How Does Spending Really Change in Retirement?

Financial planners and researchers often assume that retirees prefer steady consumption as they grow older. Social Security benefits, too, are based on the premise that people want steady inflation-adjusted benefits. 

However, several studies focused on new retirees suggest that retired households actually consume less over time.

Whether households prefer a constant, increasing or decreasing path of consumption in retirement has important implications for an understanding of retirement adequacy, according to a report published Tuesday by the Center for Retirement Research at Boston College on the findings of a recent study that used data from two longitudinal surveys to examine the consumption behavior of retired households.

CRR researchers examined retirement consumption over longer periods and used wealth and health to separate constrained and unconstrained households in order to determine whether necessity or preferences drive any declines in consumption. 

They also explored whether, within unconstrained households, those with shorter expected life spans have faster declines in their consumption preferences.

According to the results, consumption for the population as a whole declines over retirement. 

However, health and wealth constraints also matter. Higher-wealth households that self-report excellent or very good health have a nearly flat consumption pattern, declining by only 0.3% a year. 

By comparison, for those who start retirement with good health, consumption declines by about 1.1%, and for those with fair or poor health, it declines by 3.2%.

According to the report, this suggests that constraints at least in part drive observed declines for the whole population. 

Declining consumption paths may also reflect differences in life expectancy, but the results are less clear on this point, according to the report. 

It noted, for example, that healthier individuals and women have longer life expectancies. This suggests that to the extent that consumption declines reflect different mortality profiles, healthier and married households within the top wealth tercile should also have flatter consumption paths. 

In fact, the results show that these healthy unconstrained households do have flatter consumption paths. And for the population as a whole, married couples have flatter consumption profiles. 

The report concludes that wealth and health are important determinants of spending in retirement and that retirees’ consumption preference is likely much flatter than observed in the data. 

The report says many questions remain to be more fully explored. One is whether consumption profiles continue to get flatter for the top quintile or decile. 

Another is whether survival expectations matter, or whether other factors such as risk aversion or bequest motives may determine consumption paths.