David R. Henderson  

Woolsey and Sumner on QE3

Rafat Channels Tullock, Kuran,... You Might Be Signaling...

The best piece I've seen on QE3 so far is by Bill Woolsey. Even though it's long, there aren't many wasted words. So the whole thing is worth reading.

I do want to highlight one important point he makes. Woolsey writes:

The other "problem" with the Fed's policy is that it has specified that it will be purchasing mortgage backed securities. Since these are all "agency debt," that means they are already guaranteed by the U.S. taxpayer for credit risk.

Generally, Market Monetarists (like Traditional Monetarists) do not worry much about what particular assets the Fed purchases. It is the impact on the Fed's liabilities--the quantity of base money--that is important. However, this policy certainly has an odor of credit allocation--the Fed is trying to encourage people to obtain home mortgages and buy houses. Bernanke reinforced this view at the press conference by explaining how the policy is supposed to work. It is supposed to lower interest rates on home mortgages and so more mortgage lending and more spending on housing. [italics added]

Scott Sumner, replying to a similar point by George Selgin, writes:
I'm tempted to respond; and how? Most people, including George Selgin and market monetarists like David Beckworth, believe that an easy money policy during the early 2000s led to an overheated economy in the middle of the decade, and that this was one factor in the crash during the latter part of the decade. I'm not convinced, or perhaps I should say I doubt the effect was as strong as most people believe. First of all, I see little evidence that monetary policy was particularly expansionary during the early 2000s. Some people cite the low rates, but we all know what Milton Friedman said about that.

Scott makes an important point about monetary policy, one that Jeff Hummel and I have made, but he misses the point Selgin, Woolsey, and I are making here. The charge is not that it's too easy a monetary policy: the charge is that it's allocating credit to a particular sector and, therefore, is, to some extent, a credit allocation policy. That's dangerous. It's more of the central planning by Bernanke that Jeff Hummel has gone after. And to housing of all places? That's just perverse.

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (6 to date)
Ritwik writes:

All asset price driven monetary policy is distributively non-neutral.

Treasury bonds are not held by the representative household. Those that hold them in pension accounts can't borrow against them.

If some risk capital has to be bought, it might as well be housing.

With a secular decline in business investment, housing IS the business cycle in contemporary US. Ben Bernanke realises this. And thankfully, has at least gone for new mortgage origination rather than existing MBSs. It's credit allocation, yes. But no reason why it's any worse than the wealth-transfer that T-bond purchases effect.

At least this one might work, instead of channeling cheap funds into commodity speculation.

Mike Rulle writes:

It does seem like targeted central planning at first blush. But if so, it has already been occurring. 30 year mortgage rates have been declining with the 10 year treasury rate and is basically within historical spread averages. The Fed has purchased about 20% of public treasury debt (i.e., not counting the fake intergovernmental accounting debt---fake in the sense that it has not yet been issued by the government to the public, hence it does not exist---in theory laws could change and it would never be issued).

Almost 20% of our debt has been monetized---amazing. And still it may not be enough if Scott and Co. are correct. It is hard for me to accept Scott would be correct if our fiscal policies were more sound and certain. Or that Scott can be correct at all. How much QE has to occur without results before we accept that is not what is causing our NGDP shortfall? But I am digressing from your point.

Buying mortgages directly would seem to have a greater impact on mortgage rates. But it will take 2 years for the Fed to purchase about 10% of current mortgages outstanding at $40 Billion a month. The marginal difference in mortgage rates (already at historical lows) between buying more Treasuries versus Mortgages seems, well, marginal.

From a taxpayer perspective, which is preferable? I would argue mortgages because they are backed by real assets. Treasuries are backed by, well, an ability to tax more. So if we are going all in monetizing debts, I prefer owning the debt which has an asset behind it, however incrementally inflated it might be as a function of buying mortgages over treasuries.

As to your main point, we have been subsidizing housing since the crash---thus preventing a true recovery. I also, oddly, agree completely with Scott that it was not monetary policy which caused the housing bubble per se, but a variety of fiscal, regulatory and financial institutional actions and policies.

Scott Sumner writes:

David, I'm confused on several points.

1. Even if you are 100% right, it doesn't impact the argument I made to George, as I was discussing the period before the Fed started buying agency debt. Indeed I was relying to George's claim about what went wrong in 2003-06.

2. If the agency debt is 100% backed by the Treasury, as Bill claims, then it's effectively Treasury debt (or a perfect substitute.) In that case it would make zero difference (for credit allocation) which type of debt was purchased.

I'm willing to entertain the hypothesis that it is not a perfect substitute, but in that case Bill's argument seems like a non-sequitor.

I will say that I strongly oppose Fed credit allocation, and always have.

Mike, I don't think the Fed needs more bonds, it already has more than enough to hit a reasonable NGDP target. The main problem is the NGDP target is too low. If they insist on such a weak target, then (like Japan) they'll be forced to buy up endless quantities of debt.

Marc Joffe writes:

I believe we will look back in a few years and conclude that QE3 was pro-cyclical. In today's San Francisco Chronicle, business columnist Kathleen Pender reports that inventories of homes for sale in California are below their long term averages. This suggests that the recent escalation in home prices is going to gain momentum.

Like Greenspan's final discount rate reduction in 2003, Bernanke's action is being taken long after the economy reached its trough.

Due to deleveraging and demographics, potential real GDP growth is lower than it was in the 80s and 90s. Slower growth means more persistent unemployment. It is unfortunate that policymakers (even the unelected ones) don't seem to take these factors into account.

QE3 exemplifies a classic problem with discretionary monetary policy. Bernanke does not want to go down in history as the guy who presided over permanent 8% unemployment, so he feels the need to do something - anything - to push this lagging indicator down.

Monetary rules don't have emotions, aren't concerned about their reputations and are thus less prone to overreact.

Costard writes:

More important than the distributional effects of Fed asset purchases, are the distributional effects of Fed guidance during an asset boom. At that point, dovish remarks are taken as an endorsement of the bubble du jour. Speculators understood in 2005 that interest rates, whether excessively low or not, were nevertheless accommodating to the housing market; what they worried about was whether the Fed saw any "irrational exuberance" or not -- ie the future path of rates. At a certain point Greenspan's reassurances were read as, "the powers that be will never let housing prices fall."

People are absolutely wrong to view NGDP as an incorruptible figure. Renminbi devaluation and Chinese imports went a long way towards preventing inflation from showing up at the consumer level. The dollar index was falling dramatically at the time, but China's dollar peg meant that toothbrushes and t-shirts remained cheap. The Fed was fooled by this, and it seems many folks remain fooled. Both inflation and NGDP were understated, and as a consequence accommodative policy was maintained for longer than it otherwise would have been.

@Scott: "I will say that I strongly oppose Fed credit allocation, and always have."

Every monetary mechanism is going to favor some financial assets over others; oftentimes we don't know which ones until after the fact. By the time the Fed becomes aware of any unintended favoritism, it will be faced with a choice between deflating the recovery or endorsing an imbalanced growth. Experience tells us it chooses the second -- or in our current situation, eases further in an attempt to "even things out". More likely, the Fed will have to send financial markets to the moon in order to have any appreciable effect on jobs. I suppose this is the grand experiment upon which we are embarked....

I agree with Scott that talking about the credit allocative issue is the wrong way to think of this. Whatever the Fed policy does in that area pales into insignificance with, say, the GSE Act of 1992, the HUD Best Practices Initiative of 1993 and several other direct interventions. Some of which still goes on today, sadly.

But, I wish Scott had gone into more detail about his, 'Mike, I don't think the Fed needs more bonds, it already has more than enough to hit a reasonable NGDP target.'

That is the question. Why did a tripling (by the earlier QE sailings) of the Fed's balance sheet not result in more 'money' (Friedman and Schwartz's 'money stock') actually circulating through the economy? Back in the 1930s the monetary base actually increased at the same time the 'money stock' decreased by about 1/3.

It's less severe this time around, but essentially it's the same phenomenon.

Comments for this entry have been closed
Return to top