QE2: What it Did, and What it Didn't Do

QE2 is winding down, and will be completely finished sometime around the end of this month. And as I mentioned yesterday, I think that the chances for a new round of expansionary action by the Fed this year are very small. Despite what some think, however, the reason for that is not because the Fed's last effort -- what is usually, if inaccurately, referred to as "QE2" -- was a failure. In fact, the research indicates that QE2 worked as intended, and had positive, if modest, effects on the US economy.


What's in a name

Let's start by clearing up an oft-misunderstood point. The Fed did not actually engage in what we typically think of as "quantitative easing", the way that the Bank of Japan did during the early 2000s. The goal of the Large Scale Asset Purchase ("LSAP") program (the more accurate term for QE2) was never to "create money". As illustrated in the chart to the right, none of the reserves given by the Fed to banks under LSAP were actually lent out -- they were all simply held by banks as additional excess reserves, exactly as the Fed expected and intended. Since "money" is only created when banks lend out their reserves, this means that the LSAP program didn't cause any money to be created, which is why the supply of money didn't budge.(1)

So "QE2" (I'll revert to the common shorthand for convenience) was not intended to create money, and did not create money. Instead, the program was intended to help the recovery by reducing long-term interest rates. The Fed did that by selling short-term assets and buying long-term bonds -- that's really the essence of LSAP. (Since the Fed doesn't normally buy a lot of long-term bonds, this program was new and got a special name, but it's fundamentally quite a simple idea.) And the hope was that by buying some long-term bonds and thereby lowering long term interest rates, QE2 would help stimulate demand.(2)


What QE2 didn't do

Before we examine what QE2 did do, it's helpful to understand what it didn't do. A lot of the evils of the world have been wrongly attributed to QE2 in recent months. Here are some frequent criticisms:

1. QE2 has caused rampant inflation.
   
2. QE2 has artificially inflated asset prices such as the stock market.
   
3. QE2 has devalued the dollar.
   
4. QE2 simply didn't work.

Let's take these in turn.

1. Inflation: Critics say that all of the money pumped into the economy by QE2 has caused a surge in prices. Yet we know that QE2 has actually not created any additional money. So it's hard to see how QE2 has caused inflation. Furthermore, it's quite straightforward to trace the recent rise in global food and commodity prices to certain specific factors such as turmoil in the Middle East and several bad harvests for grains and vegetables (Russia, Argentina, Australia, Mexico).

2. Asset prices: Critics say that by purchasing long-term assets, the Fed has flooded the financial markets with cash, causing all asset prices to rise. A more sophisticated variant of this criticism is that by driving long-term interest rates down, QE2 has forced investors to seek higher returns in other types of assets, driving their prices up. But there are several problems with this theory. First, QE2 created no money (not to beat a dead horse or anything), so any notion that there was a broad injection of liquidity that caused commodity price inflation is wrong. Second, many of these same critics say that QE2 hasn't worked (point 4), and they point to the fact that long-term interest rates are no lower now than they were in the fall of 2010. But if interest rates weren't driven down, how could QE2 have caused asset prices to rise? (As shown below, the evidence is that interest rates actually did go down in response to QE2, but that is not generally recognized by QE2's critics.) Finally, the Fed has only purchased $600bn in long-term US government bonds. That's less than 5% of the US government bonds in existence, and probably well under 1% of the value of all types of traded assets (stocks, bonds, commodities, real estate). Even though the evidence (see below) is that QE2 was able to push down interest rates a bit, it doesn't seem likely that such a relatively tiny influx of demand into bond markets could cause a dramatic change in overall asset prices outside of the bond market.

3. The value of the dollar: Critics say that QE2 weakened the dollar by causing inflation and/or by causing a loss of confidence in the dollar. But the dollar is just as strong as it was in the middle of 2008, and only about 3% weaker than it was in Oct 2010 when QE2 started. So no systematic weakening of the dollar is apparent, which makes it unclear why the critics of QE2 raise this complaint, since they do not typically make the correct comparison against what the dollar exchange rate would have been in the absence of QE2. But more importantly, even if QE2 has caused the dollar to be weaker than it would otherwise have been, we should be thankful for that -- a weaker dollar is exactly what the US economy needs right now to help stimulate demand. It would be a good thing right now, not a bad thing.

4. QE2 simply didn't work: Some critics say that it was a waste of money -- it cost $600 billion but the economy is still lousy. First of all, QE2 didn't actually cost anything -- the Fed still has that $600bn, it's just now in the form of US government bonds. But second, yes, I agree, the economy IS still lousy... but that doesn't tell us anything about whether QE2 worked. Without QE2, it's entirely possible that the economy could have been in even worse shape. So the correct comparison to understand the effects of QE2 is to try to understand what would have happened in the absence of QE2. This, the critics of QE2 do not do. (Plus, if you think that QE2 didn't have any effects, then you should really drop criticisms 1-3, which rely on the premise that QE2 was effective.)

In short, it's pretty unlikely that QE2 caused any of the terrible things its critics attribute to it. So did QE2 actually accomplish anything?


What QE2 did do

The correct way to figure out what QE2 did is to first try to estimate what would have happened without QE2, and then compare that with what actually did happen. And while that can be a tricky exercise, that's what economic research is all about, so let's see what the research has found.

• Krishnamurthy and Vissing-Jorgensen (Kellogg Northwestern): "The Effects of QE2 on Long-term Interest Rates" (pdf). They find that QE2 was likely to reduce interest rate on long-term AAA bonds by about 30 basis points, and rates on average long-term corporate bonds and mortgage loans by about 10 basis points, compared to what they would otherwise have been. They do not estimate what impact, if any, those interest changes had on real economic activity.
  
  
• Gagnon, Raskin, Remache, and Sack (New York Fed): "Large-Scale Asset Purchases by the Federal Reserve: Did They Work?" (pdf). Using two completely different methodologies they find that QE1 (which was about 3 times the size of QE2) reduced long-term interest rates by between 50 and 80 basis points. Based on that estimate, the implied effect of QE2 would be to reduce long-term interest rates by between roughly 15 and 30 basis points.
  
  
• Hamilton and Wu (UCSD / Chicago): "The Effectiveness of Alternative Monetary Policy Tools in a Zero Lower Bound Environment" (pdf). They estimate that a $400bn LSAP by the Fed would reduce long-term interest rates by about 13 basis points; a simple linear extrapolation of this result implies that QE2 would have reduced interest rates by about 20 basis points.
  
  
• D'Amico and King (FRB): "Flow and Stock Effects of Large-Scale Treasury Purchases" (pdf). They look at the stock and flow effects of the Fed's purchases of Treasury securities in 2009 to estimate their impact on interest rates. Their estimates imply that QE2 would have had a total impact on long-term interest rates of between 15 and 30 basis points.
  
  
• Chung, Laforte, Reifschneider, and Williams (FRB / SF Fed): "Estimating the Macroeconomic Effects of the Fed's Asset Purchases". Using the Fed's own large-scale model of the US economy (FRB/US), they estimate that QE2 pushed down long-term interest rates by about 25 basis points. They also use the FRB/US model to estimate the impact that this had on unemployment and output, and find that QE2 raised GDP by about 0.75%, and employment by about 700,000.

While this is not an exhaustive compilation of the research, it's striking that the estimates are all remarkably similar, despite very different methodologies: QE2 reduced long-term interest rates by 20 to 30 basis points compared to what they would otherwise have been, and this in turn almost certainly had some positive impact on employment and output in the US.


So... Was QE2 a good idea?

Unfortunately, QE2 is not the only thing that has affected the US economy over the past 6 months, and many of the other forces at work have negatively impacted the economy. So the small but positive effects of QE2 were swamped by the other things going on such as supply-side shocks, contractionary fiscal policies, the US Treasury's program of large scale asset sales (i.e. bond sales to finance the deficit) that directly undid much of the effect of QE2 on interest rates (see Jim Hamilton for more on that), and a discouraging lack of private demand.

So yes, I would say that QE2 was probably a good idea, but it was simply not a sufficiently powerful tool to be able to single-handedly ensure a good recovery in the face of all of these headwinds. It helped a bit at the margin, and may have played a crucial role in helping the US step away from a dangerous flirtation with deflation in 2010... but the problems facing the US economy are just too numerous and substantial. My conclusion is that the only policy tool powerful enough to really help the US economy right now would be a new round of expansionary fiscal policy. Unfortunately, that's about as likely as me winning an Oscar this year.

That doesn't mean that QE3 would be a bad idea. In fact, given the new and alarming weakness in the US recovery, I think that further expansionary monetary policy would be very much worth trying. I wouldn't expect a dramatic impact, but right now the US could use every little bit of help it could get. Unfortunately, QE3 is only slightly less unlikely than a new round of expansionary fiscal policy (maybe on par with the likelihood of me winning a Grammy). And that's too bad, because every month that goes by with current levels of unemployment causes permanent long-term damage to countless individual lives, as well as to the US economy as a whole.


UPDATE: text edited slightly for clarity.


(1) For more on the distinction between what the Fed actually did and "quantitative easing", take a look at this speech by Ben Bernanke from January 2009, especially the part about halfway through.
(2) There are a number of channels through which lower long-term interest rates could possibly stimulate demand, including wealth or portfolio effects, exchange rate effects, and direct interest rate effects.