SA equity – a shifting track record
South African investors have enjoyed phenomenal returns over the last 2-3 decades, as local equity returns outpaced those of most equity markets around the world. As a result, the “cult-of-equity” has become well entrenched, with South African investors expecting double-digit growth from the local equity markets over the medium to long term (chart 1).
However, over the last 4-5 years, South African equity investors have been disappointed. Local equities have failed to live up to their expectations. After this period of lacklustre returns, investors are wondering if now is the time to buy, especially since many equity specialists are arguing that South African equities, and in particular “SA Inc” (shares specifically exposed to the South African environment), are offering compelling value.
Europe and Japan are just two examples where equity markets have exhibited periods of stagnant returns for long periods, and even economies or markets with very decent economic growth, like China, have delivered poor equity index returns for long periods. Low-return periods do happen and, while historic precedent may suggest that now is a good time to buy South African equities (on an asset class view, not specific names), we should investigate a little further.
Calculating “true” value
Currently our local equity market trades at broadly a 14 times trailing Price: Earnings multiple (PE) over the last 12 months. This means that for every R1 of earnings that companies have generated, the market has been willing to pay R14. This is close to the equivalent of the average PE multiple of 14X extending back to 1994 (chart 2).
While historic valuation metrics are crude identifiers of the worth of assets, understanding the drivers of specific valuation models or more specifically the drivers of a company’s discounted cash flows is far more important in determining if a compelling valuation is an appropriate opportunity. While valuation models can be complex, with multiple inputs, suppositions and sensitivities, we believe they can all be encapsulated into three broad questions:
- Describing stock X as cheap or market Y as expensive relative to their histories contains limited information without explaining why that specific situation exists.
- Does the sector or market deserve a different price earnings (PE) multiple than currently deemed appropriate by collective market participants and, if so, why?
- Will other market participants recognize the value that has been identified and cause a rerating of the multiple or consider that the valuation opportunity is compelling enough to buy the assets? This final point is ultimately what separates those assets where value exists versus those that are value traps.
One of the key questions currently arising from our scenario planning asset allocation process is whether investors should anchor onto a historic average PE ratio to identify when a stock or equity index (market) is cheap or expensive (Table 1)?
The above table shows a range of emerging and developed markets and some basic valuation summaries for each broad equity market. Brazil is currently trading at a 16.2X trailing PE, 2.17X Price to Book, 1.5X Price to Sales and 8.76X Enterprise Value (EV) to EBITDA. Depending on the valuation metric, and using 20 years of history, this indicates (as reflected in the second half of the table) that Brazil is marginally cheap on a trailing PE and EV/EBITDA basis but expensive on a Price to Book and Price to Sales basis. Overall, Brazil is a little expensive (0.59 average Z-score relative to its own history).
The entire sample set currently appears close to fair value on a historic basis with an average Z-score of 0.09 and median of -0.07. However, SA is cheap on every metric relative to its history (starting in 1994) and appears cheap when compared to the average of the sample set. Conversely, the US looks very expensive relative to its own history and to this broad sample set.
Using history as a guide, broad metrics might suggest a strong buy argument for South African (and Turkish) equities given these simple valuation metrics and historical observations.
However, we cannot look at history alone. We need to question whether something has changed that could challenge the previous historic precedent. Is the current anomaly an opportunity, or is there a structural shift? Small anomalies are opportunities and are created by market noise, but shifts in thinking are real warning signs. Signals are more important and they work both ways, creating longer-term opportunities for investors to make, or avoid losing, money. In the chart above, you can see that South African equities are currently either providing a opportunity compared with previous 5-10 year periods or are signalling something has changed significantly.
Exploring noise and signals
We do not believe the cashflow environment for South African equities has changed meaningfully over the last five years. Growth rates from a top-line (revenue) and bottom-line (profitability) perspective have moderated but net margins remain within recent historic parameters. Profitability, using Return on Equity as a proxy, has declined modestly but is within the range of the last decade.
What appears to have changed more significantly within the broader South African market is that the current risk-free rate has become significantly more elevated than previously, despite economic growth having moderated since pre-crisis levels.
The increase in South African bond yields has been a favourite discussion point for investors since late 2015, largely centering on the potential exclusion of South African bonds from the World Government Bond Index (WGBI). The concerns have been valid for some time although the deterioration of SA’s fiscal position has reached a point where one could consider the “fiscal cliff” is approaching, or has even occurred. The deterioration in SA’s balance sheet has reached proportions that would warrant external help if all State-Owned Enterprise (SOE) liabilities were included in the debt-to-GDP calculations, even excluding all the other liabilities classified as non-contingent. Hard policy decisions remain elusive, although policymakers realise that economic growth is desperately required. However, while policymakers acknowledge that the investment environment is drought-stricken, they have made no public acknowledgement that future growth will also be curtailed by a utility unable to provide sufficient electricity for a spluttering economy to ignite. As a result, consensus growth forecasts for the next few years remain ambitious.
The higher bond yields point to increased bond issuance by the South African government and the rising risk in the country’s credit quality. This is shown clearly in the chart above. South African real rates appear to offer value but investors are clearly demanding a greater margin of safety from South African bonds.
Less obvious to investors is the feedback loop into equities. It is true that the composition of South African equity markets has changed dramatically over the last decade and a half and does not reflect the reality of South African equity investment.
While company profitability may remain reasonable (although lower), true profitability comes from earning a higher return than the cost of capital, both in the short and the longer term. Equities by nature are long duration assets. As a result, investors looking at cashflow prospects into the future need to consider the present value of these flows and discount them accordingly. Based on the cost of equity capital for a company (or companies), we are able to make some assumptions about where the fair value of that company or asset class lies.
We highlight below two simple models for calculating the cost of equity:
1. The CAPM model is familiar to investment professionals. The model allows for the calculation of South Africa’s Real Risk-Free rate to which an observed equity risk premium (since it could be assumed that equity carries a higher degree of risk than a sovereign bond) is added, and then calculate a cost of equity, before inverting to derive a fair value PE multiple.
Applying a fair value PE multiple (derived from the cost of equity), we can establish a top-down view about the fair value of the asset or asset class in this instance.
What this model infers is that a fair value PE multiple for South African equities is about 8.92X. Relative to the historic experience, many investors assure us that about 14X is appropriate for SA. Based on that multiple assumption, and current earnings (at the time of writing), that would mean equities would need to fall by around 28% (depending on the index) to arrive at fair value. That assumes that the earnings themselves do not decline. If they do decline, there would be more downside. If earnings grow, the downside would be less than indicated above.
2. An alternative model, based on the Gordon Growth Model, takes next year’s dividends and establishes a value for the company, based on cashflows paid into perpetuity and discounting the cash flows by the cost of equity less a long-term growth rate. Growth in this model is derived by using the last five years’ dividend growth rates for the index and a proxy for implied growth using reported Return on Equity (ROE) and payout ratios.
We derive a fair value PE multiple by looking at shareholder returns (the Payout Ratio) discounted by the cost of equity less growth (similar to the Gordon Growth Model).
Using the above inputs, we can apply the “fair value” multiple to South African equity indices to establish a reasonable expectation of the fair value index level.
By applying this second model, at the time of calculation (late September 2019) the market appeared to be was reasonably efficient in its pricing, and had adjusted to the more international flavour of our equity market. The market was trading at close to fair value, but was not particularly cheap, especially given earnings risk.
Neither of the above models considers changes in a future world (aside from implying dividend growth over the next 12 months) so we would need to use quantitative and qualitative techniques to decide what earnings, asset turnover and free cashflows could be, all of which will impact earnings and hence index fair values. All of these are nuances that require skilled analysis.
In looking at rising yields, it is clear to us that South African investors are acutely aware of the impact that rating agencies, the country’s fiscal position and the low growth rate could have on the South African bond market. However, there has been little discussion of the impact of the rising cost of capital and therefore companies’ fair value multiple ratings.
We think fair value for South African equities probably sits between these two simple models. Neither model provides a compelling valuation rationale to “fill your boots”.
Taking this one step further
South African equities have generated a total return of approximately nineteen times an initial investment made in January 1999. Of that total return, 59% has come from earnings growth, 5.93% has come from dividends and the remaining 35% has come from multiple expansion, calculated here as a residual.
Assuming that the last 20 years is representative of the next 20 years, applying the aggregate historic multiple may be appropriate for establishing a fair value multiple. However, in the last 20 years SA’s cost of capital has fallen, along with global rates. Globalisation has grown Emerging Markets as an asset class. The rise of China from 2001 and the speed of its rise was unprecedented, as was the demand for the raw materials SA produces.
The flow of global capital remains strong, but technological innovation, disruption and productivity demands in a knowledge-based global economy leave SA behind a large part of the pack hunting for global investment flows. As a result, we would ascribe a very high likelihood that the future will be very different from the past. Even if earnings continue to grow at the same rate as before, we believe the multiple expansion SA enjoyed could reverse.
Reaching a conclusion
Evaluating why and how different things could be is part of our asset allocation process. Having a scenario-based approach compels us to focus on separating where noise and signals may differ. We try to focus on what may be different in future, so we understand where we might go wrong within our asset allocation and portfolio construction processes. With multiple ways of evaluating a dynamic world, we can highlight potential risks to our inherent biases and long assumed beliefs.
Through our scenarios, we are currently questioning whether South African equities offer fair value. We have provided two examples of models to highlight how South African equities may be regarded either as offering fair value or as expensive. We also highlight that with the risk to historic earnings growth and SA’s higher costs of capital – which will impair profitability – we believe there is a significant probability of a structural de-rating (over the medium term) of the fair value multiple currently being used as an anchor by many South African-orientated equity investors.
Of course, there is considerably more to generating returns than purely valuations but, in short, we believe South African equities (as an asset class) are more value trap than value opportunity.
The above article is derived from a talk by Marius Oberholzer in late October 2019 entitled:
“EYES WIDE OPEN: IS IT TIME TO SHIFT OUR THINKING ON LOCAL EQUITY”?