Don’t compare apples with oranges when assessing Mboweni’s budget
Gilad Isaacs uses fiscal multipliers calculated under different conditions to motivate for more government spending
When it emerges from the unprecedented shock of the Covid-19 lockdown, SA’s macroeconomic health will be measured by the growth rate of potential output and how well aggregate demand keeps actual output growth near the economy’s potential.
A wide range of risks are involved in how domestic and global events and chosen policies shape growth outcomes. Finance minister Tito Mboweni’s supplementary budget highlighted the risk of a rising debt-to-GDP ratio increasing further the public resources consumed by interest payments. Given the many economic, social and political uncertainties facing the post-Covid world, SA faces the additional possibility that the flows of capital that have financed spending up to now will become more volatile, with sudden net outflows squeezing South Africans’ economic welfare.
While SA’s public authorities need to weigh many uncertainties in “building back better” and expanding output in inclusive ways, the focus of Gilad Isaacs’ article (“It’s not about debt, it’s about stimulating an economy in crisis”, July 9) narrowly concerns one factor: the size of the fiscal multiplier. Critical of fiscal consolidation, he relies on claims that the “bang per buck” of fiscal spending is high.
It is true that government spending and its effects on private spending (the fiscal multiplier) is an important tool for stimulating aggregate demand in the medium term. But fiscal multipliers are difficult to calculate, with different estimates resulting from differences in model assumptions, estimation methods and time periods. The size of estimated fiscal multipliers varies with the underlying economic conditions, type of fiscal instruments, level of debt and expected changes in taxation, among others.
In particular, when fiscal deficits are large, resulting financial conditions such as rising interest rates can reduce the impact of public spending on growth. The effect the financial sector has on the fiscal multiplier changes when banks’ risk appetite and spread between borrowing and lending rates change.
For these reasons it is misleading to use multipliers calculated under different conditions to motivate for higher government expenditure now. Yet that is what Isaacs does. Selectively listing several published estimates of large fiscal multipliers to argue against the Treasury’s proposal for fiscal consolidation, he includes ours. Our estimates referred to a specific period immediately after the 2008 financial crisis that was characterised by low debt levels and competitive long-term real borrowing rates of about 1.5%, a large output gap due to the crisis-induced demand shock, and a strong recovery in lending. In our analysis we illustrated how quickly the expenditure multiplier declines if one of these assumptions is not satisfied.
Conditions now are fundamentally different. Real bond yields have started to rise as concerns about fiscal sustainability have increased. The real 10-year bond yield was close to 5% at the end of 2019 and the outlook for government debt is that it will exceed 100% of GDP over the next three years. Tax increases have not translated into the projected revenue gains. Banks have become increasingly risk averse in response to rising policy uncertainty, and concerns over fiscal sustainability have tightened credit. These all suggest, without using a complex model, that the size of fiscal expenditure multipliers has been falling. If we apply our model under these conditions, we would generate small expenditure multipliers.
If the size of the fiscal multiplier is important, then rather than citing old estimates we have to understand how the multipliers were calculated and whether the assumptions reflect current realities. For example, Schröder and Storm assume SA is a closed economy and ignore funding costs and their effect on the economy, which, not surprisingly, generates a high fiscal expenditure multiplier. Are these reasonable assumptions? Similarly, why is the average multiplier from 1980 to 2010 calculated by Stellenbosch academics relevant in the current environment?
Isaacs fails to mention that several recent studies by Charl Jooste and Harri Kemp found very small multipliers. Maybe he disagrees with their assumptions, but then he needs to explain why rather than use selective referencing to justify his view.
Economic debate based on good economic analysis is key to developing the right policy responses. Yet too often the focus is on numbers, without understanding how they are generated and whether they are based on plausible assumptions. They are accepted (right or wrong) as long as they appear to justify a certain view. This approach is unlikely to generate consensus or policies that address SA’s structural decline.
• Harris is former professor of economics at the School of Oriental and African Studies, University of London. Makrelov, a former head of economic modelling and forecasting at the National Treasury, is an economist at the SA Reserve Bank.
Would you like to comment on this article or view other readers' comments?
Register (it’s quick and free) or sign in now.
Please read our Comment Policy before commenting.