The dilemmas of Pravin Gordhan

  • On 22nd February, Finance minister Gordhan presented his annual budget to the national assembly.
  • Gordhan faced a painful trade-off between managing South Africa’s eye-watering debt situation, supporting stagnant private consumption and political sustainability in the most unequal country in the world
  • We simulate South Africa’s debt/GDP path under different assumptions, and argue that the economy still has a long way to go to achieve fiscal sustainability

On 22nd February, South African finance minister Gordhan presented his annual budget to the National Congress. In a budget hailed as “pro-poor but not populist”, Gordhan attempted to get a grip on the public finances through a combination of additional “sin taxes”, taxes on fuel, and a rise in the top marginal rate of tax to 45%. This allowed for increases in social grants and targeted support for South Africans seeking to buy a home.

Gordhan’s preference for balancing the budget through taxes on the wealthy, and maintaining social transfers, is understandable. In the last year that the World Bank published data for South Africa (2011), the country had a Gini coefficient of 63.4, making it the most unequal economy in the world. There are also pragmatic reasons for wanting to avoid hitting poorer consumers, too: South Africa’s painfully high 27.1% unemployment rate, low and falling levels of consumer confidence and declining retail sales (-2.3% MoM for December) mean that the economy can ill-afford further hits to private consumption. An environment of rising global yields leaves the SARB little room for offsetting monetary easing.

In his speech, Gordhan was cognisant of the tension between key goals: attaining fiscal sustainability, reducing inequality, and achieving growth. These three are not necessarily in conflict (higher growth would be a boon to South Africa’s public finances) but the need to avoid squeezing consumers too hard has certainly placed constraints on fiscal tightening. So how much more is needed?

To assess this, we use the following identity.(1) Define real GDP growth g(y), the real interest rate on government debt r, the deficit as a percentage of GDP d, and the debt-GDP ratio b. Then:


Clearly, prior to the budget, the public finances were in a perilous situation. In Figure 1 the “pre-budget” projection assumes that the deficit remains at 3.01% of GDP (latest print), with growth at 1.51% (a five-year average) and the real interest rate at 2.22%. The real interest rate estimate is based on the assumption that the nominal cost of government borrowing is equal to the 10 year average of 10 year government bond yields, adjusted for a 6% rate of inflation (the upper end of the SARB’s range).

So how has the budget changed things? To answer this, we input the projections for the deficit and growth from Gordhan’s speech. These are assumed to then remain constant at the final value outside the forecast period. (This is labelled “post-budget” on the graph). We can see that there is a marginal improvement in the debt profile of South Africa, but that much more remains to be done.

Finally, we examine a scenario in which the government is successful in establishing a steady path. A crucial insight is that, referring to our identity, if the real interest rate is higher than the growth rate, then South Africa needs to run a surplus in order to prevent an explosion of the debt-GDP ratio over time. We simulate a scenario in which Gordhan’s growth assumptions hold but he is successful in fully closing the deficit. This does indeed lead to a much more stable path of the Debt-GDP ratio. Note that this is still increasing, however, because real rates outpace growth. South Africa in essence can barely sustain a deficit without dangerously explosive debt dynamics.

Eliminating the deficit is a tall order, however. Can South Africa achieve a full closing of the deficit without endangering growth (and hence potentially worsening its debt dynamics)? Naturally, this is an unknown risk, and one Gordhan has, perhaps understandably, declined to take in this budget. However, for now sustainable public finances have not been met. The finance minister’s position is not one to envy.

Source: SARB, South African Treasury, IMF WEO, Macrobond, Record. The dotted black line shows the historical evolution of the debt-GDP ratio from 2007. Projections are as described in the text.(1) This is derived as follows: define b=B/(yp) where B is nominal outstanding debt, y is real GDP and p is the price level. Then where D is the nominal deficit, i the nominal interest rate paid on government debt, and Pi inflation:


The first is an assumption about how the nominal debt stock evolves; the second is derived from differentiation. It follows that:

This is rearranged to make the identity where r is the nominal rate adjusted for inflation


A framework for differentiating among Emerging Market currencies

  • How to rank the relative attractiveness of Emerging Market currencies? In this blog post we bring together various metrics that should help investors decide on the perennial question, whether or not to hedge Emerging Market currency risk.

Where currencies in growth economies able to deliver a structural appreciation for those who own them while at the same time as posing a currency risk, a key question is how to best to go about deciding whether to hedge that currency risk, if at all? Emerging Market currencies, through their cyclicality as well as long term trends, have always posed this perennial dilemma for investors.

The below table outlines a framework for considering hedging decisions in high growth / convergence economies where there is likely to be a punitive combination of a long term expected return from holding unhedged currency positions and relatively high currency hedging costs. Because of this, any hedging decision must be evaluated from multiple angles.

We begin by laying out and categorising the key criteria that we believe should be considered in any such currency hedging decision. These are:

  • The long-term appreciation prospects for the currency, here defined as a combination of the real interest rate differential and the level of undervaluation of that currency (vs US Dollar).
  • The short-term market sentiment for the currency, as measured by a (FX spot) momentum indicator.
  • The portfolio benefits in terms of a lower volatility return stream resulting from hedging, in this case, EM equity exposures (relative to the unhedged return stream)
  • The all in cost of hedging, which includes the economic (carry) cost as well as the trading (execution) cost.

Depending on which of these factors has a greater relevance for the investor, different rankings will emerge.

In the below table we take a transparent approach by combining these four decision criteria in an equally weighted fashion to arrive at an overall hedging attractiveness rank as of December 2016.

The overall framework does suggest that, for example, under these metrics, it is less punitive to hedge Korean Won exposures than Peruvian Nuevo Sol exposures, from the perspective of a US investor. There is no single, or indeed final answer to this question, but the practice of differentiating among EM currencies is one that is gaining more credence as these markets mature and become more differentiated themselves.


Real interest rates differentials are nominal rate differentials adjusted by the most recent YoY CPI numbers.

Undervaluation versus USD uses Record the Fair Value Model, which is based on econometric regressions using long term datasets.

The momentum signal uses three moving average pairs (10d-240d, 20d-120d, 10d-60d). Where all three moving averages agree on the direction of momentum, the currency is assigned to “weakening” or “strengthening”. Otherwise, it is considered to have “no momentum”.

Normalised volatility reduction from hedging is the difference in annualised volatility between the MSCI local return and the MSCI return (hedged to dollars), normalised by dividing by the unhedged volatility.

Annualised cost of hedging comprises the nominal interest rate differential, and indicative the bid/offer spreads (August 2016). This is except for MYR, which uses the Dec-2014 estimate, increased by the average percentage increase in spread across other EM pairs over the period.

All signals except momentum are normalised as follows: we transform each signal by adding the absolute of the smallest value in the universe to each. This ensures that all signal values are weakly positive. They are then normalised by dividing by largest value in the universe. For momentum, the normalisation is by assigning the value 0 to a moving average which suggests the EM currency is weakening against the dollar and 1 otherwise. This is then summed across the three moving averages and divided by three for each EM currency. The overall score for a currency is then this result divided by 2.

The overall rank is the rank of the currency by the average across these signals.

Source: IMF, Bloomberg, Macrobond, Reuters W/M.