The argument needs a number now, and the number has to be handled without rhetoric, because this particular subject attracts rhetoric.
A monetary system that targets a modest positive rate of inflation is not designed to maintain the purchasing capacity of an individual currency unit indefinitely. That is not an accusation. It is a description of the target.
For the horizon that matters to preservation, the arithmetic runs like this. At a sustained two percent, one unit retains approximately forty-five percent of its purchasing capacity after forty years. At three percent, approximately thirty-one percent. At four percent, approximately twenty-one percent.
Four things have to be said about those figures at once, so they cannot be misread.
They describe the purchasing capacity of an uninvested currency unit under hypothetical sustained inflation at the stated rates. They do not describe the worker's financial outcome, which depends on everything the worker does with the unit. They describe targets rather than realized history, which has been both gentler and harsher than the target in different decades. And they measure capacity against a targeted general consumption basket, not against the worker's own liability, which this framework says is the real benchmark.
The arithmetic makes exactly one point. A forty-year working life is long enough that the design property of the unit, its intended slow decay, is material at the horizon where deferred compensation is finally spent. Any theory of compensation that treats the unit as constant across that horizon is treating a variable as a constant.
The Response, and Why It Answers a Different Question
Now the conventional response at full strength, because it is substantially correct.
Nobody advises a worker to hold forty years of retirement savings in cash. Productive assets have historically generated real returns well above inflation across most long horizons. Interest-bearing instruments compensate their holders for expected inflation. Pensions pool longevity risk and, when indexed, inflation risk. Inflation-linked government securities solve the general basket problem about as completely as anything can. Portfolio theory exists to construct exposures whose expected real return exceeds the unit's decay with acceptable variance.
Every part of that is right, and a worker who followed it has, on the record, done well.
But it answers a different question from the one I am asking.
It answers: given that the unit decays, what should the worker do? Convert.
My question is one level down. Why must the worker transform compensation into another form merely to preserve the purchasing capacity it represents?
The conventional response takes the conversion as the natural order of things, and in the current architecture it is. But it is an architectural fact rather than a natural one, with consequences the response does not price: the preservation-motivated component of risk, the professional function transferred to the individual, the behavioral failures at every conversion boundary, and the fee drag levied on the conversion itself.
None of that proves the monetary system defective. Each is a cost the separation imposes, and costs belong in the accounting even when the arrangement that imposes them is, on balance, justified.
Every Price Is a Ratio
One observation dissolves a confusion that runs through most discussion of returns.
When it is said that equities rose ten percent, what is meant is that equities measured in a particular currency rose ten percent against that currency. The figure describes a relationship between a numerator, the asset, and a denominator, the unit, and it changes when either changes.
Nominal appreciation partly describes the asset and partly describes the unit the asset is measured in, and no reported return can be interpreted without knowing how much of it is which.
This confusion is not confined to households. Modigliani and Cohn argued that equity markets themselves had mispriced under inflation, because investors discounted real cash flows at nominal rates. Money illusion operating at the level of the market rather than the payslip.
The second paper used this diagnostically, pricing salaries in houses and portfolios in years of wages, and drew a deliberately modest conclusion: no traditional numeraire has an unimpeachable claim to be the measuring stick, and the economically relevant benchmark is the liability itself.
That conclusion stands. It does not follow that gold, housing or any other asset is the correct denominator.
What follows is narrower and, for compensation, more consequential. Every salary, raise, pension promise, retirement target, contribution and deferred award implicitly contains a monetary denominator, and the properties of that denominator are part of what the compensation is.
A pension promise of fifty thousand dollars a year for life is a promise about a numerator. What it delivers depends on the denominator at every future date on which it is paid, and the promise is silent about the denominator because the architecture treats the unit as fixed.
The Question, and the Discipline for Answering It
Naming that produces what I regard as the central design question of the paper, and I want it posed before any candidate answer, so the answer cannot be accused of having written the question.
If a portion of compensation is intended to carry the economic value of finite human time across decades, what properties should the monetary denominator of that portion possess?
The discipline that protects the answer is this: the properties are derived from the second paper's Six Tests, applied to the monetary layer, and from nowhere else. Not from monetary intuition. Not from any candidate's characteristics. Each test, asked of the unit rather than of the architecture as a whole, yields one property.
Arrival asks whether deferred value credibly reaches the horizon with the purchasing capacity the liability requires. Asked of the denominator: the unit should hold, or change predictably, in purchasing capacity across the relevant horizon, so that a promise stated in it means at delivery roughly what it meant at issue.
Assignment asks which risks exist, who bears them, and how the architecture transforms them. Asked of the denominator: the sources of the unit's purchasing-power uncertainty should be identifiable and, where possible, few, so the risk can be named and assigned rather than absorbed unnoticed. And where a source of uncertainty is a decision by an issuer whose incentives may diverge from the holder's, the rules governing that decision should be known and its exercise disciplined. Note what this asks for. Known rules, not the absence of discretion. A discretionary issuer operating under a credible, transparent and disciplined rule satisfies it as well as a fixed schedule does.
Ownership asks what the worker actually holds and what must remain true for the claim to pay. Asked of the denominator: the unit should be held as property rather than as a promise at the value layer, or, where that is not possible, its dependence on the continued solvency and good faith of a named counterparty should be minimal, because a claim's horizon is its promisor's horizon and the worker's career outlives promisors.
Journey asks whether the architecture survives the career. Asked of the denominator: the unit should persist across the employers, providers, jurisdictions and decades the career will cross, so the denominator of deferred compensation does not change with the worker's address or the provider's fortunes.
Human asks whether an ordinary person can use the architecture without unrealistic expertise or behavior. Asked of the denominator: the unit should be usable, storable and defensible by an ordinary worker with ordinary attention, or capable of being held in a governed form that meets that standard.
Conversion asks whether accumulated value can become the consumption the liability requires. Asked of the denominator: the unit should be convertible, at the horizon, into the goods and services the liability is actually denominated in, at a price that does not itself impose the sequence risk the conversion was meant to escape.
One Caution, and What Is Not on the List
A caution about the second property, because the later articles lean on it and it is easy to misread.
Issuer discretion over supply is one source of purchasing-power uncertainty. It is not the only source and not obviously the largest. Purchasing power is moved by productivity, monetary demand, demographics, fiscal policy, credit conditions, supply shocks, technological change, relative prices and global capital flows, and a unit whose supply was fixed by rule would remain fully exposed to every one of those.
Constraining discretion narrows the list of sources. It does not shorten the list to one, and any candidate that constrains discretion should be evaluated for what it does to the remaining sources rather than credited with having solved the problem.
Notice also what the specification does not include.
It does not require appreciation. A monetary denominator for deferred compensation need not increase purchasing capacity; the relevant property is the ability to preserve, predictably relate to, or be reliably matched against the liability it funds. The framework is not looking for the money with the highest return. It is looking for properties appropriate to a temporal liability.
It does not include stability at every horizon, because which horizon matters depends on the tranche being funded.
And it includes no property that only one candidate could satisfy.
Everything Available Fails Somewhere
Run the existing instruments through it.
Inflation-indexed government bonds satisfy Arrival for the general basket and Human when held inside a plan. They fail Ownership, being promises of a sovereign, and satisfy Assignment only partly, since the issuer that indexes them is the issuer whose fiscal position they depend on.
Gold satisfies Ownership and Assignment's constraint on discretion. It fails Arrival across the horizons workers actually face, as Jastram's four centuries of data showed, and strains Human and Conversion in physical form.
National currencies in interest-bearing form satisfy Conversion and Human completely, and Journey within their jurisdictions. They fail Arrival by design and Ownership by nature.
Diversified productive capital, which is not a denominator but the standard remedy for the denominator's decay, satisfies Arrival on average and fails it in sequence, and passes Ownership only in the qualified sense the second paper gave to equity.
The specification is a test that everything currently available fails somewhere.
That is the state of the problem. Next week the series turns to a recent monetary experiment that was designed, whatever else may be said about it, with several of these properties in mind, and which has to be held to all six.