The debate about the impact on US treasury yields from the Federal Reserve’s LSAP programs—often referred to as Quantitative Easing—is raging into its fourth year. In fact, now that the time series are getting long enough for more robust number crunching, I suspect academics are going to really start diving in and begin the writing of history.
Practitioners, however—both policy makers and those of us who have money on the line—don’t have the luxury of time. We are entering a critical phase right now. Why? Household leverage has been the prime impediment to a normal functioning of monetary policy. And it is now starting to get down to a point where monetary policy will start gaining traction. Not a lot of traction, because these processes are slow, but any traction at all means the days of the dreaded liquidity trap are numbered.
So we are now going to have think harder, and in more practical terms, about the counterfactual: where would UST rates be without the exceptional monetary stimuli the Fed hath wrought.
There are two basic camps in this debate, and from where I sit, neither side has it quite right.
One camp says the Fed’s massive purchases of USTs and agency mortgages have artificially lowered rates a lot. Looking at UST yields and spreads throughout the fixed income complex gives them sticker shock. Some fixed income managers are even mad. They fear that this is inducing a misallocation of resources, incenting higher government spending than would otherwise be the case, is hurting savers, and might constitute a new bubble. Many in this camp fear high inflation will follow. Their prediction for when the Fed stops buying? Pain—pain in markets, pain in the economy, and pain in the budget, stemming from the higher UST rates they assume will follow.
This camp comprises much of the professional fixed income asset management crowd, the majority of sell-side strategists, a fair number of economists (e.g. the recent op-ed by Marty Feldstein in the WSJ), and virtually all of the Policy Bears (think, for example ZeroHedge or CNBC’s Ric Santelli).
The second camp claims it is all about expectations, not physical purchases, and that QEs have actually raised UST yields relative to where they would otherwise be. Joe Wiesenthal over at Business Insider was perhaps the first to propagate this view. Others, like Matt O’Brien at the Atlantic and Matt Yglesias at Slate, have more recently laid out the same basic case: looser monetary policy from central bank bond buying raises, rather than lowers, rates because the indirect effect through expectations on future nominal GDP growth is greater than the countervailing pressures from bond purchases.
This camp comprises an increasing number of sharp-eyed financial/economic journalists, some of the more nuanced fixed income veterans, most salt water economists, and a lot of equity managers (who always seem to be on the lookout for a bullish story).
This view is always buttressed by some version of the very convincing chart shown below, in this instance lifted from Matt O’Brien:
In it, one can see very clearly that when the physical purchases of USTs and mortgages were taking place bond prices were indeed going down and yields higher.
Conversely, the moves higher in price and lower in yield happened when the Fed “wasn’t in the market”.
The rationale is simple: the Fed tended to hint at or announce QE programs when the economy and markets appeared to be weakening sharply. The chart shows that even though we were in the throes of a deep in a liquidity trap, the psychological effect of Fed support was strong enough to snap us out of slide into self-reinforcing pessimism and move us away from nastier equilibria.
Okay, that was easy. So, case closed? QE means higher rates, right?
Not so fast. Look more closely at the chart.
The idea behind large scale asset purchases, of course, is that they are supposed to drive down interest rates and facilitate the healing of bloated private sector balance sheets, in our case, in the household and financial sectors. This, in turn, would lead ultimately to a resumption of lending, once the lenders and borrowers have worked themselves back into stronger financial positions.
The theoretical debate has taken for granted that LSAPs lowers rates, and instead focused on the channel through which the purchases would achieve this. Thinking about these channels is important.
First, there is the “Flow” channel. Some academics and most markets participants have been inclined to believe that LSAPs lower rates through a flow effect, that is, through the physical purchases. The intuition here is powerful: sharply increased demand means higher prices, lower yields.
On the other hand, many academics and policy makers—including the bulk of the Fed—believe rates are lowered through a stock effect. That is, asset purchases reduce the available stock of assets, and so, for a given view, the clearing price will be higher (and the yield lower) than otherwise would have been the case. The implication is that this affects, over time, the level of yields, even if the oscillations in yields are driven by other factors, such as economic expectations.
Now, let’s go back and look at the chart again. You will see what technicians call “lower highs and lower lows”. And it’s important to note that this pattern was taking place against the backdrop of an improving economy which would normally push UST yields higher.
This to me means two important things: one, LSAPs have almost certainly over time lowered the clearing rate for UST yields—even though the impulse correlation, driven by economic expectations, has worked in the short-term in the opposite direction.
The second observation is that the “expectations effect” was of lesser amplitude with each Fed announcement, again, against the backdrop of an improving economy. In fact, after QE3, there was virtually no bump at all. This is because sentiment surrounding monetary policy has done a 180 over the past two/three years. Because the shifts in expectations were not subsequently validated by fundamentals, market participants progressively came to view effects from Fed policy as psychological and ephemeral. It went from ‘very hard’ two years ago to make the case that QE wouldn’t be inflationary to ‘fairly easy’ today. Most everyone by now has wrapped their head around the notion of “liquidity trap”.
Two conclusions can be drawn from this. One, the end of LSAPs will matter for yield levels—even if the Fed decides not to sell any of its holdings and let their book run off. So, if you think it is entirely about economic expectations you are likely to underestimate the magnitude of yield “normalization”.
Two, many investors and analysts have settled into the notion that we are in a liquidity trap, and that monetary policy here is largely “pushing on a string”. While this is still for the most part the current environment, it is finally, slowly, starting to change. Monetary policy can be very, very powerful when the soil is fertile. This is not the time to become complacent about the impotence of monetary policy. That time has passed. It may not be tomorrow, but the efficacy of monetary policy has now become, as the economists might say, a positive function of time.