Why and when to use quantitative easing
It's generally just pricey forward guidance, but QE has a time and place
Note: any and all opinions are entirely my own and do not necessarily reflect those of the Cleveland Fed, the FOMC, or any other person or entity within the Federal Reserve System. I am speaking exclusively for myself in this post (as well as in all other posts, comments, and other related materials). No content whatsoever should be seen to represent the views of the Federal Reserve System.
I received an excellent question about quantitative easing via email that I decided to turn into a full-length blog post, since I think it’s a really good thing to think about even as the Fed begins to round off its quantitative easing program before its March meeting. Last time, we looked at what reserves are really for, and how they’re more or less stuck at the Fed—they can’t leave the banking system. But, as we all probably know, quantitative easing creates reserves. In fact, it creates a lot of reserves; trillions of dollars of them! So what good does quantitative easing do, then? When should we use it, and why?
We’ll dive into that in this post. Before we do, I want to address a great point a few commenters brought up on my last post. The supplemental leverage ratio (SLR) absolutely can, and perhaps even did, stunt the effect of quantitative easing. The SLR requires banks to hold a certain ratio of capital to assets. When quantitative easing first began and reserves poured into the banking system, the denominator (which includes Treasury securities and reserves) grew massively, and banks were put into a bind that forced them to either raise more capital or cut assets. Since cutting assets was the faster and “easier” option, banks sold off Treasuries, accepted fewer deposits, and very well may have slowed loan creation.
Luckily, for the first half of the pandemic the Fed instituted a temporary rule that excluded reserves and Treasuries from being counted toward the SLR. You may remember a lot of noise around the era when this rule was about to expire last March—particular from JP Morgan. The removal of this exemption shuffled the “unwanted” excess liquidity into money markets and the Fed’s reverse repo facility, as banks and financial institutions really wanted to find a way to “get rid” of that liquidity (or at least find some place where it can earn literally any return). I’m not so sure that this is “ideal”, and I’d certainly prefer excluding Treasuries and reserves from regulatory ratios, since there really isn’t anything much safer than these two. This definitely can stall the effect of QE out there in the real world, but I’m not sure this was what Krugman’s criticism was about—his point was much more iffy, theoretically.
In any case, thanks to those commenters for bringing up that important point. Now, onward to dig into quantitative easing. I’m saving the discussion on the details of how QE affects the real economy for another time—for now, I want to discuss when it should be used, and the rationales for why we use it.
How to ease, quantitatively
I’ll begin by briefly diving into exactly what quantitative easing is and how it is done. It’s really one of my least favorite phrases in all of economics—it’s not really intuitive, it has changed meanings over time, and it still has a million meanings. I won’t get into the history of it here though, to avoid that confusion.
FedGuy has the authoritative manual on quantitative easing, step by step. The most common way it happens goes more or less something like this: somebody, some kind of a financial institution like a hedge fund, pension fund, bank, etc, has some sort of security. A primary dealer (a special financial institution that has special privileges at Treasury auctions, among other things) borrows in the repo market to raise cash to buy that asset from the original financial institution. The Fed then buys this asset from the primary dealer, and the primary dealer uses the proceeds from that sale to pay back that repo loan.
If that sounds messy, that’s because it is. At the end of the day, you wind up with that original financial institution’s bank—the bank of that hedge fund, pension fund, bank, etc—having newly issued reserves, which flow there in order to match that initial purchase made by the primary dealer. As FedGuy puts it, QE mostly happens “behind the scenes”; the Fed purchasing an asset is just the start. I’d recommend reading his post as he explains it much better than I can in this amount of space, but don’t worry if all that accounting and reserve arithmetic doesn’t add up to you. It is sufficient to understand QE as the Fed buying assets—currently, a mix of two-thirds Treasury securities, one-third mortgage-backed securities—and reserves being created somewhere down the line.
Besides reserve creation, deposits are often created too. I know I may be beating a dead horse if you’ve read my writing on QE before, but better than “created” I might prefer the word “exchanged”. Let’s say in the first story above, you’re a pension fund and you sell a 10 year Treasury to a primary dealer for $100, the Fed goes on to buy it from that primary dealer, and all that magic happens. As far as you’re concerned as that pension fund, the only change you experience is that you now have $100 in a deposit instead of $100 as a Treasury. You’re not really any richer or wealthier than before—you just had a very liquid asset swapped for a very, very, liquid asset. The only things that changed are:
You now have $100 in deposits (instead of a $100 Treasury)
Your bank now has $100 in newly-created reserves (to match that new $100 deposit you have)
The Fed now owns that $100 Treasury to “back” those newly-created reserves.
Why go through all that?
That’s quite messy, with all those moving pieces and written-out accounting. So why do we bother with such a complex system? What is the goal?
The main impetus behind using QE is the frustrating fact that the Fed only explicitly sets short term rates. The Fed’s administered rates—like the interest rate on reserves, the reverse and standing repo rates, the discount window rate—are short, super-short term rates, generally overnight. The fed funds rate, which the Fed doesn’t “set” but it “determines” by setting boundaries for it to trade within (generally speaking, the reverse RP rate binds from below and IOR binds from above), is also a super-short term rate, and typically overnight.
This can be frustrating when you’re a central bank because, while short-term rates do matter (in fact, they matter quite a lot!), long term rates matter too. The connection often starts with the 10 year Treasury bill, and flows through to corporate bonds and mortgage rates. The goal with QE is to bring down these longer term rates, which should (theoretically, at least) induce more borrowing, more investing, and overall more economic activity. As I’ll discuss shortly, the mechanics of how purchases actually do this is disputed; in my opinion, which I believe the literature broadly supports, the actual “purchase” doesn’t do too much. It’s more the forward guidance they provide.
Why use QE?
Mechanism aside—the goal of QE is, in one way or another, to bring down longer-term yields. But is it worth all the hassle? Why not just cut interest rates?
We are living in a very low interest rate environment. The Fed thinks the fed funds rate will top out this cycle at between 2 and 3 percent, more likely at the lower side of that distribution. Some markets expect even less, implying that they’ll be lucky if they reach 2%. Now, when you catch recessions early, a single rate cut or two might be very sufficient to stop it from metastasizing into a the next Great Depression. But even in a bad-enough downturn, perhaps like the pandemic-induced one in 2020, cutting two entire percentage points might not provide “enough” support. Even worse, if markets become dysfunctional (such as in 2007-20088 and March 2020), rate cuts might not even stop further snowballing into a financial crisis.
Thus enters the Fed (or any central bank). Since the Fed is the issuer of reserves, it has quite a bit of balance sheet capacity. It can pick up toxic, radioactive, and glowing assets, no matter if you hear a loud humming around them and feel sick getting too close. In the case of a crisis, the Fed is uniquely capable of picking up these assets to prevent total market dysfunction. Such dysfunction would only worsen financial conditions and make whatever underlying downturn was going on into a much bigger problem.
Outside of chaotic times, I think the Fed’s justification for QE in their statements over the last two years is helpful:
“These asset purchases help foster smooth market functioning and accommodative financial conditions, thereby supporting the flow of credit to households and businesses”
QE, the Fed thinks, can go where short-term interest rate policy can’t. By flooding the banking system with reserves and altering the private sector’s allocation of liquid securities (i.e., swapping Treasurys for deposits/reserves), it eases financial conditions and essentially enables more risk taking—often in the form of credit. I don’t mean risk taking as in, the Fed is giving money to hedge funds to go out and speculate on dumb meme stocks. No. I mean, whenever you provide credit or invest, you are accepting some risk that something might happen to your money. Your willingness to take on that risk is heavily determined by market conditions. I’ll have to save further discussion on exactly how the Fed achieves easier or tighter financial conditions, and why those financial conditions matter so much at the macroeconomic level, for another time.
This has been a long section, so let’s briefly summarize. The Fed goes through all the trouble of QE because, in some way or another, it encourages more risk taking and credit creation/flow (which, yes, can be subject to the regulatory constraints mentioned in the intro). Through some mixture of those two things, QE helps the Fed achieve the dual mandate of maximum employment and price stability.
Is QE good? Is it necessary?
These are the real questions that matter. There has been a lot of concern that QE was worsening inequality—I’m not very persuaded by this argument—but many wonder if QE is even really necessary. I’ll go into “why” QE works (and by work, I mean, have any effect) in another post, since that’s a whole universe of stuff that deserves a deeper discussion. Instead, I’ll close this post by “evaluating” QE.
The way I see it, the most important part of QE is the forward guidance it gives off. Sure, the whole mechanical process I detailed above and the resulting adjustment of liquid portfolios may have some effect out there, but a credible central bank doesn’t need to do so much QE in order to get a similar result. If we care about shortening longer-term yields, we can do that just by providing credible forward guidance on future short term rates.
In fact, I would dare say that while the Fed only “administers” those super-short-term overnight rates, it by no means only controls short term rates. If the Fed credibly said, “hey, congrats, we’re going to keep rates at zero for the next decade, come hell or high water,” there isn’t really a good reason for the expected future short rate component of the 10 year Treasury yield to be above zero unless we don’t believe the Fed.
Regardless, even when the Fed isn’t trying to, it is always shifting future short-term rate expectations by seeming more or less hawkish or dovish, with the consequence of always moving the yield curve around. If people expect higher or lower rates in the future, they have that expectation only because the Fed gave it to them! If markets are really, really not believing the Fed, QE is a good tool to sort of put your money where your mouth is. Markets understand that the Fed will not raise rates while still doing QE. (And that’s why all the talk of a rate hike next week is silly!)
Thus a “cheaper” version of QE could simply be to give credible forward guidance—even better, state-based forward guidance. “No rate hikes until x x% or y rises above y% for some time”, or something like that. However, there are occasionally times of severe market dysfunction—March 2020, for example. In that case, QE can do a lot of good by using the Fed’s balance sheet capacity to soak up toxic assets and restore market function. This isn’t a matter of bailouts—if the wrong markets freeze up, the whole system freezes, affecting you and I no matter how far removed we might be from the financial system. Easing stressed financial conditions is good, actually.
There can be a role for the “portfolio balance channel” for QE (compared to the “forward guidance channel”, which I consider much more important), but most the evidence seems to suggest it’s not really all that large. For example, some evidence suggests that a lot of those deposits that non-banks wound up with simply went back into more or less the same security—Treasury debt. To really get things going, we should want to reduce risk premiums and encourage investment into riskier assets. The linked paper found that this did happen to some extent, but in an indirect way that similarly, in my view, could have largely been accomplished via credible forward guidance. The workings of QE once its been done is a topic of much research, and deserves its own post eventually—I promise I’ll get to it eventually!
QE is thus an instrument most applicable when we’re worried about market functionality. Otherwise, we can generally get by with more credible and perhaps quantitative forward guidance. This is not to say that QE is “bad” or should only be used with extreme hesitancy. In cases where interest rates hit zero and we still feel like we need more room, or perhaps in circumstance like that of late 2019 where we feel like we need to provide more liquidity to the system, QE can absolutely be fit to the task and is unlikely to do harm. It’s really a question of efficiency. If we’ve provided credible forward guidance (and really meant it!) and cut rates but still feel like the economy needs more, it’s a good tool to have on standby.