Several readers have expressed interest in how I came up with my earnings spike potential algorithm. Essentially, this is something that has evolved over the past three weeks as a result of my desire to find companies with a high probability of making a substantial near-term move and give my CNBC Million Dollar Portfolio Challenge portfolio a chance to make a run at the finals. To make a long story short, I got the volatility I wanted, but I didn’t always get the direction right.
I would not call this a battle-tested formula. It is more like a hypothesis that continues to evolve as I get more data and continue to test and tune some of the elements. Think of it as just-in-time sausage making.
I am not an arsonist (I even missed out on youthful pyromania), but I liken this task to understanding how to get a fire started and make sure it quickly builds in intensity and spreads as rapidly as possible. For a fire, you need a starter and an accelerant; for an earnings spike, it’s essentially the same thing.
In the links below, wherever possible I have provided a favorite deep link to a free public source that includes the relevant data, calculation, graphic, etc.
Some of the more important factors I look at are:
- Implied volatility, a great initial screening tool (higher is better) – free data at iVolatility.com; if you have an account at optionsXpress, they have excellent options screens available to all
- Beta (higher is better) is another good accelerant barometer, though not as good as IV – available many places, including Google Finance
- Number of analysts (lower is better) and degree of analyst consensus (lower is usually better) – Marketwatch.com has a page that not only summarizes the analyst estimates, but also provides a “coefficient variance” number that gives you a sense of the dispersion of opinion. In many cases, the earnings and/or revenue surprise is the fire starter.
- Short ratio: days to cover (higher is better) – a classic accelerant indicator, with free data available at ShortSqueeze.com
Some secondary factors to consider:
- Price to earnings ratio (negative or n/a is best, higher is better) – available many places, including Google Finance
- Earnings history data – there is a higher probability of a surprise if there is an erratic earnings history; there is also greater potential for a high magnitude surprise if there is a consistent pattern of beating (or missing) expectations – assuming the pattern can be broken. One fun source that has earnings dates baked in to charts and post-earnings performance data available is WhisperNumber.com (more complete data for larger companies.)
- Recent analyst ranking and/or price estimate changes (none is best) – these can work both ways, but most often they reduce the probability of a surprise. Again, Marketwatch.com is a good source.
- News flow – this is a highly subjective/qualitative assessment, but there are certain types of pre-earnings news that I believe can indicate in increased or decreased likelihood of an earnings surprise. Be particularly wary of binary events, such as the pending FDA approval for a drug and the like. The best place to find the relevant information is probably by looking at company news at Yahoo Finance. I am not ready to expand upon this one at this stage, except for…
- Recent company guidance (none is best) – as with recent changes in analyst opinion, these usually dampen the surprise potential. Again, try Yahoo Finance.
- Insider transactions – these are sometimes difficult to evaluate in the context of earnings, but if an apparent transactional pattern is confirmed or contradicted by earnings, there could be an accelerant. I favor Form4Oracle as a free public source of insider transaction data.
- Technical analysis – this is good for identifying the heightened possibility of breakouts, violation of important support and resistance levels, and other factors that may act as technical accelerants. The gallery view at StockCharts.com is always a good place to start.
- Put to call ratio (higher is better) – somewhat analogous to the short data is the individual stock open interest put to call ratio, data for which is available at SchaeffersResearch.com
- Recent options activity – another subjective and difficult to assess measure, but if significant changes in open interest favor either puts or calls, this may be a tell. Not much in the way of great public data, but you may get some valuable information from Yahoo Finance.
- Company size (lower is better) – this includes revenues and market capitalization. Available many places, including Google Finance.
- Recent IPO or lack of relevant operating history (less history is better) – In general, the shorter the track record, the bigger the chance for an earnings surprise. This means that the first quarterly report or two with new management, new products, a new acquisition, etc. increases uncertainty about the result – and the potential for a surprise.
As a footnote, if you are looking for a good source for who reports when that is sortable by time of day (BMO, AMC, etc.), I like TheStreet.com’s Earnings Release calendar, where you can click on the Date/Time column to sort accordingly. If you are not familiar with a lot of the tickers/companies, then I suggest that a first pass be limited to those companies with four letter tickers whose EPS estimate and/or previous year actual EPS is negative or close to zero.
Finally, I feel obliged to remind everyone that this is a method for finding the high potential post-earnings movers, *not* the winners. I continue to play with the weightings of the various factors and ultimately your weightings should reflect your research and beliefs about the market. If you keep track of the pre-earnings data and the outcomes, you should be able to develop and tweak your own model – or at least flag some potential high fliers.