The Retirement Consumption Gap: Evidence From The HRS
Texas Tech University
Texas Tech University
Texas Tech University
August 11, 2014
Research on post-retirement asset decumulation shows that many retired households are consuming from their financial assets at a very slow rate or not at all. Factors commonly associated with this phenomenon are the complexity of decumulation decisions, uncertain longevity and medical costs, home production, and bequest motives. Using a product allocation framework and data from the Health and Retirement Study we provide evidence of a retirement consumption gap for those with financial assets in the third, fourth, and fifth quintiles. In comparison to previous studies that focus on changes in financial asset values over time to examine post-retirement consumption, our methodology provides a more dynamic evaluation of consumption behavior and quantifies the consumption available to and forgone by retirees under a variety of asset allocation strategies. When incorporating the effects of uncertain longevity, uncertain medical costs, and bequests we find that retirees with median wealth have a consumption gap of approximately 8% on average and that retirees with higher levels of wealth have a consumption gap as high as 45.6%. We discuss the implications of these findings and considerations for future research addressing this topic.
The Retirement Consumption Gap: Evidence From The HRS – Introduction
Older households are increasingly responsible for managing their own decumulation strategies. The shift from defined benefit to defined contribution plans in the United States transfers the burden of directing retirement plan drawdown decisions from professionals to individual investors. The ability to manage these decisions effectively can have an important impact on retirement wellbeing.
The life-cycle hypothesis (LC H) predicts that in the absence of labor income retirees will begin to decumulate their financial assets in an attempt to smooth the marginal utility from consumption over time (Ando & Modigliani. I963). The theoretical implication presented by the LC H is that consumption from financial assets in retirement is a requirement for maximizing lifetime utility from wealth. Given such, in the absence of a bequest motive we expect that the value of retirees‘ financial assets will diminish over time until they are consumed completely by the end of one’s life.
Much of the evidence found in the current literature is contradictory to what we expect under the LCH. Retirees generally decumulate from their savings at a very conservative rate. Rather than observing systematic decreases in financial assets during retirement. many studies find that the value of retirees financial assets is holding steady or increasing over time (Rix. 2000; Love. Palumbo. & Smith. 2008; De Nardi. French, & Jones, 2009a,b; Smith. Soto, Penner, 2009; Poterba, Venti. & Wise. 2ol la‘b; Browning. Huston. & Finke. 2013). This effect is present even under the required minimum distribution (RMD) rules associated with defined contribution and IRA accounts (Poterba. Venti. & Wise, 2011a,b), evidence that retirees are taking required distributions and reinvesting them in other financial assets (Smith & Love. 2007). These Findings that uncertainties have a material effect on decumulation decisions call into question the ability of retirees to estimate the likelihood of uncertainties and incorporate them into a complex decision framework. Browning. Huston, and Finke (2013) point to the inability to effectively incorporate uncertain life-cycle factors when making decumulation decisions as a cause for slowed decumulation among retirees who are less cognitively able. Their findings show that those with lower levels of cognitive ability are more likely to overweigh factors of uncertainty. such as longevity and medical costs. causing a drop in the rate of consumption from financial assets.
Others pose home production (Hurd & Rohwedder, 2003) and bequest motives as justifications for the lack of wealth decumulation in retirement. While home production decreases the need to spend retirement assets, assuming that an anticipated increase in home production accounts for the majority of decreased post-retirement consumption is inconsistent with trends in pre-retirement asset accumulation (Bernheim, Skinnner. and Weinberg, 200] Studies using survey data to model bequest motives as an explanatory predictor of asset decumulation find little evidence that bequest motives are strong or relevant to dissaving decisions. which is consistent with the LCH (Hurd. 1987; Hard. 2003; De Nardi. French, Jones, 2009b). These studies show that the marginal utility from consumption is much higher than the marginal utility from bequests and attribute the existence of bequests among survey participants to uncertain longevity and medical costs. Our results suggest that even after setting aside a portion of financial assets to address these concerns. consumption in retirement could be significantly higher.
See full PDF below.