Should We Plan to Spend Less as We Age in Retirement?

The idea that we’ll naturally want to spend less as we age is very seductive. It means we need less money to retire, can retire younger, and can spend a higher percentage of our savings in the early years of retirement. The latest writer to support this idea is actuary Frederick Vettese in his book The Essential Retirement Guide. This article is not a review of the entire book, but an examination of the chapter “How Spending Decreases with Age.”

All the data confirms that we do spend less as we age. The question at hand is why we spend less. If older retirees spend less because they run out of money, then it doesn’t make sense to bake reduced spending into our retirement plans. However, Vettese concludes that the dominant reason we spend less as we age is a “loss of interest in spending or physical ability to spend.” He uses this to justify “assuming that the retirement income from retirement savings ... does not have to be indexed to inflation.”

This conclusion flies in the face of my personal experience watching elderly members of my extended family suffer from non-indexed pensions being decimated by decades of inflation. But, to Vettese’s credit, he makes decisions based on statistical data and not anecdotes. So, I won’t say any more about my family’s experiences.

Vettese rejects the idea that people spend less because they run out of money by pointing to a study by Axel Borsch-Supan called “Saving and Consumption Patterns of the Elderly, The German Case,” published in the Journal of Population Economics in Bonn, Germany in 1992. I decided to look it up.

This study is based on German savings data from 1978 and 1983. Just about everyone in Germany had a defined-benefit pension, so if Germans increased their saving rate as they aged, this very likely means they voluntarily spend less. And this is exactly what the study found: the saving rate rose from about 3.5% of income at age 65 to about 10% of income for those over 80. However, the rate of spending decline is only about 0.4% per year, far less than the typical inflation rate.

Vettese provides other more recent German data showing that for each year people are older they spend about 1.5% less. However, this data isn’t scaled to income. It’s perfectly plausible that older German retirees have lower incomes than younger retirees. After all, economies grow and incomes rise in real terms. If a pension is based on your final few years of earnings, then young retirees would have larger incomes than old retirees.

Another comprehensive study of U.S. spending discussed by Vettese found that Americans spend about 1.9% less each year after age 66 (which is still less than the long term average inflation rate). The German study only allows us to attribute a small fraction of this decline to voluntarily spending less. The rest of the decline has some other explanation.

Running out of money isn’t really a possibility for anyone with a substantial indexed pension, but the many Canadians and Americans who live primarily off their own savings in retirement could easily overspend, run low on money, and be forced to spend less as they age. None of the data Vettese presented rejects this as a possible explanation for much of the spending decline we see in Canada and the U.S.

Any statistics we gather on spending decline as we age is going to mix together data from people with a wide range of reasons for changing their spending over time. Obviously, there are some retirees who spend too much during their first few years, see their mistake, and cut back on spending. I simply don’t find it plausible that the number of retirees who make this mistake is too small to be relevant. I have no data to back up this opinion, but I’ve never seen data that refutes it either.

I’m not claiming that there is no truth to the idea that people voluntarily spend less as they get older. When people get into their eighties they often seem much less willing to travel or go out at all. My claim is that there is no justification for saying that this is the only reason spending declines. In particular, there is no justification for assuming that we will all voluntarily spend less each year by the amount of inflation.

Suppose that for the next 25 years inflation matches its long-term average of about 3% per year. This would mean that costs would more than double over this period. Is it plausible that the typical retiree would choose to spend less than half as much at age 90 as he or she spent at age 65? We’re not talking about just discretionary spending here. The over 50% drop would include food and housing costs like rent or property taxes and home repair. I find the German data that predicts about a 10% voluntary drop in spending to be much more plausible.

Vettese is right when he says that the decline in spending cannot “be quantified with any precision.” The best U.S. figures we have place it at 1.9% per year. The only evidence that this decline comes from retirees simply not feeling like spending comes from German data that places this type of decline at 0.4% per year. The other 1.5% each year could be due to other causes. To assume that voluntary spending will decline at the pace of inflation is far out of line with the available data.

At first I was hesitant to criticize Vettese’s work because it is good to have a strong voice countering the usual fear-mongering from banks, insurance companies, and financial advisors about a retirement crisis. I welcome reasonable discussions about how much is enough to retire.

For my own retirement planning I’ve assumed that my spending would stay flat in real terms (which means that it rises with inflation). I could be persuaded to plan for a decline of 0.4% per year but no more than that without evidence that it makes sense.