The Insurance

The plate on subsistence left the floor in mid-air. The classical economists had pinned it to the body — the wage could not fall below what kept a worker alive and reproducing — but that floor, we saw, lies far beneath any wage paid in a wealthy society. The modern worker who turns down a job is not choosing to starve. So what, now, is the floor? What is the lowest wage a worker will accept, when refusing it no longer means death?

The answer is the worker's reservation wage: the wage below which they will say no. It is the modern, secular descendant of the subsistence floor — the bottom of the bargaining range with which this whole volume began. But where subsistence was fixed by nature, the reservation wage is fixed by circumstance, and by two circumstances above all: what else the worker might find, and what the worker can fall back on while looking. The first is a question of search. The second is a question of insurance. Together they set the floor — and, this plate will argue, they set much of the worker's power.¹

Begin with the idea itself. Every worker, offered a job, weighs the wage against their alternatives, and there is some wage so low that they prefer the alternatives — prefer to keep looking, or to live on savings, or to draw what the state provides. That threshold is the reservation wage. Above it, they accept; below it, they refuse.²

In the frictionless market of the textbook, the reservation wage is simply the next-best wage on offer; and since identical workers all command the same wage everywhere, the reservation wage equals the market wage and the whole notion collapses to a point. But the labour market is not frictionless. A worker cannot survey every job at once and step instantly into the best; jobs must be found, one at a time, at the cost of time and effort and forgone earnings. And it is precisely this friction — the brute fact that finding work takes time — that turns the reservation wage from a triviality into the hinge of the modern theory of wages.

The economics of this was worked out in the theory of job search, and its central result is elegant. Picture a worker sampling offers over time — one this week, another the next, each drawn from the distribution of wages employers happen to offer. The worker need not take the first; they may reject it and wait for a better one. But waiting is costly: another week unemployed, another week's earnings forgone. The worker's problem is therefore one of optimal stopping — accept this offer, or gamble on a better — and its solution is a reservation wage: a threshold such that the worker accepts any offer above it and rejects any below.³

What sets the threshold? Exactly the comparison the diagram draws. The value of accepting an offer rises with the wage — a better job is worth more. The value of rejecting it does not depend on the current offer at all; it is the value of remaining in search, of drawing again from the distribution next week. The reservation wage is the wage at which these two are equal — the offer just good enough that accepting it beats searching on. And now the levers are visible. Anything that raises the value of continued search — a better distribution of offers, a lower cost of waiting, and above all a more generous cushion to live on while waiting — raises the reservation wage, and the worker holds out for more. Anything that makes waiting more painful lowers it, and the worker takes less. The floor is the value of being able to wait.⁴

Here the plate pays a debt it has owed since the first. The opening plate asserted that the wage is settled within a range — a floor, a ceiling, and a contested space between — but it did not say where the range comes from. Search theory supplies the answer, and it is the same friction that set the reservation wage.

When a worker and a firm meet and find each other suitable, they create something that did not exist before: a productive match, worth more together than either party's next-best alternative. The worker brings a floor — their reservation wage, the value of walking away and searching on. The firm brings a ceiling — the worker's productivity, the most the match is worth to it. Because search is costly, these are not the same number; it is worth something to both sides to strike a deal here rather than break it off and begin again. Between the floor and the ceiling lies a surplus — the gain from the match existing at all — and the wage is simply the line that divides it.⁵

Set the wage low, near the worker's reservation, and the firm keeps most of the surplus; set it high, near the worker's productivity, and the worker does. Where exactly the wage falls is not pinned down by productivity or by subsistence; it is settled by bargaining, and bargaining is settled by power. This is the whole architecture of the volume, derived from a single friction. Take the friction away — let workers find the best job instantly and at no cost — and the floor rises to meet the ceiling: any firm offering less than full productivity is abandoned at once for one offering more, the surplus vanishes, and the wage pins to a point, exactly as competition promises. Restore the friction and the surplus reopens, the range returns, and with it the question of who gets what. The monopsony of the eighth plate is one answer — the friction lets the employer hold the wage near the floor. The reservation wage of this plate is that floor. And the safety net, we will now see, is the most deliberate way a society has of raising it — of moving the worker's walk-away point up, and the division of the surplus with it.

Why, then, does every wealthy society take the worker's fallback — the income they live on while searching — partly out of the worker's own hands and into the state's? Why unemployment insurance, paid not by the worker alone but pooled across all workers and underwritten by the public?

Because the risk of job loss is one a private market cannot insure well. A worker would gladly buy insurance against losing their job; but no private insurer will readily sell it. Those most likely to be laid off would be the keenest to buy, and the insurer, unable to tell them apart, must price for the worst — until the premium drives the safer workers out and the market unravels, the classic spiral of adverse selection. And once insured, a worker has a weaker reason to keep the job or hurry back to a new one — moral hazard, which a private insurer can monitor only clumsily. Worse, the risk is correlated: in a recession everyone is laid off at once, so the insurer faces not many small independent claims but one vast simultaneous one, precisely when its own finances are strained. For all these reasons private unemployment insurance barely exists, and the task falls to the state, which can compel participation, pool the whole workforce, and draw on the public purse in a downturn.⁶

What the state provides is, by design, partial and temporary — in the United States, a benefit that typically replaces something between a third and a half of prior earnings, for a limited spell, commonly up to about six months and in many states now less.⁷ But even a partial, temporary cushion does to the reservation wage exactly what the diagram showed: it lifts the value of searching, and so raises the wage below which the worker will not go. The insured worker can wait for a better match; can refuse a worse one; can, for a while, say no. The safety net is the most deliberate instrument a society has for setting the floor of the wage.

Insurance is not free, and the very property that makes it valuable makes it costly. The cushion that lets a worker hold out for a better job also lets a worker hold out longer than they otherwise would — and some of that holding-out is not the patient pursuit of a good match but simple delay. The evidence is plain and a little uncomfortable: workers drawing unemployment benefits search less intensively and return to work more slowly, and the rate at which they find jobs jumps sharply just as their benefits are about to run out — a spike too sudden to reflect anything but the incentive itself.⁸ This is the moral hazard of insurance, and it sets the central problem of policy: a safety net generous enough to give real security is generous enough to dull the urgency of return.

How generous, then, should it be? The question has a clean answer in principle, due to Baily and later sharpened by Chetty. Each added dollar of benefit buys something and costs something. What it buys is consumption smoothing: a worker who loses their job suffers a steep fall in living standards, and because they are risk-averse, a dollar of income in that bad state is worth far more to them than a dollar in good times — so insurance that moves income from the good state to the bad is genuinely valuable. What it costs is the behavioural distortion: the slower search, the longer unemployment, the spike at exhaustion. The optimal benefit is the one at which these are equal at the margin — where the last dollar of insurance buys just as much consumption-smoothing value as it costs in distorted behaviour.⁹

There is a subtlety here that rescues insurance from the charge of being mere sloth subsidised, and it is among the more beautiful results in the field. Part of what looks like moral hazard is not waste at all. A worker without savings, suddenly jobless, may grab the first job going not because it is a good match but because the rent is due — a fire sale of their own labour, forced by a cash constraint rather than chosen. Unemployment benefits relax that constraint, and some of the extra time they buy is spent not shirking but searching properly, finding work that actually fits. Chetty showed that this liquidity effect — insurance letting the constrained worker afford to search well — accounts for a large share of the behavioural response the crude view dismisses as moral hazard. Insurance does not only cushion the fall; it can improve the match made on the way back up.¹⁰

Step back from the machinery, and the plate's place in the volume comes clear. Everything here has been about the floor of the bargaining range — the reservation wage, what sets it, how a society lifts it. But the floor is never merely a number. It is the worker's walk-away point, and the walk-away point is the source of the worker's power.

This is the thread that runs through the whole volume. A worker who can refuse a bad bargain — because they have savings, or insurance, or a public floor beneath them — has leverage in the bargain; a worker who cannot refuse, who must take this wage or go hungry, has none. Raise the floor and you raise the worker's power: their reservation wage climbs, the share of the match surplus they can claim climbs with it, and the monopsony markdown of the eighth plate — which feeds on the worker's inability to walk away — shrinks. The safety net, in this light, is not only consumption smoothing and not only charity. It is a power-shifting institution, kin to the union of the ninth plate: where the union gathers many workers so that each can afford to say no together, insurance equips each worker to afford to say no alone.¹¹

And so the modern floor turns out to be, in the deepest sense, a choice — and a choice about power. The classical floor was fixed by nature: subsistence was subsistence, and no one decided it. The modern floor is fixed by policy — how generous the benefit, how long it lasts, how easily it is claimed — and every such decision is, whatever else it is, a decision about how much the worker may refuse, about where in the contest the wage records a society chooses to stand.¹² Reasonable people weigh the tradeoff differently, and the right point is genuinely contested: too low a floor leaves workers at the mercy of whoever is hiring; too high a floor can price the match away entirely, and a job that never forms pays no wage at all.¹³ But the volume's claim is not about where to set the floor. It is about what the floor is. It is the worker's power to walk away, rendered in money — and the wage, as ever, is decided in the space that power opens.