AN UNCONVENTIONAL ASSESSMENT OF UNCONVENTIONAL MONETARY POLICY.

• Author: Reinhart, Vincent

Beginning in late 2008, the Federal Reserve--an independent, technical agency of the U.S. government--experimented with resources cumulating to about one-fifth of annual nominal GDP. There was no settled body of research supporting the intervention. Indeed, there was not much research on the topic in the prior 35 years. But it seemed like a good idea at the time to the leadership of the Fed and proved sufficiently obscure to the media and elected classes to mostly fly under the radar relative to the scale of the commitment involved.

Ten years and $3.5 trillion of net asset purchases later, there is still no established answer to what effects the Fed's unconventional policies produced or what the legacy of those polices will be. Yet, the footprint of those policies remains considerable. The Fed's balance sheet (Table 1) is still close to $4 trillion and about 20 percent of nominal income, even after allowing some maturing investments to run off. The irony of quantitative easing (QE) is that its effects are hard to quantify. Prior to 2008, Fed officials lived in a comfortable world delimited to the basis point--actually to 0.25 percentage point--to which their interest-rate decisions always rounded. But unconventional policy, which includes adjusting guidance on interest-rate intentions and manipulating the size and composition of the central bank balance sheet, maps uncertainly into interest-rate expectations and liquidity preferences of investors.

In this article, I offer three perspectives on the Fed's unconventional monetary policy. First, I draw on my work with Ben Bernanke in 2004, in which we framed expectations on the efficacy of unconventional monetary policy. The next section revisits those arguments.

Second, there is a cottage industry, including many contributions by Fed staff, in assessing the effects of unconventional policy. (1) This article employs finance theory to identify windows where views on policy are likely to be revealed and to look for relative price discrepancies among mostly similar instruments to parse policy effects. This is the popular solution to look under the lamp post for the missing item; I widen the window of observations to demonstrate that the effect proves ephemeral.

Third, this article exploits the enormous comovement among yields along the Treasury term structure. As has been known since Litterman and Sheinkman (1991), rates along the yield curve are well described by the current level, slope, and curvature of the yield curve. The fascinating feature is that any portion of the yield curve predicts other portions, a property I use to determine if the relationship between short and long rates changed pre- and post-QE. Long-term yields were a bit lower in the QE episode than short rates predicted but much more subsequently. Both approaches suggest that QE matters, but the effects are small relative to other macro forces pulling yields lower and likely to erode as the Fed's relative footprint in markets shrinks over time.

The last section crosses over to the normative issue of the appropriateness of a large and lingering Fed footprint in markets. The revealed preference of policymakers is that they do not have sufficient confidence in market mechanisms or respect for the role of risk in directing the efficient allocation of resources. A healthier respect for both would place stricter limits on the extent to which a central bank leans against financial market volatility than was the case. The problem is that the precedent lowers the bar for future intervention and leaves the Fed operating under too large an ambit in our market economy.

What Were They Thinking?

The prior time that the Federal Reserve had a near brush with the lower bound of zero to the nominal policy rate was from 2002 to 2004. (2) The nominal federal funds rate touched the then unprecedented level of 1 percent, and replicating the ongoing "lost decade" of poor economic performance in Japan appeared a worrisome possibility. Among the work done at the time identifying policy alternatives was Bernanke (2002), Bernanke and Reinhart (2004), and Bernanke, Reinhart, and Sack (2004). When confronted by an adverse shock but pressed by the zero lower bound to nominal interest rates, a central bank could (1) signal to keep interest rates low for a sufficient time to pull longer maturity yields lower, (2) increase the amount of reserves beyond that necessary to push the policy rate to zero, (3) purchase assets in volume to influence their relative spreads, and (4) use administrative policies to subsidize lending.

The subject of this article is the third tool, also known as quantitative easing and large-scale asset purchases (LSAPs), although all these tools attempt to exploit one or more of four potential channels of influence on the economy. These channels are listed in the first column of Table 2. In particular, large-scale asset purchases might be seen as a demonstration of the willingness to keep interest rates low for a long time or "to do whatever it takes" in the memorable phrase of Mario Draghi, head of the European Central Bank. Assets are purchased through the creation of reserves and any money-multiplier effect may support lending and economic activity. If assets are imperfect substitutes, the purchase of some might influence risk premiums (in the manner discussed by Hanson and Stein 2015, and Krishnamurthy and Vissing-Jorgenson 2013).

Moreover, the willingness of the central bank to purchase government securities might encourage politicians to issue more of them, providing fiscal impetus to an economy that was flagging (as discussed in Bernanke 2002). The diplomatic term for the collaboration of monetary and fiscal policy makers is creating "fiscal space." The less delicate terms for QE that is never reversed are "debt monetization" and "financial repression."

In retrospect, only two of these mechanisms got traction, and even those had drawbacks. Rate guidance had a material effect on rates, but sending a consistent message proved hard, especially in a committee setting (the Federal Open Market Committee or FOMC) when different views are expressed (Faust 2016). By way of example, for most of the stay at the zero lower bound from 2012 on (when rate guidance was first made explicit in the Summary of Economic Projections, SEP, of FOMC participants), investors viewed the stated policy direction as more hawkish than would actually be delivered. In addition, expressing the intention to keep rates at zero turned out to be easier than signaling the onset of tightening--it is easier to give than take away...