Seigniorage

A medieval fee, modernised

The word comes from the Old French seigneuriage: the fee the local seignior, the feudal lord with a minting monopoly, charged for stamping his image on a piece of bullion and certifying it as legal coin. The merchant brought in a gram of silver; the seignior gave him back a coin worth a gram less the seignior's cut. The cut paid for the mint and the army, and it bought the lord a quiet, steady, slow-burning revenue that did not require taxing anyone visibly. It was the original line item in the public-finance ledger that no peasant would notice — until the lord debased the coinage, mixing in copper, and the fee became a tax on every silver coin already in circulation.

Modern seigniorage works the same way, dressed in different clothes. A central bank creates new money — historically by stamping notes, today by crediting reserves to commercial-bank accounts — and uses that newly-created liability to acquire real assets, almost always government bonds. The bonds earn interest. The bank's own newly-created liabilities pay either zero (currency in circulation) or a policy rate (reserves) that is typically below the bond yield. The spread, multiplied by the size of the balance sheet, is the modern central bank's annual operating revenue. Net of expenses, it flows to the government's treasury. This is seigniorage as a steady flow, in a low-and-stable-inflation economy: a quiet stream of revenue from a citizen's preference for holding currency over equivalent-value goods.

The accounting definition

The textbook formula for seigniorage is

S = ΔM / P

where M is the nominal stock of base money (currency in circulation plus bank reserves at the central bank) and P is the price level. The numerator is the change in money issued; dividing by P expresses it in real, constant-purchasing-power units. If the central bank issues $100 billion of new money in a year and the price level is 1.0, real seigniorage is $100 billion. If during the year prices rose by 2 percent, the issuer's revenue measured in end-of-period prices is slightly less: ΔM/P̄, with P̄ an average.

This is the flow definition. There is also a balance-sheet definition that asks what the government has accumulated: the present value of all real revenue extracted from money creation. Under reasonable assumptions, the present value of real seigniorage equals the real value of the monetary base outstanding — the capitalised value of running an interest-free liability. The two views agree in steady state.

A useful decomposition splits seigniorage into two pieces:

S = ΔM/P = π · (M/P)        + Δ(M/P)
            ↑                     ↑
       inflation tax          real-balance growth
       (on existing M)        (issuance to satisfy
                              rising real demand)

The first term is the inflation tax: existing real money balances lose value at rate π, and the central bank captures that loss by issuing replacement nominal money. The second term is non-tax issuance: when the economy grows and the public wants to hold larger real balances, the central bank can satisfy that demand by issuing new money without raising inflation. In low-inflation industrialised economies the second term dominates; in high-inflation economies the first term dominates and increasingly cannibalises the second as the public flees money holdings.

The Fed as a case study

In a typical recent year — say, fiscal 2019 — the Federal Reserve's income statement looked roughly like this:

Line	Amount	Source
Interest income	+ $103 B	Treasury and MBS portfolio (~$3.9 T)
Interest expense	− $36 B	Interest on excess reserves + ON RRP
Other income/expense	− $4 B	Operations
Net income	+ $63 B
Member-bank dividends	− $0.7 B	Statutory 6% on paid-in capital
Treasury remittance	+ $62 B	Real seigniorage flow

The Treasury remittance line is the visible part of US seigniorage. It exists because the Fed funded that $3.9 T portfolio entirely with two liabilities: about $1.7 T of currency in circulation (zero interest) and $2 T of bank reserves (then paying about 1.5 percent, well below the Fed's portfolio yield of ~2.5 percent). Multiply the spread by the balance sheet and you get the net income; subtract operating costs and dividends and you get the remittance.

In 2021 the balance sheet was vastly bigger (around $8.3 T, post-QE-pandemic expansion) and short rates were near zero. Interest expense was negligible, interest income was $122 B, operating costs and dividends were ~$5 B, and remittances peaked at about $107 B. Then 2022-2024 brought rate hikes that reversed the picture: the Fed paid 5 percent on roughly $3.2 T of reserves while its portfolio yielded around 2.5 percent. Interest expense exceeded interest income; remittances went to zero. The Fed continued to operate — it has no capital constraint and can run an indefinite negative-net-worth position — but recorded the cumulative shortfall as a "deferred asset" to be netted against future positive net income. By early 2025 the deferred asset stood at over $200 billion. Real seigniorage to the Treasury was, in those years, negative.

This episode is informative. Seigniorage is not simply "money the government printed" — it is the carry trade between the bank's assets and liabilities, and the carry can go negative when rates rise faster than the portfolio yield rises. The legal structure of central banks lets them ride out such episodes, but in real economic terms there are years when the issuer pays interest on its money instead of earning seigniorage from it.

Currency itself: 9.4 cents per $100 note

A subtler, slower piece of seigniorage comes from currency in circulation — physical notes and coins. The Bureau of Engraving and Printing's reported unit cost for a Federal Reserve Note is 7.7 cents for a $1 bill and 9.4 cents for a $100 bill (FY 2024 figures), including paper, ink, security features, printing, and shipping. When the Fed issues a freshly printed $100 note in exchange for the Treasury removing $100 of bonds from its portfolio, the Fed is now financing a $99.91 reduction in its interest-bearing liabilities with a $0.09 production-cost banknote. Over the life of that note — typically 22 years for a $100 — the Fed captures the foregone interest on $99.91 of avoided Treasury borrowing.

For high-denomination notes held abroad — which dominate the $100 stock — that's effectively a permanent interest-free loan from the foreign holder to the US government. The Federal Reserve Bank of New York estimates that roughly 60 percent of all US currency in circulation, by value, is held outside the United States, totalling around $1.4 trillion. Even at 2 percent real yields, that's $28 billion a year of additional implicit seigniorage on top of the formally remitted figure. This is the financial component of what Giscard d'Estaing called the "exorbitant privilege" — the dollar's reserve-currency status converted into ongoing free financing.

The Cagan Laffer curve

Phillip Cagan's 1956 paper "The Monetary Dynamics of Hyperinflation" introduced the framework that still organises the seigniorage literature. He modelled the demand for real money balances as a decreasing function of expected inflation:

m_d = M/P = exp(−α π^e)        (real money demand)

where π^e is the expected rate of inflation and α > 0 is the semi-elasticity. The intuition is direct: when people expect prices to rise, they hold less currency because currency loses purchasing power at rate π^e. In steady state with π^e = π, real seigniorage equals the inflation rate times real money holdings:

R(π) = π · m_d(π) = π · exp(−α π)

This is a hump-shaped function of π. Take the derivative:

dR/dπ = exp(−α π) · (1 − α π)

Setting this to zero gives the revenue-maximising inflation rate:

π* = 1/α

For typical estimated α values (5 to 10 per annum in low-inflation economies, lower at high inflation as habits adjust), the implied revenue-maximising inflation is 10-20 percent. The maximum revenue is

R_max = π* · m_d(π*) = (1/α) · e^(−1) ≈ 0.37 / α

which for plausible α gives a peak seigniorage revenue of 3-7 percent of GDP — substantial, but not enough to finance a runaway fiscal deficit. Crucially, beyond π*, more printing produces less real revenue: the contraction of real balances outpaces the increase in nominal issuance. The issuer is on the wrong side of the curve. The only way to extract more is to inflate faster — which only shrinks real balances further. The runaway dynamics begin.

When the Laffer ceiling meets fiscal need

Every well-documented hyperinflation has been a fiscal crisis dressed as a monetary one. The government had a deficit it could not finance from taxation or borrowing, so it issued money to cover the gap; once nominal issuance growth pushed π past π*, real seigniorage started falling even as nominal printing accelerated, and the only way to maintain real revenue was to print exponentially faster. The chase between issuance and the falling tax base is what produces the characteristic doubling-every-few-days inflation rates of late hyperinflation.

Episode	Years	Fiscal trigger	Peak inflation	Resolution
Weimar Germany	1921 – 1923	Versailles reparations	29,500 % / mo (Oct 1923)	Rentenmark, gold-asset backing (Nov 1923)
Hungary	1945 – 1946	WWII reconstruction	4.19 × 10¹⁶ % / mo (Jul 1946)	Forint introduced, currency reform
Yugoslavia	1992 – 1994	War financing, sanctions	313 million % / mo (Jan 1994)	"Super dinar" tied to DM, IMF aid
Zimbabwe	2007 – 2008	Land-reform-driven tax collapse	89.7 × 10²¹ % / mo (Nov 2008)	USD / multi-currency adoption (2009)
Venezuela	2016 – 2019	Oil-export collapse, PDVSA	~130,000 % / yr (2018)	Partial dollarisation, no formal reform

The Zimbabwean episode is the clearest modern case. From 2002 to 2008, the Reserve Bank of Zimbabwe printed money to fund government spending after the tax base collapsed in the wake of the post-2000 land reforms. As inflation accelerated, the population fled the Zimbabwe dollar — first into dollarised parallel markets, then into commodities, then into not holding currency at all. Real money balances fell so fast that the central bank's revenue from seigniorage was a small fraction of nominal printing. The note denominations spiral is the visible signature of this dynamic: in 2007 the largest note was Z$200,000; by January 2009 it was Z$100 trillion, with the note's purchasing power lower than the Z$200,000 note had been two years earlier. The currency was abandoned in 2009 and the country operated on US dollars and South African rand for nearly a decade.

Weimar Germany's path is structurally identical, separated by a century. War reparations forced fiscal deficits that the Reichsbank monetised; once inflation expectations adjusted, real money demand collapsed; the printing rate had to accelerate to keep real revenue constant; the dynamics broke down by autumn 1923, and the regime was only stabilised in November when Hjalmar Schacht introduced the Rentenmark — a new currency formally backed by mortgages on agricultural and industrial land — and the Reichsbank stopped monetising the deficit. The fiscal problem had to be solved (briefly, by the Dawes Plan reparations restructuring) before the monetary problem could be solved.

The lesson, from Cagan to Sargent and Wallace's "Some Unpleasant Monetarist Arithmetic" (1981) to modern IMF retrospective work, is invariant: seigniorage is a real revenue source, but a bounded one. A government that needs more than ~5 percent of GDP from money creation cannot get it durably — and trying anyway produces hyperinflation, not more revenue.

How the eurozone shares the take

The euro is structurally unusual: a single currency with no single fiscal authority. Seigniorage from euro issuance is generated at the ECB and the national central banks (NCBs) of the Eurosystem, but the political question of who gets the money is sensitive. Two adjustments matter.

First, the banknote-allocation key. Euro banknotes are physically issued by various NCBs (the BBk in Germany, the Banque de France, the Bank of Italy, etc.), and historically Germany has issued and held a disproportionate share simply because its banking system is the largest. Distributing seigniorage in proportion to physical issuance would hand Germany an unfair share — much of the cash issued in Germany ends up circulating throughout the euro area. The Eurosystem therefore allocates the income from currency-in-circulation to the NCBs in proportion to the capital key: each NCB's paid-in share at the ECB, itself a weighted average of the country's share in EU population and EU GDP. As of 2024, the largest capital-key shares are Germany 21.4 percent, France 16.6 percent, Italy 13.8 percent, Spain 9.7 percent, Netherlands 4.8 percent.

Second, the monetary policy income pool. Income from monetary-policy securities holdings (PSPP, PEPP) is pooled and redistributed by the same capital key. Net of the ECB's own operating costs, each NCB receives its key share and typically passes it through to its national treasury. Greek seigniorage on PSPP-held Greek bonds is therefore not retained by Greece; it flows into the pool and is redistributed by capital key — a mechanism known as Pooled Monetary Income (PMI). This is part of why German taxpayers eventually receive a slice of the income on Italian or Greek bonds and vice versa, and is one of the few structural mechanisms of fiscal-political transfer in a fiscally fragmented monetary union.

Crypto, CBDCs, and the future of seigniorage

Two structural forces threaten the seigniorage flows central banks have come to rely on. Both are visible but neither is yet dominant.

First, cryptocurrencies. Bitcoin, stablecoins, and other crypto-assets substitute for some of the demand for central-bank money. To the extent a US household holds $5,000 in a USDC stablecoin balance instead of $5,000 in $100 notes, the Fed loses ~$100/year of seigniorage on that float and Circle (USDC's issuer) captures it. Aggregate effect: at the scale of crypto adoption circa 2024-2025, US Fed seigniorage erosion is estimated in the single-digit billions per year — measurable but small relative to the ~$100 B remittance line. The largest mechanism is not retail crypto-holdings but stablecoin reserve-management: stablecoin issuers hold short-term Treasuries to back their tokens, capturing what would otherwise be the Fed's net interest margin.

Second, central-bank digital currencies. A retail CBDC is the central bank's own digital liability — functionally equivalent to currency in circulation but in digital form. From the bank's side, every CBDC unit issued is a new interest-free (or low-interest) liability on the bank, financed against the same kinds of assets that finance currency-in-circulation today. So CBDC does not by itself erode seigniorage. What it can do is reshape the deposit base of commercial banks: if households move savings from commercial-bank demand deposits into CBDC, commercial banks shrink and the central bank balance sheet grows. Most CBDC designs therefore include holding caps (typically €3,000-€10,000 retail tier in the proposed digital euro) precisely to prevent that displacement.

The structural threat to long-run seigniorage is therefore not crypto and not CBDC, but cashlessness in general. Every dollar of consumption that moves to a card, an app, or a payment rail is a dollar that does not need to sit in a wallet. The float of zero-interest currency shrinks; the natural base of seigniorage shrinks with it. In Sweden, where cash use has fallen to ~10 percent of point-of-sale transactions by value, the Riksbank's currency-in-circulation has fallen by ~50 percent since 2007 in nominal terms. The Riksbank's seigniorage from currency is now a fraction of what it was, and its proposal for an e-krona is partly a reclamation play — a digital replacement for the cash that is no longer being held.

Why seigniorage matters for public finance

Seigniorage sits in an awkward place in the public-finance picture. It is real revenue, but it is not a tax that anyone votes on; it does not appear as a line item in the federal budget; and yet it is responsive to monetary policy in ways that orthodox taxes are not. Three implications.

First, fiscal arithmetic. Standard sustainability analyses ask whether the present value of primary surpluses, plus seigniorage, covers the debt. In most industrial economies seigniorage is ~0.4 percent of GDP — small enough to be a rounding error compared with debt of 100 percent of GDP. But in emerging economies running 10-30 percent inflation, seigniorage can be 2-5 percent of GDP and is a serious financing source. The Sargent-Wallace point is that this is unstable: relying on seigniorage past the Laffer peak is impossible, and reliance even below the peak creates inflation that distorts other parts of the economy.

Second, monetary policy as fiscal policy. Quantitative easing is, in budgetary terms, a seigniorage-extracting operation. The Fed creates reserves (a liability paying the policy rate) and buys long-duration Treasuries (an asset paying the long rate). The carry — long minus short — is captured as Fed net income and remitted. QE during 2020-2021 produced record remittances; QT and the 2022-2024 rate-hike cycle reversed them. Decisions framed as monetary-policy choices have direct fiscal-revenue consequences that the legislative branch does not vote on.

Third, reserve-currency status as an asset. The US holds an effective tax base of foreigners willing to lend interest-free to the Treasury by holding dollars. That base is roughly 1.4 trillion dollars in physical currency abroad and several trillion in foreign-held Treasuries paid below-market real yields. The capitalised value of this privilege is in the high single-digit percentage points of US GDP — a structural advantage worth fighting to preserve. Most discussions of de-dollarisation and a hypothetical replacement reserve currency are, at heart, discussions about who collects this slice of global seigniorage.

Common pitfalls and misconceptions

• Confusing seigniorage with "printing money." Seigniorage is real revenue from monetary expansion; printing money is the mechanism. The two are connected but not identical — when real money demand grows naturally, you can print without raising inflation, and seigniorage equals real-balance growth, not inflation tax. Conflating the two leads to the false claim that all money creation is inflationary.
• Assuming the Fed "makes" $100 B by printing. The Fed's annual remittance is the net of interest income, interest expense, operating costs, and dividends — not a direct measure of money issued. In 2022-2024 the Fed expanded its balance sheet but recorded zero remittances because interest paid on reserves exceeded asset yields. Seigniorage flows are about carry, not about issuance volume.
• Treating seigniorage as unlimited. The Cagan Laffer curve places a hard ceiling on real revenue from money creation. Estimates put the maximum at 3-7 percent of GDP in steady state. Past that, more printing yields less real revenue. Any plan to finance a 10-percent-of-GDP deficit from seigniorage is mathematically impossible.
• Forgetting the Tanzi-Olivera effect. In high inflation, real tax revenue falls because there is a lag between when a tax is assessed and when it is paid; by the time the cheque clears, the real value has been eroded. So the higher the inflation rate, the more the government needs from seigniorage to compensate for lost real conventional tax revenue — a feedback loop that accelerates the run up the Cagan curve.
• Treating "exorbitant privilege" as free. The US captures additional seigniorage from foreign dollar holdings, but the privilege requires maintaining the dollar as a credible reserve currency. The costs — geopolitical commitments, financial-market openness, willingness to run current-account deficits — are real and large. The net value of reserve-currency status, after costs, is positive but not unbounded.
• Confusing fiat issuance with bank-deposit creation. When a commercial bank makes a loan, it creates a deposit — broad money grows — but the bank captures the lending margin, not the central bank. Only central-bank base-money creation generates seigniorage in the strict sense. Confusion of base money with broad money is a frequent source of inflated estimates.

Where seigniorage is going

Three forces will reshape seigniorage over the next decade. Higher long-run inflation targets at major central banks would expand the Cagan revenue base, but most central banks have signalled the opposite — a desire to keep π near 2 percent, leaving seigniorage as a small steady flow rather than a fiscal tool. The continued shift to cashless payments and stablecoin float will erode the no-interest-paid currency-in-circulation component, gradually shifting the central bank's funding base toward interest-bearing reserves and reducing the carry available. And the rollout of retail CBDCs — most likely in the EU, China, India and several smaller economies before 2030 — will not eliminate seigniorage but will move it from physical-cash carry to digital-account carry, with new design choices (holding caps, remuneration rules) determining the absolute scale.

The deeper question is whether the US dollar retains its reserve-currency monopoly through this transition. If a multi-polar reserve system emerges — with the euro and the renminbi each holding 20-30 percent shares and the dollar share falling to ~40 percent — the US loses a meaningful chunk of its implicit foreign-held seigniorage. The other side of the trade is the country that gains share. The numbers involved — tens of billions of dollars per year of foreign-financed government spending — are large enough that the geopolitics of currency choice is the geopolitics of seigniorage.