After QE2, what then?

Now that the end of QE2 is in sight, everyone is wondering what happens next. Is a QE3, focused perhaps on outright purchase of municipal bonds, in the offing? Or will we once again start hearing about “exit strategy”?

And how, if at all, does policy on the QE front interact with general interest rate policy? The ECB has raised its policy rate, arguably more as a signal to European political authorities that they should not continue to rely on the ECB alone to manage problems in Portugal and elsewhere, but nonetheless reminding markets that interest rates will eventually rise, even in the United States.

Inevitably, policy debate about these questions runs largely in intellectual grooves formed before the crisis. Notwithstanding the revolution in modern finance, the deepest groove remains apparently the Tobin-era portfolio choice approach. From this point of view, QE2 and QE1 were both about the Fed taking duration risk out of the market, so reducing the market equilibrium spread between long rates and short rates. According to this point of view, when QE2 ends, duration risk will remain on the Fed’s balance sheet, and hence the term spread will remain at current levels, ceteris paribus.

In the money view perspective, by contrast, the Fed is not conceived so much as a financial intermediary but rather as a kind of dealer or market maker that is able to influence market prices only insofar as it is willing to buy at a higher price (or sell at a lower price) than anyone else. Buying at a higher price means absorbing all the supply offered at that price, and that means financing the resulting asset positions either by selling something else or by borrowing. The social purpose of such intervention is to keep asset prices from getting too far away from fundamental value—if markets themselves kept prices close to values there would be no need for such intervention.

The consequence of the Fed’s interventions can be seen in the transformation of its balance sheet over the last three years. Dan has shown the asset side consequence over time. Here is the current (April 14 release) balance sheet, both sides, in trillions:

Assets Liabilities

Over the last three years the balance sheet has more than doubled, but also the composition has changed dramatically. Before the crisis, assets were almost entirely short term Treasury bills; now they are almost entirely long term Treasury plus MBS and GSE. Before the crisis, liabilities were almost entirely currency, now reserve deposits are larger than currency.

In my previous post, I emphasized how the rise of shadow banking can be seen as the counterpart of the transformation of traditional asset management. Shadow banks supply the quasi-money demanded by asset managers, while derivative dealers act as risk intermediaries, offering risk hedging to the shadow banks and risk exposure to the asset managers.

Following this line of thought, we can understand the transformation of the Fed’s balance sheet as the consequence of its intervention to backstop this new system of funding and risk intermediation.

But we can’t see it very clearly in the accounts as they are customarily presented. Instead, it is helpful to rewrite the Fed’s balance sheet as:

Assets Liabilities

All I have done is to add 1 trillion of Treasury notes and bonds and 2.3 trillion of Treasury bills to both sides; the net exposures are exactly the same. But now we can see the risk intermediation interventions clearly, and separately. The top line shows interest rate exposure; the second line shows credit exposure; and the third line shows liquidity exposures.

To see the point even more clearly, it is helpful to view the exposures as synthetic derivatives and to write the balance sheet as:

Assets Liabilities

Consistent with my previous posts, I am showing these derivative exposures as liabilities. An interest rate swap involves receiving interest at a fixed rate and paying interest at a floating rate; a credit default swap involves receiving interest at a risky rate and paying interest at a riskfree rate; a money market swap involves receiving interest at a riskfree rate and paying interest at a liquid money rate.

Before the crisis, the Fed balance sheet was essentially a money market swap of about 800 billion. Three years later, we can see that the Fed’s response to the crisis was to triple this original exposure, and to add also 1.0 trillion of CDS exposure and 2.3 trillion of IRS exposure.

All of these exposures arose as a consequence of the Fed’s dealing activity—it offered to buy risk exposures at a price that the market, given disruption, found attractive. QE1 was about buying MBS with reserves, hence both CDS and IRS exposure, as well as MMS; QE2 was about buying Treasury notes and bonds with reserves, hence IRS and MMS only.

This way of thinking about what the Fed has been doing can help us to understand better what is at stake in the end of QE2, and any eventual exit strategy. Each of the three exposures can, and should, be considered separately. This will be the subject of my next post.