Does QE Really Work? The Evidence To Date

The market's hopes and dreams for the next LSAP remain high. As gold inches higher, tail-risks priced out (expectations for extreme FX moves are considerably lower than sentiment would suggest), and US equity vol expectations (and put skews) are crushed; the equity market clearly remains 'at a premium' in its notional indices given what is sheer lunacy in earnings expectations going forward. The question every investor should be asking is not when QE or even if QE, but so-what-QE? As Credit Suisse notes, given the deterioration in US economic activity (and the extension of Operation Twist) the FOMC will probably wait until its September meeting (and remember the trigger for further pure QE is a long way off for now). The most critical question remains, will additional QE work? After all, few would argue that US interest rates are too high or that banks in the US need still more excess reserves. Two things stand out in their analysis of how QE is supposed to work (transmission mechanisms) and its results to date: QE1 was more effective than QE2, and it's easier to find QE's effect on Treasury yields than on real economic performance. Perhaps more concerning is that the potential negative effects of such unconventional monetary policy has received little attention (aside from at fringe blogs here and here).


Credit Suisse: Does QE Really Work? The Evidence to Date

Less than four weeks ago, the Federal Open Market Committee voted to continue its maturity extension program (Operation Twist) for another six months. But the sluggishness of US economic activity has the key decision-makers on the FOMC considering a more aggressive easing move.

In our view, another round of large-scale asset purchases (“QE3”) is looking increasingly likely before Election Day in November. While an announcement is possible as soon as the July 31-August 1 meeting, the FOMC probably will choose to wait until its September 12-13 meeting.

Having forecast that the FOMC will buy more assets, we address the next natural question: Will additional QE work? After all, few would argue that US interest rates are too high or that banks in the US need still more excess reserves.

To find the answer, we consulted empirical studies within and without the Federal Reserve system on the effectiveness of the previous two rounds of QE and the balance sheet neutral Operation Twist program. There are many complications that arise in the evaluation of the programs’ results, and opinions differ among economists even within the Fed.

However, a review of the empirical literature yields some common themes:

  1. QE1 was more effective than QE2.
  2. It is easier to find and quantify QE’s effect on Treasury yields than to identify and measure QE’s effect on real economic performance.
  3. QE also lowered nominal interest rates on agencies, MBS, and corporate bonds, with magnitudes differing across bond types and maturities.

Results of the studies we reviewed were less uniform on QE effects on equities, the dollar, and commodity prices.

How is QE supposed to work?

QE1 was the expansion of the Fed’s balance sheet achieved mostly through large-scale asset purchases of agency debt and mortgage-backed securities. To a considerable extent, that program was aimed at rehabilitating a particular financial market that was functioning poorly at the time. From that perspective, the report card reads favorably. Lingering issues in the mortgage market (e.g., the foreclosure confusion) are not related to the original malfunction the Fed sought to cure.

In QE1 (Nov 2008-Mar 2010), the expansion of the Fed balance sheet, and especially the provision of a large amount of bank reserves, was incidental.

QE2 (Nov 2010-Jun 2011) was not about a poorly functioning piece of the financial system. The Treasury market had been functioning just fine. QE2 was about the Fed doing large-scale asset purchases to boost aggregate demand and eventually create more jobs.

How do we get from expanded central bank balance sheets to real economic performance? There are four broad policy transmission paths economists have theorized about over the years as follows:

  • A money supply effect, which one might identify with old-fashioned (although perhaps again more timely) monetarism.
  • An interest rate effect, which one might identify with a Keynesian marginal efficiency of investment analysis.
  • A portfolio or credit channel or collateral value effect, which one might identify with the scholarly contributions of “Professor Bernanke” himself, among others.
  • A foreign exchange effect that, in current circumstances, may be more a matter of “forcibly enrich thy neighbor” global capital flows than “beggar thy neighbor” competitive devaluation.

QE Transmission Mechanisms

The Results to Date

Interest Rates – Generally speaking, the empirical evidence suggests that both rounds of QE and Operation Twist were effective at reducing interest rates on long-term Treasury securities. The estimates of reduction in the 10-year Treasury yield ranges between 20 and 110 basis points, with most estimates in the lower two-thirds of this range. Both theory and empirical evidence suggest that the reductions in interest rates primarily reflect lower risk premiums and lower expectations of future short-term interest rates.

The evidence also suggests that LSAP lowered nominal interest rates on agencies, MBS, and corporate bonds, with magnitudes differing across bond types and maturities. Since QE1 included significant purchases of MBS (totaling $1.25 trillion), it was seen, naturally, as being more effective in lowering the rates on those mortgage-related assets.

In addition, QE1 provided liquidity support to a largely dysfunctional market, and therefore its impact was probably larger than that of QE2 and Operation Twist, which were conducted in more normal environments.

Macroeconomic Effects – Research dedicated to the effects of LSAP on GDP and employment is limited, but they do generally find these programs to be effective at promoting GDP growth, though to different extents. One study (Fuhrer and Olivei, 2011) found that $600bn Treasury purchases would increase real GDP by about 40-120bps while another study (Chen et al., 2011) suggested that the effects on GDP growth are not very likely to exceed 50bps. By using Okun’s law, the Fuhrer and Olivei study also theorized that the unemployment rate would drop by 30-45bps.

It should be noted that the estimated (as opposed to observed) effects on GDP and the unemployment rate are more gradual, usually taking place over the course of about two years.

The unconventional measures were generally seen as effective in preventing deflation at the zero lower bound through the signaling effect. However, economists do have contrasting opinions on the magnitude of QE’s price impact.

One study (Krishnamurthy and Vissing-Jorgensen, 2011) found that QE1 increased 10-year expected inflation by 96-146bps and that QE2 raised it by 5-16bps. On the flip side, one study (Chen et al., 2011) concluded that the inflationary consequences of QE1 and QE2 were less than 50bps. The Federal Reserve Bank of New York noted that inflation actually trended lower over the period when QE1 was in progress, though it probably fell less than it would have done without the asset purchases.

Pros And Cons

Empirical studies of the effects of the Fed's balance sheet operations suggest that QE designed to address general economic malaise is less potent than a program targeted at a specific market dysfunction. Even QE proponents on the FOMC stress that asset purchases are not the silver bullet that will cure all that ails the US economy.

Meanwhile, the drawbacks of additional easing may be rising. In his press conference on June 20, Chairman Bernanke explained that unconventional policy has costs and should not be used without serious consideration. Among the costs he cited were (1) potentially making the Fed’s exit strategy more difficult, (2) potentially creating negative implications for market functioning, and (3) financial stability issues (about which he was very vague). The minutes of the June 19-20 FOMC meeting suggested that other Fed officials are also considering the potential limits and drawbacks of large-scale asset purchases:

“A few members observed that it would be helpful to have a better understanding of how large the Federal Reserve’s asset purchases would have to be to cause a meaningful deterioration in securities market functioning, and of the potential costs of such deterioration for the economy as a whole.”

Paraphrasing comments from San Francisco Fed President John Williams, the negative effects of unconventional monetary policy have received scant attention in the research literature and are not well understood.

Future studies should weigh these costs against the value of asset purchases for macroeconomic stabilization. In the meantime, in the face of decidedly inadequate job growth and low inflation, Fed policymakers are likely to conclude that QE3 will do more good than harm.