What the left hand buyeth, the right hand issueth

Roger Farmer has a great post showing how the maturity composition of outstanding US debt lengthened as the Fed embarked on its program of buying long dated securities.    As Summers reportedly put it, while the Fed was engaged in quantitative easing, the Treasury was doing ‘quantitative contraction’.  And surely the two arms of government should be better coordinated than that.

A few points.

The first thing to note is that what we saw doesn’t necessarily imply that the US Treasury was shifting its behaviour in order to scupper what the Fed was doing.  Its mandate to get value for money when selling debt could well imply a maturity composition of issuance that means the proportion of long-term debt brought to market rises as the term premium falls (the relative price of long-term bonds to short-term bonds rises).  So it’s conceivable that that issuance pattern amounts to running on autopilot.  In order for us to substantiate that the Treasury obstructed the Fed actively, we’d have to detect a shift in the historical relationship between the term premium and the composition of issuance.

In the UK, the Debt Management Office, an agency of the UK Treasury, and separate from the Bank of England since 1997, undertook not to change the pattern of issuance away from what would have been the case in the absence of QE.  One might have hoped for this strict form of cooperation in the States too, if you were a QE supporter.  But doing that would involve a shift in behaviour, forgoing ‘value for money’ considerations in debt sales that would lead to maturities being sold into parts of the curve where demand was strongest.  The DMO/BoE might have expected QE to reduce the term premium (especially once purchases were shifted away from the essentially neutral maturity composition of early rounds) and this would normally prompt the issuer to take advantage.  It will be interesting to see whether the DMO can be judged to have kept to this promise.  With such an extended program of purchases, I don’t see how such a promise could be evaluated.  How would the DMO or any similar agency know what it would do such a long time ahead?

Regardless, it’s also conceivable that Fed plans were not thwarted in their attempt to achieve a certain amount of stimulus through quantitative easing.  If the Fed understood the unchanging or even changing maturity reaction function of the Treasury, their purchases from the open market could be calibrated accordingly.  So long as we suppose that there are no costs associated with tilting the Fed balance sheet to compensate for the extra Treasury twisting, the Fed could do whatever was necessary to achieve its goals.

Note that some like Vissing-Jorgensen and Krishnamurthy question whether lowering the term premium through QE is a good thing at all.  They document that private yields were not materially lowered by QE.  And describe how purchases of long-term debt imposed costs by depriving the markets of a safe store of value with duration.  For these authors, even if the Treasury had been trying to frustrate the Fed, they would surely be saying ‘two thumbs way up’, since the policy would have relieved society of a costly shortage.

Some might balk at calls for coordination between the two arms of government for fear that it corrodes the independence of the central bank in its conduct of monetary policy.  These calls are not entirely vacuous, since histories of the Fed in the 1960s and 1970s seem to conclude that the Fed regularly subordinated itself.  But here what is being asked for – regardless of its rights and wrongs – is for the Treasury to subordinate itself to the Fed.  The reverse of what is usually feared.  If that were achieved (presuming it hasn’t been already) it does not follow that the Fed would have to reciprocate in some other way.

‘Monetarists’ of sorts may scoff at all this debate, arguing that this is all of second-order importance.  What matters, they might say, is how much the balance sheet expanded in line with the corresponding creation of central bank reserves.   For them, the composition is just a detail.   Of course, concerning oneself solely with total issuance and total purchases leads to a related question about co-ordination, even subterfuge, and one that certain political factions have pushed, namely, whether the Treasury’s total issuance rose because of QE, or vice versa.   However, I personally lean towards the view espoused by Woodford and others in several papers, which is that the monetary expansion part of QE is essentially redundant, no different in effect from forward guidance, and that the only independent stimulus is to be got by the twisting of maturities, or the risk transfer from private to public.  Under this view, the effects seen of quantitative easing could have been got by the Treasury issuing very short-term debt in order to take back longer term debt.

What this debate and the ambiguities in resolving it make clear is that some institutional refinement is in order.  Appropriate coordination needs to be hard-wired into the debt management function and into monetary policy, not currently in place in either the US or the UK.  My involvement in QE on the inside at the Bank of England convinced me that all parties were determined to do the right thing.  But that British kind of solution to this new (or is it old) problem is probably not the best one, because it allows composition or quantity conspiracy-theorising to live on.

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

One Response to What the left hand buyeth, the right hand issueth

  1. Pingback: QE and Fiscal Offset | askblog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s