Blog - Investing Notes

May 9, 2007 - Update: An Alternative (ISE) Put-Call Ratio Universe

In our blog entry of 5/8/07, we found that the Chicago Board Options Exchange (CBOE) daily total exchange put-call ratio is not a useful predictor of future stock market behavior. A reader proposed that the International Securities Exchange (ISE) put-call data is a more focused measure of investor sentiment. "The ISE Sentiment Index (ISEE) is designed to show how investors view stock prices. The ISEE only measures opening long customer transactions on ISE. Transactions made by market makers and firms are not included in ISEE because they are not considered representative of market sentiment due to the often specialized nature of those transactions. Customer transactions, meanwhile, are often thought to best represent market sentiment because customers, which include individual investors, often buy call and put options to express their sentiment toward a particular stock." Is this alternative put-call data a useful indicator of future stock market behavior? Using ISEE historical data and contemporaneous daily S&P 500 index closing levels from 10/1/02 (the earliest ISEE data available) through 5/4/07, we find that:

We first convert ISEE to put/call (it is call/put) for direct comparability with CBOE data.

The following chart shows the daily behavior of ISE put/call and the S&P 500 index over the entire sample period. Visual inspection suggests that there may be an inverse relationship between the two series. A relative low (high) for put/call corresponds to a relative high (low) for the S&P 500 index. However, there are no levels of put/call that are obvious short-term buy or sell signals for stocks.

To take a closer look, we examine the relationship between put/call and changes in the S&P 500 index (returns).

The following scatter plot depicts the relationship between the five-trading day (one-week) future return for the S&P 500 index and the ISE put/call. The Pearson correlation for the two series is a weak 0.08, and the R-squared statistic is close to zero. Put/call offers hardly any information for trading with a one-week horizon.

Might there be specific ranges of put/call that are predictive for a one-week trade, even though the entire series is not?

The next chart is a modified version of the prior scatter plot that orders the values of put/call to facilitate analysis of specific ranges. It also eliminates S&P 500 index return series overlap by selecting just every fifth data point. Each return value is therefore independent of other return values. Sample size is 231. While the return trendline rises very slightly with put/call, there are no ranges of put/call that reliably predict returns.

Might put/call be predictive for a one-month, rather than one-week, trading horizon?

The next scatter plot depicts the relationship between the 21-trading day (one-month) future return for the S&P 500 index and the ISE put/call. The Pearson correlation for the two series is a somewhat larger 0.16, but the R-squared statistic is unimpressive. Put/call thus offers hardly any information for trading with a one-month horizon.

Might there be specific ranges of put/call that are predictive for a one-month trade, even though the entire series is not?

The next chart is a modified version of the prior scatter plot that orders the values of put/call to facilitate analysis of specific ranges. It also eliminates S&P 500 index return series overlap by selecting just every 21st data point. Each return value is therefore independent of other return values. Sample size is only 54. While the return trendline rises modestly with put/call, variability is high and sample size is small. Trading on ISE put/call with a one-month horizon is therefore very risky.

Might put/call be predictive for a three-month trading horizon?

The next scatter plot depicts the relationship between the 63-trading day (three-month) future return for the S&P 500 index and the ISE put/call. The Pearson correlation for the two series is a more interesting 0.26, and the R-squared statistic suggests that put/call explains about 7% of overall stock market returns over the next three months.

Might there be specific ranges of put/call that are especially predictive for a three-month trade?

The final chart is a modified version of the prior scatter plot that orders the values of put/call to facilitate analysis of specific ranges. It also eliminates S&P 500 index return series overlap by selecting just every 63rd data point. Each return value is therefore independent of other return values. Sample size is just 18. While the return trendline rises with put/call such that variability does not swamp trend, the very small sample size is problematic for reliability of results.

The correlations between ISE put/call and various past S&P 500 index return intervals are roughly comparable in magnitude (but negative) to those stated above between put/call and future returns. This result suggests that investors/traders are about equally reactive and predictive with their options trades. In other words, it is about as likely that stock returns lead put/call as vice versa. This result also suggests that trading on the put/call ratio may be roughly equivalent to betting on reversion of past returns to the mean.

Since the mean ISE put/call may vary over the long run, we also tested put/call relative to its three-month moving average as a stock market indicator. However, correlations with future stock returns for this relative put/call are approximately zero at horizons of 5, 21 and 63 trading days. Correlations with short-term past stock returns are somewhat more negative for relative put/call than raw put/call. More succinctly, relative put/call appears to be more reactive than raw put/call and not predictive at all.

In summary, the ISE put-call ratio may offer a slight edge for intermediate-term trading of the overall stock market, but sample size for the critical back test is very small. Evidence suggests that it is a coincident indicator of stock returns.

For related research, see Blog Synthesis: Sentimental Journey, encompassing a broad range of equity market sentiment measures.