Saturday, August 28, 2010

More Thoughts on Bernanke's Speech

I was in the Whole Foods store buying eggplant and thinking about Ben Bernanke. That has to be weird on some dimension. Bernanke, in his speech was talking about the recent announced change in Fed policy, which was to replace MBS (mortgage backed securities) that are running off the Fed's balance sheet with long-term Treasuries. Here's his explanation:
However, more recently, as the pace of economic growth has slowed somewhat, longer-term interest rates have fallen and mortgage refinancing activity has picked up. Increased refinancing has in turn led the Fed's holding of agency MBS to run off more quickly than previously anticipated. Although mortgage prepayment rates are difficult to predict, under the assumption that mortgage rates remain near current levels, we estimated that an additional $400 billion or so of MBS and agency debt currently in the Fed's portfolio could be repaid by the end of 2011.

At their most recent meeting, FOMC participants observed that allowing the Federal Reserve's balance sheet to shrink in this way at a time when the outlook had weakened somewhat was inconsistent with the Committee's intention to provide the monetary accommodation necessary to support the recovery. Moreover, a bad dynamic could come into at play: Any further weakening of the economy that resulted in lower longer-term interest rates and a still-faster pace of mortgage refinancing would likely lead in turn to an even more-rapid runoff of MBS from the Fed's balance sheet. Thus, a weakening of the economy might act indirectly to increase the pace of passive policy tightening--a perverse outcome.
Now, as an example, suppose that $100 billion in MBS run off because people are refinancing their mortgages. Then, suppose that the new mortgages these people take out are sold to Fannie Mae, which holds them, and finances the purchases by issuing $100 billion in agency securities. What has changed? The private sector is now holding $100 billion less in reserves, and $100 billion more in agency securities. Both of these types of securities are effectively obligations of the (consolidated) federal government. Furthermore, the reserves (under current conditions) are the same as T-bills. Bernanke says that this involves an "implicit tightening." Why? This is of course the key to the whole "quantitative easing" program. The Fed thinks that shortening the average maturity of federal debt obligations in the hands of the private sector is somehow expansionary. Thus the policy action the Fed just announced: we undo the "implicit tightening" by having the Fed buy $100 billion in long-term Treasuries. Now, reserves are unchanged, and this is just a swap of $100 billion in agency securities for $100 billion in Treasuries, and those assets are essentially identical.

On my last post, Jordan sent me a link to a finance paper by Vayanos and Vila, which is a start in thinking about segmented markets, preferred habitat, and the term structure of interest rates. What I am wondering about is the following. By issuing reserves and holding long-term Treasuries, the Fed is intermediating across maturities. Of course there is an array of private intermediaries who do the same thing, including banks and bond mutual funds. If the Fed can change market asset prices by issuing short maturity liabilities and buying long maturity assets, it must be that the Fed has some cost advantage over the private sector in this type of intermediation. Outside of the circumstances of the financial crisis, what would that advantage be?

Now, go one step further. Suppose the Fed has such an advantage. What this says is that the Treasury somehow got things wrong. It issued debt of too long a maturity. But why not have the Treasury fix their own problem then. The Treasury could issue T-bills and buy up existing long-maturity debt. Now go one step further. If this is such a great idea, why doesn't the Treasury issue only T-bills on a regular basis?

17 comments:

  1. Having to constantly roll debt AND often selling a lot of it to foreign countries, especially China and Japan, exposes the US to a lot of risks.

    ReplyDelete
  2. Yes, exactly. If the debtholders lose confidence quickly, you can't roll it over, or can do so only with a very large premium. Some people think that this is not a potential problem for the United States. What do you think?

    ReplyDelete
  3. Steve: sure, this is a 'potential' problem for the US. I'm pretty sure even Krugman's said that. So far though, there's zero evidence that I can see for a quick loss of confidence. On the contrary, I had a quick look through a few recent 10-yr Treasury auctions -- they were over-bid by a factor of 3.

    Anon: I'm curious why 'especially China & Japan' might be problematic. They've shown a remarkable appetite for our debt (yeah, things can change quickly, yadda yadda). Anyway, the latest TIC data shows overall net purchases from Japan & "other Asia" declining. In fact I read the other day (WSJ maybe? can't remember) that overall foreign demand has fallen & been offset by higher domestic demand.

    Again, not much evidence that things that might happen show any signs of being imminent.

    PS

    ReplyDelete
  4. Not that I think that anything is imminent, but my Argentinian colleagues assure me that when a loss of confidence happens, it doesn't give you much warning.

    ReplyDelete
  5. I wish I could remember where I read this, but there's a quote out there somewhere that says "in international finance, the role of Argentina is usually played by Argentina."

    Another data point: CDS spreads on US 10-yr treasurys (5yr CDS) went up 4bps in the last month, currently 39 bps. The Austerian Republic of Germany is at 35.

    ReplyDelete
  6. Stephen and psummers: In my view, the US can handle a lot of debt, especially if it's mainly domestically owned and long maturity. But, the more the US deviates from this (conservative) scenario, the higher the likelihood of either default light (paying, say 80 cents on the $ or extending maturity) or default.

    Also, the fact that the $ is THE trading/reserve currency of the world, makes it much easier for the US to take on new debt without forcing the $ to zero.

    What if Europe recovers much faster and stronger than the US? Would foreign buyers of US debt substantially reduce their demand and prefer, say, German Bunds instead? Could the euro become the new trading/reserve currency? If yes, prepare the "life boats" here in the US.

    ReplyDelete
  7. Anon:

    If Europe recovers faster & stronger, that'd be good for net exports. A bit of dollar depreciation now would be good news. I'd think German corporate debt or equities might start looking better than Bunds, but whatever. "Substantial" switch? Dunno, but this scenario doesn't sound to me like a sudden attack of the bond vigilantes.

    Ditto re the euro becoming the new reserve currency. Sure, that's possible, but it ain't gonna happen overnight. I don't think you need to break out the life vest yet.

    PS

    ReplyDelete
  8. "What has changed?"

    What has changed is that Fannie expands its balance sheet while the commercial banking system shrinks its balance sheet. MBS prepayments destroy broad money as well as bank reserves. That's deflationary in terms of the broad money effect.

    In addition, Fannie either absorbs the interest rate risk associated with the MBS or transmits it to its liability holders. Either way, what's changed is that the private sector must absorb the interest rate risk instead of the Fed. That's deflationary for the private sector.

    ReplyDelete
  9. "What has changed is that Fannie expands its balance sheet while the commercial banking system shrinks its balance sheet."

    But some private sector entity has to hold the Fannie Mae debt. That could be a commercial bank or some other private entity - doesn't really matter much.

    "MBS prepayments destroy broad money as well as bank reserves. That's deflationary in terms of the broad money effect."

    So you must think that the huge expansion in base money we have had should have been inflationary. Why wasn't it?

    ReplyDelete
  10. This comment has been removed by the author.

    ReplyDelete
  11. "If the Fed can change market asset prices by issuing short maturity liabilities and buying long maturity assets, it must be that the Fed has some cost advantage over the private sector in this type of intermediation. Outside of the circumstances of the financial crisis, what would that advantage be?"

    I think that one of the key observations here is that financial intermediaries that engage in maturity transformation are risk averse. As a result, shocks to demand/supply for specific maturities are not completely transmitted across the term structure.

    As an example (one that is out of the paper I linked to a previous comment), imagine that the Fed lowers short rates through a typical open market operation. Profit-seeking financial intermediaries will go relatively longer the long end and relatively shorter the short end, causing longer-maturity bond prices to decline, i.e., yields to rise. But intermediaries do not engage in this trade to the extent that the initial shock to the short rate is absorbed across all maturities, because this carry trade is risky. By engaging in the carry trade they are exposing themselves to the risk of interest rate increase.

    The more risk averse are financial intermediaries, the less responsive the long end will be to the decrease in the short rate. During times of high risk aversion, it follows, direct intervention by Central Banks into markets for longer term bonds might be more effective policy to influence bond yields.

    I agree however, that if financial intermediaries are risk neutral, than the Fed has no "advantage" and won't be able to accomplish much by long-term Treasury purchases.

    ReplyDelete
  12. "So you must think that the huge expansion in base money we have had should have been inflationary. Why wasn't it?"

    2nd derivatives and counterfactuals.

    It was dis-deflationary in the broad money effect.

    ReplyDelete
  13. Also, I noted only the broad money effect - not necessarily the effect of broad money on CPI inflation.

    ReplyDelete
  14. Of course, in my comment I meant to say

    "Profit-seeking financial intermediaries will go relatively longer the long end and relatively shorter the short end, causing longer-maturity bond prices to rise, i.e., yields to decline."

    ReplyDelete
  15. "2nd derivatives and counterfactuals. It was dis-deflationary in the broad money effect."

    You'll need to expand on that. Do you mean we had a huge increase in the demand for base money?

    ReplyDelete
  16. Sorry, I've been atypically brief here, for no reason at all, really.

    My original comment was:

    "MBS prepayments destroy broad money as well as bank reserves. That's deflationary in terms of the broad money effect."

    MBS prepayments into Fed held MBS result in debits to both M1 and bank reserves - they "destroy" both forms of money in the process of reducing outstanding MBS balances.

    The Fed's balance sheet expansion to date has created both incremental M1 and bank reserves. (They haven't been buying MBS directly from bank portfolios.)

    I don't particularly care about the bank reserve effect - that's just a byproduct of the main action in my view. As far as M1/M2 is concerned, the effect was to mitigate a slowdown/contraction in money stock growth that would have been much worse without the intervention. I'm not a fan of monetarism at all, but the M data would have been quite alarming without the effect of Fed balance sheet intervention.

    ReplyDelete
  17. I find it more useful to stick to the quantities on the Fed's balance sheet, i.e. outside money - currency and reserves. Under current circumstances I can think of of reserves as essentially T-bills. What makes the reserves different from T-bills is that they can be converted at will into currency by the banks holding the reserves. Now, think of the price level as being determined by the quantity of currency in circulation and the demand for it. What will come into play is alternative media of exchange (transactions accounts at banks) and the willingness of banks to hold reserves. There's nothing special about the stuff we normally include in M1 or M2. All intermediation matters, not just what the banks are doing - it's only a matter of degree (maybe Tobin said something like this). If I think about things that way, I can convince myself that if some MBS run off, under current circumstances, it will work like the process I described here.

    ReplyDelete