Working Paper: NBER ID: w30396
Authors: Emmanuel Saez; Gabriel Zucman
Abstract: Recent estimates of US top wealth shares obtained by capitalizing income tax returns (Saez and Zucman, 2020; Smith, Zidar and Zwick, 2022) are close in both levels and trends except for the top 0.01% where a large discrepancy remains. We examine this difference and, using public data, quantify three main issues in Smith et al. (2022). First, Securities and Exchange Commission data at the shareholder firm level show that billionaires' equity wealth is underestimated by a factor of 2.1. Second, interest-bearing assets at the top are under-estimated by a factor of 1.6, because of an extrapolation from a small and unrepresentative sample of investment funds. We quantify this issue using mandatory filings of US hedge funds. Third, issues involving tiered partnerships and the measurement of business profits suggest that large S-corporations are undervalued by a factor of 1.2 and top-owned partnerships by up to 2.2. After incorporating these results, the top 0.01% wealth share of Smith et al. (2022) is close to the one found in Saez and Zucman (2020) and estimates of US wealth inequality are reconciled.
Keywords: No keywords provided
JEL Codes: D31; H25
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| underestimation of billionaires' equity wealth (G51) | misrepresentation of actual wealth (E21) |
| methodology of Smith et al. (2022) (C90) | underestimation of billionaires' equity wealth (G51) |
| non-representative sample (C83) | underestimation of interest-bearing assets (E43) |
| assumption of high interest rate for top 0.01% (E43) | underestimation of interest-bearing assets (E43) |
| valuation of passthrough businesses (H25) | undervaluation of large S-corporations and partnerships (G32) |
| tiered ownership structures (G32) | undervaluation of passthrough businesses (H25) |