Lesetja Kganyago: The influence of South Africa's price-setting environment on monetary policy trajectory

A public lecture by Mr Lesetja Kganyago, Governor of the South African Reserve Bank, at the Nelson Mandela Metropolitan University Business School, Port Elizabeth, 7 September 2016.

The views expressed in this speech are those of the speaker and not the view of the BIS.

Good afternoon, ladies and gentlemen, and thank you for the opportunity to share a few thoughts and ideas with you today.

South Africa's economy has been slowing since 2011 in an environment of sluggish global growth and falling commodity prices. In July, we expected the GDP¹ growth rate for 2016 to be about 0 per cent. With the relatively good second-quarter figures released on Tuesday, the Monetary Policy Committee (or MPC) of the South African Reserve Bank will be able to revise up its annual estimates at the September meeting, but the annual figure will remain low. At the same time, inflation has been elevated. Since 2011, it has averaged 5,5 per cent, and so far this year it has averaged 6,3 per cent, above the target range of 3-6 per cent.

All other things equal, the MPC would like to respond to slow and below-trend growth with lower interest rates. But all other things have been far from equal. We have had to make decisions where we don't like all the consequences because of the nature of the factors driving inflation out of the target range. I would like to discuss what those factors are and how we should think about them.

The immediate inflation problem comes from rising food prices. Two other crucial factors are the exchange rate, which I will discuss in some detail, and rigidities in labour and product markets that prevent wages and prices from adjusting in line with economic conditions. A final issue is the way in which we have operationalised the 3-6 per cent target range over time.

The main point I would like to make is that monetary policy decisions necessarily involve trade-offs, and the MPC has opted for trade-offs that serve our inflation-targeting mandate and the long-term interests of all South Africans.

Had the starting point been different - had inflation been well within the target range - we could have made different choices. But policymakers have to work with the facts they have, not the facts they'd like to have. This is my second point. As a society, we must aspire to a set of better economic outcomes than we have now. South Africa is facing the unenviable and difficult-to-solve combination of little growth, high inflation, and growing unemployment. We have to do better. Thinking through the conditions we face and the alternatives to them helps us to envision a reform agenda.

Food prices

I'll start with food prices, the factor over which we have the least amount of influence.

As you know, South Africa has been suffering from a particularly severe drought. The result has been double-digit food price inflation in 2016. This is adding about 1,5 percentage points to inflation, which is enough to push headline inflation outside the 3-6 per cent target range.

Most of the impact this year is coming from higher prices for bread and cereals as well as fruit and vegetables. A little later this year, we expect meat prices to rise more strongly and keep food inflation relatively high, even if the drought abates. The reason for this is that farmers have responded to water shortages by selling animals for slaughter and are likely to withhold supply while they rebuild their stocks. Accordingly, we expect food prices to come down quite slowly, even if we do get enough rain this year.

Higher food prices pose an interesting policy problem. They are a classic supply shock. More expensive food tends to put downward pressure on other prices because it reduces buying power and subdues demand. If this helps to check price increases, and if inflation expectations are already well anchored within the target range, then policymakers can be confident that any inflation increase will be temporary. Policy can mostly ignore these kinds of temporary shocks. But if those conditions are missing - if the initial shock generates increases in the prices of other products and factors - then a policy response may be required.

To put it in another way, a pebble can start an avalanche. If you've already got all the other conditions for an avalanche, you need to worry about pebbles.

The exchange rate

The exchange rate of the rand is another important driver of, and risk to, the inflation outlook. It has been highly volatile in recent months, reflecting rapid changes in the world economy and in South Africa. As such, I would like to discuss it in some detail.

The rand has been on a depreciating trend since early 2011. It reached the high point of R6,63 to the US² dollar in December 2010 and then commenced a long slide all the way to R16,87 to the dollar in mid-January of this year. In real effective terms, meaning the exchange rate adjusted for inflation differentials with trading partner countries, January's exchange rate was just about as depreciated as it has ever been. Only the sudden rand crash of 2001 comes close.

The rand's steady depreciation has two underlying causes. The first factor is commodity prices. In 2011, our terms of trade reached an all-time high, meaning that the prices of our export goods had never before been as favourable relative to the prices of our imports. China's economy was still growing at close to 10 per cent a year and its investment levels were at nearly 50 per cent of GDP, one of the highest levels ever recorded, which generated huge demand for commodities such as iron ore, copper, and coal. Five years later, that growth rate has come down to between 6 and 7 per cent. Investment has subsided. At the same time, many other commodity suppliers have entered the market, further suppressing prices. To take one especially clear example, iron ore has fallen from almost US$180 per tonne in mid-2011 to under US$50 in the first half of 2016.

The second factor underlying rand depreciation is capital flows. In 2010 and 2011, the US Federal Reserve was implementing quantitative easing in response to weak growth and a slow recovery in labour markets. Meanwhile, emerging markets were enjoying a strong rebound from the global financial crisis. The result was investment in search of higher yields flowing out of advanced economies into emerging markets, including South Africa.

Fast-forward a few years and the situation is rather different. US unemployment has fallen dramatically and is now close to its likely natural rate. The Federal Reserve has stopped quantitative easing and is trying to decide how much more to raise interest rates. Meanwhile, emerging markets have suffered a severe slowdown. Some of these countries have also become reliant on foreign savings to fulfil all their spending plans, meaning that they are running current-account deficits. Together, these forces have produced a reversal in capital flows and currency depreciation.

These two factors underpin the steady depreciation of the rand over the past half-decade. But to explain the shifts in recent months, we need to expand the story.

The first important episode is the sudden and intense bout of rand depreciation around the end of 2015 and the start of 2016. Commodity prices and capital flows played their role: the prices of South Africa's export commodities weakened during the second half of the year and the Federal Reserve raised its main policy rate in December 2015. Perceptions of South African risk were also important, with the rand depreciating abruptly on news that Finance Minister Nhlanhla Nene had been removed from his post.

Bond prices and credit default swaps on South African government debt also spiked, signalling investor concern about the reliability of the South African government as a borrower. Research suggests that markets were pricing government bonds as if South Africa had already lost its investment grade credit rating. Conditions deteriorated further early in the New Year, when China's economy experienced a period of stress. The Shanghai stock market index fell sharply and the renminbi was allowed to depreciate abruptly against the US dollar, triggering a bout of capital flight from China. In an atmosphere of market fear and risk aversion, the rand depreciated further, in line with many of its peer currencies.

This sell-off in emerging market currencies went too far - an excellent example of the well-known phenomenon of overshooting in exchange rate markets. As a result, emerging market currencies were primed for a rebound episode. World economic conditions began to move back in their favour in the second quarter of the year. China's growth stabilised with help from stimulus policies, and commodity prices stopped falling and even registered gains. Meanwhile, risks emanating from the Brexit convinced the world's largest central banks to either ease policy or delay further tightening. The net result was an appreciating trend in emerging market currencies. In line with the overshooting diagnosis, the countries which experienced the largest depreciations during the sell-off phase enjoyed the biggest rebounds.

South Africa joined this rally: the rand strengthened through June and July, reaching a recent high of R13,23 in early August. The South African currency benefitted from commodity movements, including from a higher gold price following the Brexit. We also saw the effects of a renewed search for yield by investors looking for something better than negative interest rates on advanced economy government bonds.

In addition, domestic factors were supportive of the rand. The major ratings agencies reaffirmed South Africa's sovereign investment grade credit rating in June. The local elections in August demonstrated the vitality and competence of our democratic institutions.

But the rand's appreciation phase would be short-lived. By late August, the rand was up to R14,70 to the dollar. Credit default swaps and bond yields also rose again, responding to heightened perceptions of political risk. On this occasion, the global exogenous factors remained positive. The rand's renewed weakness reflected local concerns yet again.

How should monetary policy respond to a problem like currency depreciation? There is a short game and a long game - and I would like to explore both.

South Africa has been playing the long game since 2000. It is about getting domestic wage and price setters to count the exchange rate less when forming their expectations of future inflation. If these economic actors watch the exchange rate, price accordingly, and thereby turn large parts of any depreciation into inflation, the overall economy and policymakers specifically are left with only bad outcomes. To minimise this effect - depreciation leading to inflation - we try to align expectations, and consequently pricing decisions by households and firms, to an inflation target that is expected to be reached over time. Economists tend to think that exchange rates go up and down and not in one direction; this has indeed been the case for the rand. But inflation weakens an economy's response to depreciation and therefore also weakens the prospect of the currency strengthening in future.

In the 1990s, the South African Reserve Bank tried to control inflation in part by managing the exchange rate. But this approach proved to be unusually difficult; it generated very large contingent fiscal liabilities as markets bet against our ability to defend the currency.³ The history of emerging market central banking is littered with cases of countries which tried controlling their currencies and ended up depleting their foreign exchange reserves and hobbling their economies with disruptive exchange controls. These stories usually finish with the central banks bowing to inevitable depreciation.

One solution to this dilemma is to get price and wage setters to use another anchor - to allow the exchange rate to depreciate in response to shocks without wages and prices taking off. This is a fundamental rationale for inflation targeting.

In recent years, we have observed that the pass-through from depreciation to inflation seems to be lower than it used to be. Many of our fellow inflation-targeting central banks have reported the same observation. This suggests that we are reaping the rewards of playing the long game. Since 2011, the South African economy has absorbed a 50 per cent depreciation of the currency without a major acceleration in inflation, and core inflation⁴ has not exceeded 6 per cent.

Then there's the short game. The South African Reserve Bank is often asked if we make monetary policy decisions based on the exchange rate. We certainly consider the impact of the exchange rate on the inflation forecast. However, given the well-established inflation-targeting regime and reduced pass-through, we do not need to react to depreciation too strongly. We can avoid unscheduled MPC meetings and instant rate hikes of hundreds of basis points.

But this doesn't mean that we can ignore depreciation entirely. Monetary policy requires some judgement as to whether currency weakness will cause inflation to accelerate to a persistently higher level that is inconsistent with the inflation target range. If we see this threat emerging, a monetary policy response is appropriate to reassure wage and price setters that longer-term inflation will remain on target.

When the rand began appreciating in recent months, some observers started speculating about a possible end of the interest rate hiking cycle and the possibility of rate cuts. The MPC explained, as clearly as it could, that such speculation was premature. South Africa's risk factors had not subsided. At the time, we believed that the rand's appreciation would not last - as indeed it did not.

Wage and price rigidity

The South African Reserve Bank has been concerned for some time about inflation persistence because of the way in which wage and price determination works in South Africa. I spoke at some length about this issue at the Labour Law Conference in Johannesburg last month and would like to repeat a few key ideas here because of their central importance to our economic circumstances.

The typical pattern is that wages and salaries, or overall remuneration, rises strongly, usually faster than prices and irrespective of where we are in the business cycle. In other words, remuneration costs rise when the economy is doing well and also when economic output is stagnant, which is where we are now. However, unlike food prices that can cause widespread inflation, rising remuneration doesn't have to become an inflation problem. If workers and firms negotiate pay and in the end agree on lower profits with higher pay, that's a distributional issue, not an inflationary one. In such a case, product prices don't rise but the distribution of returns from the sale of the product does, with workers getting more and shareholders less. Similarly, if employees become more productive and get higher pay as a result, inflation should be stable. In such a case, the higher pay is matched by higher output.

However, inflation pressures occur when remuneration goes up and firms just pass the higher costs on to consumers. With the relative lack of competition in our product markets, that's typically what happens. The end result is a sustained rise in inflation, which in turn keeps nominal interest rates high and undermines competitiveness. As long as this remuneration-price dynamic persists, our ability to lower interest rates is constrained.

To see if higher wages and salaries will become an inflation problem, the South African Reserve Bank pays close attention to unit labour costs (or ULCs), defined as average cost of labour per unit of output. Since 2010, ULC growth has ranged between around 6 and 8 per cent.⁵ Given our weaker economic growth, however, these ULCs should have moderated, resulting in less persistent inflation. They, of course, did not - or at least not by as much as we would have liked to see.

Moreover, with food prices rising and the currency depreciating, the risk of ULCs rising much more strongly has been fairly high. Higher food prices are likely to feed into ULCs with a lag, as employees over the next year demand pay increases to compensate for the above-target inflation we're experiencing this year. History has shown us that these responses to rising food prices carry through into inflation and can push inflation outside the target range for extended periods of time.

Monetary policy plays an important role in tackling this wage-price spiral. By defining and defending an inflation target range, the South African Reserve Bank can help wage and price setters to escape the spiral by giving them a focal point for their demands. But then it becomes very important for the inflation target range to hold. If wage and price setters start to believe that higher inflation is going to last, then they will adjust their demands upwards - and, in the end, they'll be right. They will fulfil their own prophecy and create higher inflation, with the more marginal, less productive jobs destroyed in the process.

Getting to the root of the remuneration and inflation problem requires more work - over and above what monetary policy can achieve. Our labour markets are quite segmented, each part operating differently from the others and each requiring different reforms to get a better mix of job creation and less inflation.

The highly skilled market reflects skills shortages, with employer competition fairly intense and unemployment low. This, in turn, generates strong salary increases, unhealthily feeding inequality and inflation. More jobs need to be created in this market, not just to reduce inequality and moderate inflation, but also to create more jobs for less-skilled workers. What this segment of the labour market needs, in the longer run, is more and better education to open the highly skilled job market to more South Africans and better match what students study with the skills required by the private sector. Temporary skilled immigrants would further ease salaries in this segment of the market and encourage economic growth directly by growing our domestic knowledge base. The possibility of pro-growth, pro-equality reform in this area should make it a policy priority.

South Africa's other two labour markets don't absorb nearly enough people, which leaves many South Africans unemployed or out of the labour force entirely. As the National Development Plan recognises, this problem has a lot to do with spatial legacies and access to job networks. Reducing the costs of transport and the costs of accessing job opportunities should lower both the supply cost of labour and, indirectly, the cost of creating jobs. Both would help to moderate wage inflation.

Institutions also play a key role in our inflation process. An important argument in academic literature, specifically by Lars Calmfors and John Driffill, connects unemployment to the kinds of collective bargaining rules used in a country.⁶ The basic idea is that a country can achieve better employment outcomes when costs take into account firm-level productivity or are guided at a very high, macroeconomic level. This is because company- or plant-specific unions have a special interest in the health of the firm that employs their members and will not want agreements that make the firm unsustainable. By contrast, very large unions have the power to shape macroeconomic outcomes. They will therefore avoid making deals which will push up inflation, drive up interest rates, reduce investment and growth, and ultimately destroy jobs.

The worst outcome is to be in the middle, between the two options. Yet our current arrangements lie precisely there, with industry or sectoral bargaining, implying that the actors (meaning unions and business associations) are big enough to inflict costs on the economy but not big enough to experience those costs themselves. Newer and smaller firms tend to have productivity levels that cannot be easily accommodated in such sectoral agreements, resulting in less job creation than we would like to have.⁷ This cost in jobs has been estimated at somewhere between 8 and 13 per cent of total employment, in large part because sectoral agreements hurt small, labour-intensive firms.⁸ South Africa has surprisingly few of these firms, relative to its peers. This helps to explain our unusually high levels of unemployment.

The implementation of the 3-6 per cent inflation target range

The last factor influencing our current position of being outside the inflation target range that I would like to discuss today is how we've gone about implementing the 3-6 per cent target range.

The original plan for the inflation target range, drafted in 1999, was to start with 3-6 per cent and then to narrow the range to 3-5 per cent.⁹ From there we would have had a pretty obvious midpoint, 4 per cent, on which to anchor inflation expectations.¹⁰

Unfortunately, the planned switch to the 3-5 per cent inflation target range was put off when the rand depreciated in 2001 and never reinstated. South Africa was left with a wide target range. In hindsight, this might have been useful. After all, it is not possible or even always desirable to steer inflation at one exact speed all the time. For instance, the standard practice with a temporary shock - such as a collapse in oil prices - is to look through the initial change in inflation and only respond if the effect is expected to be permanent. A wide target range means that these inevitable fluctuations can usually be contained within the range: inflation might move to the top or the bottom of the range but it will rarely breach the target range entirely.

Yet this isn't the way in which the policy committee has targeted inflation. Rather, we have kept monetary policy relatively loose over long periods of time, hoping for a bit more growth. We have tended to cut rates whenever inflation looked like falling below the midpoint of the target range. We have hit the lower half of the target range only for brief periods of time. Looking back, headline inflation has averaged 6,2 per cent over the entire inflation-targeting period (since 2000). Furthermore, in the event of shocks, we have repeatedly breached the upper end of the target range, with inflation at over 6 per cent for about a third of the time in the last five years. The perception of wage and price setters is that we have primed ourselves to miss the inflation target by aiming close to the top of the range, regardless of our position in the business cycle.

Trade-offs: final thoughts

By now, it should be clear why we've had high and rising inflation despite weakening GDP growth.

We've experienced a drought that has led to sharply higher food prices.

The exchange rate has been depreciating for several years and is now very weak. Some of this depreciation was appropriate and necessary, given falling commodity prices and some measure of policy tightening in the US. But the scale of depreciation we have experienced speaks to domestic challenges as much as to international ones.

Our wage and salary arrangements, along with rigid product markets, generate persistent inflation pressure, sometimes accelerating inflation but always preventing it from falling substantially.

Finally, by allowing inflation to settle at the top of the inflation target range, target misses have been more common whenever any of the underlying drivers of inflation exceeded their expected trajectory.

Imagine the counterfactuals. If wage and salary demands responded to economic conditions, they would have moderated as the economy weakened, weakening inflationary pressure. Exchange rate depreciation would have been less extreme without the domestic risk factors. With inflation expectations anchored at the midpoint of the target range, we would have had an easier time tolerating temporarily high inflation from food prices - and we would have had a clearer trigger for a monetary policy response. Indeed, inflation would probably have remained within the target range because the starting point would have been lower. In such circumstances, it might even have been possible to cut interest rates - the kind of response seen in commodity exporters such as Australia and Canada.

But we didn't have these advantages. The best response in our specific circumstances wasn't to cut rates. At a minimum, we needed to defend the upper end of the target range, ensuring that inflation expectations did not stray from the target entirely. The policy response helped to prevent investors from demanding higher interest rates to compensate for inflation risk, limiting the rise in longer-term borrowing costs.

In our second-best world, a gradual and cautious rise in rates was an appropriate means of countering rising inflation expectations. I would encourage you to be satisfied with this conclusion, but not happy. Happiness will only be appropriate after we have addressed the other problems, including the rigidities in our labour and product markets as well as the domestic risk factors that are exacerbating the rand weakness. Monetary policy can offset some of the worst consequences of these problems, but it cannot, single-handedly, get us out of this troubling mix of little growth, high inflation, and growing unemployment.

Thank you.

---