QE and its unintended consequences

In raising the possibility of QE2 at his Jackson Hole speech, Ben Bernanke mentioned two potential costs that would have to be assessed against any benefits of a QE - round #2, before the Fed makes a decision to that effect.

One had to do with the potential rise in inflation expectations due to perceptions that the Fed would have difficulties unwinding its vastly expanded balance sheet in the future.

The second had to do with economists’ insufficient understanding of the exact impact of central bank asset purchases on financial conditions (let alone aggregate demand). As Bernanke put it, “we do not have very precise knowledge of the quantitative effect of changes in our holdings on financial conditions. […]. [U]ncertainty about the quantitative effect of securities purchases increases the difficulty of calibrating and communicating policy responses.”

Here, I want to add a few more potential risks into the list: Are there any unintended consequences of the Fed’s asset purchases (and low-for-long interest rates more broadly), other than a possible spike in inflation expectations? For the sake of brevity I’ll just keep a list format, with the intention to elaborate on each one of the issues separately in future pieces.

So the first concern has to do with the potential systemic risks associated with the behavior of institutions such as insurers and pension funds in an environment of low (and falling) long-term interest rates. For starts, falling interest rates have increased the net present value of their liabilities (and have lengthened their duration). This does not have to have a negative balance sheet impact, provided that the asset side of these institutions benefits from matching capital gains as interest rates fall.

However, news reports and market talk suggest that this is not what has happened: Before the Fed’s hint of QE2, expectations that long-term rates had to go up from record low levels saw many institutions holding conservative duration exposures on the asset side of their balance sheet. The result has been a smaller gain on the asset side (with the concomitant increase in the funding gap) and a recent push towards long-duration positions to address the duration mismatch between assets and liabilities. The “systemic” risk here is that, if and when long-term yields begin to rise, the unwinding of these positions can lead to a much sharper rise in yields and volatility in the bond market.

A second concern has to do with the hedging behavior of mortgage/MBS investors. Low interest rates increase the probability of refinancings and, as a result, reduce the expected duration of MBS securities (and the underlying mortgages). MBS investors hedge against this prepayment risk by holding instruments of long duration, such as long-term Treasuries or interest rate swaps. Once again, if and when interest rates start going up, unwinding these positions can become destabilizing, as everyone enters the market in the same direction. (The impact of this is of course mitigated by the fact that the Fed itself holds a substantial chunk of the mortgage market).

Third on the list is the potential build-up of leveraged positions (or “carry trades”) “thanks” to the Fed’s promise of low-for-long interest rates. The risk is that crowded carry positions can unwind fast once Fed rates begin to rise, especially since the earlier leverage build-up has pushed asset valuations to stretched levels.

Now, unlike many pundits out there, I do not believe there is an empirically established causality from the Fed's policy rates to investors’ leverage. For example, the academic literature has failed to find a definitive link between the level of advanced economy interest rates and emerging market spreads (which would be obvious beneficiaries of carry trades). What does matter is investors’ risk appetite (e.g. proxied by the VIX).

And here is the challenge for the likes of the Fed: Economists do not have a complete understanding of how monetary policy affects investors’ risk-taking behavior (the so-called “risk-taking channel"). But this does not mean that it’s something to ignore, simply because it doesn’t fit into some well-established theoretical framework. More so since the framework supporting the Fed’s asset purchase program (the “portfolio balance channel”) has little to say about the build-up of leveraged positions and any associated risk from their unwinding.

A final concern—and maybe the least interesting—is the impact of low interest rates on the viability of money market funds (MMFs). Low-for-long nominal rates, combined with new regulations to increase the liquidity and reduce the maturity and riskiness of MMF investments has been squeezing the sector’s profitability. Add to that the fixed expense ratios of around 0.2-1% of assets, and the low rates can force at least some of the less efficient MMFs to closure.

The reason this is the least interesting problem in my view is that the economic impact is unlikely to be meaningful. Users of MMFs tend to be high-quality corporate borrowers who could easily tap bank or capital market financing. True, the costs of borrowing might be marginally higher, but this is most likely dwarfed by the impact of record-low interest rates thanks to the Fed's policy.

Still, I thought I should mention it, first because some economists have raised this issue in the past, and second because I was personally amused by Bernanke’s own take on this concern. Here is Ben in 2004:

“In thinking about the costs associated with a low overnight rate, one should bear in mind the message of Milton Friedman's classic essay on the optimal quantity of money (Friedman, 1969). Friedman argued that an overnight interest rate of zero is optimal, because a zero opportunity cost of liquidity eliminates the socially wasteful use of resources to economize on money balances. From this perspective, the costs of low short-term interest rates can be seen largely as adjustment costs, arising from the unwinding of schemes designed to make holding transactions balances less burdensome. These costs are real but are also largely transitory and have limited sectoral impact.”

That’s right… Bernanke (then-Fed Governor) responded by going philosophical, citing a supposed structural argument (something like, “MMFs are probably not that useful and zero interest rates can help eliminated them”), grounded on a much-debated theory (Friedman’s rule) to support a policy that is strictly cyclical in nature!

Anyway.. A mini-diversion from the main point, which is that the potential costs from further QE, and low-for-long yields more broadly, go beyond the possible rise in inflation expectations and the threat to the Fed’s credibility. Against this backdrop, any cost-benefit analysis should consider not only policymakers’ uncertainty over QE’s impact on aggregate demand but also their uncertainty about the potential risks for global financial stability.