22 April 2014
Based on valuation metrics and given that the impact of higher interest rates is already factored into the valuations of the SA Industrials and Interest rate sensitive counters, the downside for these SA economy-dependent stocks appears more limited than other key JSE sectors.
These are not the best of times for the SA economy, but not the worst of times for the JSE
The outlook for the SA economy is unsatisfactory. The rate of growth is slowing down and the Reserve Bank seems intent on raising interest rates that will slow growth further. This outlook does not portend well for those businesses that serve the SA economy. Investors on the JSE however can take comfort from the fact that the SA share market has become more dependent on the performance of the global economy. If we break down the earnings of the JSE All Share Index, over 60% is generated from revenues outside SA, while 70% of the daily movements on the JSE can be attributed to these global influences.
Should investors prefer global to local plays?
The issue for investors on the JSE therefore is how to allocate their exposure between the listed global and SA plays. Clearly, the worse the SA economy is expected to perform, the stronger the case for preferring the global over the SA plays.
The SA bond and money markets can be confidently presumed to have factored in at least a further 100bp increase in the key SA Reserve Bank repo rate over the next 12 months, with the expectation of more to come over the following 12 months. The stock market will also have taken this into account.
These forecasts currently weigh upon the valuations accorded the SA economy plays, especially those companies for whom the direction of interest rates has proved influential in the past, which can be identified as interest rate sensitive, such as banks, credit retailers and property companies. Yet we can also be confident that the forecasts and assumptions that have influenced the market place will be altered over time with the news flow and that the market will move accordingly.
What if interest rates turn out to be lower than currently forecast?
Should interest rates in SA increase by significantly less than currently expected, it would likely mean unexpectedly high and superior risk adjusted returns from investing in the SA economy plays on the JSE. The opposite impact would be registered should SA interest rates turn out to be even higher than currently expected. We regard the danger of upside interest rate surprises as significantly less than downside surprises for reasons to be explained.
The exchange value of the rand will determine the outcomes for inflation and interest rates
The path to lower than expected interest rates in SA would have to be opened up by a stronger rand. The path to higher than expected rates would have to follow a still weaker rand. A consistently stronger rand improves the outlook for inflation and would avoid any need to raise interest rates and vice versa. Chris Holdsworth1 has successfully replicated the inflation model of the Reserve Bank for Investec Securities and suggests that a rand weaker than R10.70 to the US dollar would lead to higher interest rates, designed to bring inflation back within the headline inflation target range of 3% to 6%. This target band for inflation is currently under threat following the rand weakness of the past 12 months – hence the expectation of higher short term rates. A rand consistently stronger than this R10.70 would keep short rates on hold.
The Holdsworth replication of the Reserve Bank econometric model gives a good sense of some of the dilemmas faced by the Monetary Policy Committee (MPC) of the Bank. The most important of these is that inflation has risen, and may rise further, even though the economy is operating at well below its potential, for want of domestic spending.
Raising interest rates would depress demand further, with little impact on the inflation rate itself. Holdsworth calculates, using the Reserve Bank econometric model, that a 50bp increase in the key short term interest rates typically leads to only a 16bp reduction in the inflation rate, seven quarters out. Therefore, should the rand breach R11.30 against the US dollar, the Reserve Bank would have to raise rates by 100bp or more to hope to get inflation below 6% by the end of 2015.
The possible influence of the SA output gap on the market outcomes
Such an increase in rates would have a highly predictable impact on spending and the so called output gap as estimated by the model. The output gap is the difference between the potential growth of the economy, estimated at 3-3.5% per annum by the Reserve Bank, and actual growth, which is currently well below this. This gap, as shown in the chart below, is currently very wide and will increase further with higher interest rates. Note too that the current GDP output gap is now at historically high levels, indicating that the economy is operating below its growth potential, for want of sufficient spending by households in particular.
1Chris Holdsworth, Quantitative Strategy, Investec Securities Proprietary Limited, Second Quarter 2014, April 2014.
A still weaker level of demand and the still fewer jobs associated with higher interest rates would be an unacceptable price for the economy to pay and the Reserve Bank might be persuaded that higher interest rates are not appropriate given the trade offs in the form of slower growth. Furthermore, as we show in our next chart, there is evidence that higher short term interest rates may be associated with a weaker rand, so the outlook for inflation, despite the lower levels of domestic spending, could deteriorate rather than improve with higher rates. Yet the market still expects the Reserve Bank to think and act otherwise. Our view is that should the output gap threaten to widen further, the Reserve Bank might think again about raising interest rates even if its inflation forecasts have not declined.
Questioning the structure of the Reserve Bank econometric model
One of our issues with the structure of the Reserve Bank econometric model is the assumption, incorporated in the multi equation model, that higher inflation leads to increases in wage inflation and in disposable incomes, adding impetus to household spending and helping to close the output gap.
This presumed relationship – more inflation and so temporarily faster and perhaps unsustainable growth – however relies on a demand side explanation of inflation. In other words, high rates of inflation reveal a state of excess demand in the economy that needs to be restrained. This, as we have indicated, is very far from an explanation of current inflation in SA.
The model also assumes that inflation expectations can be self-fulfilling and lead to higher prices even though demand pressures may be weak. There is no evidence of this. Inflation in SA leads inflation expectations, which remain very stable at about 6%. There is no statistically significant feedback loop from inflation expected to inflation.
Higher consumer prices can alternatively be attributed to reduced, or more expensive, supplies of goods and services, the result of the weaker rand as well as for electricity and other administratively set prices. Such supply side driven increases reduce rather than add to real disposable incomes and spending power, especially when formal employment is in decline or growing very slowly. A weak rand is bad news for the purchasing power of SA households. It becomes even worse news for them when accompanied by higher interest rates. Furthermore, households account for over 60% of all spending and the incentive private firms have to add to their plant and equipment – which accounts for another 15% of GDP – is derived from the spending actions of households. A return to SA GDP growth rates of a mere 3.5% pa will have to be accompanied by an increase in household consumption growth rates.
A combination of a stable rand with stable interest rates is essential for this important purpose. A lift in exports, labour relations permitting, could help stimulate more GDP growth, but would have to be followed up by a recovery in household spending growth rates.
The further dilemma for the monetary authorities is that, even if the output gap were to widen further, it is quite possible that the inflation outlook could worsen rather than improve. It would worsen if the rand were to weaken further and the so-called pass through effect of the weaker rand pick up strength. There is no predictable relationship between SA interest rate moves and the value of the rand, as Holdsworth has again confirmed (see the chart below).
The Reserve Bank – will it undergo an economic reality check?
Changes in short term interest rates may therefore help hit or miss inflation targets even as changes in interest rates have the opposite impact on the growth in domestic spending and output. This leaves the Reserve Bank with uncomfortable trade offs between inflation and growth, which may encourage the MPC to take little or no action on interest rates, as has been its inclination recently.
The MPC may recognise the inadvisability of raising interest rates in current circumstances, having seen how the rand weakened further as interest rates rose in January and strengthened after the March meeting of the MPC when the rates were left on hold.
Global forces will drive the rand and the SA markets
The inflation rate outcomes for SA will be determined mostly by the direction taken by the rand, independent of SA interest rate settings. The global forces that influence the exchange value of the rand are the state of emerging market bond and equity markets. This pattern is highly consistent. These developments in emerging bond, equity and currency markets, in turn reflect the appetite for risk by global investors and the outlook for the global and emerging market economies. The better the outlook for emerging market economies, the more upside for commodity prices, emerging market equities and their currencies.
These helpful trends for emerging markets and currencies, led by lower long term interest rates in the US were a notable feature of the markets in March 2014, when the MSCI Emerging Market Index gained nearly 3% in the month. In January 2014, the Index lost nearly 7% of its value, before adding over 3% in February followed by another strong month in March 2014.
As we show below, these trends were especially helpful to the SA economy plays on the JSE in March 2014. In Q1 2014, we calculate that our grouping of SA Industrials and, to a degree, overlapping SA Interest Rate Plays in March, provided returns of over 9%, the Global Consumer Plays generated negative returns of about 1% and the Commodity Price Plays returned 1.5%. Were the rand to surprise on the upside again, a similar pattern of outperformance by SA economy plays is likely.
The forces that could drive emerging markets higher
Over the next six months, investors in emerging markets, including the JSE, should hope for minimal pressure from US and other developed market interest rates, especially from long bond yields. They should hope that Fed governor Janet Yellen and ECB president Mario Draghi continue to emphasise deflation rather than inflation until the prospect of stronger global growth is fully confirmed. This dovish approach seems a likely one and its realisation will be to the advantage of emerging market economies and the companies dependent on them. Over the longer term, faster growth in developed economies will bring higher interest rates, as demand for global capital improves and when faster growth in emerging markets can compensate for the higher costs of finance that should follow.
Breaking up the JSE by dependence on global or domestic economic forces
The large listed companies that dominate the JSE can be broken down by the degrees of dependence on global or domestic economic forces as indicated in the figure above. There are the Commodity Price Plays, made up of resource companies that depend on the US dollar prices of metals and minerals. This category of JSE-listed companies best excludes the gold mining companies that dance to the sometimes very different rhythms of the gold price. The Commodity Price Plays include the large diversified mining companies Anglo American (AGL), BHP Billiton (BIL) and Sasol (SOL), as well as the platinum, iron ore and coal mining companies.
Then there are the SA Interest Rate Plays: the banks, the credit retailers, property companies etc whose valuations are dependent on the direction of SA interest rates. They include the major banks like Standard Bank (SBK) and insurers like Sanlam (SLM).
A further category of SA economy plays are the less interest rate-exposed SA Industrials, including the cash retailers.
The other important category of Industrial companies listed on the JSE are those clear plays on the global economy. These include Naspers (NPN), Aspen (APN), Richemont (CFR), SABMiller (SAB), British American Tobacco (BTI) and MTN, whose fortunes and valuation depend mostly on the profitability of their activities outside SA. We describe them them as Global Consumer Plays or as industrial hedges since they are hedged against the SA economy and the rand.
About 60% of the JSE All Share Index weighted by SA shareholders (the SWIX) is accounted for by a mere 10 companies, of whom only three companies may be described as heavily exposed to the SA economy: Standard Bank and Sanlam, both interest rate sensitive companies; and Sasol, a resource company, which generates much of its revenues and costs in SA but whose revenue is closely linked to the global US dollar price of oil.
2In this breakdown of the JSE we have combine the individual stocks into their various categories as follows:
Interest Rate Plays:
ABL, ASA, DSY, FSR, GRT, INL, INP, NED, RMH, SBK, SLM, BVT, IPL, MMI, MSM, PIK, SHP, TFG, TRU, WHL,
Commodity Plays:
AGL, AMS, ARI, BIL, IMP, LON, SOL, ACL, KIO
Industrial Hedges / Global Consumer Plays:
BTI, APN, SAB, NPN, SHF, CFR
S A Industrials:
BVT, IPL, SHP, TBS, VOD, BAW, LHC, AVI, SPP NPK
The case for SA economy plays – at current prices
The key question we now address is whether or not the JSE, taken as a whole, can be thought to offer the prospect of good returns for investors, at current valuations. The further issue is the valuation currently attached to the SA economy plays. Can they be considered demandingly valued at current valuations that incorporate the expectation that SA interest rates will rise over the next 24 months? Clearly there is upside for this class of shares should interest rates not rise as expected.
Yet a consideration of the downside risks to this class of shares needs consideration, as well as upside risks. Risks to the JSE All Share and SWIX Indexes, positive or negative, will impact to a lesser or greater degree on any subset of the market, including the SA economy plays.
Holdsworth shows that the JSE All Share Index is currently priced within the norms established since 1995. Cyclically adjusted or normalised earnings suggest an even less demanding rating for the market than do reported earnings. This is because the earnings from the Resource companies listed on the JSE are currently in only partial recovery mode from a deep, cyclical decline in normal earnings.
It would seem accurate to conclude that, by the standards of the recent past, the JSE was not demandingly valued at March 2014 month end, relative to trailing earnings (provided earnings from the Resource companies do in fact return to their cyclically adjusted norms). Clearly, such valuations also depend on the current state of emerging markets generally.
In our valuation models, we allow for an emerging market influence on the US dollar value of the JSE All Share Index and Financial & Industrial Index, in addition to factoring in the level of JSE earnings in US dollars and the interest rate spread between SA and the US. These valuation models, which have provided a good long run explanation of market value, suggest that the All Share Index is between 10 and 20% above the valuation predicted by the model.
Using a similar approach to valuing the S&P 500 indicates, by contrast, a high degree of undervaluation, 20% or so. This suggests that the better value for shareholders is for now still to be found in developed rather than emerging equity markets.
These currently stretched valuations of the JSE have been encouraged by the above long term average growth (of 10% pa) in Financial and Industrial rand earnings in recent years. This recent strong and predictable growth has been augmented by the exceptionally good earnings growth reported by the Global Consumer Plays. The SA Plays have also performed very well on the earnings front, especially in recent years.
Looking at Price/earnings ratios
In the figure below, we show the very different ratings enjoyed by these sub-sectors of the JSE. Global Consumer Plays now enjoy a much improved and exalted price/earnings rating compared to the other sectors.
While Naspers accounts for the largest share, 12.5% of the Top 40 companies included in the SWIX, its share of the Global Consumer Play sub-category is as high as 29%. Almost all of the value of Naspers can be attributed to its large shareholding in Tencent, a Hong Kong listed internet company. Without Naspers, which has recently traded at 65 times reported earnings, the Global Consumer Plays, taken together at SWIX weightings, would enjoy a still demanding 19 times earnings rating compared to the more demanding 24 times multiple when Naspers is included.
While the earnings performance of the SA Industrials has compared well with those of the Global Consumer Plays, these SA economy-dependent companies realise a significantly lower rating in the market. The Global Consumer Plays must be considered as growth companies while the group made up of SA Industrials and SA Interest Rate Plays can be considered as value stocks with much less demanding valuations.
Holdsworth has estimated that the trend earnings growth is 14% pa for the Commodity Price Plays, 9% p.a for the SA Interest rate plays, 15% pa for the Global Consumer Plays, 11% pa for the SA Industrials. The volatility of these earnings growth trends has been lowest for the SA Industrials at 17%, compared to 19% for the Global Consumer Plays, 23% for the SA Interest rate plays, 61% for the Commodity Price Plays and an extraordinarily high 158% for the Gold Mining Companies. Part of the case for the SA Industrials is thus their lower risk character. If these growth trends in earnings are sustained, current valuations of the SA economy plays cannot be regarded as demanding.
It may be argued therefore that, based on valuation metrics and given that the impact of higher interest rates is already factored into the valuations of the SA Industrials and Interest rate sensitive counters, the downside for these SA economy-dependent stocks is more limited than that of the Resource stocks or the Global Consumer Plays.
The upside is that the rand will be stronger than expected and interest rates lower than expected. All these sectors of the JSE will benefit from stronger global growth and strength in emerging market equities and commodity markets that would accompany any renewed appetite for bearing risks in emerging markets. But the SA economy plays as a group stand to benefit most from a strong rand and all that will follow a stronger rand, especially lower interest rates.