Recently there’s been some spread in the influence of “Modern Monetary Theory”, a set of beliefs about monetary and fiscal policy championed by figures like Jamie Galbraith and Randall Wray. It’s been given (unfavorable) attention by Paul Krugman, lengthy blog posts by Steve Waldman and Nick Rowe, ample article space at new deal 2.0 and even a platform at the New York Times. Though MMT is certainly nonsense, I’m hesitant to spend much time refuting it—since it has roughly the same stature among economists that creationism commands among biologists, there’s a sense that I’m just tilting at windmills. Nevertheless, beliefs that are ridiculed by specialists can become powerful (and dangerous) outside the mainstream, and I think it’s best to confront them head-on.
MMT starts with lengthy historical musing about the origins of money. The claim is that we can’t think about “money” independent of the special status granted to a particular form of currency by the government. Specifically, the power to levy taxes in dollars is what gives dollars value in the first place.
All this may be true. But it’s also mostly irrelevant for the conduct of monetary policy. Perhaps government, using its power of taxation, nudged us into the equilibrium where green pieces of paper are used both as a medium of exchange and a unit of account. But that doesn’t mean that taxes, at the margin, have anything to do with the demand for currency.
In fact, this is easy to refute. Until the advent of QE, the vast majority of base money was held in the form of paper currency. The role of electronic bank reserves was incredibly marginal, fluctuating around $10 billion out of a monetary base of over $900 billion. Yet these electronic bank reserves were used to conduct the vast bulk of transactions in our economy. As James Hamilton points out, circa 2008 each dollar of electronic reserves was whizzing around on Fedwire an average of 350 times every day. Before the advent of interest on reserves, excess reserves were virtually nonexistent: banks were profit-maximizers, and there was no need to hold even a cent more than the 10% reserve requirement on checking accounts. The only question was whether to hold reserves in the form of “vault cash” (mainly in ATMs) or electronic accounts at the Fed.
In short: pre-2008 demand for base money—the asset that the government creates—consisted overwhelmingly of demand for cash.
But certainly no more than a trivial fraction of taxes are paid in cash. Employers don’t FedEx their withholdings to the IRS in envelopes filled with $100 bills; corporations don’t send a fleet of armored trucks when their bill comes due. Even Grandma probably sends a check. (In fact, if you’ll forgive my ignorance, I’m not positive that we’re even allowed to pay the feds in cash.)
Yet there has to be some reason why individuals are holding cash: it pays zero interest, which in normal times is well below even the safest alternative assets. And certainly there’s a demand for cash to conduct petty transactions: it’s how I pay at food trucks and Chinese restaurants too stingy to accept Visa. But this isn’t enough to justify the vast amount of cash in circulation, which at nearly $1 trillion amounts to over $3000 for every man, woman, and child in America. (That’s a lot of food trucks!)
In reality, of course, much of this cash is held abroad—estimates derived from formal cash shipments hover around 50%, but given the vast sums smuggled out of the country under questionable circumstances, the actual fraction is probably far higher. Cash is useful for tax evasion, illegal transactions, and storing value in a country without a reliable banking system. You’ll notice, however, that none of these explanations have anything to do with the MMT theory of money, in which the government creates demand for money by levying taxes (except insofar as taxes create demand for tax evasion!).
And even if (against all evidence) you accept the MMT view of money, it doesn’t become fiscally relevant. Suppose that you’re trying to patch a $20 trillion hole in the long-term federal budget. You raise taxes by $300 billion (in real terms) a year. At a real interest rate of 2%, this works out to $15 trillion. You’re most of the way there! Now consider the effect of your tax increase on demand for base money. As I’ve stressed, it’s abundantly clear that there is no significant effect, but let’s indulge MMT and make the ludicrous assumption that money demand will increase one-for-one with the rise in annual tax revenue. At a nominal interest rate of 4%, your annual seignorage income increases by $12 billion; the present value of the long-term increase is $600 billion. Not much compared to $15 trillion you raised from taxes, or the $20 trillion you owe—you’re still deep in the hole.
In fact, $600 billion is exactly 4% of $15 trillion. Even in the utterly implausible case where annual tax increases lead to a commensurate rise in the demand for money, the extra revenue from seignorage is just a fraction of the revenue you raised from taxes in the first place—this fraction being the nominal interest rate.
To sum up: MMT is wrong about money. Even supposing that it’s right, its impact is fiscally marginal. The government does have a budget constraint.