Four articles in, this argument needs a discipline applied to it, and the discipline is a question I should have to answer in public.
If the monetary unit had zero expected inflation, and its safest interest-bearing forms delivered a reliable zero real return, would workers still require preservation architecture?
The answer is plainly yes.
They would still face longevity. Disability. Sequence risk on whatever growth assets they held. Uncertain future consumption. The behavioral failures the second paper documented at length. The need for pooling, which requires promises. The intergenerational transfer problem. Career fragmentation and portability. Productive-investment decisions. Taxes. Institutional failure. And the conversion of a stock into an income at the horizon.
Every one of the Six Tests would still apply, and most of them would return the same answer.
Preservation architecture is not a consequence of monetary decay, and nothing in this series attributes the preservation problem to the money. Anyone reading it that way is reading it wrong, and I would rather draw the boundary myself than let someone else draw it for me.
So What Does the Money Actually Add?
Three things. Only these.
A preservation burden that can rise above zero. Under stable money the worker could stand still by holding the safest claims. Under the conditions described last week, sometimes they cannot, which is why preservation-motivated investment exists at all.
Purchasing-power uncertainty within the unit. A worker holding a perfectly diversified portfolio of nominal claims still bears the risk that the unit's future purchasing capacity differs from expectation, and that risk cannot be reduced by holding more of the same unit. It is not diversifiable from inside.
A denominator problem in measurement. Under stable money, nominal and real compensation coincide, the ruler holds still, and the profession's habit of measuring in the unit costs nothing. Under decaying or uncertain money, the habit conceals a variable.
Those three are the whole of this paper's addition to the preservation problem. They are enough to make the money a design variable. They are not the preservation problem itself, and the rest of the series should be read with that boundary in view.
The counterfactual carries one further implication, which I record without resolving. If a unit whose safest claims reliably maintained liability-relative purchasing capacity would remove the preservation-motivated component, then the size and complexity of the architecture are themselves partly contingent on the unit's properties. Such a unit would not eliminate longevity, pooling, conversion, governance or behavior, and the architecture would remain necessary as a whole. But it could reduce the financial transformation required merely for preservation, and the stack might become simpler at some layers.
This paper evaluates candidates against the architecture as it stands, because today's liabilities fall due inside it. It records the contingency and leaves it there.
Now the Defense
Any paper that arrives at a criticism of the monetary unit owes the unit its strongest defense first. The defense of elastic, discretionary money is strong, and it should be stated in the vocabulary of the economists who built it.
Elastic money exists because rigid money failed in ways that were catastrophic and well documented, and because monetary theory concluded a managed money could do better than a commodity anchor. Friedman's own position was that a rule-following authority would outperform gold.
Under metallic standards, when the demand for money rose faster than the commodity stock, prices fell. Falling prices under fixed nominal debts raised the real weight of every loan, and bankruptcies, bank failures and further contraction followed. Fisher formalized the debt-deflation spiral. Friedman and Schwartz documented that the Great Depression's severity was inseparable from a monetary contraction the authorities failed to reverse. Eichengreen showed that the interwar gold standard transmitted deflation internationally and constrained the response, so that the countries which abandoned gold earliest recovered soonest. And Bagehot's doctrine, that a central bank must lend freely against good collateral in a panic, presupposes a bank that can create the reserves to lend.
Modern discretionary central banking is a direct answer to that record. It provides monetary-policy flexibility, so a demand shock can be met with easier money rather than unemployment. Countercyclical response. Credit elasticity, so the payments system expands and contracts with the real economy rather than with mining output. Crisis management, demonstrated in 2008 and 2020. And nominal adjustment, the capacity of modest inflation to let relative prices and real wages move without the nominal cuts that money illusion makes so costly.
The Defense Made on Compensation's Own Terms
That last capability deserves particular attention in a compensation paper, because it is a defense of positive inflation built out of facts about pay.
Akerlof, Dickens and Perry argued that because workers resist nominal wage reductions, a small positive rate of inflation allows real wages in declining firms and sectors to fall without the layoffs that would otherwise be required. Inflation, on this view, greases the wheels of the labor market.
The argument is contested in its magnitudes but not in its mechanism, and the mechanism is a fact about compensation. Positive inflation exists partly because of how pay behaves.
A compensation theorist who criticizes the decay of the unit has to reckon with the possibility that the decay is, in part, a service rendered to compensation's own rigidities.
I do not resolve that in either direction. I note only that a service rendered to the labor market at the moment of employment is paid for by whoever holds nominal claims across the decades afterward, and that the payer and the beneficiary are frequently the same worker at different ages.
The Same Person at Different Ages
That observation is where this connects to the theory the trilogy has been building, and it is the idea in this article I would most want to survive.
Elastic monetary architecture can benefit the worker-as-employee, whose job is protected in a downturn by the countercyclical response and whose real wage can adjust without a nominal cut. It can simultaneously impose uncertainty on the worker-as-saver, whose deferred compensation is denominated in the unit whose flexibility protected the job.
The employee, the debtor whose mortgage is eroded by inflation, the saver, the investor, the pensioner and the retiree are not different people with competing interests.
They are one person at different points in time.
A monetary architecture that serves that person at one age and burdens them at another is redistributing within a single life. Whether the redistribution is on balance favorable depends on the person's debt, saving, career, and the realized path of the regime, and I make no claim about the sign.
What compensation theory can say is that the trade is intra-personal and inter-temporal, that it runs through the unit in which pay is denominated, and that it is invisible in every instrument the profession uses to describe what a worker was paid.
Call it intertemporal monetary incidence: the same monetary architecture imposes different benefits, costs and risks on the same individual at different stages of life, and compensation, being a transfer through time, is the channel through which the difference travels.
This is additional to the conventional distributional analysis of who gains and loses across different people. What it adds is that compensation itself crosses the stages across which the incidence changes.
Applying the Principle to the Money
Now apply the second paper's principle. Risk-bearing is an architectural choice even when nobody consciously makes it, and every architecture reduces some risks while assigning the residual.
Elastic monetary architecture genuinely reduces the risk of deflationary collapse, of liquidity crisis, of nominal rigidity producing mass unemployment, and of a payments system hostage to a commodity's supply. The twentieth century's second half was, for advanced economies, materially more stable than its first in ways that owe a great deal to those reductions. That is a real achievement and this argument does not diminish it.
The residual risk the architecture assigns is purchasing-power uncertainty for holders of long-duration nominal claims. Under a discretionary standard, the future purchasing capacity of a unit depends on future policy, shocks and political constraints that the holder neither controls nor fully anticipates. And the second paper's price-level asymmetry means that under inflation targeting a loss, once realized, is not restored by design.
The bearer of that residual is whoever holds nominal claims across time. Bondholders. Depositors. Pensioners with unindexed benefits. And workers whose deferred compensation is denominated in the unit.
The point is not that the assignment is wrong.
It is that it is an assignment. It lands on the holder of long-duration claims. It is nowhere reported in the compensation architecture that produces those claims. And it therefore belongs inside preservation architecture as one of the risks the Assignment Test has to name.
Next week: every compensation figure carries a denominator, so what properties should that denominator have? The answer gets derived from the Six Tests, before any candidate is named, so the answer cannot be accused of having written the question.