Some thoughts on South Africa's exchange rate

The weaker rand exchange rate raises alarms in the monetary policy sphere. The rand depreciated by 23% against the US$ from January 2012 to December 2013. This has implications for people wanting to take a holiday in the US or purchase anything that has a $ tag assigned to it. Apart from the worrying tourist, the central bank has to keep an eye on the exchange rate as it could lead to higher inflation. A final issue related to the rand is the trade deficit narrowed – and why it has not narrowed.

In essence three economic questions arise: (i) Is the rand permanently weaker (i.e. what has happened to the equilibrium exchange rate), (ii) what is the effect of the weaker rand on inflation (i.e. what is the exchange rate pass-through) and finally (iii) how do imports and exports respond to the rand.

I will summarise the results here in case you don’t have time to read the entire section.

·    The rand’s equilibrium value to the US$ is R8.50. The rand is thus undervalued and should start appreciating (given that SA or the rest of the world don’t undergo some massive structural change).

·    Rand appreciations are shorter lived than depreciations. Depreciations are usually more persistent over time.

·    The exchange rate pass-through to inflation is time varying; i.e. it changes over time and across different policy decisions. The current exchange rate pass-through is low, however, depends on low and stable inflation.

·    The trade deficit did not narrow since the rand depreciation. Both export and import volumes increased. Imports simply increased more than exports. Some researchers argue that this is due to a J-curve effect. I, however, argue that the response of certain imports is inelastic to the exchange rate. This means that if SA imported exchange rate inelastic imports, then any rand depreciation would have a very small effect on decreasing imports.

11.    Some (important) preliminary statistics of the exchange rate

Figure 1: The rand is not alone – Other developing countries’ exchange rates have also depreciated

Source: Bloomberg

South Africa’s rand is not the only developing country exchange rate that depreciated; The Indian rupee along with the Brazilian real depreciated by similar margins. The Chinese yuan remained at similar levels only because of its massive reserves – the Chinese actively intervene to maintain a stable currency. 

The top row of Figure 2 illustrates the distribution of depreciations vs. appreciations. Notice that large depreciations occur quite frequently. This is supported by the fact that depreciations are more persistent. Appreciations are small the majority of time. There are, however, periods of massive appreciations. These occur at low frequencies as appreciations are abrupt. Row two of Figure 2 shows that the r/$ exchange rate is much weaker than it historical averages (comparing the average of 1990 until 2013 vs. the average since inflation targeting). We also see that the rand is weaker than an HP-filtered exchange rate (a measure of the long-run exchange rate).

Figure 2: Appreciations are abrupt while depreciations are persistent

Source: South African reserve bank and own calculations

It can be argued that the rand fluctuates around its steady state (equilibrium) if one can identify the factors that determine it. The figure above illustrates that the rand does not deviate too much from such a steady state (assuming that this steady state is represented by the smoothed line).

This paper shows that interest rate differentials, productivity differentials, openness and the government’s budget balance determine the equilibrium value of the rand. Thus any deviation from the rand away from equilibrium will tend towards equilibrium later. The author finds that it takes just under a year for the rand to move back to equilibrium.

  

The paper above shows that:

·        * If SA interest rates increase by 100 basis points relative to US interest rates then the exchange rate will appreciate by 0.8 percentage points.

·        * If SA GDP growth increases by 1 percentage point relative to its trading partners then the exchange rate will appreciate by 1.3 percentage points.

·         *A 1 per cent increase in the government deficit will appreciate the currency by 1.5 percentage points.

Figure 3: The equilibrium exchange rate in 2011Q4 was R8.15 to the US$

Source: de Jager

22.      On exchange rate pass-through to consumer prices

How will the SARB react to the weaker exchange rate? This depends on what they believe will be the impact on inflation. If the exchange rate depreciation translates into higher consumer price inflation then the probability of an interest rate hike increases. So not only are consumers concerned about higher prices due to the exchange rate, but also higher interest rates from the SARB to curb inflation.

Econometric analysis that measures the exchange rate pass-through to inflation over time shows that it has decreased since implementing inflation targeting (see Figure 4). The highest pass-through occurred in 2002 and 2003 at a time when inflation was very high and volatile and the exchange rate very weak. Currently the pass-through is close to zero. This means that the weaker exchange rate hardly translated into higher consumer prices. In Figure 4 the exchange rate pass-through is highest over four quarters. Exchange rate shocks almost dissipate completely over three years.

  

Figure 4: Pass-through has been lower since inflation targeting

Is it reasonable to assume that low pass-through will persist? If this can be assumed then we need worry very little about the possibility on an increase in interest rates due to exchange rate pass-through. But to make this assertion we need to know what causes high vs. low pass-through.

We use the pass-through coefficients in Figure 4 as a dependent variable and analyse likely factors that influence it. The most important determinants of exchange rate pass-through is the extent of mark-ups and inflation (see Table 1). A 1 per cent increase in mark-ups reduces exchange rate pass-through by 2.71 per cent on average. This seems counterintuitive at first but makes sense given that monopolies are able to absorb price shocks. Perfectly competitive firms that operate at the margin have no choice but to pass-on higher prices when a shock occurs. High and volatile inflation increases exchange rate pass-through by large margins. This highlights the importance of a credible monetary policy: If firms expect inflation to be contained by the monetary authority it knows very well that high inflation will be reduced by an increase in interest rates. These firms trust the SARB to do its job and they are thus not that much concerned about the exchange rate shock and could absorb it. This story is however dependent that depreciations do not occur for much longer periods than one would normally anticipate.

Table 1: Low pass-through depends on low and stable inflation

Variable	Mean	Standard deviation
Market concentration (PCM)	-2.71	0.43
Neer	-0.01	0.04
Neer-squared	0.00	0.01
Inflation	1.00	0.77
Inflation-squared	0.02	0.04
Output gap	-1.23	0.99
Government debt (gov)	0.06	0.14
Variance inflation	5.39	1.53
Variance Neer	-0.74	0.23

33.      Some imports do not respond that strongly to an exchange rate depreciation

Imports have grown by 24.8%, 2.71 and 14.3% during the first three quarters of 2013 while exports grew by 11.3%, 5.1% and 13.5% over the same period. The consequence is that the trade deficit worsened and payments to SACU countries increased.

So by how much do imports respond to exchange rate depreciations? To control for income effects we used G7 growth as a proxy and to control for long-run vs. short-run effects we estimate these elasticities in a dynamic panel setup (we use Pesaran’s (1999) PMG methodology). Aggregate imports decrease by 0.8 per cent given a 1 per cent exchange rate depreciation. But we are only interested in import components. We see that building equipment, petrol, metals and transport equipment have a less than one per cent decline when the exchange rate depreciates. Coincidentally minerals, vehicle equipment, machinery and appliances contributed the largest shares of import in 2013. This analysis helps explain why imports have not decreased in line with the exchange rate depreciation.

Figure 5: Recently SA has imported more inelastic goods

44.      So what do we make of all this?

The concerns about the devalued rand are probably overstated in the media. Exchange rate pass-through to inflation is still low (partly thanks to credible monetary policy and firms absorbing the exchange rate shocks); the rand has deviated from its equilibrium – it is expected to return to its equilibrium (i.e. close to R8 against the dollar). As a negative, the exchange rate has not reduced the current account deficit. Imports have increased in response to the rand depreciation despite muted household consumption growth. While it is difficult to say whether this is a permanent trend (it could reflect fixed contracts that were signed over a specified period), the likely future appreciation will deter import growth – probably causing the current account deficit to persist.

55.      Appendix: The probability of an appreciation is not too low

In one of the previous sections we were interested in whether the exchange rate will depreciate indefinitely. We mentioned that an indefinite depreciation is not realistic if the structural fundamentals of our economy are maintained. These structural fundamentals, as mentioned above, are informed by the interest rate differential, GDP growth and the government’s budget balance. We estimate the probability of an oncoming appreciation by using a probit model. Our dependent variable is a binary variable that equals 1 in case of a rand depreciation and zero when the rand appreciates. Unfortunately some of the data is only in quarterly frequency with the latest update in 2013Q3. We can use this as a gauge on how useful the model is in picking up the likelihood of an appreciation. Figure 6 shows that the probability of an appreciation from 2013Q3 was less than 20%; the rand has not appreciated since then. On the right hand side of Figure 6 we examine how probable a depreciation is on a scale of interest rate differentials (assuming that all the other variables are evaluated at its mean except for growth). 

Figure 6: The probability of an appreciation vs. depreciation?