Working Paper: NBER ID: w29559
Authors: Victor Duarte; Julia Fonseca; Aaron S. Goodman; Jonathan A. Parker
Abstract: We develop a machine-learning solution algorithm to solve for optimal portfolio choice in a lifecycle model that includes many features of reality modelled only separately in previous work. We use the quantitative model to evaluate the consumption-equivalent welfare losses from using simple rules for portfolio allocation across stocks, bonds, and liquid accounts instead of the optimal portfolio choices, both for optimizing households and for households that undersave. We find that the consumption-equivalent losses from using an age-dependent rule as embedded in current target-date/lifecycle funds (TDFs) are substantial, around 2 to 3 percent of consumption, despite the fact that TDF rules mimic average optimal behavior by age closely until shortly before retirement. Optimal equity shares have substantial heterogeneity, particularly by wealth level, state of the business cycle, and dividend-price ratio, implying substantial gains to further customization of advice or TDFs in these dimensions.
Keywords: Portfolio Choice; Lifecycle Models; Machine Learning; Welfare Losses
JEL Codes: C61; D15; E21; G11; G51
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| type of portfolio allocation rule used (G11) | resulting welfare outcomes (I38) |
| wealth levels (D31) | optimal equity shares (G12) |
| state of the business cycle (E32) | optimal equity shares (G12) |
| dividend-price ratio (G35) | optimal equity shares (G12) |
| age (J14) | divergence between TDFs and optimal behavior (G40) |
| typical TDF portfolio (G11) | consumption loss (E21) |