Election night 2016: Red balloons rose at Trump Tower, tears fell in Chappaqua. The next day, and in the following weeks, extreme emotions overflowed into the financial markets as exuberant Republicans bought equities and dismayed Democrats shifted into bonds.
The cumulative effect was numerically slight—an estimated market-wide $3.3 billion decrease in demand for equities in the first six months post-election—but for the four MIT-Sloan School researchers who measured and analyzed it, the detection of politically-driven trading cast fresh doubt on conventional wisdom about investor behavior.
To those researchers, authors of “Belief Disagreement and Portfolio Choice” (NBER Working Paper 25108), Democratic investors didn’t seem to behave rationally. In theory, they too could see that Donald Trump’s tax and regulatory policies might lift the stock market, and would invest accordingly.
But they didn’t, even though it would have been in their economic best interest. In the year after the election, the Dow Jones Industrial Average grew 28.5%, from an already record-breaking base.
The study was based on trading activity by individual retirement account holders at an unnamed major financial institution. Their average household income was $101,600 and the median was $78,000. The average investor in the study had $156,500 in investable wealth, of which 81% was in retirement accounts. The researchers did not include the wealthiest or poorest 10% of the population.
“Our main finding is that (likely) Republicans increase the exposure of their investments to the US stock market relative to (likely) Democrats following the election. Democrats increase their relative holdings of bonds and cash-like securities. This result is not driven by differences in returns but by active trading over a six-month horizon following the election, and the relative change in equity shares is more than twice as large among previously active investors,” the paper said.
Maarten Meeuwis, Jonathan A. Parker, Antoinette Schoar and Duncan I. Simester, all of the MIT Sloan School of Management, co-authored the paper.
In an email to RIJ, Schoar wrote, “We provide evidence that people with different political affiliations take different portfolio actions that are not driven either by the different effects of the election on their own economic circumstances, such as cities and towns or jobs or tax burdens, but instead actively trade assets in response to the changes in their relative views about the national economy.”
Ironically, “Republicans report becoming relatively more optimistic about the national economy after the election, but they do not report becoming relatively more optimistic about their own economic situations,” the researchers wrote.
In most theoretical models of financial markets, participants typically agree about the probability of different states of the world, the researchers wrote. But, in practice, they showed, different people have different “models of the world,” and make investment decisions on that basis.
The researchers were somewhat surprised, and even troubled, by the finding “that people save and invest and react differently to public events not because of their own situations (like their age, their city, their job prospects), but because they fundamentally take different views from others,” Schoar told RIJ.
“This finding is worrisome, in that people should take into account that they may be wrong and so learn from the behavior of others and from objective information, which is the foundation for most asset valuation models and asset management advice.”
Another of the paper’s authors, Jonathan Parker, told RIJ, “Most investment advisors give the advice: Keep your politics out of your portfolio. The downside of disagreement is that people are not pooling risks and diversifying as much as they could.” We vote in different directions, he noted, but “with stocks, we can all bet in the same direction. We don’t have to bet against each other at all.”
The findings were based on the review of data provided by a large unidentified investment institution, which provided the researchers with “anonymized” information about individual client trading behavior before and after the election. One of the authors was a paid consultant to the financial institution.
Likely political affiliation was determined using publicly available data on individual campaign donations during the 2015-2016 election cycle, aggregated to the zip code level. The researchers restricted attention to contributions from individuals to political action committees with at least $20 million in donations and associated with the two main parties or with their presidential nominees.
The wealthiest and poorest 10% were excluded, and only one retirement investor (RI) per household was included. The survey included heads of households between the ages of 25 and 85 in retirement saving accounts (excluding defined benefit plans and Social Security).
The sample captured 40% of the US population and 47.3% of retirement wealth. According to the study, the richest 10% of Americans own about 50% of U.S. retirement wealth, which the Investment Company Institute estimates at $16.9 trillion in IRAs and defined contribution plans ($28.2 trillion if public and private defined benefit plans and annuities are included).
The $3.3 billion shift in overall portfolio allocations was small, the researchers wrote, because they observed differences at the zip code level, which would include both Republicans and Democrats, and because there is very little active trading in retirement accounts. The changes in asset allocation were larger among households that actively re-allocate their wealth.
© 2018 RIJ Publishing LLC. All rights reserved.
