A comfortable income is supposed to be insurance against misfortune. When the marriage ends, when the job disappears, when health turns — surely it is better to face those days with money than without. The General Social Survey, which has asked Americans about their happiness since 1972, lets us check. It records not just who is rich and poor, but who is divorced, widowed, unemployed, in fair or poor health, and whose finances have recently been slipping.
Because the survey is a snapshot rather than a diary, we cannot watch the same person before and after a setback. What we can do is compare four groups at once: higher-resource people with and without an adversity, and lower-resource people with and without the same adversity. If money cushions the blow, the happiness gap between the divorced and the married should be smaller at the top of the income ladder than at the bottom.
Mostly, it is not. Each setback costs roughly a quarter to a third of a step on the survey's three-step happiness scale — not too happy, pretty happy, very happy — and it costs the affluent almost exactly as much as it costs the poor.
Adversity drops happiness by similar amounts for rich and poor
Survey-weighted mean of the outcome (scale 1–3) for each group, with and without the adversity. Steeper line = bigger penalty. Hover any point for details.
The lines fall in near-parallel. Fair-or-poor health costs about 0.3 steps of happiness whether your family income sits in the top or bottom third of the country — its toll is class-blind. Widowhood, once you compare people of similar ages, is equally indifferent to the size of the estate. Unemployment hurts the affluent at least as much as the poor.
Two real differences survive statistical adjustment for age, sex, and survey year — one in each direction.
Divorce is the one blow money visibly softens. Among people in the bottom income third, the divorced and separated sit 0.31 steps below the married on the happiness scale; in the top third the gap is significantly smaller, by about 0.07 steps — roughly a fifth of the penalty erased. The same cushion appears when resources are measured by a college degree instead of income (0.05 steps, also statistically solid). Divorce is, among other things, a financial event: it splits households, lawyers are billed, and one income must do the work of two. It makes sense that this is where a buffer of money helps most.
Unemployment does the opposite — it punishes the privileged. On financial satisfaction, losing work costs a college graduate 0.60 steps against a working graduate's baseline, versus 0.39 for workers without a degree — the educated penalty is half again as large, and the difference is one of the strongest in these data. People with more to lose, lose more: a higher standard of living to maintain, and an identity more tightly tied to work. The happiness penalty of unemployment also tilts against graduates, though that tilt is within the margin of error.
The cushion, measured: how the penalty differs for the better-off
Difference in the adversity penalty between higher- and lower-resource groups, after adjusting for age, sex, and survey year. Bars right of zero: the better-off lose less (money cushions). Bars left: the better-off lose more. Whiskers are 95% confidence intervals.
Switching the outcome to financial satisfaction flips the picture. When finances have recently worsened, the better-off report a larger drop in financial satisfaction than the poor — about 0.67 steps for college graduates against 0.63 for everyone else after adjustment, with the gap statistically reliable. A decline from comfort registers as loss; a decline from hardship is more of the same. Psychologists would call it a reference-point effect, and it shows up here at population scale.
What this does — and doesn't — show
These are cross-sectional comparisons, not lives followed through time. People are not randomly assigned divorce or unemployment, and adversity can move people between income groups: losing a job lowers income itself, which is why every result is shown twice, once with income terciles and once with a college degree — a resource marker that a layoff cannot take away. The two versions agree.
Selection cuts both ways. If hardship-prone people sort into low incomes, the low-resource penalties may be overstated; if wealthy people exit bad marriages faster, the divorce cushion may partly reflect who stays divorced. And the happiness scale has only three steps, so penalties are compressed; a 0.3-step drop is roughly half a standard deviation of the national distribution — a large effect by the standards of this literature.
The honest headline is the null: across fifty years of data, the cushion that wealth buys against life's worst ordinary events is real in one place, absent in most, and inverted in the rest.