I have been following the fortunes of various ETFs over the past few weeks, and especially the closure of ETFs which implemented specific strategies - many of which featured volatility, momentum and high and low beta in their nomenclature.
That specialized ETFs failed in beating common indexes is disturbing. These products were supposed to bring to retail portfolios some of the strategies (and by inference returns) previously only available to the institutional and wealthy investor classes using high powered hedge fund managers.
Nevertheless, I retain an interest in some of these strategies. What is the use of studying computational finance and not giving something like this a go? The "High Beta" strategy is one which is attracting my attention. With the market surging unexpectedly but regularly, my view is that "High Beta" should be a pretty good trading strategy requiring a modicum of quantitative analysis and minimal additional fundamental research. Consequently, it should deliver cheap alpha.
So, given these advantages, what needs to be done to develop a "High Beta" strategy and get results commensurate with expectations, and of course better than a specialist ETF?
I have organised an analysis along the following lines:
- Classifying "High Beta".
I understand that "High Beta" is defined in relation to volatility, with "High Beta" stocks being more volatile than an index. I think there is more at stake than that. To attract my attention, a "High Beta" stock should exhibit, as much as possible, smooth increases in alpha, not just on odd occasions. Therefore, my view is that the best candidates for "High Beta" investing are those stocks which exhibit consistency of out performance in times of a rising index. I can then focus my decision making on broader movements of the index while enjoying reduced timing risk on my selections.
In the above example, GOL does not follow the consistency principle as well as CAT and is therefore, a more difficult proposition for successful timing. In effect, the high volatility of GOL is a hindrance, rather than a benefit and not a stock to include in my "High Beta" strategy.
Sorting through thousands of stocks to find a group which have the "right" characteristics is a daunting task. My basic methodology rests on analysis of risk/reward ratios. I have found that adding this step in the analysis makes for easier statistical classification and analysis.
The following graph, shows the risk/reward ratio for Caterpillar versus SPDR S&P 500 Index (SPY) as a typical example. Ideally, I would aim to find stocks which follow the "High Beta" line as closely as possible. The increasing slope of this line increases the likelihood of significant out performance on upward moves. The better the fit of a stock to the "High Beta" line, the greater the reliance on index moves to generate alpha (rather than stock specific information). That is our aim. Computationally, it is a relatively easy task to accomplish.
Two years ago, I would not have considered the concept of market timing in any of my strategies, and I do not claim any skill in the process. However, talk that "Buy and Hold" is dead should at least alert one to the possibility that ground rules may be changing. If it is accepted that "High Beta" implies out performance in rising markets and under performance in falling markets, then it also implies that additional attention to market timing is required in implementing this strategy.
I have made frequent reference in past articles to an "Indicator" based on the SPDR S&P 500 ETF and the iShares Barclays 20 Year Treasury Bond (TLT). This is not my work, but I have found it a very worthwhile piece of analysis in terms of market timing, especially since 2009.
It is a simple concept. When the indicator is positive, the market tends to a general upwards move. These periods are shaded in the following graphs. The rule is very clear. A long only investor seeking to limit risk should seek market exposure only when the "Indicator" is positive.
The results appear quite consistent for SPY. At the individual stock level, there is more variation. Nevertheless, in the above example, if Caterpillar was to be classified as "High Beta", then the only time to have been long the stock was in the periods covered by the shaded area. In this period, the proposed timing system would have yielded 3 profitable trades out of 4.
There are many articles in SA citing the need to limit emotion in investing, and to trade to a plan. I have tested "Indicator" with this in mind, and it has produced favourable results. My view is that applying this rule makes investing easier, but not easy. It is still a fact that predicting the future is impossible.
In an article posted on 22 June 2012, I selected 10 "High Beta" stocks based on my classification system. I refer to this to avoid suggestion of selection bias in evaluating any results from this analysis.
Business Support Services
Commerical Vehicles & Trucks
At the time, I was advocating a core investment strategy of low volatility stocks, with the above selection as insurance against unexpected upward moves in the index. How has this selection fared in the period since posting?
The "Indicator" turned positive on 13 August 2012, which is the trigger for implementation. In the period since, about 5 of the above stocks have done well enough while 5 have basically flat lined or declined. Ameriprise Fin (AMP) and Dow Chemicals (DOW) are examples of stocks from each group. Results are as shown in the graphs below.
The graphs show the historical context for their selection, which is consistency of returns when "Indicator" is positive. In the fourth period, AMP has followed historical precedent while DOW has not.
In terms of generating alpha, the results of the total portfolio shows around 4% of out performance on the SPY, and the best five around 10% of out performance.
I draw a few conclusions from these results.
Most importantly in my view, is confirmation of the disclaimer "past performance is not an indicator of future performance". A particular stock, such as DOW, can and does fall out of classification. In terms of generating alpha, I was expecting consistent alpha once the portfolio was implemented, but that has not proven to be the case. From time to time, it doesn't look like a successful strategy.
The danger here is that the strategy might be dumped because of near term under performance even though, over the 3 month term, out performance of between 5-10% over 3 months is acceptable.
In the same period, Power Shares S&P 500 High Beta ETF (SPHB) produced alpha of about 2%.
IS THERE PROFIT IN "HIGH BETA"?
So, is a "High Beta" strategy worth considering now?
Since 13 August 2012, the returns to "High Beta" have been muted compared to 2009. In general, out performance in absolute terms has been less and fewer stocks have participated. My sense is that only about 50% of stocks which could have been classified as "High Beta" would have retained that classification in the period.
I remain satisfied that the timing analysis works and needs to work. If one were to pick a starting point based on the "Indicator" being negative, then there is some evidence that "High Beta" outperforms the SPY. However if this strategy were to be initiated when "Indicator" is markedly positive, it might prove quite difficult for "High Beta" to out perform.
Holding "High Beta" stocks as "Indicator" turns negative is a recipe for underperformance. This may explain where some ETFs fail. There are a few options when this market situation occurs, defer to cash, or short the same stocks that were previously held long. As at 13 November, "Indicator" was negative.
TIPS FOR BUDDING HEDGIES?
Tips for budding hedge fund managers, and perhaps those reviewing ETFs employing this strategy:
- Appropriate classification of stocks is a given. The important part of the analysis is to consider which of these stocks are most likely to follow historical precedent.
- From visual inspection, "High Beta" investing is a game of "Snakes and Ladders". To avoid the "Snake" part of the equation, a timing strategy is essential.
- Timing is most likely an on/off switch. In the "off" mode, the strategy must include graded responses depending on circumstances. From conservative to aggressive the choices might be cash/index/short.
Disclaimer: The content in this document is provided as general information only and should not be taken as investment advice. The contents in this explanatory paper shall not be construed as a recommendation to buy or sell any security or financial product, or to participate in any particular trading or investment strategy. The ideas expressed on this site are solely the opinions of the author. The author may or may not have a position in any security referenced herein. Any action that you take as a result of information or analysis on this site is ultimately your responsibility.