure 17 above where the corporate tax rates were increased without recycling the revenue.

Figure 15 Cumulative deviation from the baseline in the main GDP expenditure components levels (revenue not recycled)

Industry Winners and Losers

The three government industries would perform better than any other industry, because their labour cost is relatively lower, by design. A handful other industries who are very reliant on exports also perform better than in the baseline, but the majority of all industries lose with the contractionary government policy.

Discussion

Enough was said above about contractionary government policies, so we would rather do something with the savings that the government build up through paying lower nominal wages than in the baseline.

Case 2: Government revenue is recycled

We ran a second simulation where we used all the savings from the government wage bill to subsidize the sale of Construction Services, to all users thereof. This should have similar results to our subsidy on all investment done in Section 1 above, because Construction Services forms about half of all investment goods, and  Figure 16 shows that this is exactly the case. The big winner is Investment that grows to 5% above the baseline towards the end of the simulation period. This is to be expected because all Construction Services are subsidized, so Investors’ cost of production is decreased. Real GDP benefits from the higher levels of Investment also. However, Households are again the worst off. Many people work for the government and they are being paid less than in the baseline, so their purchasing power is decreased. They cause a decrease in total demand for many goods in the economy, which leads to decreasing real wages and price levels, which benefits Exports.

Figure 16 Cumulative deviation from the baseline in the main GDP expenditure components (revenue recycled)

Figure 17 shows how a subsidy on Construction Services, funded by lower government wages, would distort the supply side of the economy. Investment is growing fast, which leads to the economy becoming capital intensive: the capital stock is growing while employment is decreasing. Even though GDP is higher than the baseline, Households suffer because employment is decreasing, and the employed work force have to face lower real wages.

Figure 17 Cumulative deviation from the baseline in the Supply Side components of GDP, as well as real wages and the CPI. 

Discussion

This forth macroeconomic policy intervention should stress two important points from above: (i) decreasing government expenditure through lower real wages would lead to a contraction of the economy with mostly negative results on both the macro- and industry levels, and (ii) the choice of recycling mechanism is a matter of paramount importance.

Conclusion

The three macroeconomic interventions modelled above used three different methods of increasing the government’s spending power, namely (i) increasing the debt-to-GDP ratio (borrowing money), (ii) increasing the personal income tax rates on the rich, (iii) keeping government wages constant in real terms, rather than giving annual increases.

One fact that should be clear after reading the three sets of simulation results, is that increased savings by the government is generally bad for the economy if the government does not spend the savings properly. Keeping the increased revenue in or under the sofa is contractionary economic policy which decreases total economic activity; it is deflationary and mostly inefficient.

One of the three macroeconomic interventions showed very positive results, namely, a wage subsidy on youth labourers. The subsidies on investment and construction services also have positive GDP results, but these two recycling schemes would only benefit households in the longer run, and would therefore be unpopular for the government to implement in the short run, but which will yield positive economic infrastructure development to enable future production expansion and GDP growth.

This report should not be interpreted, read or understood as a fixed set of recommendations, but rather as a catalyst for dialogue on proposed interventions.

Bibliography

UN-DESA. (2018). Handbook on Supply and Use Tables and Input Output-Tables with Extensions and Applications. Available at https://unstats.un.org/unsd/nationalaccount/docs/SUT_IOT_HB_Final_Cover.pdf

[1] Constant elasticity of substitution

[2] They are not published on an annual basis, but rather every five or ten years, or so.

[3] One percentage point is very small and if only those above R600 000 were chosen the additional revenue would not get very far in terms of wage subsidies.

 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

This report has been published by the Inclusive Society Institute

The Inclusive Society Institute (ISI) is an autonomous and independent institution that functions independently from any other entity. It is founded for the purpose of supporting and further deepening multi-party democracy. The ISI’s work is motivated by its desire to achieve non-racialism, non-sexism, social justice and cohesion, economic development and equality in South Africa, through a value system that embodies the social and national democratic principles associated with a developmental state. It recognises that a well-functioning democracy requires well-functioning political formations that are suitably equipped and capacitated. It further acknowledges that South Africa is inextricably linked to the ever transforming and interdependent global world, which necessitates international and multilateral cooperation. As such, the ISI also seeks to achieve its ideals at a global level through cooperation with like-minded parties and organs of civil society who share its basic values. In South Africa, ISI’s ideological positioning is aligned with that of the current ruling party and others in broader society with similar ideals.

Email: info@inclusivesociety.org.za

Phone: +27 (0) 21 201 1589

Web: www.inclusivesociety.org.za