Any Investor Response to Presidential Polling Data?

Do presidential election polling data have any effect on stock returns? In other words, do investors act on the survey-indicated election prospects for the two major party candidates? Presumably, investors would tend to enter (exit) stocks if they thought the candidate with policies more (less) favorable for equity valuation were gaining ground. Using Gallup Daily polling data and contemporaneous daily closing levels of the S&P 500 index from June 6 (when the major party nominations solidified) through October 14 (91 trading days), we find that:

The following chart compares the daily Gallup polling data gap in election prospects between Senator Obama and Senator McCain to the level of the S&P 500 index over the entire sample period. Each day’s gap is the result of surveys of registered voters conducted over the prior three days. Visual inspection indicates no systematic relationship between polling data and stock market behavior, although the recent market plunge coincides with the buildup of a solid lead for Senator Obama.

For greater precision, we compare daily changes in polling data and daily stock market returns.

The following scatter plot relates the same-day S&P 500 index return to the daily change in the Obama-McCain polling gap. A positive (negative) change in the gap means that Senator Obama (Senator McCain) is gaining ground. The Pearson correlation between the two series is -0.19, hinting that investors tend to sell (buy) when Senator Obama (Senator McCain) is gaining ground. However, the correlation is very weak, and the R-squared statistic is just 0.03, indicating that variation in the daily polling gap explains only 3% of daily stock market returns. Given the sample size, evidence this weak does not support a belief that investors act on daily presidential election polling data.

In case there is a delay in investor reaction, we also test the relationship between next-day S&P index returns and the daily change in the polling gap. The Pearson correlation for this delayed-response scenario is 0.00.

We also test the weekly relationship between the two series, with results (for a very small sample) weaker than those for daily data.

As a further test for any relationship, we rank the daily polling gap changes and calculate the average same-day S&P 500 index returns for:

• The 39 trading days on which Senator Obama gains ground.
• The 19 trading days on which there is no change in the gap.
• The 32 trading days on which Senator McCain gains ground.

The following table summarizes the results, which suggest some investor aversion to Senator Obama. However, when we remove the two most extreme outliers from the sample, the average same-day S&P 500 index return for “Obama gaining” is -.05% and that for “No Change” is -0.8%, indicating subsample sizes too small for reliable inference and little support for a belief that daily presidential election polling data influence daily stock returns.

In summary, limited evidence indicates that daily presidential election polling data have very little or no effect on returns for the broad stock market.

You May Also Enjoy...

• Google Search Data as a Measure of Investor Attention
• Monthly Returns During Presidential Election Years
• UBS/Gallup Measurement of American Investor Optimism
• Presidential Politics and Industry Returns
• Technical Analysis Tested on Long-run DJIA Data