The equity market environment reminds me of the opening line to Charles Dickens’ A Tale of Two Cities. For most equity investors, 2008 represents one of the worst years on record. Yet, the massive decline in equity prices recorded across global markets since mid-2007 has dished up some exceptional value. Stocks are cheap if measured by multiples of their earnings, book value or dividends. Value stocks are particularly cheap. Given this result, the current environment may turn out to be the best of times for equity investors who have the patience to see out the turbulence and tenacity to take advantage of the bloody environment.

However, one aspect of the current setting that demands particular attention is the deceptive attraction of valuation multiples. To be sure, market multiples, such as the earnings yield and dividend yield are extremely attractive if compared to long-term averages. The JSE, for instance, currently offers a trailing dividend yield of 4.5 percent and an earnings yield of 10.5 percent – figures that are some 30 percent above the long-term averages suggesting a wide margin of safety and ample upside in price. However, the attraction in these historical multiples derives from the fact that the denominator in each case, namely price, has collapsed, but the numerators, namely dividends and earnings, are yet to fall as they correspond to the more robust business environment of early and mid-2008. Once this lag has worked itself out, earnings and dividends will fall, making it clear that the current ratios are deceptively flattering. This begs at least two questions as we head into South Africa’s earnings season. First, by how much are earnings likely to fall during the economic decline? Second, based on the lower earnings that we can expect, is there still value on offer for equity investors?

In trying to answer the question of the extent to which earnings are vulnerable to the economic slowdown and the degree of adjustment necessary to arrive at a realistic forward valuation multiple, people conventionally turn to analysts’ forecasts. But, such predictions tend to lag the results and are more often backcasts based on analysts’ reactions and hindsight than reliable projections. The evidence to support this point is well established in international literature. However, a recent example may help support the point. The US economy officially went into recession in December 2007, but analysts took some time to work this reality into their earnings’ forecasts’ (see Figure 1).

Notwithstanding analysts’ reactionary bias, even if the average forecast was reliable, there is often a high degree of dispersion in forecasts which further undermines their use in investment decision making. Given the high level of uncertainty around commodity prices at the moment, this point is particularly valid with regard to extant forecasts for South African resource stocks, which comprise 50 percent of the JSE/FTSE’s All Share Index. For example, ten analysts contribute to INet’s consensus forecast for Anglo American plc. At the start of February, the lowest expected earnings per share for 2009 is 1180c and the highest is 6220c. The consensus forecast for the stock stands at 4300c. Depending on which figure you use, the forward price-earnings ratio for the entity ranges from an incredibly attractive three times to a demanding 18 times. In any event, similar examples abound, with the wide forecast range clouding investors’ decision making.

So if reliable forecasts are evasive, what can give us guidance regarding future earnings? As a firm disciple of the value investing philosophy, the force of mean reversion steps immediately into the frame as a powerful tool for earnings guidance.

In this regard, it is noteworthy that current returns on equity (or ROE, which is measured as bottom line earnings divided by shareholders’ equity) are exceptionally high – and unsustainably so. For instance, if we take Central Eastern Europe, Middle East and Africa (CEEMEA) as a category (of which South Africa makes up 35 percent), then the current ROE of about 20 percent is well-above the long-term average of 15 percent. For the CEEMEA ROE to regress to the mean, it is necessary for earnings to decline 26 percent (other things remaining the same). Earnings must fall 57 percent to take the ROE down to its fifteen-year low (see Figure 2). This gives a sense of the extent to which earnings could decline in CEEMEA markets.

Some sectors show ROEs that are particularly high. Energy (gas and oil), materials (resources), industrials (heavy industry, including building and construction and motor) have the most to give up in terms of ROE regressing to the mean. That is to say, it is in these sectors that earnings arguably are most vulnerable. Information technology, telecoms, consumer durables and utilities have ROEs that look ‘more normal’, implying that earnings may already have begun the process of normalising in those sectors. By contrast, ROEs in the consumer cyclical, health and financial sectors now stand well below average. These sectors, in other words, might already be pushing through the trough of the earnings curve.

The ROE theme in CEEMEA is echoed in South Africa. Current ROEs for the market are above the long-term average. On this front, it is reasonable to expect a long-term ROE in South Africa of around 20 percent. Yet the ROE on resources (materials and energy) is well above this average (34.5 percent). Industrials are marginally above the average. Financials, in line with CEEMEA, are comfortably below the long-term average.

If we allow the force of mean reversion to apply to earnings, over 12 months we can expect the following performance for South African earnings: resources -45.0 percent; financials +12.5 percent; and industrials -9.5 percent. If we apply the JSE/FTSE All Share Index’s super-sector weights (50 percent resources, 30 percent industrials and 20 percent financials), this implies that earnings in South Africa need to decline by 22.9 percent to restore the long-run ROE.

Under this assumption, the current price-earnings ratio of the JSE of 9.5 times climbs to 12.3 times and the dividend yield moderates to 3.5 percent. Still, if these revised ratios are applied to a five-year return model it implies that the JSE can be expected to return 17.5 percent per annum off the revised valuation base. Under this assumption, an investment of R100, if left alone for five years, would grow to R225. Inside of this terrain, value investors will quickly identify greater opportunity. Financial stocks in South Africa, for instance, trade on an earnings yield of 14.8 percent and offer a dividend yield of 6.8 percent. If the force of mean reversion applied above holds, then these multiples will become even more compelling as 2009 evolves.

A shortened version of this article was published in The Weekender