An Edge in Forecasting Volatility (HOG and NTAP)

"Volatility is a symptom that people have no idea of the underlying value."

- Jeremy Grantham

Volatility is making a come-back. For 2013, and the first 3 quarters of 2014, volatility was so muted it reached levels not seen since the pre-crash levels in 2006 and 2007. Then we had the October market pull back where the VIX reached a two year high of 22 and it has been in an upward trend since. Where is it headed?

Just about everyone is expecting the first rate hike from the Fed in 9 years to come within the next couple of quarters, and volatility in other asset classes (see Oil and Currencies) is having a significant impact on U.S. Equities. Protecting against volatility, forecasting volatility, and for many, trying to take advantage of increasing volatility are now important topics once again for Equity investors.

When looking at single stock equities the most common inputs for forecasting volatility are calculations for historical volatility and implied volatility. But the problem is, while historical volatility is indicative of future volatility, it can also differ greatly from future volatility. The main reason is that the drivers from recent history are unlikely to be the drivers of changes in price going forward. The forecast is incorporating information that is limited.

Historical volatility typically takes into account the last few years or maybe five years of volatility, but in the current environment we’re in (low volatility six year bull market) is that the best gauge for forecasting volatility in stocks for the coming month, quarter, year? The other thing lacking is discriminating information about direction. Historical Volatility and Implied Volatility are expressed as symmetrical cones, but as we all know the markets are anything but symmetrical. In short, historical and implied vol are solid guides but incomplete.

Let’s throw down the gauntlet here. EidoSearch can help! At the center of what we do is quantifying the value residing in historical return distributions of like patterns and conditions to today in order to generate probabilities that can give you a competitive edge in forecasting returns and volatility.

In plain English, by taking a stock’s current trading behavior or price trend, we find similar instances through decades of history and all different market conditions within a peer group. By capturing similar instances historically, we collate the actual historical return distributions of the stock’s price trend to formulate projections on the range of probable forward returns. We’ve done 5 million predictions going back to 2006, from 1 week to 6 months forward in length, validating the accuracy of our return predictions.

Here are a couple of examples of forecasted return probabilities, for Harley Davidson (HOG) and NetApp (NTAP), that have a very different magnitude and directional skew than implied volatility suggests for the next 1 month.

One month forward projection for HOG, using a 3 month price trend and finding 86 similar instances of this pattern in history in their peer group dating back to 1979. Implied volatility, calculated for 1 month forward, is the overlay in ORANGE.

One month forward projection for NTAP, using a 3 month price trend and finding 99 similar instances of this pattern in history in their peer group dating back to 1980. Implied volatility, calculated for 1 month forward, is the overlay in ORANGE.

Have a great week.