SEE THIS IN ACTION -->| Crash Peak Recovery Strategy | Rebound Strategy Vindicated |

ALSO SEE  THESE      --> | Avoid the crash in the first place | Pick the bottom of a JSE crash |
                                   | Understanding the Stock Market Cycles |

In this entertaining and fascinating research note, we make some interesting observations about the recent and 2nd largest JSE crash since the Great Depression, for a view to having a strategy in place the next time a correction rolls our way. In our ground-breaking research titled "UP/Down Volume and 90% Lowry's Days" we show you how to time the trough (bottom) of significant corrections and market crashes. This page you are currently on deals with what SHARES you must buy once you have decided the correction is over.

There is a well known Wall Street legend : "In a significant bear market that you feel is nearing the bottom, buy the shares that have fallen the most in value since the peak. Also buy the shares that fell the least in value."

We will put this to the test by ranking the JSE by share growth between the peak on 22 May 2008 and the trough of 20 Nov 2008 and then splitting the resulting list up into 38 portfolios of 10 shares each and measuring subsequent portfolio growth to 6th Jan 2009 (the peak of the "bounce" from 20th Nov 2008). The graphic below shows the periods we will be testing.

The chart below shows each of the 38 portfolios' growth during the bounce versus the peak-to-trough fall. We see that portfolio's 1 and 2, with those 10 shares having shown the most decline during the crash (-87% and -77% respectively) posted the highest growths of 24% and 42% respectively.

So whilst we can confirm the wisdom of buying the shares that fell the most, we cannot confirm the wisdom of buying the shares that fell the least, since portfolios 37 and 38 posted rather sub-par gains during the bounce, certainly below the ALSI's 29% gain. Note that only portfolio 2 managed to out-pace the gain of the ALSI, so it seems the wisdom lies in buying the 10 shares that declined the 2nd most as opposed to the most.

For your interest, here is the portfolio of shares that fell the 2nd most (shares ranked 11 through to 20 in size of fall) that would have earned you a 42% return in 7 weeks (over 300% CAGR). If we take the average growth of 42.63% and divide it by the standard deviation of 37.46% we get a Sharpe ratio of 1.14 which shows a reasonable correlation between extent of fall and rise.

Here we attempt to establish relationships between size (market capitalisation) and extent of fall and subsequent rise. The graphic below is a repeat of the previous exercise, but with the JSE ranked by market-cap and split into 38 portfolios: 

We make an interesting and important observation, namely that the larger cap shares seem to enjoy the extent of the rise the most (portfolios 1 to 14, ranging from R2.0Bn and up) as fund managers climb back in. The portfolio with the 10 biggest shares on the JSE, enjoyed the biggest bounce at 35%. The rest of the market remains fairly neglected and some shares even continue their decline.

Another interesting observation is that the sell-off, represented by the red line, seems fairly broad-based, with no discernible trend other than portfolios 1 and 2 of the largest cap shares "enjoying" significant declines of 50% and 39% respectively, and the smallest of shares represented by portfolio 38 (MCAP < R20M) seeming to come out of the sell-off relatively unscathed at -5%, probably due to their illiquid nature having saved them (however they did not feature in the bounce at all either, actually declining a further 5%)

Here is the contents of portfolio-1, which showed the highest return of 35% in the 6-week bounce period till Jan 6th 2009. Note the low standard deviation which leads to a much higher Sharpe of 1.81 versus the previous exercise which had 1.14. This shows that market cap is a very strong correlator for a subsequent rise.

We saw in the previous sections that "extent of decline" and "market capitalisation" were both good indicators for extent of subsequent rise out of a bear market. What if we were to combine the two and attempt to invest in the "biggest shares that declined the most"? We will examine two approaches to doing this.

In this exercise we will add each JSE share's "fall rank" and "market cap rank" together to get the desired "combined score" and then sort the JSE by combined scores and divide into 38 10-share portfolios. (Just like we do with PE and ROC in our magic formula strategies!) The results are shown below:

We see that by incorporating the fall rank into our overall scoring for a share gives us a "reasonably smooth" bottom-left to top-right curve for the "FALL" line. We also see more differentiation in the portfolios, with portfolios 1-4 dominating the rise landscape. In addition, this method boosted our portfolio-1 return to 51.5%, versus the Market-cap methods 35.5% and the extent-of-fall methods 42.63%.

But more importantly, look at how this 3rd "magic formula" method combined the high returns of the 1st method coupled with the high correlation of the 2nd method to give high returns and high correlation as shown by the analysis of portfolio-1 below. The incredible Sharpe of 2.43 shows that this method was an exceptionally strong correlator or "picker" of strong rising shares in the bounce.

As is characteristic of "magic formula" type strategies (as borne out by our extensive backtests on this type of methodology), the strong correlation would allow you to get away with only having 5-share portfolios without impairing returns or introducing volatility. In the above example, only picking the first 5 shares in portfolio-1 would have yielded a 49.89% return with a standard deviation of 10.48 which would have given you an eye-popping Sharpe of 4.76!

In this approach we isolate our universe to the JSE TOP-100 shares by market capitalisation, which according to current levels is all shares above R4.0Bn. As we saw in Section-2 these shares are likely to dominate the rise landscape. We then simply sort this universe by their growth in the crash and divide the resulting list up into twenty 5-share portfolios. The results are shown below:

We see a clear correlation between fall and subsequent rise in this graph, with portfolio-1, with the 5 shares that fell the most (-78.52%), rising the highest (70.78%). There is also clear differentiation between the portfolios with portfolio-1 almost double the return of portfolio-2 and from then on a bumpy but steady decline in returns. This is a definite confirmation of the " the shares that have fallen the most in value since the peak" part of the myth!

The shares in portfolio-20 that rose the most (23.98%)during the crash, only rose a further 15% in the bounce. This gives some credence to the "...also buy the shares that fell the least" part of the myth, but we are sure you would prefer 70% returns to 15% returns. However we do concede that whomever was an owner of the 5 shares in portfolio-20 on 22 May 2008 would have been very happy indeed as his/her shares were completely immune to the crash, having as a group grown a total of 39.68% during the entire period under review. For your interest the shares in Portfolio-20 are shown below (Note that FALL actually means RISE in this table):

But let's get onto the shares we are really interested in, the ones in portfolio-1, that got hammered the most during the crash, but also rose the most in the recovery. They are shown below:

So it would seem that picking the Top-100 shares and ranking by price-fall produces the more superior returns (70.78%) to the rank-combo 5-share method (49.89%) however the Top-100 method has a Sharpe of 2.49 versus the rank-combo 5-shares' eye-popping 4.76. This would indicate to us you were better off with the rank-combo method from a risk adjusted return perspective.

We only wish we had this type of research available during the crash. The conventional wisdom is clear : the next time there is a  market crash and you think you're getting close to the bottom, buy the biggest shares that fell the hardest to reap the greatest rewards!

Read our "UP/Down Volume and Lowry's 90% days" paper to see how to pick the bottom of major corrections.

Who said you could not make money out of a stock market crash! The next time there's a crash we will definately be following this strategy.

As if our prayers documented in section 4 above were answered, the market suffered another 21% decline to a trough on 3 March 2009. So we are going to put this research to good use by applying it!
We are going to create two "Bouncers" portfolios, at as 3 March 2009, based on methods A and B discussed in Section-3 and track them. Method-A is "Combo Bounce" and Method-B will be dubbed "TOP100 Bounce".

The portfolios details are below - they look scary, but we're convinced the theory will stick! So convinced in fact, that we are throwing R50,000 into each one.

You can see the progress of these two live portfolios at the SCOREBOARD menu.


B. TOP-100

SEE THIS IN ACTION -->| Crash Peak Recovery Strategy | Rebound Strategy Vindicated |

ALSO SEE  THESE      --> | Avoid crash in the first place | Pick the trough/bottom of a JSE crash |
Make a Free Website with Yola.