Household Portfolio Rebalancing and Equity Market Fluctuations

Featured in print Digest
W32001 This figure is a two-panel line graph titled Monthly US Equity Flows and Returns during COVID-19. The y-axis on both panels is labeled, net purchases a percentage of assets in US equities. It ranges from negative 0.4% to 0.4%, increasing in increments of 0.2.  The x-axis on both panels is time, ranging from October 2019 to July 2020. Both panels have line graphs that show two lines: Flow to US equities and Scaled equity return. The left panel is titled "Investors with total assets less than 3 million dollars." In this graph, the "Flow to US equities" line begins at 0, spikes to 0.4%, drops to -0.3% in March 2020, and then rebounds to 0.4% in June 2020. The "Scaled equity return" line starts at 0.1%, dips to -0.3% in March 2020, spikes to 0.4% in April 2020, and then declines to around 0.2% by June 2020. The right panel is titled "Investors with total assets exceeding 100 million dollars." In this graph, the "Flow to US equities" and "Scaled equity return" lines are nearly inverted. They hover around 0 until January 2020, when they diverge. The "Scaled equity return" line drops to -0.2% in March 2020, while the "Flow to US equities" line spikes to 0.2%. In April 2020, they invert, with the "Scaled equity return" line spiking to 0.2% and the "Flow to US equities" line dropping to -0.2%. Both lines then converge towards 0. The note of the figure reads, “Scaled equity return” shows the aggregate return on the US stock market, rescaled to math the time-series mean and volatility of flows. The source line reads, Source: Researchersʼ calculations using data from Addepar.

How investors adjust their portfolios in response to movements in asset prices and other shocks is a key input to asset pricing models, yet data limitations mean there is relatively little evidence on these behavioral responses, particularly for high-net-worth households. In Asset Demand of US Households (NBER Working Paper 32001), Xavier Gabaix, Ralph S. J. Koijen, Federico Mainardi, Sangmin Oh, and Motohiro Yogo leverage a new dataset on households’ portfolios to explore portfolio rebalancing behavior.

The data were provided by Addepar, a wealth management platform that provides investment advisors with data analytic tools. The researchers analyze data from January 2016 to March 2023. The security-level dataset includes information on a broad set of households across the wealth distribution, including 990 ultra-high-net-worth (UHNW) households, defined as those with portfolios worth more than $100 million, and it covers a broad set of asset classes at a very high frequency. Addepar collects information on indirect holdings (e.g., through mutual funds, hedge funds, and exchange-traded funds) as well as direct holdings, and the data cover public and private asset classes. The Addepar data provide a larger sample of high-net-worth households than US household surveys, such as the Survey of Consumer Finances, and can therefore be used to study their behavior.

Data from the last decade show ultra-high-net-worth households bought equities during market declines while high-net-worth households sold.

The researchers present summary information on portfolio composition and the way it varies with investor net worth. The fraction of the household portfolio held in cash varied relatively little across the wealth distribution, while the portfolio share of municipal bonds increased with portfolio size and the fraction invested in corporate bonds and global equities declined.

For most investors in the dataset, portfolio flows to risky assets are positively correlated with returns, indicating procyclical investment decisions. For the typical investor in the dataset, a 10 percent rise in equity values was associated with a shift of about 0.1 percent of household portfolios toward equities. Disaggregating the investor population, however, revealed important heterogeneity in rebalancing behavior across the wealth distribution: UHNW households are countercyclical. Given the larger value of their portfolio holdings, when rebalancing patterns are weighted by portfolio value, the net effect of a rise in equity returns is an outflow from equity, albeit a small one. The researchers conclude that it is “unlikely to be an important stabilizing force.”

The researchers attempt to explain the factors that drive changes in household asset allocation patterns. They find that three factors that provide exposure to the long-term equity risk premium (defined as the difference between expected equity returns and long-term Treasury yields), the credit premium (the yield spread between corporate bonds and Treasuries), and the municipal bond premium (the gap between the yields on taxable fixed income securities and municipal bonds) explain approximately 81 percent of the variation in rebalancing across the 13 asset classes they consider.

— Linda Gorman

The researchers acknowledge the Ferrante Fund, the Center for Research in Security Prices at the University of Chicago and the Fama Research Fund at the University of Chicago Booth School of Business for financial support.