Wednesday, January 12, 2011

What's the Treasury Up To, Part II

Here is an update on this post, this one, and this one. Apparently, before fall 2008, standard Treasury procedure was to target the balance in the Treasury's balance in its General Account with the Fed at about $5 billion each day. The Treasury would then allow its balances in Treasury Tax and Loan Program (TTLP) accounts to fluctuate as funds flowed into and out of the Treasury. The idea was to both make life easier for the Fed, which in normal times targets the daily quantity of reserves to hit its fed funds target rate, and to earn interest on its cash balances. Targeting the fed funds rate would be more difficult if the Fed had to account for a fluctuating Treasury reserve balance, and interest on TTLP accounts with private depository institutions was significantly higher, pre-crisis, than the zero interest the Treasury would receive on the balance in its General Account with the Fed.

After fall 2008, however, Treasury procedures changed. You can find information on TTLP account balances here. The data is a bit awkward to look up, as there is a file for each date, but what this indicates is that the Treasury's TTLP balances are now fixed at a token $2 billion. The flip side of this is that it is the balance in the Treasury's General Account with the Fed that fluctuates. Why the change? First, of course, the Fed is not currently in the business of trying to hit a fed funds target. With more than $1 trillion in reserves in the system overnight and the interest rate on reserves pinning down overnight interest rates, the Fed appears to be unconcerned about the very large variance in the quantity of reserve balances held privately. Second, TTLP balances are essentially earning zero now, so there is no interest rate advantage for the Treasury in holding cash balances with the private sector.

Now, a key question is whether the fluctuations in private reserve balances matter. Clearly the Fed does not think so, and presumably banks would be kicking up a stink if this were actually disruptive. What this suggests is that dumping $600 billion in reserves into the system may not have any effect. The effects of QE2, if there are any, would then have to come solely from an effect on the term structure of interest rates from changing the relative supplies of long and short-maturity Treasury securities.


  1. Good find and good analysis. I had not realized that they made this information public, nor that TTL balances have hovered around $2 billion for the last few years.

    I disagree that the purchases have had no effect. Coming from my trading experiences of the last months, I'd say that 1. the Fed announcement that cash from expiring agency debt would be used to purchase treasuries; 2. hints that QE2 would start, and 3. an actual QE2 program, have all coincided brilliantly with the stock market bottoming and its continual rise. And also with a bottoming in the TIPS spread. [and also a great runup in the Canadian dollar].

    Now obviously there could be many reasons for these changes, but I think QE2 is a big one. Why it would have such an effect is my question.

  2. 1. I can't claim credit for figuring this out. Someone else led me to the information.
    2. Yes, I think this has to matter, at the least through a shortening of the average maturity, and therefore liquidity, of the outstanding consolidated government debt. This has to ultimately increase prices.

  3. "Yes, I think this has to matter, at the least through a shortening of the average maturity, and therefore liquidity, of the outstanding consolidated government debt. This has to ultimately increase prices."

    I don't think I understand this process. Have you written a blog post on this? Or can you send me to an outside link that explains the process?

  4. JP,

    No, this is just my evolving thinking on the matter. I haven't really fleshed this out entirely, but it seems to me that, in thinking about how the price level is determined, we have to consider the whole range of consolidated government liabilities (from currency, reserves, and T-bills to 30-year government bonds) and how that debt is intermediated, traded, and used as collateral.

  5. Given that you've been in the economics game for a while, aren't you already supposed to have a fully developed theory of price level determination? What about old monetarism MV=PQ, or the new monetarist theories?

    When you've fleshed it out, it would be interesting to see your conclusions.

  6. Sometimes the theory is worked out, sometimes there is good empirical work to help us sort out the theory, sometimes we have neither - only ideas in our heads. I don't think old monetarism - i.e. the simple-minded quantity theory of money - is very useful here. New monetarism (the name we gave to a particular research program) is what I think about, but in that research program we have not addressed seriously issues to do with the maturity structure of the debt and the role of collateral in financial transactions. New monetarist models do, however, offer a tractable way of thinking about liquidity, which is a the heart of the issue, I think.