Friday, May 14, 2010

The ECB and Inflation Control

Earlier this week, the ECB announced that it would be buying more euro-denominated government bonds. Presumably the ECB will do what it typically does with government debt purchases, which is to buy the debt of EMU members in proportion to their sizes. However, this is clearly aimed at bailing out Greece. The EMU has decided that it wants to keep Greece as a member, and the central bank is therefore going to take on Greek debt.

Some people, including Jean-Claude Trichet's interviewer seem worried that the additional injection of outside money by the ECB will be inflationary. As I have argued previously, with respect to Fed policy, these fears are unwarranted.

To understand this, we'll need to review ECB policy institutions and recent developments on the ECB balance sheet. The ECB operates on a channel system, which has some similarities to the Canadian system. There is a marginal lending facility, for central bank overnight lending to financial institutions, and a deposit facility allowing for the holding of reserves. Current interest rates are 1.75% at the marginal lending facility, and 0.25% on the deposit facility, and these interest rates bound the overnight interbank lending rate. As well, there is additional lending through refinancing operations, at maturities of 1 week, 1 month, and 3 months. The current 1-week lending rate is 1% (see here for details).

Now, if we look at the most recent ECB balance sheet, note that, as in the US, the ECB is holding a large quantity of reserves, though not as much as a fraction of liabilities as the US. With a roughly 2 trillion Euro balance sheet, the ECB holds 282 billion Euros in its deposit facility. Consistent with this, in the May 2010 monthly bulletin, Chart 11, page 34, you can see that the overnight interbank rate is essentially being determined by the interest rate on the deposit facility (just as in the US).

Now, with a positive quantity of reserves in the financial system, and the overnight interest rate determined by the interest rate on reserves, purchases of government debt by the ECB are essentially irrelevant. The ECB is simply swapping reserves for government debt, and this will have little effect on anything, including the inflation rate. The only issue is the maturity mismatch on the ECB balance sheet - the ECB is borrowing short and lending long.

Now, in Trichet's interview, he seems a little confused about this. Like Ben Bernanke, he seems to think that he has to "lock up" the reserves in the deposit facility so that they will not escape and cause inflation. He says:
The additional liquidity that we are providing through the purchase of government bonds will be withdrawn again. Interest-bearing time deposits are an appropriate way to withdraw this liquidity.
This is the same type of term-deposit facility that the Fed appears to be going ahead with. The idea is quite lame. When there is a positive quantity of reserves in the system, the central bank can control inflation by setting the interest rate on reserves. There is no need for a term-deposit facility - the central bank just has to pay more for the reserves this way.

Here is another point of confusion:
HB: Do you now start the same what the central banks in the United States and the United Kingdom have already been doing for a while?

Trichet: That is not comparable. What the Federal Reserve and the Bank of England have done was “quantitative easing”. They were injecting liquidity into the markets and that with the explicit goal of augmenting the overall liquidity. As I said already what we are doing through the Securities Market Programme is not quantitative easing.
Trichet seems to think that he's different because of the term deposit facility. Baloney. Behind the smoke and mirrors, the ECB program is intended to subsidize Greece, just as the Fed's MBS purchases subsidize the housing market.


  1. This comment is not about the fundamental point in your discussion, which is clear and compelling, as far as I can tell. It is about a component of your argument. Let me re-write one of your paragraphs in more generic form: "...with a positive quantity of reserves in the financial system, and the overnight interest rate determined by the interest rate on reserves, purchases of outside (third-party) debt by the central bank are essentially irrelevant. The central bank is simply swapping reserves for outside debt, and this will have little effect on anything, including the inflation rate." However, could large quantities of excess reserves complicate the timely implementation of monetary policy? I was wondering if you see any merit in the ideas discussed in this piece: Incidentally, a system of non-transferable term deposits (60-day maturity, say) with no possibility of early withdrawals would appear especially helpful for solving the "behind the curve" (potential) problem. Think of this, what if the Fed would have bought all those MBSs with term deposits instead of with reserves; maybe we would have a lot less nervous monetary policymakers out there (in the US).

  2. Maybe in the end how central bank injects money, says term deposit of 3 months, doesn't matter. Banks which receive term deposit can create new liquidity, for example repo with the term deposit as collateral. Isn't it the meaning of banking?

    At the time with excess reserve in the financial system, it implies no additional borrower in the market can afford an interest rate higher than the deposit facility (risk adjusted), so does the Greece's euro bond. So ECB looks subsidying Greece by making an exclusive arbritrage opportunity for Greece gov't to buy low (its euro bond) and sell high (deposit facility in ECB). As long as ECB doesn't lower the interest rate of deposit facility, all the injection will go back to ECB rather than leaking to the economy. No inflation pressure. Do I get it correctly?

  3. I'm not following.

    Is it not as simple as this: The ECB, in a stark departure from established policy, is now directly monetizing member state debt. (Nothing wrong with this, in principle...after all, the Fed directly monetizes treasury debt).

    This change in policy (apart from the eyebrows it raises in the financial community, now worried about whether the ECB might depart from other "established" principles, like their inflation target) should have no effect on long as the Greeks, and other states, are expected (and ultimately do) pay back their debt.

    The question, however, is whether one actually expects Greece to repay its debt (in full). Suppose that Greece does not repay. How would such an event not be inflationary? And if the prospect of such default is real, why shouldn't this new ECB policy raise inflation expectations?

    One other question. I am wondering how the policy in question can simultaneously not have an effect on (expected) inflation, while at the same time constituting a subsidy to Greece. Who is paying the subsidy in your story above?

  4. Once again it seems the correct answer is "it depends". The key point of Steve's argument is this:


    Hence, as long as the government debt is not defaulted upon, everything works. Meaning, that everything depends on the fiscal policies.

    If Greece, and Portugal, and Spain, and Italy, and Belgium, ..., clean up their fiscal houses and start running primary surpluses everything works fine. This is completely analogous to Bernanke's bet: if the American households manage to save what's needed to pay their mortgages and not default (more than expected) his purchases of MBS will turn out to be a pure "liquidity injection" and it will not cause inflation.

    That's where the game is still being played, on the fiscal side. Clearly, by doing this the ECB is taking a bit bet (i.e. that Greece and company will not default), and as long as markets do not fully agree with this bet, this action raises inflationary expectations.

    It is also clear who's subsidizing whom.

    IF the ECB is correct the "losers" are the private investors that would have liked to charge a higher premium on, say, Greek debt and must now accept a lower interest rate because the ECB is also financing that debt. I am not sure this is a subsidy, but if you like to see it this way then those investors are "subsidizing" the Greek government. But, as long as they purchase that debt at whatever interest rate the ECB manages to determine, I am not sure I would really use the word "subsidy". They just earned less money than they hoped to.

    IF, instead, the ECB is wrong, then it will be usual stuff. The average European consumer/taxpayer will have subsidized Costas' consumption and leisure.

    Just wait and see. This week it may be Portugal's and Italy's turn to feel the jitters ..

  5. P.S. Sorry, I did not realize that the code running these blogs treats <<>> as an "erase" command. The statement by Steve I meant to quote, in place of the <<>> you now see was:

    The ECB is simply swapping reserves for government debt, and this will have little effect on anything, including the inflation rate. The only issue is the maturity mismatch on the ECB balance sheet - the ECB is borrowing short and lending long.

  6. "The only issue is the maturity mismatch on the ECB balance sheet - the ECB is borrowing short and lending long."

    This maturity mismatch is to some extent inflationary.

  7. The maturity mismatch is inflationary only if the ECB holds the long-term assets for a long enough period of time that short rates have to go up. If they go up enough, the ECB can be losing on the intermediation activity, whether they sell the long-maturity assets or not. They make up the losses by printing more outside money, which of course is inflationary.

  8. The maturity mismatch is also inflationary because it strengthens the credit channel of monetary policy.