JSE Power of Value,Part-III

PREVIOUS : Power of Value on the JSE, Part-II

We had already shown in previous research (Price:Book, an all time classic) that in the crash of 2003, low price:book stocks outperformed high ones substantially over the ensuing 5 year bull run to 2008. By replicating the Brandes Institute exersise for the JSE for multiple 5 year periods, we were hoping to provide strong confirmation of trends discovered in our 2003 tests.

We only had data going back to 2000, so we could only create four sets of deciles to measure average 5 year growth, namely 2000-2005, 2001-2006, 2002-2007 and 2003-2008. We did however also decide to measure 1,2 and 3 year average growths for each of these decile sets. Decile sizes ranged from 25-27 shares depending on year. We are pleased to perform a first for the JSE with our results shown below:

We see startling confirmation that low price/book deciles substantially outperformed high price/book deciles, for 1,2,3 and 5 year growths, with significant outperformance by decile-10 which contained shares with lowest price-to-book ratios of an average of 0.37 and less. The results on the JSE are amplified enormously compared to other larger stock exchanges where similar research has been conducted. The graph below compares annualised compound growth achieved for each decile and growth period. Note how decile-10 has 1,2 and 3 year annualised growths almost identical at 66%:

Try and understand what the above graph is saying: "If you chose any starting date between the 4 year period of 2000 and 2003, and ranked the JSE by price-to-book and built a portfolio of the 25-27 lowest price-to-book shares, your portfolio would have grown on average 871% in 5 years, or 57.6% annual compound!" Put another way : "If you started in 2000, built a portfolio of the 25 lowest price:book shares and sold them 1 or 2 years later, and repeated the process you would have earned 66% average growth per annum for each of the four times you repeated the process!"

One last point : remember that the entire test period under review is 8 years (2000 - 2008) and thus includes the 2002/3 crash and the 2008 crash, which makes the above achievements even more remarkable.

NEXT : The Power of Value, Part-IV

Was the bottom on 27 Oct?

Was the bottom of the current bear market reached on 27 October? You will note that most of our candidate scans were run as of 29 October, as we felt this was a fair guess to the bottom of the market from which we obviously want to pick candidate stocks. But there is a Chinese proverb that goes : "Man who try to pick bottom get very smelly fingers!"

Our intent at PowerStocks is not to pick the bottom, but to get to a point close to the bottom either on the way down or on the way up. We want to figure when this bottom is close as that's when our research has shown the best value is available. Look at the below table on past JSE bear markets for the last 21 years (click for detailed view.)



The current bear has pulled the ALSI down 44.7%, slightly shy of the 45.9% in the '87 crash. The PE on the ALSI is down 50% to 8.3 (incidentally the same as the PE at the bottom of the last bear market) and the Dividend Yield is up 104% to 4.9%. High yielding stocks are going for a song.

Another interesting point is that those wiley investors that bought on 27 Oct would have shown a 15% increase by 5th Nov, merely 9 days later! Although earnings may drop in the ensuing months as business conditions tighten, markets are mostly forward looking and in all likelyhood the bulk of this has been priced in already. As Warren Buffet eloquently puts it "If you wait for the Robins to arrive, you will have missed Spring." Similarly, if you wait for all the good news before climbing in, you will have probably missed a good buying opportunity.

We will continue to peg 29th October as our theoretical "Bottom" in ongoing research and stock screens unless the ALSI goes significantly below 18,000. On that point, it is interesting to note that the trough on 27th October is 144% higher than the trough on the last crash - makes you think doesn't it?

Finally we note with interest that the current fall is very similar to the fall of '98. We are down 44.7% now and were down 43.8% then. We have been falling for 158 days versus 144 days then. The crash of '98 may be a better candidate for testing our theories we devised around the much longer bear market of 2002, and that's exactly what our next scope of research work will be. We will re-run the PE, Price:Book and Piotroski valuations on the crash of '98 and compare these to growths achieved by stocks in the subsequent 45 month bull. We can then compare our findings to the crash of 2002 to see if they are consistant!

Piotroski JSE Performance

As we saw from the previous post, the theory is you only apply the Piotroski scoring methodology to "undervalued" stocks with low price-to-net-asset-values (Price:NAV). The theory goes that the growth of these "distressed" stocks in the medium term (5 years) is closely correlated to the Piotroski "F" scores they achieved at the start of the assessment period.

PowerStocks put this to the test on the JSE, ranking the 5 year average growth of stocks together with their Piotroski "F" scores as at 28 April 2003 (trough of the last big bear market). We performed the tests with all JSE shares listed at that time and then only on those shares with Price:Book ratios of less than 1. The results are shown below:


The results are rather remarkable, confirming Piotroskis' theory for the JSE. Note how Piotroski is a great bear market screen for ANY stocks (shown by the blue bars) but worked especially well for "undervalued" stocks (shown by the red bars) when F>6. Note how portfolios of shares selected with F>6 significantly outperformed the ALSI and the group average growths.

Note how the average growth of the group P2Bk<=1 is much larger than the average growth of the entire group (All), confirming the theory we postulated that undervalued share portfolios are much more likely to outperform coming out of bear markets than other stocks (See Price-to-Book : An all-time classic)

We can state that historically, Piotroski "F" scores greater than or equal to 7 for "undervalued" shares showed significant out-performance to the group average (of undervalued stocks) and the all-share index (ALSI). It also showed a lesser but still significant out-performance for all stocks regardless of their price:book ratios.

The most important observation we can make here is that even though Price:Book has been shown to be a stong, safe selection criteria for potential growth stocks, Price:Book coupled with the Piotroski valuation significantly enhances portfolio performance. The almost 1,200% performance of the portfolio of "undervalued" shares with F greater than or equal to eight is double the group average of undervalued stocks (588%) and triple the growth of the market-cap based ALSI index (350%).

NEW : The Piotroski screen is our top performing portfolio we track on the JSE since 20 Nov 2008. You can view its performance coming out of the 2008 JSE crash over here.

NEXT : Piotroski JSE Candidates

Introduction to PE Ratio

The PE is a valuation ratio of a company's current share price compared to its per-share earnings. The P/E is sometimes referred to as the "multiple", because it shows how much investors are willing to pay per Rand of earnings. Made popular by the late Benjamin Graham, who was dubbed the "Father of Value Investing" as well as Warren Buffett's mentor, Graham preached the virtues of this financial ratio as one of the quickest and easiest ways to determine if a stock is trading on an investment or speculative basis. A low PE indicates value and conversely a high PE indicates "expensive".

Another great investor, Martin Zweig commented : "The data going all the way back to the 1930s show conclusively that stocks with low price/earnings ratios outperform stocks with high price/earnings ratios over the longer term". Zweig's view was that although stocks with high PE's have been known to perform well, when these stocks miss forecasted earnings they get punished much worse than lower PE stocks and thus he would avoid them altogether. However Zweig also avoided stocks with "too low" PE's as he considered that something must be fundamentally wrong for these stocks to warrant such PE's.

Of course, a low PE may indicate a "cheap" share but it will not necessarily indicate a "good quality" or "value" share. Generally speaking, PE ratios of stocks need to be compared to those of stocks in similar industries, however for the purposes of the PowerStocks PE study, since we assume the stocks to be all "depressed" or heavily sold off we will be simply seeking to correlate the growth of all shares related to their PE's reported during the end of the bear cycle.

The graph below shows the growth achieved by JSE stocks since the trough of the last big JSE bear market, relative to their trailing PE's as at 28 April 2003. Note that only 136 stocks that averaged more than 3 trades a day during the month of April 2003 were considered (98 stocks were considered not liquid enough)



Sure enough we find a strong correlation between PE and share price growth. More significantly, those stocks with PE's between 2-4 significantly outperformed the ALSI (All share Index) as well as all the other PE groupings. True to Zweigs affirmations, those PE's that were "very low" significantly underperformed the market.

There were 11 stocks with PE ranging from 2 to 4 and 28 stocks ranging from PE of 4 to 6, making altogether 39 stocks with a PE betweeen 2 and 6 which averaged 508% growth. As a general observation we can safely state that historically, stocks with PE's between 2 and 6 outperformed all other stocks.

NEXT UP : PE Ratio JSE Candidates