E-piphany

Monday was as-expected…which is to say, it was quiet. Stocks ended the day flat; fixed income lost a little more steam but the 4/32nds decline in the June Note contract was not exciting. The 10-year Treasury now yields 3.72%: still firmly within its recent range, but slipping a little to the high end of that range.

It will bear watching how yields develop over the next month as the Fed ends asset purchases. Tonight Brian Sack, head of the New York Fed’s Open Market Desk (SOMA), spoke to the National Association for Business Economics Policy Conference in a general reprise of his comments last December. His remarks tonight are at this link. My generally laudatory comments on his December talk are linked to here. 

I really like the way that Dr. Sack explains how the asset purchases helped the market. In this speech, he says:

In my previous speech back in December, I discussed in detail the channels through which these market effects may arise. By removing large amounts of duration and prepayment risk from the market, the Fed’s asset purchases reduced the volume of risk that the market had to hold, which lowered the risk premia on those assets. Put differently, the purchases bid up the prices of those assets and hence lowered their yields. The lower levels of yields would be expected to boost other asset prices as investors substitute into alternative asset classes. These patterns describe what researchers often refer to as portfolio balance effects.

Where I disagree a bit with Sack is when he talks about the impact of the eventual exit. In both the December talk and this evening’s remarks, he expresses great confidence that the Fed can sell a couple of trillion dollars of securities into the market, as long as they do so slowly, or effectively shrink the balance sheet through massive reverse repurchase agreements. Recognizing that he doesn’t really have the option to say “Holy geez, this could get ugly,” I still think it’s worth pointing out that if it turns out to be the case that the Fed can keep rates low by buying lots of securities but can also sell them without impact (or add liquidity by buying them and drain liquidity by doing reverse repos), then we have a financial perpetual-motion machine.

I am generally skeptical of perpetual motion machines.

I rather suspect that when the Fed goes to sell a couple of trillion dollars’ worth of MBS, Treasuries, and agencies – and especially if they wait until the economy is growing nicely, as Dr. Bernanke has suggested they will – while the Treasury is also raising trillions in new money plus rolling trillions in maturing securities…the result may not be as graceful as Dr. Sack suggests. In several years in the early ‘Aughts…2001 and 2003, if memory serves…the bond market in October got absolutely shellacked when “convexity-related selling” overwhelmed a dealer community that was thin due to year-end risk budgets at some major dealers. The scale of that convexity selling was on the order of a couple hundred billion, and it moved thin markets a hundred basis points.

Yes, the Fed is not talking about doing this specifically in thin conditions at year-end…but the Volcker Rule being contemplated would make every day year-end for dealers, from the standpoint of their ability to take proprietary risk. These two policies are like elemental Sodium and water. They won’t mix well.

There is no data today, so I wanted to return very briefly to a comment that someone posted to this blog recently. As a reminder, I write this thing mainly so that I can receive (considered, and considerate) feedback on the thoughts and ideas, and link to greater ideas from greater minds. It also helps me avoid sloppy thinking when I have to explain my thoughts prosaically, and when a chance exists that someone will call me on the carpet for an error.

One of my favorite examples of this came not following a blog post but during a talk. I once spent a couple of days traveling to visit some investors with the great Bob Shiller. The company I was with was working with Dr. Shiller’s company on a product and we wanted to get some feedback from some important investors; people won’t open their doors for me, but Dr. Shiller always gets a meeting (especially in August 2007, when this happened). Well, at each meeting I would deliver the reasoning for the product we were working on jointly. And I would say something like “when spectacles were first invented, there weren’t lots of prescriptions. There was just one, and it probably improved your vision but it wasn’t made for you” and this led me into the meat of the pitch. 

After delivering this line a couple of times, we were at another meeting and, after delivering the line I happened to glance over at Dr. Shiller, who was looking at me intently and curiously. He interrupted me, asking “Is that true? Was there really only one ‘prescription’?”

Well, I looked back at him and tried desperately to say with my gaze, “Not now. I am delivering a pitch. It’s not even important. It’s plausible. And it makes the relevant point.” But he didn’t get the impact of my gaze, and pressed me further on the point, whereupon I had to sigh and say “I think so. I’m not 100% sure, but I wouldn’t think that prescriptions preceded the invention of spectacles. But I don’t know.” Which was awesome, to have to deconstruct my analogy right in the middle of a pitch meeting. (It turns out that the answer is that we don’t know because no one is sure when spectacles were first invented, but that’s not the point dammit!)

So when someone raises an objection to something I say, then I will often go and check my facts (and again, unlike in the Shiller case this is actually a key reason I am writing this blog). Several days ago, someone mentioned that the BLS, in making the oft-maligned “hedonic adjustments” that correct the CPI for changes in the quality of the sampled items, “never makes negative hedonic adjustments.” (I hadn’t claimed otherwise, but I was discussing the importance of hedonic adjustment and how it factors into people’s perceptions of the CPI).

It turns out that isn’t true, and in fact the biggest component of consumer inflation is hedonically adjusted to reflect lower quality over time. The BLS hedonically adjusts the housing inflation figures to account for the aging of the housing stock. If you hate the concept of adjusting the computer price index (down) for quality even when the sticker price doesn’t change, then you ought to recognize that in the case of housing, the effect works in the opposite direction because the housing stock ages and so, if no adjustment is made, over the years a rental property that goes from a mansion to a decrepit hovel as it falls into disrepair, but carries the same rent, should show inflation since you’re getting less housing for your money. 

The effect is about 0.3% or 0.4% per year on the housing part of the CPI, which doesn’t sound like much; housing, however, is about a third of the CPI so this effect adds 0.1% or so per year to the CPI. At the current level of core CPI, that is actually reasonably significant! You can read more about the update to the “Housing Age-Bias Regression Model” here.

As I have said before, I am convinced that of all the statistics put out by our government, the one that has probably been studied most meticulously and therefore does what it is supposed to do the best is CPI.