Saturday, April 1, 2017

Learning from Investment History

      Perhaps the most important lesson I have learned from studying investment history is not to trust what I learn from investment history.  Some recent research I just completed with Rob Arnott and Vitali Kalesnik of Research Affiliates underscores this lesson.  The work relates to the relation between the political party in power and average stock market returns.

      In a provocative paper, Santa-Clara and Valkanov (2003) presented evidence that stock market returns are much higher under Democratic presidents than Republican presidents.  Their work was updated by Pastor and Veronesi (2017), who find that the effect is even stronger when the data are extended through the end of 2015.  Given the strength of the results, Pastor and Veronesi go on to develop a model based on time varying risk aversion to explain the pattern.  They hypothesize that when risk aversion is high, voters are more likely to elect a Democratic (left leaning) president; when risk aversion is low, they elect a Republican (right leaning) president.  Because, risk aversion is higher under Democrats, this results in a higher equity risk premium, and, therefore, higher average returns.

     But there is reason to be suspicious.  Two key events are responsible for much of the differential return under Democratic and Republican presidents.  Specifically, the fact that Republicans were president during the two great crashes beginning in 1929 and 2008 and, not surprisingly, Democrats were president during the subsequent recoveries explains a majority of the differential.  Had the order been reversed, the effect would be largely eliminated.  This suggests the finding may be serendipity.

      To see if the results for the US were chance events, we turned to international data.  Our study covered Australia, Canada, France, Germany and the UK because we had good data on the political parties and stock market returns for those countries.  We examined whether the US pattern - high returns when the "liberal" party was in power relative to the "conservative" party - held in the international context.  For the five countries as a group, we found no relation between the the party in power and stock market returns.  From our perspective, history once again threw investors a curve ball.  What appeared to be a significant relationship between between the market and politics based on US data turned out to be a mirage.

No comments: