Thursday, March 10, 2016

Dow Theory Special Issue: Putting the Dow Theory under Stress-Test (IV)

Will future secular bear markets allow the Dow Theory to outperform buy and hold? Will future secular bear markets be so obliging to the Dow Theory as past ones?

The preceding two posts of this saga (here and here) dealt with the Dow Theory performance under secular bear markets. I concluded that I thought that future secular bear markets would resemble past ones in that the “swings” were going to be of sufficient magnitude so that to allow the Dow Theory to be profitable. Today, I will explain why I believe that the future will be similar to the past, and hence, secular bear markets will not be characterized by small swings (i.e. not reaching 20% from top to bottom) but rather ample ones which may include even 50% swings.

First of all, I believe in trends. Trends persist. It has been amply documented that the market is not a random walk (study, for instance, Michael Covel's book "Trend Following"). Trends exist because there is a tendency to overreact for psychological and fundamental reasons. Hence, trends based on fundamentals (i.e. economic crisis) will last longer than one expects and when coupled with psychology (bandwagon effect, etc.) there will likely be overshooting 

Secondly, the very rationale for pricing stocks warrants big swings. Basically, the price of stocks depends on two factors: The expected earnings and the discount rate (PER). An adverse economic environment, results in decreasing (or slowing at least) earnings. However, as a side effect, reduced earnings result in lower PER (high capitalization rates). In other words, people are less willing to pay for earnings. Thus, an earnings reduction of 50% may perfectly entail at the same time a PER reduction of 50%, a double whammy. Hence, if the stocks market was at a level of 100 with earnings at 5 units and PER at 20, an earnings reduction of 50% may result in the following stock market price:

            100 -50% earnings reduction = 50 units.
            50 – 50% PER reduction (PER of 10) = 25 units.

Which amounts to a total decline of 75%. This is precisely what happened in the 1930’s. There was a drastic decline of earnings, which at the same time triggered a massive PER reduction.

If you think it carefully, PER reduction makes fully sense. Less earnings means less money (wages too, as companies lay off) to be spent in the stock market (less demand). Furthermore, hard economic conditions increases the supply of stocks (people now in dire straits which is forced to sell, rich people getting less dividends who have to sell to maintain their lifestyle, etc.).

Furthermore, I feel that current and future times will encourage extreme swings, as central banking creates bubbles which subsequently pop (as it may be happening right now as a write these lines). Credit expansion and contraction further exacerbates the ebb and flow of earnings and PER.

Thus, it is not outlandish to predict an S&P 500 level of just 500. A dramatic reduction of earnings by 50% would see to that, as PER would also likely decline by 50%. However, I distrust those pundits predicting that it will occur tomorrow. I am not interesting in fundamentally calling the top of stocks; I have the Dow Theory to help me if such a dramatic decline à la 1930 happens again. 

Thirdly, on an empirical basis (which may not necessarily be a guide to the future, but helps), primary bear market signals as determined by Schannep’s Dow Theory entail, on average, a further -13% decline until a bottom is made. Since we trade, and live or die, by such primary bear/bull market signals, the past record seems to suggest that it is likely that the average swing (as detected by the Dow Theory) amounts to ca. 20% from top to bottom (-13% subsequent decline after the signal + ca. 6-7% from the top until the signal is flashed). Swings of an average magnitude of 20% are sufficient to be profitable traded according to the Dow Theory. More about the average decline following a primary bear market signal, with updated spreadsheet, here.

Forthly, even if we ignore the Dow Theory, and we just stick to a “bear market definition consisting of a simultaneous decline of -16% both on the S&P 500 and the Industrials (more about it, on Schannep's website "", and his must-read, master piece book, "The Dow Theory for the 21st Century"), we know that in past instances, a subsequent -13.20% decline followed the declaration of a bear market.

For all this reasons, I feel it is very likely that future secular bear markets will “oscillate” with a range sufficiently ample to be profitably traded under the Dow Theory, especially “Schannep’s", which is very reactive.

However, in spite of all the foregoing, an astute reader could be asking himself: All this is nice and dandy. However, the right “overlay” for Dow Theory signals (especially , those of Schannep) are the cyclical bull and bear markets, not the secular ones. A secular bear market may easily “contain” several cyclical bull and bear markets, which at the same time “contain” primary bull and bear market Dow Theory signals. Will such cyclical bull and bear markets allow the Dow Theory to outperform, and more importantly, avoid big drawdowns?

The answer to this rather abstruse but important question will be given in the next chapter of this saga (with a clarifying graph to be provided). In the meantime, digest the present post, and ponder about this vital, question.

The Dow Theorist.

No comments:

Post a Comment