MMT

The public was taught that a government is a household with a printing press. Modern Monetary Theory says that picture reverses the order of operations for a country that issues its own currency.

MMT is a description of how fiat public finance already works, plus a policy claim about what that description allows. Its main developers are Warren Mosler, Bill Mitchell, L. Randall Wray, Stephanie Kelton, and Pavlina Tcherneva. The claim is not that deficits are free. It is that solvency is the wrong constraint, and inflation is the right one.

What the textbooks taught

Introductory macro, from Mankiw-style principles courses through AP Macro, still walks students through a sequence that looks like personal finance.

The government wants to spend. It collects taxes. If taxes fall short, it borrows by selling bonds. Borrowing adds to the demand for loanable funds, which raises the real interest rate, which crowds out private investment. The deficit is a stock of obligations that future taxpayers must service. “Printing money” to cover the gap is treated as a separate, inflationary last resort, governed by something like the quantity theory: too much money chasing too few goods.

That story has three load-bearing pieces.

The household analogy. A family must earn or borrow before it can spend, because it uses a currency someone else issues. Textbooks transfer that logic to Washington, London, or Tokyo.

Loanable funds and crowding out. Savings are a pool. Government borrowing competes with firms for that pool. Higher deficits mean higher rates and less private capital formation. The formula students memorize is simple: deficit up, demand for funds up, real rate up, investment down.

An exogenous money supply. Older chapters still teach a money multiplier. The central bank injects reserves, banks lend a fraction out, deposits multiply, and the money supply is largely a policy dial. Taxes and bond sales are how the treasury “gets” the money it then spends.

The moral that follows is familiar. Balanced budgets are prudent. Deficits are a burden shifted onto children. The binding question in any spending debate is “how will you pay for it?”

How MMT says the system works

MMT starts from a different institutional fact. A monetary sovereign issues the currency it taxes and spends in, owes its debts in that currency, and does not promise to convert it into gold or a foreign currency at a fixed rate. The United States, Japan, the United Kingdom, Canada, and Australia qualify. Texas does not. Greece, inside the euro, does not. A country that has dollarized, or pegged its currency, does not.

For that issuer, the sequence runs the other way.

Spending creates the dollars. When the Treasury spends, the Federal Reserve credits the reserve account of the recipient’s bank. New dollars appear as liabilities of the government and assets of the private sector. Nothing in that keystroke requires a prior pile of tax receipts. Kelton’s shorthand is that the issuer does not need to “get” dollars from taxpayers or from China. It is the source of them.

Taxes delete dollars. When a tax payment clears, reserves are debited and the dollars leave private circulation. In the MMT account, taxes do not fund spending. They create a permanent need for the government’s currency, because only that currency settles the tax bill, and they drain purchasing power so spending does not spill into inflation. Mosler’s line is that federal taxes do not pay for anything. They remove money after it has been spent.

Bonds are an interest-rate tool, not a funding necessity. After spending credits reserve accounts, banks hold more reserves than they want. Left alone, the overnight rate falls toward whatever the central bank pays on reserves, or toward zero. Selling Treasury securities drains those reserves and supports the policy rate. MMT treats bond issuance as monetary plumbing that accompanies deficits, not as the thing that makes the spending possible. A sovereign that can coordinate with its central bank cannot be forced into default on debt denominated in its own currency. It can always credit the accounts.

The real constraint is resources. Bridges, nurses, steel, electricity, and unemployed workers are finite. If the government buys more of them than the economy can supply at current prices, it bids prices up. MMT’s inflation story is closer to “spending past productive capacity” than to “the debt ratio crossed a line.” Spare capacity, in this view, is what makes a larger deficit safe. A hot economy with no slack is what makes it dangerous.

Sectoral balances make the accounting explicit. Following Wynne Godley, the private sector’s surplus, the government’s balance, and the foreign balance must sum to zero. In a country that runs a current-account deficit, the only way the private sector can net save is if the government runs a deficit. A government surplus, in that arithmetic, is a private-sector deficit. MMT treats the routine call for balanced budgets as a call for the non-government sector to lose financial assets.

The policy piece that usually travels with the theory is a job guarantee: a standing offer of public employment at a fixed wage. Advocates argue it anchors the price of labor, hires people the moment private demand falls, and releases them when private demand returns, so full employment and a nominal anchor arrive together.

Why “how will you pay for it?” is the wrong question

The sentence that ends most spending debates is not an economic question. It is a household question applied to an issuer of currency. A family asking “how will we pay for it?” is asking where the dollars will come from, because the family cannot create them. A monetary sovereign already can. Once that is granted, the sentence stops describing a constraint and starts hiding the one that matters.

What a government actually buys is not dollars. It buys hours of labor, kilowatt-hours, tons of steel, hospital beds, acres of land, and the output of factories that are either idle or already busy. Dollars are the ticket used to move those things. The ticket can be issued. The things cannot. Kelton’s reframe, and the one MMT keeps returning to, is that the relevant question is not how the spending will be financed but whether the real resources exist to absorb it without bidding up prices.

That is a harder question, which is why the pay-for version survives. “How will you pay for it?” has a clean answer: raise a tax, cut another program, or issue a bond. It can be scored by the Congressional Budget Office and settled in a committee room. “Do we have the welders, the transformers, the concrete plants, and the unused workers?” cannot. It requires looking at the economy rather than the ledger. A country can write a check for a high-speed rail line on a Tuesday and still fail to build it for a decade, because the constraint was engineering capacity, rights of way, and skilled labor, not the Treasury’s balance.

The same logic runs in reverse, and this is the part textbooks rarely dwell on. Leaving resources idle is also a cost. An unemployed electrician, a shuttered mill, and a half-empty freight yard are productive capacity the economy already paid to create and is now refusing to use. In the MMT account, a deficit spent into that slack does not “take” resources from the private sector. It activates resources the private sector was not using. The pay-for question treats every new public outlay as if the economy were already at full employment, which is the special case, not the normal one. Crowding out is real when the welders are all busy. It is a diagram about a situation that often does not obtain.

Inflation is what it looks like when the question was skipped. If the government, households, and firms all try to buy the same scarce inputs at once, prices rise. That is the signal that nominal spending has outrun real capacity. The correct response, on this view, is to release the pressure: tax purchasing power back out, delay the project, or free the bottleneck, not to discover that the government has “run out of money.” The 2021–22 episode is the illustration critics and proponents fight over. Demand was pushed up while ports, chips, energy, and labor supply were constrained. Prices moved. The scarce thing was not dollars. It was goods and workers. A pay-for debate conducted in 2019 would not have forecast a semiconductor shortage.

None of this makes public spending wise. Resources used for one thing are unavailable for another, and a government can waste them just as a firm can. A bridge to nowhere consumes steel and labor that could have gone into a bridge to somewhere, or into private construction. The opportunity cost is real. It is just not a financial cost in the household sense. It is a use-of-capacity cost. MMT’s claim is that collapsing that distinction into “how will you pay for it?” produces two recurring errors: austerity when people and plants are idle, and complacency when they are not, because the ledger still looks fine.

So the sequence MMT wants in place of the textbook question is short. What is the real goal. What resources does it require. Are those resources idle, available at a tolerable price, or already claimed. If they are idle, the financial operation is the easy part. If they are claimed, no financing scheme repairs the shortage. The dollars were never the limit.

Debt and taxes

Textbooks treat the national debt as a family credit card and taxes as the paycheck that pays it down. The government spends more than it collects, borrows the difference, and hands the bill to future taxpayers. Interest compounds. The debt-to-GDP ratio becomes a danger gauge. The responsible move is to raise taxes or cut spending until the ledger balances. That picture is coherent for a household, a firm, or a state government. It is the wrong picture for the issuer of the currency.

Under MMT, a federal deficit is not money the government has failed to raise. It is money the government has created and left in private hands. When the Treasury spends $100 and taxes back $80, the private sector is holding the remaining $20 as bank deposits or, after the usual reserve drain, as Treasury securities. The “debt” is the record of dollars that were spent and not yet taxed back. It is a liability of the government and an asset of whoever holds it: households, pension funds, banks, foreign central banks. Retiring it means deleting those private assets, either by taxing them away or by the government spending less than it taxes. A campaign to “pay off the national debt” is a campaign to drain net financial savings from the non-government sector.

That is why MMT treats the debt-to-GDP ratio as a poor target. Japan has carried a very high ratio for decades without a funding crisis or a spike in yields, because the debt is in yen and the Bank of Japan sets the yen interest rate. Greece suffered a genuine funding crisis because its debts were in euros, a currency it does not issue. The textbook habit of ranking countries by debt ratios mixes currency issuers and currency users into one list, then draws one moral from both. For an issuer, the binding questions about the debt are different. Who holds it. In what currency is it denominated. What interest rate does the central bank choose to support. Is the spending that created it still sitting inside the economy’s productive capacity.

Interest is not a solvency test either, in this account. The government pays interest by crediting the same reserve accounts it credits for any other spending. A higher rate raises private incomes for bondholders and, if large enough relative to slack, can itself add to demand. It does not bring the issuer closer to an empty account. The political choice is whether that interest income is a transfer the country wants to make, not whether the country can “afford” it.

Taxes, on the MMT account, do three jobs, and funding is not one of them.

They create demand for the currency. People accept dollars because taxes, fees, and court judgments are due in dollars. That is the chartalist core of the theory: the currency has a floor under it because the issuer imposes an obligation only that currency can settle. Without that obligation, acceptance is a convention. With it, acceptance is a legal requirement for anyone who owes the state.

They remove spending power. After the government has spent dollars into the economy, taxes delete a portion of them so the remaining purchasing power does not chase a fixed supply of goods. A tax increase in a hot economy is an inflation tool. A tax cut in a slack economy is a demand tool. The size of the tax bill relative to spending is a thermostat setting, not a receipt for services rendered.

They shape the distribution of income and activity. A tax on high incomes, on carbon, on land, or on imported goods moves resources and wealth even though it is not filling a Treasury vault. MMT does not claim taxes are pointless. It claims their point is demand management, distribution, and behavior, and that judging them by how much “revenue” they raise answers a question the issuer does not face.

The textbook sequence runs the other way. Taxes are collected first. Spending is limited to what those taxes, plus prudent borrowing, will allow. A proposed program is serious only if it has a pay-for. Under MMT that sequence is backwards as a description of operations and misleading as a test of affordability. The government spends by marking up accounts, then taxes by marking them down. A program can be fully “paid for” on paper and still be inflationary if the workers and materials are already employed. It can have no pay-for at all and still be non-inflationary if the resources are idle. The debt that results is the accounting residue of the second case, held by the private sector as savings.

None of this makes the debt irrelevant. A large stock of interest-bearing liabilities concentrates income among bondholders. Foreign holders earn a claim on future American output. And if spending really has outrun capacity, the inflation shows up whether or not a matching tax was written into the bill. The error MMT points to is treating the debt as a sum the country must someday scrape together in dollars it already issues, and treating taxes as the scrape.

Where the two stories actually diverge

The disagreement is not about arithmetic identities. Gregory Mankiw, in a 2019 skeptic’s guide, granted that a sovereign issuer cannot be forced to default in its own currency. The fight is over what follows from that.

Textbooks say the government is financially constrained and must tax or borrow before it spends. MMT says the constraint is inflation and real capacity, and that tax-and-borrow is a story retrofitted onto operations that already run in the other order.

Textbooks say deficits raise rates and crowd out investment through loanable funds. MMT says the central bank sets the policy rate, and that large deficits have often coincided with low rates because the deficit itself supplies the savings the private sector holds as Treasuries. Japan’s long period of high debt and near-zero yields is the example advocates reach for.

Textbooks treat “printing money” as a distinct, reckless alternative to borrowing. MMT says spending already creates money, and bond sales merely swap one government liability (reserves) for another (securities). The inflation risk is in the spending relative to capacity, not in which liability is left outstanding.

Textbooks teach the public to ask who pays. MMT wants the public to ask what is idle. The first question produces pay-fors, debt-ceiling fights, and austerity when unemployment is high. The second produces a test: are there unused workers and factories, or is the bid only rearranging a fully employed economy?

What MMT does not say

It does not say a government can spend without limit. Proponents are explicit that past full employment, more spending raises prices. It does not say state governments, cities, or euro members can do this. They are currency users. They really can run out of money. It does not say exchange rates are irrelevant. A country that imports its food and fuel in dollars can lose real living standards if its own currency falls, even if it never misses a nominal payment. It does not replace politics. Congress can still refuse to spend, and a central bank can still refuse to cooperate.

Mainstream critics argue the theory still fails as a guide. Lawrence Summers called it voodoo economics and warned that it understated inflation from large fiscal expansions. The Mercatus Center’s critique lists a weak inflation model, too much faith in fiscal authorities as inflation fighters, and too few safeguards on debt. The 2021–22 inflation surge is the episode skeptics point to: demand was pushed hard while supply was constrained, prices jumped, and the tool that actually cooled them was the Federal Reserve raising rates by more than five percentage points, not a carefully timed tax increase. Mankiw’s narrower point is that an accounting identity is not a causal lever. Expectations, the exchange rate, and the interest rate still sit between “the government can credit the account” and “the public will accept the currency at stable prices.”

Why the household story stuck

The textbook version survived because it is intuitive, because it was built when currencies were tied to gold or to each other, and because it disciplines politicians. Under a gold standard, a government really could run out of the thing it had promised to redeem. That world ended for the dollar in 1971. The curriculum did not fully follow. “How will you pay for it?” also remains a useful political brake, whether or not it describes the plumbing.

MMT’s counter is that the brake is attached to the wrong wheel. A currency issuer cannot bounce a check in its own unit of account. It can absolutely bid prices up, misallocate labor, and weaken the currency. The argument worth having, on this account, is about spare capacity and inflation, not about whether the Treasury has enough of its own IOUs left in the drawer.