I remember twenty years ago (long before these stats were tracked with near precision using quantitative techniques) it became abundantly clear to me that accurately predicting earnings (the occupation of all professional financial analysts) did not help one make extra money (‘alpha’ in industry parlance). This excerpt from A Maverick Investor’s Guidebook (Insomniac Press, 2011) is a true story:
“Once upon a time, I made a speech to an audience of young aspiring financial analysts, hoping to point them towards the maverick way. Using data from an ancient article published in the Financial Analysts Journal, I pointed out a disparity in the expectations of investment analysts. Companies that were expected to have outstanding earnings growth actually underperformed (by a very wide margin) those companies that were expected to continue to report disappointing earnings. You may want to reread that sentence.
The study simply determined that optimistic earnings forecasts by research analysts usually turned out to be wrong, and the stock prices went down after it happened. Alternatively, analysts were generally too pessimistic about some companies, and after they were proven wrong, the stocks went up and by a lot. Opportunities are created when expectations are either too pessimistic or too optimistic. When a company simply keeps on delivering what is expected of it, the news is generally benign (having no significant effect) in terms of its impact on the stock price.” A Maverick Investor’s Guidebook (Insomniac Press, 2011).
This realization helped the performance of portfolios I managed early in my career immensely. I’d deliberately seek out companies whose stocks were washed out and tried to identify trends that would change/improve earnings and future expectations.
Jump ahead twenty plus years and companies’ efforts to manage earnings, and incentives for analysts to be as close as humanly possible to getting it right mean it has become almost impossible to earn the excess return that used to be available immediately following the earnings announcement – even if it is a surprise. The scatter diagram illustrates that although there seems to be a positive correlation between announced earnings surprises and returns, the statistical fit is simply awful. In English – not worth a bet.
“There exists indeed a positive relationship meaning that a higher earnings surprise leads to higher excess return and vice versa. The problem is just that the R-squared is only 0.09. In other words, the earnings surprise only explains nine percent of the variation in the excess return. So even if you were the best earnings forecaster in the world you would not be able to make consistent trading profits.” Business Insider, July 22, 2013.
The only approach that might work is hoping to identify what I call earnings shocks! Of course this supports good old fundamental analysis rather than quantitative techniques. Predicting massive errors in expectations requires much experience, although data mining can help get a grip on measuring true (if incorrect) expectations or consensus thinking.
An example might help. Below is the chart of Canadian forest products company Canfor. Largely forgotten over the past several months and with lumber and paper products far from the focus lists of investors and investment bankers, look what happens when earnings are shockingly better than expected.
Trying to predict huge potential errors (that may result in shocks) is controversial. There is great personal risk since if you get it right there’s little glory but if wrong you are put under a microscope (remember Kerkorian and his Ford position?), even if there is very little in the way of actual financial risk. There’s also a bit of an embedded contradiction – similar to what modern day physicists – trying to model the universe – are dealing with. In order to be successful, you have to accept that the end result is based on any number of probabilities and unknown expected returns. This involves some assumptions and (like all good mathematicians) we hate assumptions.
BUT, if you have to make them then you make them. For instance, my own (still waiting for my Nobel Prize) model assumes:
- The probability of making money betting on a positive earnings surprise diminishes as the number of investors/analysts that are optimistic grows.
- The reward for betting on an a positive earnings shock increases the more pessimistic and disinterested investors/analysts become and vice versa.
Many would argue that it is impossible to predict anything. However, at the time I wrote A Maverick Investor’s Guidebook (summer of 2010) I predicted a huge market rally. Why? Simply because an optimistic earnings expectation was nary to be found. Analysts’ forecasts were widely dispersed as well. My model suggested the potential reward for betting on positive earnings ‘shocks’ must be large because investors weren’t at all interested in the stock market at the time. You can see (with a magnifying glass) that the S&P 500 earnings did “shock” for quite awhile following the financial crisis, until expectations began to be more aligned (like today) with reported earnings (fewer shocks).
- 2010 – Beware of too much pessimism
- 2013 – Beware of too much optimism
Where is the potential for excess returns over the next 6 months to a year? You should be able to figure it out based on this quote I found today.
Over the past month, analysts covering the materials sector have lowered their 2013 estimates by 4.6%, while those covering energy and tech have cut theirs by 1.7% and 1.5%, respectively. In the case of materials and tech, much of that has come in the past two weeks during the peak of earnings season, with materials estimates down 3.4% and tech estimates scaled back by 1.3%, according to data compiled by Morgan Stanley. On the other hand, estimates for the telecom and financial sectors are up. MarketWatch (WSJ) July 30, 2013.
Expectations for materials, energy and tech are eroding, increasing the probability that once the dust has settled there will be some profitable earnings shocks.
For a fun read, click on this pic to visit my latest blog “How to tell the difference between investing and gambling.”