Wednesday, September 1, 2010

The Right Kind of Helicopter Drop

Some observers say the Fed is out of ammo, that any further attempts by it to stimulate nominal spending is like pushing on a string--it's futile.  This understanding ignores the fact that the Fed has yet to use all of its big guns and that these guns were found to be highly effective in ending the Great Contraction of 1929-1933.  Moreover, Fed officials including Bernanke believe the Fed could do more if it wanted to do so.  So the "Fed is pushing on a string" folks are simply wrong.  Still, it is always useful to consider exactly how the Fed could stimulate total current dollar spending. Ricardo Caballero does just that in his recent proposal to have the Fed do a helicopter drop via the U.S. Treasury Department. His proposal is very explicit in how it would work and with a few minor tweaks I believe it could be effective in stabilizing aggregate demand. Here is Cabellero:
[T]he Federal Reserve has the resources but not the instruments, while the US Treasury has the policy instruments but not the resources. It stands to reason that what we need is a transfer from the Fed to the Treasury...what we need is a fiscal expansion (e.g. a temporary and large cut of sales taxes) that does not raise public debt in equal amount. This can be done with a “helicopter drop” targeted at the Treasury. That is, a monetary gift from the Fed to the Treasury.
I would tweak this proposal in two ways.  First, I would do fiscal expansion via a payroll tax holiday.  Second, I would announce that this payroll tax holiday would be contingent on hitting an explicit nominal GDP or price level target.  Thus, as long as the target was not being met the payroll tax holiday would be in effect. I really like this proposal for the following reasons:
  1. The money is sent directly to the public; it bypasses the credit-clogged banking system and puts into the hand of the spenders.
  2. There is no increase in the public debt, thus there is no Ricardian Equivalence problems.  
  3. It is politically feasible: the Republicans get a payroll tax cut and the Democrats get fiscal expansion.
  4. It is radical enough to work.  To change expectations there has to be some shock-and-awe break from the current policy of allowing declines in inflation expectations, core prices, and nominal spending. This should do it.
The biggest drawback to this proposal is the issue of how the Fed could unwind the monetary expansion at a later date. This program would have the Fed increase its liabilities (i.e. the monetary base) without any offsetting increase in Fed assets (e.g. treasury securities).  Having these assets available would be important for the Fed down the road if, say after the economic recovery, it needed to pull back some of the money created through this program.  Caballero suggest the Fed could use some of its new tools (e.g. Fed term deposits ) or add contingency conditions that would require the Treasury to return money to the Fed. None of these solutions would be painless.  Felix Salmon suggests a way around this problem is simply to front-load the seigniorage (i.e. Fed profit)  returned to the Treasury.  It is unclear, though, how well this would work.   Seigniorage is limited and thus the Fed could not unconditionally commit to an explicit NGDP or price level target with it. It would therefore be difficult shake deflationary expectations. One soultion might be to have the Fed simply buy treasury securities directly from the U.S. Treasury instead of giving it a "monetary gift" via a helicopter drop. As long as the Fed held  securities there would be no increase in the amount of publicly held debt. Some of the debt may ultimately leak bank into the public domain if the Fed used it to reverse some of the monetary expansion.  As long as the leakage was not too much the same benefits outlined above would apply.


  1. A very interesting and useful proposal. I hope it gets plenty of
    attention. We need entrepreneurship from our policymakers

  2. ECB,

    Sometimes it takes a crisis to think outside the box, and that includes my thinking too. I never would have suggested such an arrangement before, but given the current mess we are in I find it appealing. Hats off to Caballero who I have criticized in the past. Hopefully, now I am getting radical enough for you!

  3. Yes, this is 21st century Austrian economics...its in the Popper spirit of conjecture and trial, applied to public policy. We need to experiment to break the logjam.
    But long term, institutional reform is required to restore competitive capitalism rather than the suboptimal crony capitalism we have today. Its an open question whether that will happen. What was that Zingales book, "Saving Capitalism from the Capitalists"

  4. "There is no increase in the public debt, thus there is no Ricardian Equivalence problems."

    I think this is an illusion. The ultimate effect on the public debt will be exactly the same as if the Treasury financed the purchases by issuing (and subsequently rolling over) short-term T-bills (initially at zero interest). Once the recovery happens and the Fed tightens, the public debt is going to go up, because the Fed will have to start paying more interest on reserves (thus reducing its contribution to the Treasury) or use some other method to contract the money supply.

    Caballero's proposal is fiscal policy disguised as monetary policy. Having the Fed pay for it is an accounting gimmick, since the Fed would be accommodating public borrowing anyhow, and the Fed will eventually undo that accommodation in either case (or else it won't undo it in either case). Outside money is effectively a government liability in the same way that T-bills are, and when interest rates go up, the interest on either one will have to rise.

    This is not to say that I disagree with the proposal, but I don't think it's really monetary policy; it's just a way of selling fiscal policy by pretending there is no increase in the public debt. (Or alternatively, one could say that this is monetary policy, and that it would still be monetary policy if the Treasury did it alone. Under today's circumstances one can regard the Treasury's ability to issue short-term debt as an ability to create money.)

  5. Well done professor. Perfectly succinct solutions are often better than complicated wordplay.

  6. This is excellent commentary. From here you need to get concrete--how much? Should the drops be in monthly installments until we see a robust recovery?

    A big bang? Are we taking hundreds of billions or more than a trillion. Lay out a plan--start people talking about a program, a plan of action.

    Call yourself a "growth and jobs hawk." Identify timid Fed officials as "Neville Chamberlains who quiver in front of excessive inflation dears."

    Worry about the exit strategy later.

  7. Scott Fullwiler already exposed "Helicopter Drops" as fiscal operations quite well.

    Monetary operations are a sham. Its time for people to start calling for Fed/Treasury to be explicitly cohered. This whole Fed kabuki theater is getting quite tiresome to observe.

  8. David,

    I agree with Andy Harless. There is not a lot of unique substance to this proposal in the sense that the combined fiscal/monetary effect could be achieved through other means and the fiscal cost reduced by the same amount. Any proposal to create a money financed deficit requires the systematic expansion of excess bank reserves (unless it’s done in currency - except that currency is likely to be cashed in for the same amount of deposits and reserves). And any systematic expansion of excess reserves requires the Fed to pay interest on reserves at a rate at least equal to the lower bound of the fed funds target range. For example, as Andy says, the same fiscal cost effect could be achieved here by Treasury’s issuance of short term bills.

    As to the proposal itself, the one described in the article is not sufficiently complete. Any proposal to create money financed deficits in a system that features institutionally separate treasury and central bank functions requires specified accounting entries corresponding to specified transactions. In the normal course, the bifurcated system requires consistent inter-institutional accounting, so in the case of the proposed “gift”, we therefore really don’t know what we’re dealing with until the accounting detail is fleshed out.

    I can think of at least 5 different ways of achieving the author’s objective of a “gift” in this case. They are all permutations of central bank asset liability management:

    1. The purest form in my view would be a Fed credit to treasury’s deposit account with the central bank, offset by a debit to Fed capital. The government deposit ultimately gets spent to create new M1 and new bank reserves. The net effect is that the money financed deficit is offset by a negative capital position for the central bank.

    2. Another way is for the Fed to buy debt directly from the government, crediting treasury’s account with the bank. Again, new M1 and new bank reserves are eventually created by government spending.

    3. The Fed could permit treasury to spend its account into overdraft, resulting in new M1 and reserves.

    4. The Fed could match incremental government debt issuance with QE. The net effect is an incremental money financed deficit.

    5. The most radical implementation would include the formal merging of the Fed and Treasury. The government can simply monetize expenditures at will, creating new M1 and bank reserves. Accounting for the asset side of the consolidated balance sheet is a moot issue because the combination is essentially a nominally mismatched, negative equity entity.

    Of course, only one of these alternatives is currently legal - # 4. Changing the legality would be a significant obstacle in any political environment, let alone this one.

    My own interpretation of the author’s intention is that it effectively amounts to a variation on # 2. I think of it as a “private placement” of structured treasury debt, where the structure includes repayments that are contingent on the government budget position at the time. A further refinement might include the additional condition that repayments are also contingent on central bank profits at the same time, with repayment of the debt secured by those profits instead of the central bank remitting same to treasury as per normal course.

  9. DB: I thought you might be interested to see yours and Caballero ideas are discussed in comments to this post:

  10. Caballero is following the time honored tradition of reasoning as-if one was some omnipotent monetary god.

    This ignores the many institutional features of the Fed that significantly restrict the number of things the leader of the Fed might do. Not a single one of Caballero's proposals is permissible under the Federal Reserve Act.

    These hail Mary plans that are becoming so common nowadays need to be formulated within a context that incorporates the actual legal and institutional restrictions of real-life central banks. Else they are nothing more than theorizing.

  11. David, I agree 90% with Andy. This fiscal/monetary hybrids don't really solve any of the problems associated with either policy in isolation. It is often said that a helicopter drop is a foolproof way of creating inflation. In fact, it will inflate if expected to be pemanent, and it won't if not expected to be permanent. But that's also true of OMOs, so what have you gained from adding fiscal policy? The gain is that it is considered more likely to be permanent, but that's precisely because it seems rather reckless, and that's also the reason the government won't do it in the first place.

    I'm not against the idea, but I think there are more pragmatic ways of getting the job done.

    I said I agree only 90% with Andy, because it actually is a litte bit monetary, not just fiscal. The monetary part is the commitment to raise the inflation target, so in the long run you don't pull all the extra money out of circulation.

  12. Andy, JKH, and Scott:

    Yes, I see that when the Fed tightens it will ultimately impose some fiscal cost on the taxpayer. My (maybe naive) assumption was that the fiscal cost would not be as large as the actual "monetary gift" to the Treasury. Here is why: (1) Fed will have pushed us to a higher price level/inflation rate as noted by Scott and (2)the Fed need not sale many of Tresuries back (under my latter scenario above) if it used something like Fed time deposits. Yes, these too imply less money being sent back to the Treasury and thus a fiscal cost, but would it not imply a lower current fiscal costs than selling the Treasuries back outright? If so, then there would be far less Ricardian equilance issues. Heck, if we throw in JKH's suggestions too the Fed might be able to hold on to securities long enough for them to reach maturity.

    I am for simplicity so if we can get it down the standard way--QE, explicit nominal target, eliminate IOR--I would go that route. Still, I think it is useful to think about other potentially innovative ways to get the job done if the standard approach is not implemented.

  13. Benjamin:

    Great questions... do you have any suggestions?