Sunday, April 4, 2010

LSAPs: A Tale of Overkill Gone Too Far

With the Fed’s quantitative easing (QE) completed last week, I thought it might be a good time for stock-taking: Did QE achieve its intended objectives? And could the Fed have done things better?

By QE I mean of course the “Large-Scale Asset Purchases” (or LSAPs) of GSE debt, mortgage-backed securities (MBS) and US Treasuries. These were first announced in November 2008, expanded in March 2009 and concluded in March 2010.

So let’s start with the intended objectives first. In the case of the purchases of MBS ($1.25 trillion) and GSE debt ($200 billion), the objective was clearly stated at the November 2008 Fed statement:

“[..] to reduce the cost and increase the availability of credit for the purchase of houses, which in turn should support housing markets”

In other words, at the height of the crisis, the Fed decided to provide enormous support to a specific sector (housing), in the context of its efforts to “improve conditions in financial markets more generally.” As I argued at the time (here and here), by effectively getting the Fed into the credit-allocation business, the MBS purchases were an overreach of its mandate and a potential threat to its independence. They were also unnecessary, as we’ll see below.

Nonetheless… There is no doubt that the Fed achieved its stated objective: Mortgage yields have shrunk since the purchases were announced and the reason is clear: When the Fed buys up an enormous share of the MBS market, it bids up MBS prices and lowers their yield—full stop.

So far so good. But QE was not just intended to affect the price (/yield) of the assets purchased. Another critical objective was to affect the price of other risky assets such as equities and corporate bonds through the so-called portfolio balance effect.

The clearest description of how the portfolio balance channel was thought to work was provided by the NY Fed’s Brian Sack last December (my emphases):

“[T]he purchases bid up the price of the asset [being purchased] and hence lower its yield. These effects would be expected to spill over into other assets that are similar in nature, to the extent that investors are willing to substitute between the assets.

[..] With lower prospective returns on Treasury securities and mortgage-backed securities, investors would naturally bid up the prices of other investments, including riskier assets such as corporate bonds and equities. These effects are all part of the portfolio balance channel.”


The problem with this argument is that it treats Agency MBS and Treasuries (i.e. the things the Fed bought) as assets that are “similar in nature” with corporate bonds and equities. However, as highlighted in a 2004 paper by Takeshi Kimura and (the Fed’s own) David Small, this may not be true.

The reason is that the returns of equities and corporate bonds (other than high-grade) tend to be positively correlated with an investor’s consumption stream; whereas the returns of “safer” assets such as US Treasuries (and, arguably, GSE(/government)-backed MBS) tend to have a low or negative correlation with private consumption.

This means that, in a portfolio context, such “safer” assets provide a hedge against a drop in consumption during bad times (i.e. during a recession and uncertainty about labor income).

And for this reason, QE operations that remove “safer” assets such as Treasuries (or Agency MBS) from the market give rise to portfolios that are heavily “overweight” pro-cyclical assets (like equities and high-yield corporate bonds) compared to investors’ optimal allocation. In response, investors might actually shed pro-cyclical assets to rebalance their portfolios, raising their risk premium. In other words, the Fed’s LSAPs in their 2008-2010 form may have had the opposite spillover effects from what the Fed had wanted to achieve.

What are the lessons here?

The first thing to note is that, even at the zero bound, the Fed can achieve a great deal without LSAPs. This is via (1) the commitment to keep the policy interest rate at near-zero levels for an extended period; and (2) the ability to extend unlimited amounts of liquidity to the financial system to safeguard financial stability, thus helping address the root of the crisis (like the ECB did).

In my view, it was these two tools (along with the bank stress tests and recapitalizations) that were instrumental for the rebound in financial markets since mid-March 2009. Specifically, the commitment to low-for-long rates helped to:

• bring down the level of private long-term interest rates by reducing the long-term “risk-free” rate;
• lower credit spreads, since lower interest rate levels mean lower debt servicing costs and, thus, lower expected probabilities of default;
• improve the outlook for credit conditions and aggregate demand, which in turn has reduced risk aversion (and, thus, lifted asset prices).

Meanwhile, the liquidity (and recapitalization) programs helped improve financial stability and lift asset prices by reducing systemic risk and volatility.

Against this backdrop, the Fed’s asset purchases can be seen as a supplementary tool, in the context of policy overkill. However: As argued above, for LSAPs to lower the private long-term cost of capital they have to directly target pro-cyclical assets such as equities, high-yield corporate bonds and other “toxic” stuff; not Treasuries and Agency MBS!

Now, you’ll probably say that for someone who cares about the Fed’s financial independence (and I clearly do), this is a toxic proposition. Not really! The Fed already took a large amount of risk on its balance sheet with its involvement with Bear Stearns, Citi and AIG, as well as with the MBS purchases (unless it holds them *all* to maturity).

Once you’ve gone through that route, the case for focusing on Treasuries and MBS alone is weak at best; and it’s also counterproductive and inconsistent with its Fed mandate, as argued above.

What matters for the portfolio balance channel to work is that the Fed removes large amounts of pro-cyclical assets from investors’ aggregate portfolio; the Fed does not need to target a specific asset class.

In fact, had it bought a mix of equities and corporate bonds (indiscriminately, e.g. by buying equity and bond indices) it would have made tons of money AND would have also avoided the fear of causing market havoc when trying to offload these assets later—a dominant concern in the Fed’s and investors’ minds when it comes to unwinding the MBS purchases.

Call me a purist, but I continue to see the Fed’s MBS and Treasury purchases a mistake; I did back then, and I still do now. All I can hope is that researchers can reach consensus about this before the next crisis!

1 comment:

bb said...

i wouldn't call myself a purist, but would rather point at the untold reason to target agency paper: the realisation in the whole world that those monsters with 50 times or more leverage are bust, the ability of the government to finance its growing fiscal deficit, to originate new home loans in falling prices environment (10% cash down is just way too much for the average folk), and lastly, to support prices of housing, their stated objective.

i have the gut feeling it is just a matter of time before they resume their MBS purchases.