Download PDF version

In my last project analyzing household wealth using the Survey of Consumer Finances (Cohen 2017), I found that cross-sectional regression analyses predict age to be the strongest predictor of wealth. The relationship makes sense. People save money over time and are more likely to inherit familial property later in life.

So what are the typical differences between in household finance across age groups?  What counts as “rich”, “middling”, or “poor” for a twenty-something versus a fifty-something?  And what do these differences tell us about household finances?  In this analysis, I examine differences in net worth across age groups in the 2022 data. It looks at the size of wealth differences across age groups and the incidence of low net worth over the life course.

Data and Methods

I use data from the Survey of Consumer Finances, a triennial survey of several thousand U.S. households (with a high-wealth oversample) (Federal Reserve 2023). In this analysis, we will examine the distribution. Readers can see the details of my acquisition and wrangling of this data in the folds of this document or by consulting its associated R Markdown file. I discuss the data and its analysis at greater length in this post.

Here, the focus is on the household net worth, which is the market value of a household’s assets, less the costs of settling their outstanding debts. This can be treated as a metric that captures a family’s ability to come up with money. The metric is less useful for making fine distinctions between households with similar wealth levels and better for assessing differences on the order of hundreds of thousands or millions of dollars.

We are examining how the distribution of net worth varies according to the age of the household head. In the SCF, the household head is the individual or partners who are described as “economically dominant” in a primary economic unit. It includes the person(s) who cover the household’s basic living costs for themselves and all economic dependents who live within that household. We group households according to the age of the SCF’s “reference person”, whom the Survey considers to be the male in a mixed-sex pair and the older person in a single-sex couple. This convention is likely maintained to create a constancy over time in the measurement of the SCF concept of “reference person”. My own expectation is that the choice of reference person does not strongly affect the wealth accumulation differentials that we observe, but I would test different ways of counting the reference person (i.e., the oldest in any relationship, the mean age when people are paired) if there were interest.  I chose to group them by decades, mostly so that the research findings conform to commonplace terms used to discuss personal finance and age.

Analysis

The table below describes the distribution of net worth across our age groups. The columns represent age groups, and the rows represent different percentile net worth values for each age group.

Percentile Net Worth Values by Age Group

	<30	30s	40s	50s	60s	70s+
10th Percentile	-$13,278	-$3,274	$736	$4,086	$5,420	$9,536
25th Percentile	$1,122	$15,552	$31,326	$69,840	$78,254	$99,192
50th Percentile	$21,054	$107,752	$189,746	$306,886	$394,394	$367,860
75th Percentile	$117,796	$323,606	$600,252	$1,074,428	$1,106,748	$1,035,780
90th Percentile	$280,300	$750,544	$1,388,340	$2,711,830	$2,962,016	$2,976,530
95th Percentile	$411,880	$1,223,800	$2,704,490	$5,382,202	$6,639,864	$5,875,320
99th Percentile	$2,128,300	$5,026,540	$7,670,380	$15,530,000	$18,921,600	$18,552,600

First, note that the distribution of household wealth is much more compressed before age 30. Many young people are still in school, which often involves accumulating debt without earning money. Many of the poorer households in this age category will ultimately end up among the wealthier later in life. Households that enter the workforce immediately may get an earlier start accumulating property, but my sense of the data and evidence is that advanced training has much higher returns.

Differences open up over people’s 30s and 40s, and eventually stabilize in the 50s onward. A household’s 30s and 40s may be a period in which they experience the run-up of their career and property accumulation, setting them on the trajectory into which they will settle in the 50s onward. Households that spend their 30s and 40s with the privilege of creating a successful business, ascending quickly and going far in their professional field, getting an early start on their retirement savings, or having successful first forays into real estate. Households accumulate earning power in part through business or professional success, and a growing portfolio of money-making or price-appreciating property.

The distribution of wealth seems to steady when we get beyond people’s fifties. It does not mean that people’s personal fates are static. Some people will run up professional success and savings in their 30s and 40s, and then find themselves pressed out of the workforce when they hit their 50s. Others will have earned more steadily and will be able to stretch out their working years into their 70s. But the general distribution of wealth remains pretty static in the aggregate beyond the 50s. Another reason that wealth may stabilize after the 50s is that people are most likely to receive intergenerationally-transmitted wealth by this decade. These transmissions may also propel the run-up of wealth in people’s 30s and 40s, as the incidence of parental death rises.

I close the analysis with a visualization that resembles the age-weight charts at the pediatrician. It offers a way to look up wealth levels over the life course. I can create a year-by-year one if there’s interest.  The first version is on a log scale because we would not be able to see clear differences in non-millionaires when we include multi-millionaires in the figure.

Keep the log scale in mind when interpreting the data.  Wealth levels get larger in multiplicative rather than additive terms as we move up the y-axis.  The differences look pretty evenly spread out on this scale, but this is the figure without the logged axis (showing only up to the 90^th percentile):

The graph gets much taller when we include 95^th and 99^th percentile scores.  All of this is to clarify the fact that the run-up of wealth that occurs during people’s working years differs on multiplicative scales.