Farrell Sentiment Index

Elmwood Data's picture

 

 

In this chart and analysis, we compare the S&P 500 (black line) to the American Association of Individual Investors (AAII) weekly survey. The AAII reports their respondents in three categories: bullish, bearish, and neutral. To analyze these AAII statistics, we create the “Farrell Sentiment Index,” defined as the number of bulls, divided by the number of bears, plus .5x the number of neutrals = Bulls/(Bears+.5Neutrals). This data can be quite volatile week to week, so we opted to convert this formulaic data series into two different types of charts.  First is using a slow MACD (24,52,18) to help analyze the accelerations and decelerations in sentiment.    In the second we created a chart converting the Farrell Index into 30-day (blue line) and 90-day (red line) simple moving averages.  

The first chart shows the Farrell Index as a MACD data series.   You can see that when the MACD spikes up to +.5, as seen on the left axis, this excessive bullishness has marked the point of a market correction, though not always immediate.   February 1, 2012 marked the last time the MACD spiked up to this level.  We also ran a scenario analysis and every time the MACD moved 2 standard deviations above trend (Feb 1st), the subsequent returns for the S&P were -.04% 1 month later, -1.66% 3 months later, -13.49% 6 months later, and -18.55% 12 months later.   Conversely, there also appears to a decent buy signal -.4 when bearishness gets too extreme, but at present we are a ways away from that level.

 

 

The next chart shows the same comparison of the S&P versus the Farrell Index, but here we converted the Farrell into a 30 day simple moving average (SMA) and a 90 day SMA.  This gives another good look of how over/under extended near term (30 days)sentiment has risen/fallen relative to a longer period (90 days).   The further the short term SMA, or blue line, extends from the longer term SMA, or red line, serves as a warning sign.   Right now the short SMA is 32% higher than the long SMA, which by historical standards is quite large and should serve as a caution signal.   We also ran a scenario analysis on this differential, so that we tested for when the short SMA is 2 standard deviations above the longer SMA.  When this event transpires, (as it currently is at this level) the subsequent returns for the S&P were -1.55% 1 month later, -.99% 3 months later, -11.55% 6 months later, and -14.78% 12 months later. 

 

Sentiment clearly is volatile over time, but it can also act as a leading indicator to the market. Large spikes up and down often signal stock market reversals.  At the same time, sustained trends such as when the Farrell 30-day SMA crosses above/below the 90-day SMA,  can act as buy and sell signals as well.   These recent readings, however,  suggest that a correction is not far off.