The Evolution of Leakage and Retirement Asset Flows in the U.S.

Changes in Retirement Savings during the COVID Pandemic

“…We find little change in contributions in 2020, which increased from the previous year at a rate in line with recent trends, a stark difference from the drop in contributions that occurred during the Great Recession (Argento, Sablehaus, and Bryant, 2015;Goodman, Mackie, Mortenson, and Schramm, 2021). This may have occurred because the effects of the COVID-19 pandemic were disproportionately worse for workers at the bottom of the income distribution, who save at much lower rates than middle-and high-income earners, while the earnings shocks during the Great Recession affected middle-and high-income workers to a greater degree than during COVID-19 (Larrimore, Mortenson, and Splinter, 2022).…”

“…Both papers also find that early withdrawals increased during the Great Recession, although to a smaller extent than the change in contributions. Goodman, et al (2021) find that the share of contributions made by working age Americans that exited the retirement saving system increased modestly during the Great Recession, from a base rate of around 22 percent to 26 percent in 2009.…”

“…3 While the panel is unbalanced, it is representative of the population of taxpayers within each year from 2003 to 2020, and individuals can enter the panel by immigration, birth, or receiving a tax form, and they can exit the panel by emigration, death, or failing to receive a tax form. These base data have been used in several prior papers (Mortenson, Schramm, and Whitten, 2019;Goodman, et al, 2021;Larrimore, Mortenson, and Splinter, 2022).…”

“…However, within the category of workplace pension plans, we are not able to distinguish between distributions made from defined contribution (DC) plans and those made from defined benefit (DB) plans. In a prior paper Goodman, et al (2021) applied an algorithm to classify 1099-R distributions as DB or DC based on various factors. However, this algorithm is unable to classify data from 2020 because it requires several years of "burn-out" (i.e., the last several years of data are used to classify earlier distributions but cannot be classified themselves).…”

“…Several studies have found that the likelihood of taking early withdrawals increases significantly with negative shocks including divorce and job loss or wage reduction (Argento, Sablehaus, and Bryant, 2015;Amromin and Smith, 2003;Goodman, et al, 2021;Brady, 2019), the latter of which applies directly to the COVID-19 employment shocks. In relation to the Great Recession, Argento, Sablehaus, and Bryant (2015) also find that contributions to defined contribution (DC) plans declined significantly, and Goodman, et al (2021) find that net contributions did not recover their inflationadjusted 2007 levels until 2014. Both papers also find that early withdrawals increased during the Great Recession, although to a smaller extent than the change in contributions.…”