Five Facts About Beliefs and Portfolios

Stephen Utkus is Visiting Scholar at Wharton Business School and Fellow at the Center for Financial Markets and Policy at Georgetown University. This post is based on a recent paper, forthcoming in the American Economic Review, by Mr. Utkus; Stefano Giglio, Professor of Finance at Yale School of Management; Johannes Stroebel, David S. Loeb Professor of Finance at NYU Stern School of Business; and Matteo Maggiori, Associate Professor of Finance at Stanford University Graduate School of Business.

Why do investors allocate their portfolios as they do? What causes them to change their minds and trade in their portfolios? These are critical questions in the field of macro finance. And central to answering them is understanding the role of expectations: what investors believe about future market returns and risks, how their beliefs vary with time, and how those beliefs come to influence real-world portfolio choices.

Among financial economists, there has been a growing interest in measuring investor expectations in a systematic way using surveys. However, surveys have been hindered by too qualitative a set of measures or by the lack of data showing how surveyed investors actually act on their beliefs.

In order to tackle such concerns, we have created a new research initiative on investor beliefs that combines administrative and survey data for a panel of individual investors at Vanguard. Vanguard is a well-known global asset manager; it is also large provider of investment services to U.S. retail investors. The survey, conducted every other month since February 2017, asks investors about short- and long-term expected stock market returns. It also elicits the subjective distribution of expected returns, allowing us to examine beliefs about low-probability events like a market crash. We also survey expectations on economic growth. These survey results are paired with portfolio allocation, trading, demographic and digital attention data.

As an example, Figure 1 presents one-year expected U.S. stock market returns, spanning from February 2017 to present. As of the most recent wave in February 2021, the average expected one-year U.S. stock market return was 5.3 percent

From our new dataset, we are able to draw a number of empirical insights about the role of expectations in portfolio decision-making. Our first result is that, for the average investor, beliefs are tied to how she allocates her portfolios, but quantitatively the sensitivity is low. Investor A may expect one-year stock market return of 5% and Investor B, an otherwise identical investor, 10%. But despite this markedly different outlook, these investors’ allocation to equities will only differ by just over five percentage points. We link this low sensitivity to a number of frictions—such as capital gains taxes and limited investor attention—that may slow the transmission of beliefs to portfolios.

Similarly, changes in beliefs have minimal effects on investors’ willingness to trade in their portfolios, at least on average. However, once an investor decides to trade, beliefs are important for both the direction and magnitude of trading.

Another surprising finding is the persistence of individual pessimism or optimism in beliefs. The pessimists tend to remain pessimistic; the optimists, optimistic. This data-driven finding runs counter to the popular narrative that market fluctuations are driven by relatively homogenous investors whose beliefs vary widely over time.

Figure 2 below captures another interesting statistic in our survey: the probability of a stock market crash. The mean value from the most recent wave is 5.4%. But it has varied notably over the period, especially during and immediately following the Covid mini-crash of 1Q 2020, when it jumped to an all-time high.

These highlights capture only a small flavor of the findings in our two papers on the role of investor beliefs, and we believe that there is much promise in future research in this area. From an academic perspective, we anticipate that results like ours will help build better macro-finance models of investor behavior, leading to a more nuanced understanding of what determines the aggregate level of prices and volume of trading in financial markets. The wide heterogeneity in beliefs we uncover, and the persistence of beliefs over time, suggests underexplored areas for future research.

From a practitioner perspective, we can foresee a day when eliciting investor expectations over various horizons—and calibrating them with data on other investors’ beliefs—will become a critical part of the investment advisory process. Whether it is within an individual household or with members of an institutional committee, calibrating beliefs with others could be a novel way to develop a shared understanding of portfolio outcomes and risks, thereby improving investor decision-making.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.