Cowen on Central Banks and Liquidity Traps

corporate bonds, discount window, liquidity traps, Marginal Revolution, Tyler Cowen

corporate bonds, discount window, liquidity traps, Marginal Revolution, Tyler Cowen

These are challenging times for monetary economists like myself, what with central banks making one dramatic departure after another from conventional ways of conducting monetary policy.

Yet so far as I'm concerned, coming to grips with negative interest rates, overnight reverse repos, and  other newfangled monetary control devices is a cinch compared to meeting a challenge that nowadays confronts, not just monetary economists, but economists of all sorts.  I mean the challenge of  getting one's ideas noticed by that great arbiter of all things economic, Tyler Cowen.

Last week, however, Tyler may have given me just the break I need, in the shape of a brief Marginal Revolution post entitled,  "Simple Points about Central Banking and Monetary Policy."

Tyler's "simple points" are these:

Central banks around the world could raise rates of price inflation, and boost aggregate demand, if they were allowed to buy corporate bonds and other higher-yielding assets.  Admittedly this could require changes in law and custom in many countries[.]

There is no economic theory which says central banks could not do this, as supposed liquidity traps would not apply.  These are not nearly equivalent assets with nearly equivalent yields.

Tyler isn't one to traffic in banalities, so it's no surprise that his claims are controversial.  Why so? Because the prevailing monetary policy orthodoxy, here in the U.S. at least, insists that, rare emergencies aside, the Fed should stick to a "Treasuries Only" policy, meaning that it should limit its open-market purchases to various Treasury securities.  For the Fed to do otherwise, the argument goes, would be for it to involve itself in "fiscal" policy,  because its security purchases would then influence, not just the overall availability of credit, but its allocation across different firms and industries.  So far as the proponents of "Treasuries Only" are concerned, Tyler's remedy for deflation would create a set of privileged or "pet" corporate securities, analogous to, and no less obnoxious than, the "pet banks" of the Jacksonian era.

All of which is good news for me, because I'm prepared, not only to side with Tyler in this debate, but to offer further arguments in support of his position.  For I took essentially the same position in a paper I prepared for Cato's 2011 Monetary Conference.  In that paper, I first counter various arguments against having the Fed purchase private securities, and then proceed to recommend a set of Fed operating-system reforms involving broad-based security purchases.  I figure that, with a little luck, Tyler may find those arguments and suggestions worthy of other economists' attention.

Here is a quick summary of my paper's arguments and suggestions.

Concerning the "pet corporate securities" argument, to give it that name, I find both it and the anti-Jacksonians' original complaint against pet banks equally unpersuasive.  If those state banks to which Jackson distributed government's funds yanked from the  second Bank of the United States were "pet banks," just what, prey tell, was the B.U.S. itself while it held all of the government's deposits, if not a single (and correspondingly more odious) government "pet"?

Likewise, if purchasing corporate bonds means favoring particular corporations, and venturing thereby into "fiscal" policy, isn't "Treasuries Only" not itself a means of shunting scarce credit to one particular economic entity — in this case, the federal government — at the expense of all the others?  Is there not, indeed, something positively Orwellian about the suggestion that, by buying Treasury securities, the Fed steers clear of "fiscal" policy?

Though they never heard of Orwell, the Fed's founders would certainly have considered such talk perverse.  Far from seeing "Treasuries Only" as a means for keeping the Fed and the fisc at arms length, they took precisely the opposite view: so far as they were concerned, to allow a central bank to purchase government debt was to risk having it become a tool of inflationary finance. Consequently they favored a "commercial paper only" rule, or rather a "commercial paper and gold only" rule, with a loophole allowing purchases of government paper only for the sake of stabilizing the Fed's earnings at times of low discount activity.  Like all loopholes in the Federal Reserve Act, this one was not left unexploited for long.  Yet it was not until 1984 that the opposite, Treasuries Only alternative took force.  Nor is Treasuries Only the norm elsewhere.  The ECB, in particular, ordinarily accepts euro-denominated corporate and bank bonds with ratings of A- or better as collateral for its temporary open-market operations.

The operating system reform I recommend in my paper involves replacing both the discount window and the anachronistic and unnecessary primary dealer system with an arrangement resembling the Term Auction Facility (TAF) created in December 2007, at which the Fed auctioned off credit to depository institutions against the same relatively broad set of collateral instruments, including corporate bonds, accepted at its discount window.  To assure competitive allocation of credit among bidders offering different types of collateral, the facility could make use of a "product-mix" auction of the sort Paul Klemperer developed for the Bank of England.  To rule out subsidies and limit its exposure to loss, the Fed could also follow the Bank of England's example by setting minimum bid rates for the various types of eligible collateral, reflecting predetermined penalties or "haircuts."  Finally, to allow emergency credit to be supplied as broadly as possible, and therefore in a manner fully consistent with Walter Bagehot's last-resort lending principles, the Fed could open its auction facility to various non-depository counter-parties, including money market mutual funds.

Besides making liquidity traps relatively easy to avoid, as Tyler suggests, adopting such an alternative system would have many other advantages.  It would reduce the systemic importance of  present primary dealers.  It would guard against the risk of having the Fed gobble-up collateral that's essential to private-sector credit creation.  It would allow a single operating system to meet both ordinary and emergency demands for credit.  Like the TAF, it would avoid the "stigma" of discount-window lending.  In fact, it would dispense entirely with the need for direct lending to troubled financial (and perhaps some troubled non-financial) institutions.  Most importantly, it would render any sort of ad-hoc central bank lending during financial crises otiose, and by so doing would bring the Federal Reserve System one step closer to being based on the rule of law, instead of the arbitrary rule of bureaucrats.

Addendum (2/24/2016; 2:33PM): As always happens when I suggest some marginal reform of the Fed's constitution or conduct, I am receiving responses to this post suggesting that free banking is a better solution.  But free banking itself doesn't solve the problem of achieving a rule-based dollar.  Doing that requires taking steps to reform the Fed, especially by moving away from the present arrangement in which the Fed is capable, especially during a crisis, of arbitrarily expanding the scope and nature of its activities, to a fixed, rule-based arrangement that is nevertheless capable of preserving overall monetary stability, even in the midst of a crisis.   So long as we have a crisis-prone currency, we will never make progress toward free banking.

  • Keith Weiner


    I would like to make two points:

    1. Buying bonds does not cause rising prices, it causes falling interest.

    2. There is always a credit gradient. Large corporations can borrow at much lower rates than mom and pop small businesses or startups. By imposing a policy of buying corporate bonds (but presumably not mom and pop or startup bonds), the Fed would steepen this gradient. The primary issue is not whether the playing field is level between Corporate A and Corporate B (though that is an issue of course) but further stacking the odds against startups.

    • George Selgin

      Keith, you make a false dichotomy. Buying bonds lowers the yield (interest), but raises the prices, of the bonds purchased. And it also increases, ceteris paribus, the money stock, and hence prices generally. This shouldn't be controversial.

      The fact that the Fed would necessarily favor corporate bond issuers over other borrowers cannot be denied. But it is not clear why the conventional Treasuries Only alternative should be regarded as a better option, for as I pointed it also involves stacking the deck, this time entirely in favor of the government. That's an even stiffer "gradient." Is it what you favor?

      • Keith Weiner

        Yes, the bond price varies strictly as the inverse of the interest rate. Buying bonds pushes up bond prices, which pushes down interest rates. Early monetarists starting at least with Wicksell in the 1890's and Fisher and others observed that prices did not correlate with quantity of money, but with interest rates. The quantity theory of money is not too controversial yet–but it should be. Since 2009 we have had rapidly rising quantity of dollars and yet we don't have the rapidly rising prices that nearly everyone predicted. Now we have freefalling commodity prices.

        I don't favor a Fed that favors the government. I favor a free market. I don't want any central planners, most especially I want no central banks.

        • George Selgin

          Yes, fine. Me too. But there's a question: so long as there's a fiat dollar, what assets should the Fed hold to match the liabilities on its balance sheet? It is no solution to any problem to merely display one's impeccable free-market credentials.

          • Keith Weiner

            Agreed, there's no point in saying, "I am more free market than thou."

            No matter what the Fed buys, it suppresses the interest rate. The artificially high bond prices in that market bleed through to bond (and other asset) prices elsewhere. Shale oil producers were able to sell junk bonds with lower yields than what prime credits had to pay just a few years prior.

            I don't want to give the Fed advice on what the "best" asset to buy is, lest I give the impression that I think it is good to buy that. No matter how the Fed centrally plans, it caused harm.

            That said, I think it causes more harm to push up some corporate bonds but not others, than it does to simply push up the "risk free" asset.

  • Tom Humphrey

    Glad you mentioned Walter Bagehot in support of your position. Bagehot believed that during financial emergencies few restrictions should be placed on types of assets upon which the central bank might lend, or the kinds of borrowers it should accommodate. Besides the conventional eligible commercial bills and government securities, acceptable collateral would include "all good banking securities," and even "railway debenture stock." Last-resort loans, Bagehot claimed, should be made available "to merchants, to minor bankers, to this man and that man", indeed to any and all sound borrowers. To Bagehot, the "amount of the advance is the main consideration . . . not the nature of the security on which the advance is made, always assuming the security to be good."

    The Fed's "Treasuries only" restriction is a far cry from the Bagehot imperative.

    • George Selgin

      Tom, your telling me that I've got my Bagehot right is music to my ears!

  • Mike Sproul

    A related point, also much-neglected:

    Central banks should not hold assets denominated in the same money that they issue, as it creates a potential for inflationary feedback:
    1) The dollar loses value.
    2) The Fed's bonds, being denominated in dollars, lose value
    3) The loss of assets causes the dollar to fall still more
    4) which causes the bonds to fall more, etc.

  • Jim Brown

    Frankly I am a shocked at your proposal, Professor. What about the "slippery slope" of the government buying (i.e., owning) business assets? If corporate bonds are fair game for fiat money, why not stocks? Bernanke has already proposed private asset purchases, whereby the Treasury buys private assets financed by fiat money from the Fed. The BOJ and the PBOC (especially the latter) already own huge quantities of publicly traded equities. Please explain why central bank purchases of corporate securities does not put us on a collision course with fascism (govt control or business) or socialism (govt ownership of business). Thanks.

    • George Selgin

      The proposal is simply to make the same securities presently accepted as collateral for discount window loans ones that can be purchased via auctions. That list is well-defined.

      Of course any change can be regarded as a "slippery slope" in the sense that it can be followed by other, less desirable changes that treat the first change as a precedent. But bear in mind that Treasuries Only was itself an innovation, and not one without undesirable consequences, one of which was the introduction of a distinct inflationary bias in central bank conduct. For most of the early history of central banking, the received (and not unsound) wisdom was that central banks should stick to commercial paper, while avoiding government debt like the plague. Nor is there the least indication that by acquiring commercial paper early central banks gained control of the business whose paper they acquired.

      And although I don't accept the claim that commercial paper guarantees automatic regulation of the money stock, I don't consider it at all inevitable that allowing central banks to buy certain private securities much least to their favoring particular firms. On the contrary: it should be relatively easy to rule that out by means of properly-designed legislation. Remember, finally, that the whole idea is to eliminate the present, perceived necessity for direct central bank lending to troubled firms.

  • Dan Thornton

    There are two issues associated with what the Fed should purchase when conduction open market operations: risk (default risk and market risk) and credit allocation. With respect to credit allocation, the issue is the efficiency of financial markets. If financial markets are pretty efficient, as I tend to believe, what the Fed purchases will be relatively unimportant "in equilibrium." So I am not opposed to the Fed purchasing and index of private securities. Indeed, absent Treasuries, that is what it should do. However, because I believe that financial markets are "pretty darn efficient," I see no reason why it should because it will have the same effect with no risk if it purchases shorter-term Treasuries. Of course, if markets are not as efficient as I tend to think they are, the "purchase only short-term Treasuries" doctrine avoids the credit allocation problem as well. It seems both arrogant and dubious to think the Fed can allocate credit better than the market can. Indeed, increased efficiency is the primary reason that markets exist.
    In any event, if the Fed had done this I doubt it would have had much effect for another reason; namely, the Fed would have likely purchased liabilities of firms who are sitting on massive amounts of cash. Moreover, even if the Fed purchased a broader range of securities, nothing would prevent it from impounding the funds in excess reserves in order to prevent excessive money growth, as it has done the past 7 years.
    I agree completely that the "primary dealer system" is anachronistic. Indeed, it has been so for some time. Why does the Fed keep it? One answer is "tradition." Of course, there is a less favorable bureaucratic possibility.

  • Ray Lopez

    "Central banks around the world could raise rates of price inflation, and boost aggregate demand, if they were allowed to buy corporate bonds and other higher-yielding assets. " – as Selgin says later ("rare emergencies aside") in the USA it already is permitted for the Fed to buy commercial paper. So this proposal by Cowen is not that radical. More radical is Sumner's proposal for the Treasury to send money directly to households for them to spend until such time that a NGDP target is met. What is Selgin's reply to that?

  • Dallas Weaver

    It is strange how the real growth in the economy is dominated by the silicon valley type companies that don't have any bonds to purchase (successful innovative companies have high profits and little to no debt) and don't even want to be the FBI's or NSA "pet". The old established non-growing companies do have lots of debt as they try to buy "growth" and these are the ones that will become "pets".

    If one wants to help the economy, investing in this group of "pets" seems very counter productive and anti-innovative and would kill future growth as the "pets" can save themselves from disruptive competition with cheap money to buy out future rivals.

    If you asked the government elite back in the late 50's whether they should bet our future on RCA and the other Vacuum tube manufactures that owned the electronics market and the politicians of the day, and who were spending millions on R&D in solid state physics, they would have said YES, they will be the future. However, an unknown geotechnical firm called Texas Instruments plus the growth of other unknowns in silicon valley put all the big vacuum tube companies with their massive R&D out of business. New ideas can't float through old vacuum tube management.

    The thought of our bureaucrats picking the right "pets" is preposterous.

  • Nick_Rowe

    "Likewise, if purchasing corporate bonds means favoring particular
    corporations, and venturing thereby into "fiscal" policy, isn't
    "Treasuries Only" not itself a means of shunting scarce credit to one
    particular economic entity — in this case, the federal government — at
    the expense of all the others?"

    Yep. But why not have the central bank just buy the index of bonds, or stocks?

    • George Selgin

      That's not a bad idea, Nick. But my plan is intended to allow hard-up banks (and perhaps other firms) to pay a premium for liquidity by accepting bigger haircuts on those eligible securities they happen to possess. The idea is to make the discount window and other direct lending redundant. The "index" scheme doesn't offer this potential.

    • Tom Barson

      The answer to Nick's question must pay some attention to the composition of the index. "Just" buying an index is easier in some markets than others.

      I recall (from a ways back, so this might be stale info) that, practical corporate bond indexes, the kinds actually used by index funds and ETFs, are statistical. Since corporate bonds don't trade that smoothly, the indexes are biased, I believe, towards easily traded issues. This probably doesn't matter too much for the incremental trades in corporate bond ETFs that we see from private and institutional investors, so a rise in the index fund effectively lowers rates for everybody.

      But if a massive player like a central bank were on the scene, its purchase or sale of index funds might drastically affect the supply and price of the small number of issues chosen by the fund to represent the larger index. Conversely, if the index fund were forced to hold the entire index, transaction costs would increase sharply and the scarcity of smaller issues might exaggerate the price moves of the index or affect the liquidity of the index fund.

      Not saying do it or don't do it, just pointing out a need to step carefully…

  • Gary Anderson

    Well, if you want to help the rich, buy assets. I guess it is better than negative bond rates and a cashless society. But it still picks the rich as winners and everyone else as losers. Very unamerican.