On Empty Purses and MMT Rhetoric

Federal Reserve, U.S. Treasury, government debt, Modern Monetary Theory

Federal Reserve, U.S. Treasury, government debt, Modern Monetary TheoryA couple posts ago, I criticized in general terms Modern Monetary Theorist Stephanie Kelton’s suggestion that the expense of the proposed Green New Deal need not make it “a drag on the economy” since the Fed could always foot the bill.

That post steered clear of the technical subtleties of Modern Monetary Theory, focusing rather on some of the bolder lessons its proponents like to draw from it. Today I'll instead consider one of those subtleties, as found in the following passage from Professor Kelton’s article:

As a monopoly supplier of U.S. currency with full financial sovereignty, the federal government is not like a household or even a business. When Congress authorizes spending, it sets off a sequence of actions. Federal agencies, such as the Department of Defense or Department of Energy, enter into contracts and begin spending. As the checks go out, the government’s bank — the Federal Reserve — clears the payments by crediting the seller’s bank account with digital dollars. In other words, Congress can pass any budget it chooses, and our government already pays for everything by creating new money.

What's so subtle about that? Actually the subtlety consists, not of any detail Professor Kelton supplies, but of one she chooses to omit. For while she scrupulously has the Fed “crediting the seller’s bank account with digital dollars,” she unaccountably forgets to mention that, to complete the clearing transaction in question, the Fed must also debit the Treasury’s own account, known as the Treasury General Account, or TGA, for short.

What difference does this make? Plenty, actually. For it means that there is, after all, a practical limit to how much Congress can spend, and that limit really isn't all that different from those faced by ordinary households or businesses: when the rest of us write checks, our banks also end up debiting our checking accounts by corresponding amounts as the checks are returned to them for payment. That matters because at some point, if we spend too much, our bank balances will be depleted, and our checks will start bouncing.

Professor Kelton insists, on the contrary, that Congress “is not like a household or even a business.” In the government's case, she suggests, the Fed will fill any breach "by creating new money." In other words, the government doesn’t have to worry about its checks ever bouncing, because the Fed will cover any shortfall.

But that just ain't so. Indeed, when it comes to being able to rely on its bank to cover its expenditures, Congress is in one crucial respect more constrained than ordinary households and businesses are. That’s because, although most bank depositors enjoy certain overdraft privileges, and the Fed once granted similar privileges to the Treasury, in 1981 Congress itself permanently eliminated the Treasury’s overdraft privileges. Consequently, if Congress is to avoid running out of money, it can’t write checks in amounts exceeding the balances in its TGA account.

Furthermore, as Eric Tymoigne, another well-known (and especially thoughtful) Modern Monetary Theorist, explained in a blog post written several years ago, even when the Treasury did enjoy overdraft privileges on its TGA account, it made only very limited use of them, and never did so “because it was running out of money.” In those days, and indeed until the 2008 crisis, besides its TGA (“Treasury General Account”) balance at the Fed, the Treasury also maintained substantial balances in TT&L (“Treasury Tax and Loan”) Service accounts at various commercial banks. Although the Treasury occasionally overdrew its TGA account, thereby borrowing a corresponding amount from the Fed, it never borrowed more than it had available in its TT&L account balances. Instead, it borrowed in anticipation of anticipated receipts to avoid making temporary withdrawals from those TT&L accounts that would otherwise have necessitated countervailing Fed open-market operations.

In short, while they were still permitted, the Treasury’s TGA overdrafts served, not to make life easier for Congress, but to make it easier for the Fed. Although in principle the Treasury might also have employed its TGA overdraft privileges to cover revenue shortfalls during “national emergencies,” it never did so. Moreover since 1935 it could have done so only up to what was, by the 1970s certainly, a paltry $5 billion limit. As Tymoigne also explains, although during the ‘60s Congress mulled the possibility of expanding the Treasury’s overdraft limit, the idea “never went anywhere because the use of the overdraft for fiscal purpose [sic] was seen as inflationary and unsound.” Instead, after first limiting them in 1979, Congress ended up withdrawing the Fed’s TGA overdraft privileges altogether two years later.

None of this means, of course, that the Fed plays no part in helping Congress to pay its bills. Most obviously, its seignorage earnings (the revenue it earns on its portfolio, minus its operating and interest expenses) are part of the general government’s revenues. And so long as it maintains any particular interest-rate target, it’s bound to limit the government’s nominal borrowing costs correspondingly. Nevertheless it doesn’t follow that Congress can spend willy-nilly without facing any risk of running-out of money.

Might Professor Kelton's claims be at least partially vindicated by the existence of some mechanism through which Congressional disbursements that reduce the TGA balance automatically lead to corresponding, additional Fed security purchases? No such luck. It's true that, under the pre-2008 system, disbursements from the TGA account tended to be expansionary, because they shifted high-powered base dollars into the commercial banking system. Once there they tended to flow into the fed funds market, reducing the effective fed funds rate, on their way to ultimately promoting disproportional expansion in the quantity of bank lending and deposits. But for that very reason the Fed routinely resorted to open-market security sales to offset such shifts.

At one point in his otherwise very informative post Tymoigne himself appears to suggest that there is, after all, some merit to Professor Kelton's assumption that Congressional disbursements trigger corresponding money creation. He writes:

Even though direct [Fed] financing was discouraged and later forbidden, the Treasury uses Fed’s monetary instruments to spend. Thus, one way or another, the Treasury will get financed by the Fed because only the Fed supplies the funds that the Treasury uses. While TT&Ls are used to receive bond and tax proceeds, they are just a tool that allow to smooth the impact of tax and bond proceeds on reserves. Ultimately all the proceeds go to the TGA, which drains reserves. As a consequence the Fed has to ensure that banks have enough reserves to allow the Treasury to be able to make the transfer from TT&Ls to TGA.

But while it’s true that, prior to the recent crisis and other things equal, transfers from the Treasury’s TT&L accounts to its TGA account (as opposed to  disbursements from the TGA itself)  called for offsetting Fed open-market bond purchases, such purchases add, not to the government's bank balances, but to the balances of bond dealers. To raise a like amount of revenue for itself, the government would have to sell a like amount of additional bonds. That is, it would have to raise extra money by going further into debt. Tymoigne’s claim that “the Treasury will get financed by the Fed because only the Fed supplies the funds that the Treasury uses” is a non-sequitur. The Treasury will get financed only if it either taxes or borrows more — though the Fed may assist it in doing the last of these by increasing its own purchases of Treasury securities.

In today's environment of abundant excess reserves, in contrast, it’s no longer necessary for the Fed to offset movements of high-powered money either into or out of the TGA, because such movements no longer translate into increased bank borrowing or lending on the fed funds market, and a corresponding tendency for the effective fed funds rate to differ from its assigned target.

I realize that, in pointing to a seemingly minor error within the large body of writings that make up Modern Monetary Theory, I might be accused of nit-picking. But I plead not guilty, for such minor errors are an important source of MMT's popular appeal. It's often by dint of them, rather than any genuinely innovative or profound insights, that Modern Monetary Theorists succeed in turning otherwise banal truths about the workings of contemporary monetary systems into novel policy pronunciamentos that are as tantalizing as they are false.


*After writing this I discovered this excellent, 2012 J.P. Koning post covering much the same ground. Since the current MMT fad has made the error J.P. points to more important than ever, I hope the repetition will be pardoned.