December 3, 2023 – When I wrote my post on the “Die With Zero” philosophy in October, I dug through the Survey of Consumer Finances (SCF). I used the detailed wealth distribution data to study the extent of asset overaccumulation late in retirement. The Federal Reserve releases the SCF only every three years, and just a few weeks ago, we got another survey covering 2022 and providing a wealth of information – pun intended. Quite amazingly, in 2022, for the first time in history, the average household net worth crossed one million dollars, now standing at about $1,060,000. Of course, wealth is unequally distributed, so while we may all be millionaires on average, the number of millionaire households is much smaller.
Then, what’s the percentage of millionaires? Is it a tiny elite, like the wealthiest 0.1% or 1%? I remember reading years ago that the share of millionaires was in the high single digits. So, I was surprised that more than 18%(!) of households were millionaires in 2022. That’s across all households and all age groups, and it is significantly higher for older folks. Also, the overwhelming majority of millionaires are homeowners. Homeownership can’t be such a terrible investment after all.
Since I did all that work, writing a Python program to dig through all those datasets, I thought I might as well write a blog post and share the results with you. Let’s take a look…
A few notes on the Survey of Consumer Finances
Note that the $1m average net worth is not per capita but per household. So, with an average household size of about 2.5 members, the per-capita net worth is “only” about $400k. So, individually, we’re not even close to reaching millionaire status, but $400k is also very impressive because that’s the net worth of every resident, old or young, working or not.
Also, the Federal Reserve publishes aggregate data on household balance sheets as part of the “Financial Accounts of the United States (Z1)” dataset. Those releases are quarterly (though with several months’ delay). Even before the new SCF data release, crossing the $1m average net worth threshold was a forgone conclusion; we have about 131 million household units in the U.S., and the total net worth has been above $131t since 2021. For example, in Q2 of 2023, the most recent data available, the total net worth reached just under $146 trillion, so the average net worth has increased to over $1,130,000.
What’s the use of the 2022 SCF data then? The SCF goes further and interviews a large sample of individual households. Thus, we can gauge the distribution of wealth that would get lost when looking at aggregate data. Please find more info on this landing page at the Fed. The summary is available in HTML and PDF formats.
So, let’s get to the SCF data now…
Findings
The average net worth is above a million dollars. But it’s unequally distributed!
While the average household net worth is above $1m, the median is only $192,700 (all figures rounded to the nearest $100). The Gini coefficient, a measure of inequality (0 means all wealth is equally distributed, while 1.0 would imply one person owns everything and everyone else has zero), is notoriously high in the USA at 0.83, higher than in most other developed nations. Surprisingly, though, Sweden ranked higher, at 0.881 in 2020.
We can also study the distribution in more detail. I calculated the net worth at different percentiles of the distribution. Here are the cutoffs you’d need to make it into the bottom or top 1%, 5%, 10%, and 25%. The bottom 1% and even 5% of households have negative net worth numbers. If you own only $450, you’re already better off than the bottom 10%. Meanwhile, to belong to the top 10%, you’ll need almost $2m. Just under $4m puts you in the top 5%, and you’ll need to reach the “eight-figure club” at over $13m to get into the top 1%.
Notice that the numbers above are the exact cutoffs to get into specific percentile groups. We can also calculate how much of the total net worth is owned by particular population percentiles. In the chart below, I plot the 2022 Net Worth Lorenz Curve, i.e., if I were to rank all U.S. households from low to high on the x-axis, then the Lorenz curve tracks the cumulative share of those households on the y-axis. The Lorenz Curve would be a straight 45-degree line if the net worth were perfectly equally distributed. But in reality, the Lorenz curve is the convex curve below. For example, in the Lorenz Curve below, we can see that the bottom 50% of households own only about 2.2% of the total net worth. And for the quantitative geeks, the Gini Coefficient is derived from the Lorenz Curve. Specifically, we calculate the Gini coefficient as one minus two times the blue-shaded area underneath the Lorenz Curve.
Some additional observations: the top 0.1% of households own 14.7% (=100%-85.3%) of the total net worth. The top 5% already own 61%, thus a majority of the total wealth. And the top 10% of households own almost three-quarters (100%-26.6%=73.4%) of the nation’s wealth. Pretty mind-blowing numbers! But take solace in the distribution being even more lopsided in Sweden!
Percentage of millionaire households:
I remember years ago learning that about 10% of households were millionaires. Now that share has increased to 18%. What’s more, the percentage of “5x-millionaires” is an equally impressive 3.7%, so about one in 27 households. But also notice that about 7.9% of all households have a zero or negative net worth! In fact, if we look at the Lorenz Curve above and zoom in to the lower left corner, one would see the curve dip below zero for a while. (Side note for the math geeks: the Gini coefficient could theoretically exceed 1.0 if enough households had a negative net worth!)
Age is a significant determinant of net worth!
This should not be too surprising: people accumulate wealth as they age, so households led by older individuals are significantly wealthier. In the chart below, I plot the mean and median net worth by age group and the top and bottom 10% and 25% cutoffs. The mean household net worth increases for every age up to the 65-69 cohort and slowly decreases again. Only for the 50+-year-olds do you get an average net worth in the seven figures. I also noticed that for older households, the 75th percentile is already above one million dollars, so more than a quarter of the senior households are millionaires. And the mean net worth is well above $1.5m for all age groups 60+.
Side note: I should stress that the SCF is purely a snapshot of the 2022 population. We cannot translate the SCF into time series paths of actual households. So, for example, in 2022, the 60-64 and 65-69 age groups had an average net worth of $1,675m and $1,837m, respectively. It does not imply that today’s 65-69-year-olds grew their net worth by $162k over the last five years. But the general pattern, i.e., saving and accumulating over the typical career years from your mid-20s to mid-60s and then slightly decumulating over your retirement years, is still valid. However, we’d have to employ panel data, i.e., tracking a large panel of households over a longer time horizon, not just in one single snapshot.
The rise in net worth is not just due to inflation.
Naysayers will object that the rise in net worth is simply due to rampant inflation. But that’s not the whole story. If we plot the average and median net worth numbers over the entire set of 12 SCF surveys from 1989 to 2022, we get the charts below. Notice that I plot both the nominal net worth (i.e., in current dollars) and the real figures in CPI-adjusted dollars, measured in 2022 dollars. So, for example, in 1989, the average net worth was $184,900 in nominal terms, which is equal to $436,600 in 2022 prices. By definition, nominal and real dollars are identical in the base year 2022.
According to the charts below, average and median net worth numbers increased substantially, even when adjusting for inflation!
However, it is noteworthy that the real median household net worth took until 2022 to pass its 2007 level again. The average real net worth already reached a new high in 2016.
And again, we can also track the share of millionaire households over time, both in nominal and real dollars. The blue line (percentage of millionaires in nominal dollars) went from 3% in 1989 to 18% in 2022. Eroding purchasing power over time makes it much easier to surpass the million-dollar threshold. But even adjusting for inflation, that share of millionaires has more than doubled since 1989 (8.2% to 18.0%).
Why America should not be ashamed of its Gini coefficient
If I want to put a positive spin on the unpleasant wealth inequality stats in the U.S., I would again point to the net worth chart by age group: Some of our inequality is due to the natural wealth accumulation lifecycle. For example, within my age group (45-49), the wealth Gini coefficient is lower: 0.769. Americans are very good at building assets, thanks to their entrepreneurial spirit and generous tax incentives, like tax-advantaged retirement plans and capital gains deferral.
In other countries, wealth accumulation is not as common thanks to people relying more on government-run retirement systems, like in most of Western Europe, especially in my native Germany. Let’s see how that works out for their retirement planning! I’d rather take my chances with an S&P 500 index fund than with the German government. However, I was surprised that even in Germany, the Gini was relatively high at 0.788 in 2020, according to Wikipedia. Germany has much less mobility and a lot of “money nobility,” i.e., sticky, inherited wealth, while most American millionaires are self-made. So, especially in the FIRE community, we should not be ashamed of a high Gini. As long as we maintain wealth mobility, we should celebrate wealth inequality because it’s a symptom of self-made affluence. In an old post in 2017, I once calculated the Gini Coefficient of ten different ERN household net worth snapshots during my accumulation phase and found a Gini of 0.62. Just from the life-cycle effect.
Wealth inequality has decreased (slightly) since 2019.
It’s also worth pointing out that the Gini coefficient decreased in 2022 and now stands at the lowest level since 2007, though still far above the Gini in the 1990s. The Global Financial Crisis apparently caused a sizable bump in inequality, and we’re now slowly walking it down again. That makes sense because many middle-class households had their net worth tied up in real estate, which in many places didn’t recover until after the pandemic. Most affluent families in the top 10% of the wealth distribution likely held assets that recovered much faster: publicly traded equities, private businesses, and sometimes even multi-family real estate, which did much better during and after the crisis, more on that later. Also noteworthy: If I calculate the Gini in my age group only, we are almost back to the Gini levels in the late 1990s.
The rich are getting richer. But what about the rest?
In a growing economy, you’d hope that the gains from growth will reach all parts of the population, not just the super-rich. How are we doing in the U.S.? You hear often that the middle class is not just missing out on the gains but even falling behind. That’s not exactly true. If we compare the wealth distribution in 1989 with 2022, most percentiles gained ground. True, the 1%, 5%, and 10% lowest percentile had negative to zero net worth figures. The 1% poorest got deeper into debt. But the middle class is getting richer, albeit modestly slower. The middle two quartiles, ranging from the 25th to 75th percentiles, also gained between 77.6% and 107%. Quite intriguingly, the exact middle, i.e., the median, grew the slowest (77.6%), while the lower and upper cutoffs, folks closer to the lower middle class and upper middle class, did slightly better, though not as well as the heavy-hitters in the 90th percentile and above:
Individual results may vary!
One issue that always rubs me the wrong way when folks compare historical net worth numbers is that the time series of the net worth mean and median (or any other points in the points in the net worth percentile distribution) are not really that meaningful. As mentioned above, the typical household should experience much faster net worth growth than the economy-wide median household because of the lifecycle pattern of wealth accumulation. For example, imagine we look at the age 30-34 cohort in 1989. Those same households are 33 years older in the 2022 survey and will fall into the 60-64 and 65-69 cohorts. Well, not all of them because some folks might have died, divorced and married a head of household with a different age, moved to another country, etc. But for most families, it should be safe to assume that they moved through age groups over time and now ended up in those two age cohorts in the 2022 survey. So, let’s see how their net worth numbers compare to 1989; see the table below.
Quite intriguingly, by this measure, the median had a higher growth rate than the higher percentiles, just under 1400%, i.e., almost 15x. Of course, in dollar figures, the net worth growth of the higher percentiles was much larger, but relative to the starting point, the middle class did quite well. In fact, the lower end of the middle class, which I loosely define as the 25th percentile of the net worth distribution, had the fastest net worth growth. And just for the record, there is no guarantee that the median household in 1989 is now still the median. Some might have fallen below the median, and some might have advanced into the higher percentiles. But the distribution above the 25th percentile experienced substantial growth, much more than when looking at economy-wide figures. The lower end of the net worth distribution in the aggregate number will always look so poor due to the never-ending supply of “poor” people, a.k.a., people in their 20s with low-to-no wealth and a pile of student loan debt. But the path of actual Americans over their lifecycle will look much better!
Also, notice the tremendous rise in the share of millionaires, growing by more than 10x (slightly above 900%) from 2.7% in 1989 to 27% in 2022. And the growth in 5x millionaires is even more impressive, from 0.05% in 1989 to around 5-7% in the two cohorts in 2022, which is more than a 100x growth in the share.
More on Millionaires
Let’s look more carefully at the composition of household net worth numbers. Obviously, millionaires have higher net worth numbers, but can we spot any differences in their balance sheets? Where do millionaires invest their money? Do they own homes? Is their investing style more aggressive?
So, I slice the 2022 population into non-millionaires and millionaires. And then, within the millionaire’s category, I further distinguish between “ordinary” and “multi-millionaire” households, specifically, households with a net worth of $1-5m vs. $5m+. And I’m aware that there are different definitions/cutoff values of multimillionaires. I’ve seen people use cutoffs at $2m, $5m, and $10m. On the one hand, I wanted the cutoff to be significantly above the $1m mark, and on the other hand, there aren’t enough households with a $10m+ net worth in the sample, so I settled on the $5m cutoff.
Let me show you three tables. First, the 2022 average values in different subcategories of assets and liabilities:
And then the same table, but I display all values as a percentage of each group’s average net worth:
And third, the percentage of households that own a positive amount of the different subs:
What do we learn from the data here? A few things stick out:
- Millionaires are more heavily invested in financial assets. If we combine the categories Investment Funds + Stocks + Bonds + Retirement accounts, then about 40% of millionaire household net worth falls into that category. For non-millionaires, the four categories add up to only 24%. Also, the share of households owning any “higher-return” assets is higher among millionaires. For example, 47.5% of all millionaires and 62.6% of multimillionaires own stocks outright, while only 15.1% of non-millionaires do. And half of the latter might be GenZs living in their mom’s basement, trading Gamestop; I’m just kidding.
- A lot of non-millionaire household net worth is tied up in their home. Specifically, more than 106% of non-millionaire households’ net worth is in the nonfinancial asset bucket, mostly in homes and vehicles. The gross value of the primary residence comprises over 80% of non-millionaire household net worth, compared to 19% and 11% for millionaires and multimillionaires, respectively. Even home equity (home value minus mortgage) comprises over 50% of the non-millionaire household net worth. But only 15.1% and 9.2% of millionaires’ and multi-millionaires’ net worth, respectively. So, it seems that non-financial assets, especially homes, are crowding out the other investments of non-millionaires.
- 95% of millionaires are homeowners. This is not a typo. Even though the net worth share of the primary residence is low for millionaires, the share of millionaires that own a primary residence is about 95%. And that number is pretty uniform among both sub-categories, i.e., millionaires and multi-millionaires. In contrast, only about 60% of non-millionaires are homeowners. 66% is the overall homeownership rate.
- Millionaires are likely to be business owners. 35% of millionaires own a business. And that share increases to 55% for the multimillionaire category. So, to strike it really big in the net worth world, you probably achieved that through owning a business.
- And a lot of millionaires’ net worth is tied up in that business. The average value of the business is only $8,300 for non-millionaires but in the six-figures for ordinary millionaires and over $5m for multi-millionaires.
- Millionaires still have debt. That was a bit of a surprise! 52.3% of all millionaires and 45.3% of multimillionaires have a mortgage on their primary residence, while “only” 40% of non-millionaires do. Of course, only 59.7% of non-millionaires have a home. So, conditional on owning a home, millionaires have only about a 50% chance of having a mortgage, compared to about two-thirds for non-millionaires.
- Millionaires (somewhat) splurge on vehicles. The average value of vehicles in millionaire households was almost 3x that of non-millionaires (73.4k vs. 26.2k). Multimillionaires splurge even more and own $127.2k worth of vehicles, on average. Of course, relative to their net worth, that’s a drop in the bucket. Non-millionaires have 12.4% of their net worth tied up in a depreciating asset, a much more significant percentage than affluent households. Also, there is a much higher incidence of vehicle loans among non-millionaires.
- Millionaires prefer good debt over bad debt. As I pointed out in the previous paragraphs, the affluent still have some debt, including mortgages and even car loans, but a much lower incidence of credit card debt.
Sidenote: the high homeownership rate among millionaires is not just due to age. We know that both net worth and homeownership go up with age. So, is the higher homeownership rate among millionaires possibly due to their age? No! Even if we bucket the population by their age, millionaires and multimillionaires have a higher incidence of homeownership; please see the chart below. One exception is the very young age category (0-29 years), where the $1m-$5m category has a slightly lower homeownership rate than the non-millionaires. I suspect these are the young, nouveau rich in NYC and SF who are still renting. But in all other age cohorts, millionaire households are far more likely to own than non-millionaires.
Is there a homeownership conundrum?
Are we getting conflicting signals? On the one hand, homeownership seems to be a signal of economic success because there is a 95% homeownership rate among millionaires. On the other hand, it appears that a home is something of an albatross on the balance sheet of non-millionaires.
Could it be possible that homeownership is a terrible investment after all, as proclaimed by some personal finance influencers, i.e., Ramit Sethi and others? Are millionaires financially successful not because but despite owning a home? Maybe a house is like all the other money pits, like boats, vacation homes, etc.? So, there may be a correlation, but the causality goes the other way around: rich people can afford a poor investment. But I don’t think that’s the case. On my blog and in real life, I have always been consistent with my philosophy. I always like to point out two critical issues: An investment choice and a budgeting choice!
1: Investment. A primary residence is likely a good investment if you compare apples to apples. The total return of a home includes the implicit rental income, i.e., the benefit of not having to pay rent. Mathematically illiterate influencers who compare real estate price returns with stock total returns are deceiving themselves and their readers and listeners. Please see my post “How To “Lie” With Personal Finance – Part 2 (Homeownership Edition),” item 1.
I indeed concede that a house may have a real expected return a bit below that of an equity index fund. But once you consider the equity volatility and the tax benefits from housing, i.e., tax-free implicit rental income and tax-free long-term capital gains of up to $500,000 for couples in the U.S., you have a desirable return profile. Especially in the context of safe withdrawal rates, a paid-off home helps alleviate Sequence Risk.
2: Budgeting. A home being a good investment does not imply that a bigger home is always better. See again that “Lie with Personal Finance” post, item 4: Overconsumption is not a good investment. We must distinguish two decisions every household must make: 1) rent vs. own and 2) size and value of the home. You may indeed make the right investing decision in the rent vs. own dimension by purchasing a 5,000 sqft McMansion. But it’s a terrible budgeting decision in dimension 2 for most middle-class households! I always use the following analogy: Imagine Delta Airlines crunched the numbers and decided owning a Boeing 737 airplane to serve the route from Atlanta to Nashville is financially superior to leasing that same airplane. Would this imply that buying a Boeing 777 or 787 is an even better investment? No, because that plane would be too large for a puddle-jumper route that asks for a 737. The investing decision is separate from the budgeting decision.
Thus, I posit that the millionaire balance sheet data validates both points. First, while we cannot precisely ascertain any cause vs. effect direction, the almost 100% homeownership rate among the most financially successful households and the much lower rate among non-millionaire households, plus the simple IRR calculations factoring in all the costs and benefits of renting vs. owning and showing that housing often has a pretty good IRR, definitely support the idea that homeownership is generally a sound financial decision.
Of course, today’s home values and mortgage rates seem a bit high, but at least historically, homeownership has been helpful. Then again, today’s equity valuations also seem very unattractive relative to bond yields, so if you prefer to remain a renter in today’s economy and invest in the stock market instead, you may not be too impressed with the results either.
Second, the millionaire balance sheet data supports enjoying homeownership in moderation. Millionaires own relatively modest homes. Clearly, millionaire homes are more expensive on average than middle-class homes, but millionaires own much smaller homes relative to their net worth than non-millionaires. There seems to be a sense that large houses are holding back the middle class in their financial success because homeownership in excess impedes the accumulation of other assets, like stocks, bonds, and other high-return vehicles like investment funds held directly or in retirement accounts. To thread the needle, it would be ideal for households to own a modest home, much smaller than what your realtor and banker indicate you can afford. Then, invest the excess cash flow in high-return assets, ideally equity index funds.
Conclusion
I hope that in the FIRE community, where many of us are already millionaires or are striving to become one, people would find my little data analysis about net worth and millionaire stats valuable. I learned from the data that the millionaire club isn’t as exclusive as it used to be; almost one in five U.S. households is already in the seven-figure net worth club. And nearly one in three in the older cohorts! I also found additional evidence to support my theory that homeownership is helpful for your financial picture – if used in moderation.
Thanks for stopping by today! I look forward to your comments and suggestions.
Title picture credit: pixabay.com
Technical Appendix
I get slightly different mean and median net worth figures from the SCF, usually within 0.1-0.5%. Maybe the SCF researchers are using different functions/methods. Here’s what I used:
For weighted averages, I use Python’s numpy function, and for the Quantile values (e.g., median, 99th percentile, etc.) I use the DescrStatsW tool from statsmodels. Please see the screenshot below for a sample code generating a median of $192,700 and a mean of $1,059,470. That’s slightly different from the SCF report, i.e., $192,900 and $1,063,700 for mean and median, respectively. It’s close enough for government work, I guess, but it’s still puzzling. If anyone has any insights as to what’s going on here, please let me know! Quantile estimates can vary slightly due to different interpolation methods (e.g., closest, linear, cubic spline, etc.), but the weighted mean calculation should be standard.
You can download the above Python code here.
You should download the (very large) zipped STATA datafile from the FRB website and put the folder with the file into the “Data/” subfolder. For your entertainment, I also posted the real and nominal net worth numbers over time and by age group in these two Excel Files: