Should the Fed purchase Treasuries only?
By Scott Sumner
George Selgin has an interesting post discussing the question of whether the Fed should buy only Treasury securities, or a cross section of marketable securities. This issue came up at a recent Cato money conference, where Joe Gagnon suggested the Fed might want its balance sheet to include a relatively comprehensive portfolio of public and private sector assets. George has advocated a similar policy, but also quoted Marvin Goodfriend favoring a “Treasuries only” policy:
When the Fed substitutes an extension of credit for a Treasury security in its portfolio, the Fed can no longer return to the Treasury the interest it had received on the Treasury security that it held. In other words, when the Fed sells a Treasury security to make a loan, it’s as if the Treasury issued new debt to finance the loan. Credit policy executed by the Fed is really debt financed fiscal policy. …
In effect, Fed credit policy works by interposing the United States Treasury between lenders and borrowers in order to improve credit flows. In doing so, however, the Fed essentially makes a fiscal policy decision to put taxpayer funds at risk. In the event of a default, if the collateral is unable to be sold at a price sufficient to restore the initial value of Treasury securities on the Fed’s balance sheet that was used to fund the credit initiative, then the flow of Fed remittances to the Treasury will be smaller after the loan is unwound. The Treasury will have to make up that shortfall somehow, namely, by lowering expenditures, raising current taxes, or borrowing more and raising future taxes to finance increased interest on the debt.
Think of the argument this way. If the Fed buys equities with newly created money, it’s logically equivalent to the Fed buying bonds, giving the bonds back to the Treasury, having the Treasury destroy them, and then having the Treasuring borrow money with newly created bonds in order to buy stocks, and then giving the stocks to the Fed. Do we want the Treasury to borrow money to invest in the stock market?
At first glance, Goodfriend’s argument seems pretty appealing to me, and indeed it’s similar to an argument I often make about FDIC. That is, the FDIC creates a situation where bank deposits at US commercial banks are logically equivalent to depositors loaning the money to the Treasury, and then having the Treasury re-lend the same funds to banks at the same interest rate. That’s not good.
George is skeptical of an argument by Goodfriend (and also Charles Plosser) that a Treasuries only policy is more fiscally neutral. He notes flaws in Plosser’s argument that limiting the Fed to the purchase of Treasuries would reduce the risk of the Fed enabling fiscal expansion. In George’s view, a Treasuries only policy actually favors the federal government, and the Fed should adopt a more balanced approach:
Three aspects of flexible OMOs would contribute to their neutrality. First, they would have the Fed standing ready to purchase, in its routine open-market operations, not just Treasury or agency securities but a much broader set of assets, consisting of all those marketable securities that can presently qualify as collateral for the Fed’s discount-window loans. Second, they would be undertaken not with a score or so of “primary dealers” only, but with numerous counterparties, including all banks that might be eligible for discount window loans as well as all those counterparties presently taking part in the Fed’s overnight reverse repurchase (ON-RRP) operations. Third, flexible OMOs would be undertaken using a version of Paul Klemperer’s “product-mix” auction procedure specifically designed to prevent the Fed from favoring any particular securities yor counterparties.
Consider three options:
1. Treasuries only.
2. Treasuries preferred, other assets if necessary.
3. A balanced portfolio of marketable assets.
I’ve always tended to favor option #2. Option #1 is out of the question, as the need for central banks to hit their target is vastly more important than the question of which assets to buy.
The difference between #2 and #3 can be framed in two different ways. Some classical liberals may prefer the Fed not buy private sector assets, as that smacks of socialism. George Selgin frames it differently, arguing that a Treasuries only policy favors the government, and that a balanced portfolio is more neutral. There are some practical issues with deciding exactly which such assets to purchase, but I’ll leave that question to others and instead focus on a different issue.
Update: I’d like to clarify that George does not favor having central banks purchase equities and real assets. See his comment below.
It seems to me that the best argument for #3 is not “neutrality”, but rather than it might make monetary policy more potent. Just to be clear, the Fed ought to be able to hit any reasonable target with a Treasuries only policy. But for various reasons, the Fed may not be willing to pursue that target aggressively enough. For instance, they may worry about buying up too large a share of Treasury debt outstanding.
I’d like to see a new global norm established where it would be perfectly acceptable for central banks to buy a balanced global portfolio of assets. In this regime, the decision by the Bank of Japan to purchase $1 trillion worth of global bond and equity index funds would not be viewed as “currency manipulation”. In that case, the BOJ would have no trouble hitting its 2% inflation target. And here’s the irony that most people miss—under my proposed regime the BOJ balance sheet would be far smaller than it’s current level of roughly 100% of Japanese GDP. That’s because at 2% inflation there will be less demand to hold zero interest Japanese base money, as compared to the roughly 1% annual inflation under Abenomics. The BOJ would not have to buy the foreign assets, due to the “Chuck Norris effect”. The willingness and ability to buy these assets would by itself create the desired inflation. At most a few symbolic purchases would be needed, to prove credibility.
In the end, I’m relatively agnostic on the Treasuries-only dispute. To the extent I sympathize with allowing the purchase of other assets, it’s not because of any vision of “neutrality”, rather it’s because for me the paramount issue is macroeconomic stability. A Treasuries only policy could make that more difficult to achieve.