Scott Sumner  

"Helicopter drops" are a bad idea at any time

PRINT
Eminent Domain Bleg... Huemer's Moderation...

Market monetarists tend to prefer using monetary policy to stabilize the path of NGDP, not fiscal policy. We are skeptical of the efficacy of fiscal policy, and also view it as being inefficient. Now David Beckworth has suggested that a helicopter drop approach could be brought into play if monetary policy was hampered by a liquidity trap.

I have several problems with using fiscal stimulus at the zero bound:

1. Fiscal stimulus is inefficient and might not work.

2. Fiscal stimulus is not needed. As Mishkin said in the number one money textbook, monetary policy remains "highly effective" at the zero bound.

I should first explain that most people use "helicopter drop" as a metaphor for combined money/fiscal expansion, as when the Fed creates money and deposits it in each citizen's bank account (or accounts at the Fed itself.)

The fiscal part of this plan is inefficient for the same reason as it would be inefficient to give people money when not at the zero bound---the deadweight cost of taxation. Some people respond that the funds are created out of thin air, so there is no deadweight cost. That's wrong for two reasons. First, the funds will probably have to be partly removed when the economy exits the zero bound, in order to prevent high inflation. That requires distortionary taxes. Yup, there's still no free lunch. And even if the funds never had to be removed, if we are at the zero bound forever, there is an opportunity cost; the same funds could be used to reduce distortionary taxation.

That raises the second issue. Let's assume helicopter drops are not perfect, might they still be better than mass unemployment? Of course, and I'd favor them if they were needed. But they are not; monetary policy is still highly effective.

1. The best solution to the zero bound problem is to adopt a policy of NGDP targeting, level targeting.

2. Now it's theoretically possible (although unlikely) that the country might be at the zero bound even if they had an optimal NGDPLT policy. In that case they should buy more Treasury securities.

3. Now its theoretically possible (although very unlikely) that a central bank might be at the zero bound even with NGDPLT, and even after buying all the Treasury debt out there. In that case they should buy agency debt and safe foreign government bonds.

4. Now its theoretically possible (although exceedingly unlikely) that a central bank might be at the zero bound even with NGDPLT, and even after buying all the Treasury debt, agency debt, and safe foreign government debt out there. In that case they might want to consider:

a. Set a higher NGDP target growth rate.
b. Negative interest on reserves.
c. Both

5. Now there is a very tiny, tiny, risk that we've gotten past step 3 and step 4 options are not feasible for political reasons. Should you do fiscal stimulus in that case? No! Or at least not the sort of fiscal stimulus most economists imagine (tax cuts, infrastructure projects that are otherwise not worth doing, government consumption, etc.) Of course some of those projects may be worth doing, but in that case it's not fiscal stimulus, it's just ordinary government spending, worth doing even when not at the zero bound. I'm talking about the sort of stimulus that wouldn't ordinary pass cost/benefit analysis, were it not for the assumed benefit of reducing the unemployment rate. That's not worth doing, even if all the preceding initiatives leave us stuck at the zero bound.

Instead the government should buy riskier assets such as index funds of stocks and bonds. Thomas Piketty tells us that the government can expect to earn 5% in real terms in the long run, and since governments live forever, they can take the long view. Keynesians will insist this is actually fiscal policy, but I don't buy that. Fiscal policy has a cost; the deadweight loss associated with distortionary taxes required to finance debt in the long run. That's not true (on average) for purchases of stocks and bonds, although obviously it might be true in a few cases. The point is that the government should never do the sort of fiscal stimulus that is expected to lead to efficiency losses from taxes. If "something should be done," it's purchase AAA bonds, then other bonds, then stocks, etc.)

Of course all this is "angels on a pin" stuff, in the real world we'd never get beyond step 2. So no need to worry about creeping socialism.

PS. I do realize that the logic of Piketty's figures imply that countries like Switzerland and Singapore might want to borrow a lot of money at near zero rates and buy so much wealth that they can eventually live off just the 5% income generated, and free their citizens from taxes. Better yet they should have done so in March 2009. But that utopian thought is better left for another post.


Comments and Sharing





COMMENTS (19 to date)
Josh writes:

Apologies, not related. Sumner destroys Piketty again:

http://www.bbc.com/news/magazine-28472884

CMA writes:

What about heli drops conducted by fed independantly without purchasing any asset?

Say the fed provides a depository and pmt system for the money it creates for all people which is elemental if providing a money supply anyway.

Increasing supply of money to all people evenly at a moderated pace at regular intervals will induce them to spend and this will increase ngdp. Imagine the average person including the unemployed, students, low wage workers or people with little to average wealth. They wont have a high demand for money and as they offload their money the economy picks up. On top of that peoples wealth increases which improves balance sheets and eligibility for credit. As people spend the return on capital picks up and investors also pick up investment spending.

Scott Sumner writes:

CMA, As I explained, that's a really bad idea. The same growth will come from monetary policies that don't needlessly give away tax revenue, forcing higher distortionary taxes in the long run.

BC writes:

Why is it preferable for the central bank to buy, in order: treasuries, agency debt, safe foreign debt, risky stocks and bonds? If the central bank instead bought a random basket of assets or a portfolio of assets that most closely resembled the "market portfolio", then wouldn't that cause the least impact/distortion on assets' relative prices to each other, or is that not a desirable goal?

CMA writes:

"First, the funds will probably have to be partly removed when the economy exits the zero bound, in order to prevent high inflation. That requires distortionary taxes."

The fed can conduct OMO's to remove the excess money if we leave the zlb, we dont need distortionary taxes under heli drops either.

"CMA, As I explained, that's a really bad idea. The same growth will come from monetary policies that don't needlessly give away tax revenue, forcing higher distortionary taxes in the long run."

How does heli drop by fed needlessly give away tax revenue? Why is money in hands of government more important than in hands of people. Are you becoming a little socialist here?

acarraro writes:

How can be tax cuts inefficient? If you argue taxes are distorting, reducing taxes must reduce distortion, no?

And it would seem rational to increase the amount of public infrastructure in a depressed economy as more projects would pass a cost benefit analysis as there are lower financing costs (and potentially lower labour costs).

Are you arguing that demand for money is somewhat mean-reverting? So that you expect the central bank will need to drain reserves at some point? If that's the case don't you have a substantial time consistency problem? If I expect the demand for money will revert, and the amount of money will go back to the previously expected one, I have little incentive to modify my demand for money.

It seems to me you need future shocks to be random, which means there is no expectation of higher future taxes after a monetary injection.

Nick Rowe writes:

Scott: for simplicity assume a linear tax system: Ti=a+tYi where Yi is an individual's income. (The constant term "a" could well be negative.)

1. There will be some years when we want to reduce 'a' (i.e. make it more negative). (And other years we will want to increase 'a'.)

2. There will be some years when we want to increase Ms. (And other years when we want to reduce Ms).

There will be some years when we want to do both 1 and 2. (Unless they are perfectly negatively correlated.) You could say that those years are years when helicopter money is optimal.

Here's another way of looking at it: every year the government gives some people money (welfare cheques, etc.). And almost every year, the central bank earns seigniorage profits which it transfers to the government. Because money is fungible, you *could* say that almost every year the government does a helicopter drop.

acarraro writes:

By your own admission you believe that the chance of hitting the zero lower bound is pretty low when you adopt NGDP targeting. It strikes me as suboptimal to refuse support for your main policy just because someone else has a slightly different view of the optimal strategy in a tiny corner case.

Surely at some point it just doesn't really matter. Do you really care if someone believes you need fiscal intervention once every 10,000 years... You'll never have the evidence to decide one way or another anyway...

Surely raising the probability of adopting your preferred choice is worth the tiny cost...

Jesse writes:

If one admits that the economy is not completely efficient, I think it's conceivable that distributing directly to consumers or businesses or anyplace else for that matter first may be the most effective method of monetary easing. If the economy is efficient, the money should find it's way to the optimal place anyway, right? I think the political ramifications of pulling the money back is the more fundamental problems with helicopter drops.

Scott Sumner writes:

BC, Of course this would never happen with sound monetary policy. But if a central bank was silly enough to end up in this situation, it might want to minimize risk and/or transactions costs. I suppose alternative arguments are also possible. But again, it's a moot point. With good monetary policy there is no need to buy risky assets.

CMA, If they've done a helicopter drop, how do they do open market sales?

You said:

"How does heli drop by fed needlessly give away tax revenue? Why is money in hands of government more important than in hands of people. Are you becoming a little socialist here?"

I have no idea what this means. Money in the hands of the government?

acarraro, You said:

"And it would seem rational to increase the amount of public infrastructure in a depressed economy as more projects would pass a cost benefit analysis as there are lower financing costs (and potentially lower labour costs)."

Yes, but that has nothing to do with fiscal stimulus, as I explained in the post. That's just classical cost/benefit fiscal policy. Fiscal stimulus is assumed to be fiscal policies that only can be justified under the assumption they reduce unemployment.

Regarding the deadweight loss of tax cuts, you need to think of the long run effects. A stable tax rate over time is less distortionary than one that rises and falls, yielding the same long run revenue.

Yes, if rates are at the zero bound I do expect them to rise. If not, there'd be no reason to have government bonds, all spending would be financed by printing money. Seems too good to be true? That's because it probably is.

Nick, Yes, if you define "helicopter drop" as years when both the base and the national debt increase, then I see your point. But that's not really what people mean by helicopter drop. They are talking about fiscal policy that cannot be justified on cost benefit grounds (either because tax rates are too unstable, or projects that don't meet cost/benefits criteria are enacted) still being enacted because it is assumed to have the external benefit of reducing unemployment. I'm criticizing that "external benefit" assumption. Classical public finance theory suggests you want to run bigger deficits in recessions, and that infrastructure projects might be more desirable in periods when rates are low. We all accept that fact. In the tax area the models usually suggest holding MTRs pretty stable over the cycle, as I recall.

acarraro, You misunderstood me. If the Fed said they'd do 5% NGDPLT plus helicopter drops at the zero bound I'd jump for joy. I'm not refusing anyone's support. That would be much better than current policy. My point is it would be even better not to have the helicopter backup.

Jesse, If having the government just give people money is a good monetary policy, then it's also a good fiscal policy. And states should be doing it too. In fact it's a terrible idea, it suggests a sort of intellectual bankruptcy in governance. But even if I am wrong, the fiscal authorities should do that regardless of whether the Fed happens to be printing money at that moment. And then if the Fed does start printing money, it should not be given out for free, but rather used for more productive purposes, like reducing distortionary taxes (or for infrastructure, if you are a liberal.)

David Beckworth writes:

Scott,

Here are a couple quick replies.

First, Note the my proposal was not really about a liquidity trap, but about insuring against monetary policy incompetence. Helicopter drops only would be tried if the central bank failed to do its job--the very thing that happened in the 1930s and over the past five years. Sure I would love the ideal, but this is a pragmatic approach that incentives good behavior.

Second, your efficiency argument does not hold up here. My proposal is tied to a NGDP level target. Thus, if the situation arose where the helicopter backstop had to be used because of central bank incompetency it would only do so to the extent needed to restore the targeted growth path. That is, there would never be a need to unwind it since all it did was maintain a NGDP level target. It would be no different than the CB restoring the NGDP to its targeted path growth via a permanent open market operation.

Another way of saying this is under my proposal the helicopter drop would be permanent if needed. This is far different than the helicopter drop you are imagining in your critique above.

acarraro writes:

I still don't understand how you prove that a helicopter drop requires changes in taxation later on.

It seems to me this requires:
1) changes in the demand for money are temporary and require unwinding
2) government would use taxation rather than seigniorage or reduction in welfare to cover the gap

You say:

And then if the Fed does start printing money, it should not be given out for free, but rather used for more productive purposes, like reducing distortionary taxes (or for infrastructure, if you are a liberal.)

Surely you must believe that people are the best judges of what is best for them? What policy would you favour if you found some form of mineral wealth under federal land? An equal distribution of that newly discovered wealth or a policy of lowering marginal taxes? As an utilitarian, shouldn't you support the first one?

Rafael writes:

"The fiscal part of this plan is inefficient for the same reason as it would be inefficient to give people money when not at the zero bound---the deadweight cost of taxation. Some people respond that the funds are created out of thin air, so there is no deadweight cost. That's wrong for two reasons. First, the funds will probably have to be partly removed when the economy exits the zero bound, in order to prevent high inflation. That requires distortionary taxes. Yup, there's still no free lunch. And even if the funds never had to be removed, if we are at the zero bound forever, there is an opportunity cost; the same funds could be used to reduce distortionary taxation."

Imagine the following scenario:

1. Fed purchases t-securities either directly form the Treasury or immediately after in the secondary market.

2. The Fed/Treasury credits private bank accounts. In doing so, excess reserves are created as bank assets while bank deposits are created as new liabilities of the banking sector.

3. The Fed would do this as a last resort the NGDP growth path.

4. As Mr. Beckworth states:
"That is, there would never be a need to unwind it since all it did was maintain a NGDP level target. It would be no different than the CB restoring the NGDP to its targeted path growth via a permanent open market operation."

5. This isn't so different than a tax cut financed in the same way except this type of heli-drop would reach more people who are not in the current labor force at that time.

6. Less bureaucracy needed to carry this out. Money would be distributed to citizens evenly allowing them to decide how to use it rather than a government agency or contractor/corporation. This strikes me as the most "free" market solution over cutting taxes as it effects a greater number of Americans in an even fashion.

So when would the distortionary taxation be a factor? What if the taxation occurred in the same way as the drops but in reverse order? Why would increased taxation even be needed as long as NGDP is on target?

If needed:
1. The Fed/Treasury would reduce the same private accounts which were first credited during the heli-drops.

2. The Treasury buys back the t-securities from the Fed.

3. Both of these actions drain Fed excess reserves and bank deposits from the banking system. The banking system's balance sheet is reduced. The Fed's balance sheet is also reduced. The treasury bonds are retired.

4. If the original purchases are permanent then the Treasury would just issue new t-securities right to the Fed as the old ones mature.

Ralph Musgrave writes:

Scott Sumner's claim that taxes are "distortionary" is complete nonsense. Taxes CAN BE distortionary (e.g. a tax on red cars but no on cars with a different color).

But taxes can equally well be very non-distortionary: e.g. the same percentage tax on everyone's income.

Moreover, market monetarism is distortionary: it involves printing money and chanelling it just into the pockets of the asset rich, not into Main Street in general.

Floccina writes:
the same funds could be used to reduce distortionary taxation.

One way to do a helicopter drop would to stop collecting FICA and matching FICA on the first 15K. This would cut a distortionary tax.

CMA writes:

"CMA, If they've done a helicopter drop, how do they do open market sales?"

The fed can contract the MB with things like RRP, term deposit facility, the ECB can issue debt certificates. The fed should be able to issue securities like the treasury to mop up MB.

ThomasH writes:

The question remains, what is the optimal fiscal policy if we assume that the Fed will not target NGDP or NGDP futures but were to act as if it were constrained by a ZLB.

You are persuasive that NGDP targeting is robust against sub-optimal fiscal policy (spending does not increase to take advantage of decreasing shadow prices of inputs into capital projects during recessions), but what fiscal rule is robust against sub-optimal monetary policy?

Scott Sumner writes:

David, I'd prefer that helicopter drops not have been done over the past 5 years. That's why I keep pushing for first best policies--there's no reason for the Fed not to do first best policies. And I'd add that I'm not at all confident that helicopter drops would work. Combined fiscal/monetary base expansion has been tried in many countries since 1990, and failed just about every time. NGDPLT would be far more effective.

Now if you combined helicopter drops with NGDPLT it would cause much less harm, as it wouldn't be needed. But it's still not the best policy.

I don't follow your argument that the monetary injections would be permanent. If the demand for base money rises from 5% of GDP to 20% of GDP in a zero interest rate scenario, then it will fall back to 5% when interest rates rise. So if the money is not withdrawn at that point you'll have hyperinflation.

acarraro, Almost all the libertarian economists I know would prefer to cut distortionary taxes---it moves us closer to the libertarian ideal (and both policies let people spend the money as they wish.) Also see my answer to David.

Rafael, See my previous answers.

Ralph. I've never heard anyone claim that flat rate income taxes are non-distortionary. Care to explain.

You said:

"Moreover, market monetarism is distortionary: it involves printing money and chanelling it just into the pockets of the asset rich, not into Main Street in general."

This would be absurd even if you replaced "market monetarism" with "monetary policy." As written it's nonsensical on many levels. I don't even favor "printing money" I favor NGDPLT, which would involve massive reductions in the money supply.

CMA, And where do they get the money to pay interest on these "securities?" Let me guess; they print more money.

Thomas, It depends on the precise way that monetary policy is suboptimal. What is the regime? Are there any real world central banks that follow that regime? Certainly not the Fed. (see my new moneyIllusion post.)

CMA writes:

"CMA, And where do they get the money to pay interest on these "securities?" Let me guess; they print more money."

Is there a problem with that? Any increase in MB can just be offset by further issuance of fed liabs.

Btw it is less likely that fed would need to mop up a large quantity of MB in first place because if fed directly does heli's without purchasing assets then money expansion is permanent. Under asset purchase its questionable because fed has to constantly offset income from the assets it purchases.

Comments for this entry have been closed
Return to top