VoxEU Column Labour Markets

Savings after retirement: A survey

Rich US retirees are known to spend their last years living it up in retirement hubs such as Florida. This column presents new evidence from the US suggesting that, in fact, those with high incomes run down their assets more slowly than implied by the basic life cycle model. Uncertainty over when they’re going to die and the possibility of high medical expenses – along with altruism and bequest motives – are important in understanding their low rate of spending.

More than one-third of total wealth in the US is held by households whose heads are over age 65 (Wolff 1998). This wealth is an important determinant of their consumption and welfare.  For example, Scholz et al. document that, with the notable exception of people in the bottom lifetime income decile, net worth is a major source of funds.  A striking feature of the wealth data is that retired US households, particularly those with high income, de-cumulate their assets at a rate slower than that implied by the basic life cycle model. Understanding why they do so is important to understanding how savings would respond to potential policy reforms. 

Most explanations for why assets fall so slowly can be split into two categories. The first set of explanations emphasises the health-related risks that the elderly face late in life, such as uncertain lifespans and medical spending. Elderly households may be holding onto their assets to cover expensive medical needs at extremely old ages. In fact, the observed patterns of out-of-pocket medical expenses, which rise quickly with age and income during retirement, coupled with heterogeneous lifespan risks, can explain a significant portion of US savings during retirement. The second set of explanations emphasises altruism. Individuals may receive utility from leaving bequests, or from not burdening their children with long-term care obligations.  These two sets of motivations have similar implications for savings in old age, making it difficult to disentangle their relative importance.   

In De Nardi et al. (2015), we discuss promising research that attempts to resolve this problem by looking at additional features of the data.  We point out the importance of going beyond total saving to look at housing, purchases of annuities and long-term care insurance, and government-provided medical insurance, especially the means-tested support provided by Medicaid.

Descriptive evidence

Figure 1, taken from De Nardi et al. (2010), displays median assets, conditional on birth cohort and permanent income quintile, for older singles. Figure 1 presents asset profiles for survivors – each point represents median assets for all the members of a particular income-cohort cell that are still alive at a particular date. Median assets are increasing in permanent income, with 74-year-olds in the highest permanent income quintile holding about $200,000, and those at the lowest permanent income quintiles holding essentially no assets. Individuals with the highest income tend to hold significant wealth well into their nineties, those with the lowest income never save much, and those in the middle de-cumulate their assets at a moderate rate. Thus, even at older ages, richer people save more, a finding first documented by Dynan et al. (2004) for the whole life cycle. 

Figure 1. Median assets, conditional on birth cohort and permanent income quintile, for singles.

Figure 2. Out-of-pocket medical expenses by age and permanent income.

Slow asset de-cumulation of the rich and old

The slow asset de-cumulation of the high-income elderly may be driven by health-related risks. Figure 2, also taken from De Nardi et al. (2010), plots the out-of-pocket medical spending of each permanent income quintile. Permanent income has a large effect on average medical expenses, especially at older ages. Average medical expenses are less than $1,000 a year at age 75 and vary little with income. By age 100, they rise to $2,900 for those in the bottom permanent income quintile and to almost $38,000 for those in the top permanent income quintile. Mean medical expenses at age 100 are $17,700, which is greater than average income for that age.

An important reason why those with low income spend less on medical services is that they are more likely to be covered by Medicaid, a means-tested government health insurance programme.  For example, De Nardi et al. (2013) show that those in the bottom permanent income quintile have a Medicaid recipiency rate of 70%, whereas those in the top quintile have a recipiency rate of 5%.  We show that Medicaid transfers rise rapidly with age, as people enter nursing homes.

Like medical spending, longevity increases with income.  Table 1, also taken from De Nardi et al. (2010), presents predicted life expectancies at age 70. Individuals in the top permanent income quintile typically live three and a half years longer than those in the bottom quintile.  Moreover, for rich people the probability of living to very old ages, and thus facing very high medical expenses, is significant. For example, we find that a healthy 70-year-old woman in the top quintile of the permanent income distribution faces a 14% chance of living 25 years, to age 95.

Models of savings behaviour

Can the risk of living long and having high medical spending explain the lack of asset de-cumulation shown in Figure 1?  De Nardi et al. (2010) show that when single retirees face the medical spending risks shown in Figure 2 and the longevity risks shown in Table 1, a model of optimal savings decisions can fit the savings profiles shown in Figure 1, even in the absence of bequest motives. De Nardi et al. (2009) use another version of this model to show that changes in the risk of living long and having high medical spending can generate significant changes in saving.

De Nardi et al. (2010) estimate their model with and without a bequest motive and find that both versions of the model closely fit the savings patterns in Figure 1, as well as the distribution of observed bequests.  When bequest motives are removed, modest changes in utility function parameters yield larger precautionary savings motives, allowing the model to fit the wealth data almost equally well. This shows the difficulty of distinguishing precautionary savings motives from bequest motives on the basis of wealth data alone.  Both motivations encourage saving and both motivations are strongest for the rich –bequests are typically modelled as luxury goods to match the data and precautionary savings motives are strongest for rich people who rely less heavily on means-tested government insurance. As Dynan et al. (2002) note, many people are likely driven by both motivations. These identification problems mean that modest differences in model specification or sample selection may lead to significantly different findings regarding the importance of bequest motives.

De Nardi et al. (2015) survey research on the role of altruism and bequest motives.  Much of this work utilises additional information beyond wealth (or saving) data in order to improve identification. Hurd (1989) and Kopczuck and Lupton (2007) find that the presence or absence of children is not important to determining either the existence or the strength of bequest motives. Lockwood (2014) matches additional data on purchases of long term care insurance. His key idea is that matching the data without bequest motives requires strong precautionary motives, which implies that the demand for long term care insurance is very large. However, the data show that purchases of long term care insurance are low.  In the absence of insurance market frictions, the only way to simultaneously match savings and low purchases of long term care insurance is to have modest precautionary savings motives and a significant bequest motive.  Using a complementary argument, Inkmann and Michaelides conclude that the life insurance holdings of UK households are consistent with bequest motives.  Ameriks et al. (2015) match the responses to ‘Strategic Survey Questions’ that involve hypothetical trade-offs between consuming long term care and leaving bequests. The hypothetical wealth splits chosen by survey respondents help identify the relative strength of bequest motives. Their results, based on samples of wealthy retirees, suggest that precautionary motives are at least as important as bequest motives. De Nardi et al. (2013) match Medicaid recipiency rates and transfer amounts. Matching the Medicaid data bounds the medical expense risk and the strength of the associated precautionary saving motives generated by their model. To match observed assets holdings in this environment, the model attributes part of savings to bequest motives. More generally, papers that bring more aspects of the data to bear tend to find more evidence in favour of a bequest motive, at least for richer families.

In addition to precautionary savings against health-related risks or bequest motives and altruism, elderly households may dissave slowly in order to continue living in their own homes.  Nakajima and Telyukova (2012) show that elderly households decumulate their housing assets more slowly than their financial assets (a fact also pointed out by Yang (2009)), and emphasise the difficulty of borrowing against housing wealth.


The elderly run down their savings much more slowly than implied by a basic life cycle model with a known date of death. The literature suggests that uncertainty and heterogeneity in the length of life and medical spending, along with altruism and bequest motives, are important to understanding the slow decumulation of retirement wealth. Identifying the relative importance of the bequest and precautionary saving motives is a priority, because the consequences of policy reforms hinge on their relative strength. For example, the savings impact of public health insurance, which reduces the need to save against longevity and medical spending risk, is tied closely to precautionary motives. Looking at additional features of the data, especially those not matched in the calibration or estimation of the model, is a key assessment tool. Recent studies have brought additional information to bear in promising ways.


Ameriks, J, J Briggs, A Caplin, M D Shapiro, and C Tonetti (2015), “Long-term care utility and late in life saving”, Vanguard Research Initiative Working Paper.

De Nardi, M, E French, and J Jones (2015), “Savings after retirement: A survey”, NBER working paper 21268.

De Nardi, M, E French, and J B Jones (2010), “Why do the elderly save? The role of medical expenses”, Journal of Political Economy 118(1): 39-75.

De Nardi, M, E French, and J B Jones (2012), “Medicaid and the elderly”, Working Paper 17689, National Bureau of Economic Research, June.

De Nardi, M, E French, and J B Jones (2013), “Medicaid insurance in old age”, Working Paper 19151, National Bureau of Economic Research, June.

De Nardi, M, E French, J B Jones, and J McCauley (2015), “Medical spending on the U.S. elderly”, mimeo, University College London.

Dynan, K E, J Skinner, and S P Zeldes (2002), “The importance of bequests and life-cycle saving in capital accumulation: A new answer”, The American Economic Review 92(2): 274–278.

Dynan, K E, J Skinner, and S P Zeldes (2004), “Do the rich save more?”, Journal of Political Economy 112(2): 397–444.

Hurd, M D (1989), “Mortality risk and bequests”, Econometrica 57(4):779–813.

Inkmann, J and A Michaelides (2012), “Can the life insurance market provide evidence for a bequest motive?”, The Journal of Risk and Insurance 79(3): 671–695.

Kopczuk, W and J Lupton (2007), “To leave or not to leave: The distribution of bequest motives”, Review of Economic Studies 74: 207–235.

840 Reads