Life-Cycle Consumption Patterns at Older Ages in the US and the UK: Can Medical Expenditures Explain the Difference?

Abstract

In this paper we document significantly steeper declines in nondurable expenditures in the UK compared to the US, in spite of income paths being similar. We explore several possible causes, including different employment paths, housing ownership and expenses, levels and paths of health status, number of household members, and out-of -pocket medical expenditures. Among all the potential explanations considered, we find that those to do with healthcare-differences in levels and age paths in medical expenses-can fully account for the steeper declines in nondurable consumption in the UK compared to the US. What can explain a difference of this magnitude? An obvious starting point is to examine age paths of income to access the extent to which consumption expenditures are tracking age paths in household income. But the second panel in Figure 1, which plots cohort averaged paths of household income at older ages in the two countries, demonstrates that, if anything, incomes decline at a slightly faster rate in the US than the UK. This therefore seems unlikely to be the major reason for a flatter spending profile in the US. In this paper we investigate other possible reasons that may explain the dramatically different patterns of nondurable consumption of older ages in the two countries by investigating differences in both inter and intra-temporal consumption for households around and beyond retirement age. The set of factors that we explore in this paper include: differential cohort effects in the two countries that may distort average life-cycle age profiles, differences in timing of retirement in the presence of separabilities with employment, differential paths of housing expenditures possibly driven by institutional differences in housing markets between countries, level and path differences in health status and mortality, and finally levels and volatility of medical spending in the US as deteriorating health with age leads to higher spending there while this is not true in the UK because of the National Health Service (NHS). Our empirical strategy is to first quantify cross-country differences in three potential factors-employment, housing status and health-and look for any immediate differences that might explain the differential consumption paths observed in Figure 1. We find,however, that most of these variables evolve in a similar way in both countries although there are some notable differences in tenure status. But even in the absence of differences in profiles, these three factors could play a role in explaining the different shape of spending profiles if there are differences in the nature of non-separabilities between these variables and consumption expenditures across the two countries. Hence, we move on to look for evidence of such non-separabilities by examining their effect on within-period budget allocations in a simple demand system estimated in each of the two countries. We find evidence that the relationship between health expenditures and mortality and employment is much stronger in the U...

Negligible senescence: An economic life cycle model for the future

“…The predicted pattern of medical care expenditure and biological aging is remarkably consistent with the observed patterns of increasing budget shares for medical care over the lifetime (Banks et al, 2016), and with the evidence on the dynamics of hazard rates observed in supercentenarians (Barbi et al, 2018). In particular, note that our model of deficit accumulation with uncertain lifetime does not change the main predictions from the standard health deficit model that medical care increases with age, a prediction in line with observable life time pattern of medical care (e.g.…”

Is the basic life-cycle theory of consumption becoming more relevant? Evidence from Portugal

“…To the best of our knowledge, this is the first paper to report estimates of a life-cycle consumption profile for Portuguese households allowing for time and cohort effects. We find a much flatter life-cycle consumption profile than the profiles reported in the literature for other countries-see, e.g., Alessie and Ree (2009) for the Netherlands, Banks et al (2016) for the UK, and Gourinchas and Parker (2002), Aguiar and Hurst (2013) and Fernández-Villaverde and Krueger (2007) for the USA.…”

“…In this sense, our results are closer to what the life-cycle theory of consumption predicts. However, consumption of older households, namely those over sixty years old, tends to decrease, as has been reported in the literature-see, e.g., Banks et al (1998) and Banks et al (2016).…”

“…The pseudopanel includes a measure of consumption (the median consumption of the cohort in each year), besides cohort (defined by the year of birth of the reference person) and time (year of the survey) dummies-details are provided in section 3. Other studies that employ a similar procedure are Zhou (2012) for China, Jappelli (1999) for Italy, Alessie and Ree (2009) for the Netherlands, Banks et al (1998) for the UK, Attanasio and Weber (1995), Attanasio (1998), Gourinchas and Parker (2002), Aguiar and Hurst (2013), and Lim and Zeng (2016) for the USA, and Banks et al (2016) for both the UK and the USA.…”

The Lifetime Medical Spending of Retirees

“…Several papers (De Nardi et al 2010, Kopecky and Koreshkova 2014, Ameriks et al (2015 show that health care costs that rise with age and income explain much of the U.S. elderly's saving behavior. 1 Differences in medical spending risk are also important in explaining cross country differences in the consumption (Banks et al, 2016) and savings decisions (Nakajima and Telyukova, 2018) of elderly households. More generally, the literature on the macroeconomic implications of health and medical spending is growing rapidly.…”