Editor: we redid this post for newer data.
Today we set out to figure out the correlation between income and net worth in the United States.
One of the arguments we’ve made in the past on Don’t Quit Your Day Job… is that, well, wealth and income are not the same (even though we tax income and are uncomfortable with our wealth distributions). While argumentatively it makes sense that there would be some correlation between income and net worth – without massive luck in the form of a lottery win, unplanned inheritance, or a huge stock gain it’s impossible to hit $10,000,000 in wealth on minimum wage – there is a lot of grey area in the middle where things don’t follow a convenient equation.
These grey areas contain a lot of exceptions – the aforementioned unplanned inheritances, personal finance bloggers spending under $25,000 a year (is this a big demographic?), 20-something year old investment bankers and young doctors with student loans, pro athletes, musicians, actors, early employees at tech companies, lottery winners, and all manner of other non-traditional earners and savers. And hanging above it all is age – not only do the youth have less years of savings, home appreciation and compounding (and, recently, fewer years of student loan paydowns…), conflating wealth and income also ignores a huge number of retirees of all wealth levels with negligible income.
Income is better analogized to acceleration and wealth to speed – just as a car accelerating quickly from start means something completely different than a powerful car accelerating quickly on the highway, the two quantities are related in a way… but are not the same.
The Relationship Between Wealth and Income
There’s a lot going on here, so let me try to break it down. On the x-axis, you’ve got household incomes, as defined in the SCF (think ‘taxpayer units’). On the left axis, you’ve got a measure of wealth (see the SCF methodology for details). The color of the dot represents the age of the ‘head of household’, where darker dots represent households headed by younger folks. The ‘size’ of the dots is the number of households which are supposed to be represented by the sample, so even though there appears to be a large number of dots in the upper right… they don’t mean as much as the lower left dots do.
Now, obviously (since I’m using log scales) this isn’t a linear relationship you’re looking at – but I wanted to give you some correlation statistics anyway (and our friend Jim from Free By 50 asked):
Weighted correlation: 0.5729566
Weighted r^2: 0.3283
Unfortunately, our goto curve fitting site ZunZun is down – but I think the picture itself conveys the information pretty well.
Methodology/Source on the Correlation Between Income and Net Worth
All numbers calculated from 2013 Survey of Consumer Finances microdata. We took the lazy option again, and dumped all 5 imputed data sets into one. The graph above is all five of those data sets in one. Read the take on this (excellent, since we used it successfully) Github Repo for R users here.
For informational purposes, again, we’re happy with the results. We aren’t in college anymore – but if you want to give us some grad students, we’ve got a large checklist of things we’d like to try!
Correlations, Causations, Etc.
Now, (recognizing we’re abusing linear regressions) from the strict viewpoint on correlations, an R^2 of .33 isn’t too significant to begin with. However, it is positive, so, yes, people with higher incomes do tend to have higher net worths.
But before assuming that “high incomes cause high net worths” or similar, note that we don’t know for certain from this alone that there is a connection (although we can theorize reasons why that could be the case – #1 being that generally the affluent save more), nor can we say with absolute certainty which direction that connection might point. However, it’s likely that there is a connection (correlation isn’t causation, but it doesn’t DISPROVE causation). If the relationship isn’t income causing higher net worth, perhaps traits which allow people to earn higher incomes also allow them to amass higher net worths… motivation, intelligence, choice of major/specialization/profession, social skills, penchant for saving, health, etc.
Let’s discuss the reverse option, (or, perhaps feedback?) for a second as well:the case that higher net worths cause higher incomes. There are a few arguments which make logical sense in this ‘opposite’ direction:
- Family wealth might lead to nepotism – the rich procure better jobs for their kids
- Stronger negotiation skills – psychologically easier to bargain with a larger net worth to fall back on
- General confidence – even when not negotiating, more financial confidence may lead to more raises, opportunities, etc.
- Ease of movement – more wealth means easier to take opportunities, interviews, etc.
- Less stress – having more wealth is less stressful, allowing more stressful jobs to be taken (generally higher paying)
… and maybe some of them apply, some of the time? However, I would agree that, generally, “high income comes first”… even if the connection is a lot looser than it may appear before looking into the statistics.
Of course, this still isn’t the full story. Millionaire teachers and middle/low income financial bloggers, bankrupt athletes, lottery winners, folks with large inheritances, lucky investors and the like are the outliers. Young professionals with large amounts of student debt, and retirees with a lot of wealth and little income are the others – I can probably account for the latter group, but not the first group. Because this is such an interesting topic I think we should try to narrow it down a bit by age, shouldn’t we? (Stay tuned!)
Is the correlation between income and net worth strong enough? Did you expect a stronger relationship? Should we still tax income and wealth in the same way? How do you think this data will look if we divide into ages or age groups? How awesome is my acceleration/speed income/wealth analogy?