DeLong is Right!

Brad DeLong quoted Krugman's view on what Obama should do, and then "added" some additional steps.

And he was mostly on target.

His first proposal was close to perfect:

1. Apply a full-court press to the Federal Reserve to get it to target nominal GDP to close the spending gap, for it is fear of risk that nobody will spend to buy what you make and confidence that your purchasing power is safe in cash that is holding back businesses from spending money to hire people.

All I would add is that rather than target a growth rate over a short time horizon (say, 10% next year,) to close the spending gap, this should be part of a regime change to a target growth path for nominal GDP. The target growth path should be substantially higher than the current level of nominal GDP, and getting to it should require something like 10% nominal GDP growth over the next year. But, the growth rate for that path should be slower. I prefer 3%, though 4% or 5% would be tolerable.

The new growth path regime is important because it won't create the impression that actual nominal GDP growth over the next year is not the first step to a high growth rate, high inflation growth path.

2. Apply a full-court press to the Federal Reserve to get it to engage in more quantitative easing--into taking more risk onto its own balance sheet, for it is an unwillingness on the part of Wall Street to hold the risk currently out there that is making it very difficult for a wide range of risky spending projects to get financing.

I think that the Fed must promise to do as much quantitative easing as necessary to get nominal GDP to target. I don't see any value to having the Fed purchase assets with yields that are as negative as the cost of storing currency. And so, after the yields on T-bills are driven down that low, they will have to move up the yield curve, and purchase Treasury bonds and bear more interest rate risk. If long term Treasuries have yields driven down that low, then they will have to purchase securities with more credit risk. (My argument here assumes that the Fed not only stops paying interest on reserves, but also sets charges on at least excess reserves equal to the cost of storing currency.)

Unlike DeLong, I don't see having the Fed take more risk on its balance sheet as a good thing. It rather may be a necessary evil that follows from making deposits generally, but especially reserve balances at the Fed, convertible into government issued, zero nominal interest, hand-to-hand currency. It is reasonable to assume that moving to a cashless payments system, a temporary currency suspension, or privatizing hand-to-hand currency are off the table. And so, the Fed will have to accept that it might have to hold risky assets.

If people believe that the only way for nominal GDP to rise to target is if the Fed holds risky assets, and they believe that the Fed will never do this, then this would make it very difficult to get nominal GDP to target.

However, it is quite possible that the not only can nominal GDP be returned to target without the Fed bearing much interest rate risk, once spending recovers, the private sector might be willing to hold "riskier" assets. Or perhaps it might be better to say that there will be less risk involved in holding most assets if the Fed gives up on its current regime--the one that worked so well during the Great Moderation but turned into a disaster after 2008.

Regardless, full employment of resources is possible with an allocation of resources that involves less risk. If such an allocation of resources involves a slower growth path of potential income, and even if people choose to work less, that shouldn't be considered a bad thing. The notion that "we" need to take more risk, so expected returns will be higher, so real output will be higher, so real wages will be higher, so people will be motivated to work more, seems wrong to me.

3. Quantitative easing does not have to be done by the Fed: the Treasury can use residual TARP authority to take tail risk onto its own books as well, and should be doing so as much as possible.

I agree that this would work. With T-bill yields below the interest rate the Fed pays on reserve balances, or in a slightly better scenario, where both are slightly negative and equal to the cost of storing currency, then T-bills are a perfect substitute for money on the margin. In effect, the Treasury is creating money by issuing additional T-bills.

Presumably, the Treasury, using the TARP authorization, could sell T-bills and use the proceeds to purchase risky assets. In effect, the Treasury is creating money.

Like I explained above, I don't really think the Fed needs to purchase risky assets to get nominal GDP back to target, and so going through this back door approach with TARP, isn't really necessary. Once nominal GDP is back to target, we may find that the private sector is willing to bear plenty of risk.

I suppose the advantage of using the TARP authorization to sell T-bills and fund risky assets is that the Obama administration can do this, and doesn't have to convince the Fed to change its ways. I would add that the Treasury can effectively "create money," by simply funding the national debt with T-bills rather than Treasury bonds. As longer term bonds come due, the Treasury can refinance with T-bills. Or they could even sell T-bills, purchase T-bonds, and retire them. With T-bills being (near) perfect substitutes for money, this increases the quantity of money.

(I opposed TARP from the beginning, and would rather not use any residual TARP funding for anything. But, can a "full court" press on the Fed get the Fed to come to its senses?)

4. Expansion does not require that the federal government spend: using Treasury (and Fed!) money to grease the financing of infrastructure and other investments by states would pay enormous dividends.

I suppose this means that TARP, and perhaps the Fed, should purchase state (and local?) bonds. Basically, The Treasury, using its TARP authorization, can sell near zero T-bills and use this to lend to local government. The Obama administration can do this without any Congressional action or convincing the Fed to do anything. (As above, I don't favor doing anything with TARP funds.) I agree that if the Fed does run out of Federal government bonds to purchase at interest rates higher than the cost of storing currency, then high quality state and local bonds would be a good place to go next. I don't really think this will be necessary, but having the Fed explain that it is willing to do this if necessary will help.

5. For the Treasury Secretary to announce that a weak dollar is in America's interest right now would not only boost exports, but it would immediately lead to a shift in monetary policy in Europe toward a much more expansionary profile--which would be good for the world.

I don't favor having anyone announce that a weak dollar is best, any more than I favor making announcements about the level of interest rates in the future. What should be announced is the target value of nominal GDP, and it should be explained that if this results in a lower value of the dollar, then so be it. Similarly, if keeping nominal GDP on target means that interest rates will be low, then that is the way it will be.

It is important to avoid a scenario where people think that in order for nominal GDP to get up to target, the dollar would have to fall, and that the U.S. would never let that happen because it is committed to a "strong dollar." If lots of people believe that nominal GDP cannot get to target, increasing it will be much more difficult. Having the Secretary of the Treasury say that a weak dollar is a good thing would help avoid that mistake, but what actually happens to the dollar is going to depend, among other things, on what other central banks do.

What was missing from DeLong? Getting the Fed to reduce the interest rate paid on reserves. How low? To something slightly less negative than the cost of storing currency. For that matter, the target for the Federal funds rate should be reduced to that point as well. Sure, it is unlikely to be enough, but that should be done, along with the rest. (And just forget about TARP.)