Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Friday, November 24, 2017

Tyler Cowen on Central Bank Cryptocurrencies

I enjoy reading Tyler Cowen and have learned a lot from his columns. So before I criticize his most recent effort, I want to thank him for all his fine contributions! Unfortunately, I think he drops the ball a little bit on his most recent effort. No worries--we all do sooner or later. What follows are some thoughts that came to my mind as I read his most recent article entitled "Cryptocurrencies Don't Belong in Central Banks." My (unedited/uncensored thoughts are recorded in blue...Tyler, take note: "uncensored"= risk-taking central banker!) 
Should central banks embrace cryptocurrencies, or even pioneer their own? In a nutshell, no. Crypto assets are an unusual innovation, still in flux and often poorly understood. Trying to centralize them in a bureaucracy is exactly the wrong way to go. 
"Crypo assets are an unusual innovation" -- this is a claim that makes little sense without first defining what is meant by the term. Most people have in mind some notion of a distributed (shared) append-only ledger of information updated and maintained by members of a community through some consensus algorithm that makes use of cryptography in securing information. I'm not sure how "unusual" this idea is where fintech is concerned (advances in information management systems have been happening for a long time). And, as I explain here, the concept of distributed ledgers updated via communal consensus is an ancient idea. 
"Trying to centralize them in a bureaucracy" makes no sense at all if the very concept depends on a decentralized record-keeping arrangement. And in any case, a central bank that experiments with such a protocol is not "centralizing" all crypto assets--just its own crypto asset. What is inherently wrong with this? The fact that a central bank is a "bureaucracy?" 
Yet China’s central bank claims it is working toward a blockchain-based digital currency. Singapore has already experimented in this direction. The phrase “Fedcoin” is sometimes bandied about, though I’ve seen no concrete sign of the U.S. Federal Reserve jumping on this bandwagon. In its recent quarterly review, the Bank of International Settlements asked central banks to consider whether cryptocurrencies might make sense for them.
The "Yet" beginning the paragraph does not belong there. It presumes we've accepted the premise laid out in the previous paragraph which, as I have suggested, makes little sense. 
Central banks, however, are intrinsically conservative bureaucracies. They shun bad publicity, and they don’t like to be “out there” ahead of the curve. They don’t want their names connected with potential mishaps -- because they value their credibility and their political capital so highly. That’s appropriate, because central bank independence is typically fragile.
Almost everyone shuns bad publicity. No one wants their names connected with potential mishaps. It's not that central banks don't like to be "out there"--it's that they are usually prevented from being "out there" by government legislation. Of course, central bankers should be "conservative" in the sense of respecting the legislation that is designed to govern them.  
Given that background, (which is wrong) should we foist a new and potentially risky responsibility on them? Central banks will feel some anxiety at having to manage a crypto project. (central bankers will always feel anxiety). To conserve their political capital, they will take fewer risks elsewhere, such as unorthodox monetary policy or larger balance sheets. (what justifies this bald speculation?) Yet the response to the 2008 financial crisis shows a certain amount of central bank risk-taking is needed. (those conservative central bankers again, I guess). I’m worried about central banks taking on unnecessary risky projects, thereby rendering them too cautious in other areas. (there's no need to worry about this in my opinion.)
 An additional reason for skepticism stems from the nature of crypto assets. The word “cryptocurrency” is far more common than “crypto asset,” but it’s a misleading term. Bitcoin, for instance, is used only rarely in retail transactions, and for all its success it isn’t becoming more important as a medium of exchange. Bitcoin thus isn’t much of a currency in the literal sense of that term. There is a version of bitcoin, Bitcoin Cash, that changed the initial rules to be better suited as an exchange medium, but it isn’t nearly as popular. (I don't disagree, but I also don't understand the point of this paragraph.)
If you think of these assets as “cryptocurrencies,” central bank involvement will seem natural, because of course central banks do manage currencies. Instead, this new class of assets is better conceptualized as ledger systems, designed to create agreement about some states of the world without the final judgment of a centralized authority, which use a crypto asset to pay participants for maintaining the flow and accuracy of information. (Good, I agree with all this.) Arguably these innovations come closer to being substitutes for corporations and legal systems than for currencies. (Except that in monetary theory, we understand money as a ledger -- a record-keeping system. See, e.g., Kocherlakota, 1998). 
Put in those terms, an active (rather than merely supervisory) role for central banks in crypto assets is suddenly far from obvious. Consider other financial innovations: Does anyone suggest that central banks should run their own versions of ETFs or high-frequency trading? Is there a need for central banks to start managing the development of accounting and governance systems?
Um, yes, yes, and yes. Central banks are already like bond ETFs (assets are bonds, liabilities are reserves and currency). High-frequency trading--like Fedwire, you  mean? And I'm not sure why better accounting and governance systems should not be employed by central banks if it makes sense for other institutions?
Finally, bitcoin and other crypto assets are still in the midst of rapid evolution, with basic questions still unanswered. Should bitcoin “fork” to allow for greater speed in processing transactions? Is the future going to favor bitcoin, the Ethereum platform, or something else altogether? How many initial coin offerings make economic sense, as opposed to being bubbles? Should initial coin offerings be used to fund startups? How many crypto assets should survive in the long run? Can blockchains be used to record and settle the transfer of property titles? Are there any circumstances when it should be possible to revise transactions on a blockchain?
Yes, there are always questions concerning the outcome of fintech. This has been happening for hundreds of years. What is the point of stating this obvious fact and why should the prospect of evolving information technology discourage central banks from experimenting with it? What if such an attitude was adopted when email first appeared on the scene? 
In general, I think the central banks in the world’s developed economies have done a pretty good job. But consider a simple question: Would any central bank have had the inspiration or taken the risk of initiating the bitcoin protocol in the first place?
Let's assume that the answer to this question is no. The only thing that follows from this is that we should not rely on central banks as our only source of fintech. But we already do not have such a reliance. So what is the point being made here? Let's experiment away, I say! 
Well, maybe not. My own sense of the discussion concerning central banks and cryptocurrencies is that people are confounding two conceptually distinct issues. 
The first issue is whether central banks should open the digital component of their balance to the general public (say, the way the U.S. Treasury does at https://www.treasurydirect.gov/). And why not? After all, central banks allow anyone in the world to hold the paper component of their liabilities. This idea is actually very old--it is related to the question of whether a central bank (or post offices) should permit people to open book-entry utility accounts.  This remains a good question and there are many issues to sort out, including what impact such an innovation would have on banks. But it has nothing to do with "cryptocurrencies" except to the extent that cryptography is used to secure communications--something that is already widely employed.
The second issue is whether central banks should issue "digital cash." The digital reserve money associated with the scheme in the paragraph above is not cash-like in the sense that it is not a bearer instrument. People have to identify themselves when they open accounts at the U.S. Treasury. Presumably, they'd have to do the same thing if they opened up a digital money account at at post office or central bank. Digital cash is something more akin to paper money. As far as I can tell, the technology to create digital cash has been around for a long time. That is, there is nothing technological that prevents a central bank (or anyone) from issuing numbered accounts (think of the good ol' anonymous Swiss bank account, for example). Even PayPal could issue digital cash--if it was legally permitted to, which it isn't. So, the question here is whether a central bank should issue digital cash. The wallet-to-wallet debit/credit activity could be done in-house via a central book-keeper, or the activity could be delegated to some third parties (e.g., Bitcoin miners). I don't really see central banks getting into this business, but they will have to think about how private-sector digital cash may impinge on their ability to conduct monetary policy. (Note: the innovation with Bitcoin is not digital cash per se, rather the innovation has to do with P2P digital cash--debit/credit operations that do not rely on a central book-keeper.)

Sunday, August 6, 2017

A monetary-fiscal theory of inflation

On December 17, 2015, the FOMC has raised its policy rate (IOER) from 25bp to 50bp. It has since raised the IOER rate three more times to 1.25%. Many on the committee seem convinced that further rate hikes are needed (in addition to actions designed to shrink the Fed's balance sheet, which is already shrinking relative to the size of the economy). What is the source of this enthusiasm for monetary policy tightening, given that the unemployment rate is close to target, and given a PCE inflation rate that has been undershooting the Fed's 2% inflation target for several years now?

The short answer is the Phillips curve. Or, to be more precise, a belief in the Phillips curve theory of inflation. The basic idea is that at very low rates of unemployment, competition for workers will lead to higher wages, with the associated costs passed on to consumers in the form of higher prices. Even if this wage pressure has been largely absent to date, it will (form sign of the cross here) eventually happen, and it's better for the Fed to get ahead of the curve, rather than risk having to raise its policy rate abruptly (and disruptively) in the future.

But what if the Phillips curve theory of inflation is not the best way to guide our thinking on the matter? What other theory might we turn to for guidance? Binyamin Appelbaum of the New York Times discusses a number of alternatives here (which I review in my previous post). In addition to the Phillips curve theory, he mentions explanations that I labeled: [1] Monetarist, [2] Expectations, [3] Internationalist/Technology. I mentioned in my previous post that I'd return to examining the Monetarist view, which I think is too often given a short shrift. I explain below how the Monetarist view is consistent with [2] and [3]. There is also the question of what the Monetarist view implies for policy. While the Phillips curve view has turned doves into hawks, I argue below that the Monetarist view should turn hawks into doves (given the present state of the economy).
 
Many people feel Monetarism has been discredited because economists who employed the theory to predict the inflationary consequences of QE were proved embarrassingly wrong. But this is along the lines of viewing scissors as a lousy tool because many barbers have used scissors to give awful haircuts.

To be fair to their critics, Monetarists sometimes overstate the role of money supply in determining the price-level and inflation. But let's also give credit where credit is due. We know how to create inflation. Just look at Venezuela today. No one can take seriously the notion that inflation is very high in Venezuela because the unemployment rate is far below its natural rate. Moreover, we know how to stop inflation. Tom Sargent's The Ends of Four Big Inflations showed us how it was done in history. Our understanding of these episodes revolve around Monetarist explanations that also take seriously fiscal considerations. Why can't the same theory be used to understand the present low-inflation environment and help guide policy? I think it can.
 
By the way, I've worked this all out in an open-economy version of the model I describe here. But nothing I say below hinges on this specific formalization; the basic idea is much more general. The two essential elements are: [1] safe government debt is a close substitute for central bank money; and [2] the demand for government money/debt can wax and wane over time (perhaps in St. Louis Fed regime-switching style).

The first property is important for understanding the economic consequences of open-market operations like QE. In the old days, when U.S. treasuries were yielding (say) 10% and Fed reserve liabilities were yielding 0%, an open-market swap of money for bonds could be expected to have a big effect. The same size open-market swap in a world of 1% reserves and 2% treasuries is not likely to have as great an impact. In the extreme case where reserves and treasuries have identical yields, open-market operations are not likely to have any effect at all--apart from inducing banks to accumulate excess reserves (in place of the treasuries they would have otherwise held). I think this is the main reason for why large-scale asset-purchase (LSAP) programs have had much smaller effects than what many had expected.

The fact that bonds become close substitutes for money when their yields are similar explains how the supply and demand for bonds can influence the inflation rate. Normally, we think of an increase in the demand for bonds as lowering bond yields. This is correct. But what happens when those yields approach the corresponding yield on interest-bearing money? (In the old days, when interest on reserves was zero, this limit was called the zero-lower-bound). An increase in the demand for bonds in this case must manifest itself in other ways. One way is for the price-level to fall. That is, a market-mechanism for expanding the real supply of nominal bonds is for the price-level to fall. One way this manifests itself is as China selling its goods for less USDs to acquire the USTs it so desperately wants.

The second property is important for understanding how inflation can fall even in the face of a growing supply of money/bonds. Admittedly, a bit of religion is required here, but I'm not sure what else to believe in. Suppose we can observe the supply of oil. We see a sudden increase in the supply of oil. At the same time, we see the price of oil rise. While the demand for oil is not directly observable, I think it's fair to say that most people would conclude that the (unobserved) demand for oil must have increased by more than the (observed) increase in supply. I want to apply the same thought-organizing principle to the price of money and bonds.

The story is familiar to those who point to declining money (and debt) velocity. In my formal model, I have a parameter that indexes the growth rate in the demand for real money/bond balances (where money and bonds take the form of USDs and USTs, respectively). In the open-economy version of my model, I have a "money demand growth regime" originating from the foreign sector. In the model, this regime translates into persistent U.S. trade deficits, representing the foreign sector's desire to acquire USD/UST at an elevated pace. There is in fact considerable evidence suggesting a large and growing foreign appetite for U.S. money/bonds over the past decade. Japan and China have each accumulated about one trillion dollars in USTs, for example. Moreover, it is known that USTs play an important role as exchange media (collateral) in credit derivatives markets and the shadow banking sector. Lately, the demand for such securities has been enhanced by a variety of regulatory reforms targeting the banking sector.

Bringing these elements together, the story that unfolds goes something like this. For years, several forces have conspired to elevate the (growth in the) demand for USD/USTs, driving yields ever lower. The financial crisis and associated "flight to quality" phenomenon served to exacerbate this secular force (with subsequent regulatory reforms adding to it further). Given an historically normal pace of money/debt expansion, these forces would have been hugely deflationary. The effect of the large increase in USTs following the crisis was to counteract this deflationary effect. But the U.S. debt-to-GDP ratio has essentially flat-lined since 2013. In the meantime, demand for the product continues to grow. With bond yields very close to the Fed's IOER rate, the result is persistently low inflation. And it's no surprise that now, after years of low inflation, that inflation expectations remain subdued.

No doubt some of you will find holes in this story, some inconsistencies perhaps, with past episodes or other countries. But I'm not claiming that this is the story; I'm simply suggesting that it may be an important part of it. And to the extent that it is, what does it imply about the current configuration of monetary and fiscal policy?

In my model, raising the policy rate in the face of stable or declining inflation has the effect of increasing the attractiveness of government money/bonds. The model highlights a portfolio substitution effect where savers redirect resources away from private capital spending (including expenditure on recruiting activities) toward money and bonds. The effect is contractionary. Is this really what we want/need right now? Moreover, in my model, the effect a higher policy rate on inflation depends critically on how the fiscal authority responds. (As Eric Leeper and others keep on reminding us, every monetary policy action must have a fiscal consequence.) A higher policy rate will increase the carrying cost of the debt, and Fed remittances to the U.S. treasury will decline. How will this added fiscal burden be financed? If the government makes no adjustment to its tax/spend policies, then the treasury will be forced to increase debt-issuance at a more rapid pace--an effect that is likely to increase the inflation rate (a result consistent with the so-called NeoFisherian view). Alternatively, if the government goes into austerity mode, cutting expenditures and/or raising taxes, the effect is likely to be disinflationary. This is all based on standard Monetarist thinking--we do not need the Phillips curve (which, by the way, exists in my model via a Tobin effect).

To the extent that the forces I've described above are present in reality, the analysis here calls into question the need for monetary policy tightening too rapidly at this time. Low unemployment does not necessarily portend higher inflation. And keep in mind that other measures of labor market activity, like the prime-age employment-to-population ratio, are still below their historical norms. Of course, this does not mean that monetary policy makers can afford to ignore the threat of inflation. While the worldwide demand for U.S. nominal debt instruments has been robust for a long time now, this "high U.S. money demand" regime is not likely to last forever. When the growth in money demand abates, the consequence is likely to manifest itself as higher inflation expectation (and bond yields)--much like what we observed following the November 2016 presidential election in the United States, except on a much larger scale. A good policy framework should make provisions for these and other contingencies, including sudden changes in the structure of fiscal policy.

Let me sum up and conclude. An elevated demand for U.S. dollars and treasuries has put downward pressure on bond yields and the inflation rate. Both the Fed and U.S. Treasury have partially accommodated this elevated demand. The result is a PCE inflation rate averaging about 1.5% since 2009, only 50bp below the Fed's official 2% target. The economic losses (or gains) associated with this "missing" 50bp of inflation going forward are difficult to quantify, but it's difficult for me to imagine that they are very large (and especially at this point in the recovery dynamic, where inflation expectations appear roughly consistent with the actual inflation rate).

But suppose that I am wrong and that it would be desirable to raise the price-level path back to its pre-2008 trend (something that would require a few years of inflation running above 2%). Is this even economically feasible? Does economic theory and experience provide a recipe? The answer is yes: have the central bank monetize the deficit until the price-level hits its target. (If the price-level never rises, then the government can enjoy a perpetual free lunch, cutting taxes and paying for goods and services with newly-issued non-inflationary money.)

But don't hold your breath for this to happen anytime soon. The constraints in place are not economic, they are political. Many public officials and the people they represent are growing uncomfortable with historically high debt-to-GDP ratios and large central bank balance sheets. They see the large supply of government debt, but they cannot see the large demand for the product driving yields down. Instead, they interpret low interest rates as enabling a large supply. And so, political pressure is presently running in the direction of austerity and smaller central bank balance sheets. Of course, if this is what the people want, this is what the people should get. But then, let's not spend so much time fretting over a 50bp miss on inflation, or bemoan the apparent lack of a coherent theory of inflation.

*******

PS. This post was motivated in part by Noah Smith, who tweeted:


I discuss the case of Japan in greater detail here: The failure to inflate Japan

Thursday, August 3, 2017

Where's the inflation?

The PCE inflation rate in the United States has been consistently below the Fed's official 2% target for many years now. Equally persistent are the forecast errors of those who have expected inflation to rise to its target level (and possibly beyond).

What accounts for the missing inflation? In a recent NYT article, Binyamin Appelbaum mentions four theories of inflation: (1) Monetarist, (2) Phillips Curve, (3) Expectations, and (4) Internationalist. Let me briefly describe and comment on these four views.

Monetarist. The price-level is determined by the supply of money relative to the demand for money. Inflation (the rate of change in the price-level) is therefore determined by the rate of growth of the money supply net of the rate of growth in the demand for money (often proxied by the growth rate in real GDP).

This theory has been discredited by conservative economists using it to forecast impending inflation following the large increase in the supply of money, as measured (say) by the Fed's liabilities. (Whether the theory deserves to be discredited is another matter, to which I will return in a subsequent post.)

Phillips Curve. This theory, held by Janet Yellen and other FOMC members, is based on empirical evidence like this:
That is, it appears that the inflation rate is negatively correlated with the unemployment rate, which is
used widely as a proxy for aggregate demand. The interpretation of this data goes something like this: As the aggregate demand for goods and services picks up, firms are motivated to hire more workers. As the unemployment rate declines, workers are able to demand higher wages. These costs are then passed on to consumers through higher product prices.

While the story sounds plausible, it is not without problems. It seems sensible to suppose that the bargaining power of workers is improved when the unemployment rate is low. And growing worker productivity is an important source of economic growth. Both of these forces suggest that the real (inflation-adjusted) wage should rise when unemployment is low (or falling). But why should rising real wages result in higher wage and price inflation? The answer is not immediately clear.

Another problem associated with this view seems is the propensity of its adherents to take the statistical evidence of the Phillips curve as prima facie evidence of their theory of the Phillips curve. In fact, economists have known for a long time that there are many other mechanisms that might generate a negative relationship between the unemployment rate and inflation. The Tobin effect, for example, asserts that the direction of causality runs in the opposite direction: higher inflation induces a portfolio substitution out of government securities into private investment (including recruiting investment), which leads to lower unemployment.

Finally, as one can see from the data, the Phillips curve slopes down. Except for when it doesn't.  (This is not entirely fair as it is what one would expect from an inflation-targeting central bank adjusting its policy rate judiciously in response to various shocks.)

Expectations. There are several variants of this view. One is that the rate of inflation depends on the expected rate of inflation and that inflation expectations are largely indeterminate in the sense that they can become a self-fulfilling prophecy. After almost a decade of low inflation, what else are individuals supposed to believe? Eventually, low inflation expectations get baked into lower wage settlements and lower pricing decisions, which results in low inflation.

This story sounds plausible. But it suggests an inertial aspect to expectation formation that may have less to do with recent experience and more to do with how individuals expect policy to evolve in the near future. This latter possibility has been demonstrated convincingly by Tom Sargent in The Ends of Four Big Inflations.
 
Internationalist. As explained by Binyamin, this view holds that low inflation across the developed world is due to the rise of the developing world. The threat of outsourcing keeps a lid on domestic wage pressures, while a flood of cheap goods from foreign countries helps to keep domestic product prices down, both directly (because we pay less for imports) and indirectly (because the threat of competition from imports induces domestic producers to keep prices low).

This view was also expressed as a reason for low inflation by Janet Yellen. Quoting the article, "She and other officials also have noted that the weakness of the global economy allowed the United States to import foreign goods at low prices." There is some evidence suggesting this is true. The following diagram plots the PCE inflation rate (blue) against the inflation rate associated with the import price deflator:
What are we to make of all this? According to Adam Posen, "policy makers are sailing without the guidance of a convincing model."  This sounds right to me, but not because a convincing (or at least, semi-plausible) model is absent. In particular, I can think of a model that is broadly consistent with observation. Moreover, it's a model that's not inconsistent with any of the theories described above.

I'll describe this model in my next post (stay tuned, I won't keep you waiting too long).



Wednesday, April 19, 2017

The St. Louis Fed's Macroeconomic Outlook

St. Louis Fed president James Bullard recently gave this speech on the U.S. macroeconomic outlook. The key themes of his talk were:
  1. The U.S. economy has converged to a low-growth, low-safe-real-interest-rate regime, a situation that is unlikely to change dramatically over the near future;
  2. The Fed can afford to take a wait-and-see posture in regard to possible changes in U.S. fiscal and regulatory policies;
  3. The U.S. policy rate can remain relatively low for now and that doing so is consistent with the dual mandate;
  4. Now may be a good time for the FOMC to consider allowing the balance sheet to shrink in nominal terms.
What does Bullard have in mind when he speaks of a low-growth "regime?" The usual way of interpreting development dynamics is that long-run growth is more or less stable and that deviations from this stable trend represent "cyclical" mean-reverting departures from trend. And if it's "cyclical," then it's temporary--we should be forecasting reversion to the mean in the near future--like the red forecasting lines in the picture below.
This view of the world can lead to a series of embarrassing forecast errors. Since the end of the Great Recession, for example, you would have forecast several recoveries, none of which have materialized.

But what if that's not the way growth happens? Suppose instead that growth occurs in decade-long spurts? Something like this.

This view of the development process does not say we're presently stuck forever in a low-growth regime. It simply suggests that we have no idea when the economy will once again embark on a higher (or heaven-forbid, lower) growth regime and that in the meantime our best forecast is for continued low-growth for the foreseeable future.

A reader suggests plotting the annualized ten-year growth rate quarter-by-quarter. Here is what it looks like:
What determines a growth regime? Government policies may play a role. Or perhaps it's just the way economies grow. There is no God-given rule which says that productivity growth must at all times proceed in a straight line. Here is the San Francisco Fed's measure of total factor productivity:
Note that the most recent productivity growth slowdown occurred well before the financial crisis.

The notion that the economy has converged to a low-growth regime is also evident in a variety of labor market measures. The prime-age unemployment rate is essentially back to its recent historical average, for example.
Measures of prime-age employment and participation still have a way to go, but arguably not very much.
Next, what does he have in mind when he speaks of a "low-safe-real-interest-rate regime?" Bullard associates the "safe-real-interest-rate" with the expected real rate of return on (nominally) safe U.S. treasury debt (which he labels "r-dagger"). Operationally, he uses the one-year U.S. treasury yield minus a measure of year-over-year inflation (e.g., the Dallas trimmed-mean inflation). Below I plot "r-dagger" using year-over-year PCE inflation. I also plot an hypothetical "r-star" interest rate which (as suggested by theory) should track the expected growth rate of real per capita consumption expenditure.
The (theoretical) real interest rate (as measured here by consumption growth)--the blue line--is on average high in high-growth regimes and low in low-growth regimes (the 1950s provide an exception). The r-dagger interest rate appears to move broadly with r-star (the early 1980s provide a dramatic exception).  The gap between r-star and r-dagger could be interpreted as a risk-premium (or a liquidity premium). The secular decline in r-dagger since the early 1980s reflects a number of factors. Inflation expectations fell and became anchored under Volcker. And since at least 2000, there's been an ever-expanding global demand for safe assets which are used extensively as collateral in shadow banking, as safe stores of wealth in emerging economies, and as objects that fulfill growing regulatory requirements (Dodd-Frank and Basel III). Evidently, Bullard does not believe that the appetite for these safe assets is likely to dissipate any time soon.
 
As for inflation, headline PCE inflation has only recently ticked back up close to the Fed's official 2% target. Nominal wage growth has also ticked up recently, but remains rather muted. The growth in real wages remains low--which is consistent with the U.S. economy operating in a low-growth regime.
Market-based measures of long-run inflation expectations appear well-anchored. Below I plot the 10-year breakeven inflation rate (expected inflation 10 years out) and the real yield on the 10-year U.S. treasury (blue line).
Given these observations, what's the rush to raise the policy rate?

At the same time, Bullard is suggesting that it might be a good time to think about reducing the size of the Fed's balance sheet. He notes that recent FOMC policy is putting upward pressure at the short end of the yield curve (via recent policy rate increases) at the same time putting downward pressure at the long end of the yield curve (via the long-term securities purchased in the LSAP). Bullard notes that "this type of twist operation does not appear to have theoretical basis." In fact, it's not clear what policy should aim for (if anything at all) in terms of influencing the slope of the yield curve.

Nevertheless, there are some good reasons to shrink the balance sheet (I provide a reason for keeping it large here). First, if there is indeed a shortage of safe assets, why is the Fed buying them up (replacing them with reserves that only depository institutions can access directly)? Ending the reinvestment program would release additional safe assets for the market, the effect of which would be to increase yields on safe assets (a good thing to the extent higher yields represent diminished liquidity premia.) Second, ending reinvestment (especially in MBS) would be a good move politically. One concern about ending reinvestment seems centered around the possibility of creating another "taper tantrum" event. But it seems unlikely that disruption in the bond market would occur if the policy change is communicated clearly and with plenty of advance notice.

Saturday, April 15, 2017

Sectoral and Occupational Trends in the U.S. Labor Market

The labor market is back to normal. Or so we are told. Here's the prime-age unemployment rate for the United States beginning in 1960.
Above we see the familiar cyclical asymmetry, in which the unemployment rate spikes up sharply at the onset of a recession and declines gradually during the recovery. As of today, the prime-age unemployment rate is at 4%, close to its recent historical norm. 

Other measures of labor market activity, however, tell a slightly different story. Here is the prime-age employment-to-population ratio. Employment took a big tumble during the recession--especially for males--and its taken a lot longer to recover than unemployment. In fact, we're not quite back to pre-recession average of about 80%. 
The reason employment has recovered more slowly than unemployment is because significant numbers of prime-age workers left the labor force in the aftermath of the Great Recession. This pattern has reversed only recently.
By these standard measures of aggregate labor market activity, the U.S. labor market appears close to "full employment." Nevertheless, there is still much anxiety concerning the present state and future course of the U.S. labor market. Much of this anxiety stems from the perennial concerns regarding the impact of international trade and technology on future employment opportunities. 

To assess the nature of these concerns, we have to move beyond the standard aggregate measures of labor market activity, which have remained relatively stable since 1990 in the face of growing trade deficits and technological progress. So if trade and technology are going to have an impact on employment opportunities, it's likely going to happen at the sectoral and/or occupational level. 

Of course, a changing allocation of human labor across sectors of the U.S. economy is nothing new. Here is the share of employment in agriculture and manufacturing since 1815. 

In 1815, 80% of employment was devoted toward the production of food. And for those of us who have never worked on a farm, here's how Charles H. Smith describes the venture in 1892:
“It’s about the meanest business I have ever experienced. It’s all fact—solemn fact – no romance, no poetry, no joke. It does seem to me that all this sort of work ought to be done by machinery or not to be done at all.” Source: Farm Life at the Turn of the Century.
As we can see, Smith got his wish. Over the centuries, machines have replaced most of human labor in the production of food. Was this a welcome development? Most of us today would likely answer in the affirmative. However, sectoral transformations like this also come at a cost, borne by those who are compelled to do the adjusting. A shovel that helps ease a manual burden is one thing. A combine harvester that renders a certain kind of labor redundant is another. When the demand for a certain type of labor in a given sector is diminished, what are workers to do? 

One thing they can do is employ their labor in the production of the machines that replaced their labor on the farm. And so youngsters left the family farm and immigrants flowed increasingly into the communities that offered employment in manufacturing and other sectors. I'm not sure how much of an improvement it was to substitute the drudgery of farm work with the drudgery of factory work, but it was probably an improvement in net terms (however small). The relatively high levels of comfort most Americans enjoy today did not come about until the latter half of the 20th century. And by that time, manufacturing employment (as a share of total employment) began its long secular decline. 

So, what accounts for the decline in manufacturing sector employment? According to Dean Baker, the U.S. trade deficit has a lot to do with it. 

Since 2000, employment in the U.S. manufacturing sector has dropped from about 17 million to 12 million workers. The sharp decline started with the "China shock" in 2001 (also a recession year). On the other hand, Germany has also experienced a decline in manufacturing sector employment while running trade surpluses. Arguably, the trade surpluses muted the secular decline in manufacturing employment in Germany, while the U.S. trade deficits exacerbated the process of structural readjustment. The trade deficit has almost surely had some impact on U.S. manufacturing employment, but it's hard for me to escape the conclusion that most of the sectoral reallocation of labor has been driven by technology. The next diagram tells the story. 
But it hardly matters to an individual worker whether they are displaced by a foreigner or an automaton. Where are the new jobs to be found and what are they paying? Are all of our good, high-paying manufacturing jobs being replaced by lousy low-paying service sector jobs? There is an element of truth to this. As Daniel Alpert points out, since the cyclical peak in 2007, the U.S. economy has added close to 7 million jobs. About 60% of these jobs appeared in the "low-wage, low-hours" sectors that account for about 36% of all private sector jobs; see table below. 
The picture is not quite as bleak as Alpert paints it. Andrew Spewak (my trusty RA) took it upon himself to produce this table. 
What this decomposition shows is that there is in fact substantial net job creation in high-wage/high-hour jobs in the service sector. The bleakness is heavily concentrated in manufacturing (which we know is in secular decline) and in construction (which is not in secular decline, but which hit a peak just prior to the housing crisis). 

Let's dig a little deeper and explore the task-based view of the labor market developed by Daron Acemoglu and David Autor (see here). According to this view, goods and services are produced by a set of tasks that can be performed by various inputs, like capital and labor. Tasks differ along two important dimensions. The first dimension measures the relative importance of "brains v. brawn" in performing a task. This is labeled "manual" v. "cognitive" in their analysis. The second dimension measures the extent to which a task can be described by a set of well-defined instructions and procedures. If it is, then it is possible to have an automaton execute a program to perform the task. Acemoglu and Autor label these "routine" tasks. If instead the job requires flexibility, creativity, on-the-fly problem-solving, or human interaction skills, the occupation is labeled "non-routine." 

Here's how some broad occupational classes fall into the four implied categories:
Of course, no classification scheme is perfect, but you get the idea. And here's how the employment shares for each category behave since 1983:
This graph could be labeled "The Decline of Routine Work." We have seen how the machines have substituted for manual labor. Robots are increasingly doing the same thing for many cognitive tasks. Here's a male/female breakdown of the phenomenon:




The decline of routine labor has been associated with "polarization"--that is, a "hollowing out" of America's middle class. This is because routine jobs, whether manual or cognitive, generally pay "middle class" wages.
 
The share of employment allocated to these middle class jobs is declining. Where is this middle class going? Largely to jobs in higher wage category, associated with non-routine/cognitive occupations. But there's also a moderate increase in the share of employment in the lowest wage category, in those jobs associated with non-routine/manual occupations.

This is already a long post, so I don't really have space to ponder the question of Whither Human Labor? It seems likely that an increasing number of tasks presently labeled non-routine will become routine (still lot's of room in manufacturing it seems, see here). We have seen this with robo-advisors in the financial sector and the emergence of artificially-intelligent doctor apps (see here) among many other places. How might the future unfold? Here's one take, by Andrew McAfee: Are Droids Taking Our Jobs?



Saturday, April 8, 2017

Fiscal over monetary policy?

The Economy May Be Stuck in a Near-Zero World (Justin Wolfers).

Justin does a good job describing how many economists view the role of monetary and fiscal policy in the post Great Recession world of low interest rates and low inflation. I am curious to know where I agree and disagree with what he says. So, here goes.

[T]he real (inflation-adjusted) interest rate consistent with the economy operating at its full potential has fallen...from around 2.5 percent to 1 percent, or lower.

I think this is true. I also do not think it's surprising that the "natural rate of interest" (r*) fluctuates and that its trend path may shift over time. Indeed, I'd be surprised to learn this was not the case. According to standard macroeconomic theory, r* should follow the trend in consumption growth. The basic idea is simple. If the economy is expected to grow rapidly, people will want to save less (or borrow against their higher future income) in order to smooth their consumption. Collectively, their efforts to consume more and save less puts upward pressure on the real interest rate. The converse holds true if pessimism reigns: people will want to save more, to make provisions against a bleak future. Collectively, the effect is to depress the real interest rate.

Of course, we cannot observe r*. But theory suggests it should be roughly proportional to consumption growth. We can observe consumption growth. Here is what the growth rate of real (inflation-adjusted) consumption of nondurables and services in the postwar U.S. looks like (series is smoothed):

If you're trained in the art of haruspicy, as most of us appear to be, then you'll divine all sorts of patterns from the picture above. You might see a 2 percent trend growth rate with a break down to 1 percent (or lower) in either 2000 or 2007. You might even detect a decade-long era of low growth in the 1970s.

Combined with the Fed's 2 percent inflation target, this implies...

In "normal times," the nominal interest rate -- the neutral interest rate plus inflation -- has fallen from around 6 percent to 3 percent. That creates a serious problem for the Fed. Here's why: Most recessions can be cured by lowering rates by several percentage points. When interest rates were closer to 6 percent, the Fed could lift the economy with plenty of leeway.

This is textbook stuff, which is not the same thing as saying it's correct. My own view on the matter (which is not necessarily correct either) is that the Fed is largely constrained to follow what the market "wants" in the way of real interest rates. It's not that the Fed "cures" a recession by lowering its policy rate -- the Fed is accommodating market forces that would have driven the real interest rate lower even in the absence of a central bank. Rightly or wrongly, the Fed acts to "smooth" these interest rate adjustments in the short-run. But at the end of the day, the trend path of r* is beyond the control of the Fed.

Yes, but what if r* is so low that the effective zero lower bound (ZLB) on the short-term nominal interest rate (the Fed's policy rate) prevents the Fed from accommodating what the market wants? With 2 percent inflation, the real interest rate can only decline to -2%. What if that's not low enough? Then something else has to give--for example, the unemployment rate will rise and remain elevated for as long as this unfortunate situation persists--a secular stagnation.

Perhaps the answer lies outside the Fed. It may be time to revive a more active role for fiscal policy--government spending and taxation--so that the government fills in for the missing stimulus when the Fed can't cut rates any further. Given political realities, this may be best achieved by building in stronger automatic stabilizers, mechanisms to increase spending in bad times, without requiring Congressional action. 

In this spirit, Justin recommends a mechanism that automatically increases funding for infrastructure programs when economic growth slows. I personally don't think this is a terrible idea. (Though, I'd rather that infrastructure be geared more to long-term needs.) But no doubt it's probably easier said than done.

Sometimes though, when I sit back and reflect on this line of thinking, it strikes me as rather odd in a couple of respects.

First, is the ZLB really a significant economic problem? If it is, then why not abolish it as recommended by Miles Kimball? Would permitting significantly negative real rates of interest solve our problems? I don't think so. I'm inclined to think of a low r* as symptomatic of more fundamental economic forces. And eliminating any (real or perceived) gap between the market interest rate and r* is probably small potatoes (see here).

If r* is low, then we need to ask why it is low. There's no shortage of possible explanations out there (low productivity growth, demographics, etc.). If we somehow decide we'd like to see it higher, the solution is likely to be found in growth-promoting policies. (Whether we want growth-promoting policies is a completely separate matter, by the way. Personally, I think more attention should be paid to policies that encourage social cohesion, which may or may not be consistent with higher growth. But this is a column for another day.)

Second, I think the world has indeed changed for discretionary monetary and fiscal policy, but in a way that almost no one talks about. Quite apart from any possible changes in r* (which we cannot measure), the real rate of return on U.S. Treasury (UST) debt--what my boss James Bullard calls "r-dagger"--has been declining for over 30 years (diagram taken from here).


One interpretation of this pattern is that USTs were initially a flight-to-safety vehicle with the disruptions that occurred in the early 1970s (so real yields declined). With the breakdown of Bretton Woods and fiscal pressures (Vietnam war, War on Poverty, etc.), however, inflation became un-anchored. The high real yield on nominal UST debt reflected a growing inflation-risk premium in the early 1980s (when inflation was high and volatile). Subsequently, as inflation declined and inflation expectations became anchored (thanks to Volcker and a terrible recession) the inflation risk premium declined over time. Since about 2000, a China trade shock and other factors led to a growing world demand for USDs and USTs. R-dagger (r+) remains extremely low even today--reflecting the "liquidity premium" that the market now attaches to UST debt.

Moreover, the distinction between USDs and USTs is much diminished in financial markets. In the old days, when the Fed wanted to move interest rates through an open-market swap of USD for UST, it meant something. But today, it means almost nothing, since interest-bearing reserves are a very close substitute for interest-bearing treasuries. In short, U.S. treasury debt is essentially "money" as far as financial markets are concerned (USTs circulate as such in repo markets, for example).

The implication of all this for monetary and fiscal policy is quite interesting. The fact that the yield on USTs is less than (our estimate of) the natural rate of interest suggests that the policy rate is presently too low -- not too high (as is suggested by standard ZLB concerns). The most direct way to raise interest rates (i.e., eliminate the liquidity premium on USTs) is for the U.S. treasury to issue debt at a faster pace. One way to do this is through Justin's automatic infrastructure funding plan that kicks in when liquidity premia on USTs are elevated (bond yields are low). Another way is to have automatic (temporary) tax cuts kick in. Yet another way (though far less desirable) is to have the Fed increase the interest in pays on reserves. Politically this is dynamite, but from an economic perspective, it forces (ceteris paribus) the treasury to issue debt at a faster pace (because it lowers Fed remittances to the treasury). Yet another way is to have the Fed sell some of its treasury holdings (since treasuries are sometimes more liquid than reserves in financial markets--i.e., only depository institutions have direct access to reserves).

Depending on which view one adopts, the recommended Fed policy action matters a great deal (at least, in principle, if not quantitatively). If the interest rate is too high (ZLB view), then it should be lowered, or the inflation target raised. If the interest rate is too low (liquidity premium view), then it should be raised, through asset sales or some other mechanism.

On the other hand, the recommended Treasury policy action seems robust across the two views: the treasury should expand its debt at a faster pace (via tax cuts or increased spending, or some combination). This seems like a promising development from the perspective of competing theories. If a policy recommendation follows from many different perspectives, we become more comfortable with the idea of actually implement them. Of course, there are some caveats to consider, which I discuss here. But enough for today.

Wednesday, March 1, 2017

A reply to Lawrence White

People are generally not accustomed to the idea of a central bank altering the maturity structure of outstanding government debt in a manner that might confer a financial benefit to the government (and by extension, to the citizens it represents). The purpose of my previous post was to make people aware of this possibility, using the recent experience of the Federal Reserve's quantitative easing (QE) program as an example.

The gist of my story went something like this. Suppose that the Fed makes a one-time purchase of a 10-year treasury bond yielding a risk-free 2.5% annual nominal coupon. Suppose that the purchase is financed by "printing" reserves. (Note, this asset purchase and method of finance describes the basic nature of QE). Suppose further that the interest paid on reserves remains below 2.5% for the duration of the bond. Then the Fed makes a profit off the rate of return differential in each of the 10 years it holds the bond on its books. If I understand my critics correctly, none would dispute the financial benefit associated with this Fed intervention as far as the public purse is concerned; at least, as long as it is somehow known ex ante that the short rate will remain below the long rate in the manner assumed.

One objection to this story asks why the Fed is needed to convert a 10-year bond into a lower-cost short-term instrument when the Treasury itself could have just issued a short-term debt instrument to begin with. That's a good question. I'm not sure what factors determine the Treasury's choice over the maturity structure of its debt. The key question to ask is whether the structure chosen is optimal from a social standpoint. The benefit associated with the Fed intervention I described in my post assumes that it is not. What justifies this assumption? No one really challenged me on this. Another objection relates to the optics of the enterprise from a political perspective. Even if the intervention makes sense from an economic perspective, paying interest on reserves held by big banks with the effect of reducing Fed remittances to the Treasury just looks bad (even though it's equivalent to the Treasury paying those big banks for holding treasury debt).

The main objection to my post seems to rest on the claim that I have ignored "duration risk." Let me quote directly from Lawrence White, who was kind enough to take the time to critque my post:
When the Fed borrows short from the banks to lend long to the Treasury, it does not do so costlessly. Duration transformation carries a risk of capital loss, also known as duration risk. Suppose the yield curve shifts up, both short and long interest rates rising together. The Fed will experience a decline in present value of its assets that will swamp the smaller decline in the present value of its liabilities. Such an event is not unknown: in 1979-81 it rendered insolvent about two-thirds of US thrift institutions, who were financing 30-year fixed-rate mortgages with 1- and 2-year deposits. In cash-flow terms, such an event would mean that the Fed would quickly have to start paying higher interest rates to borrow, while its asset portfolio continues to pay low yields and roll over much more slowly. The Fed’s annual net transfer to the Treasury might even go negative. Smaller or negative transfers from the Fed to the Treasury would mean a sudden jump in the present value of the public’s ordinary tax liabilities. Net interest income from playing the yield curve is not a free lunch.
To assess the merit of this critique, let's consider the following thought experiment. First, imagine that there is no central bank (End the Fed wins). Next, imagine that the Treasury wants to borrow money. One option is to issue the 2.5% 10-year bond described above. But another option would be to issue very short-term treasury debt, rolling it over for 10 years. If the Treasury behaved in this manner, it would in effect be replicating the Fed intervention I described above. There would be obvious cost savings to the government and taxpayer, at least, assuming that short rates remained below 2.5%. So what could possibly go wrong?

Well, the risk is that the short-term rate might rise and stay on average over 2.5% for the next 10 years. In that event, which is certainly a risk to consider, the Treasury will have made a loss relative to locking in its debt at the long-term rate of 2.5%. But from the perspective of the taxpayer, it matters not one iota whether this prospective loss is incurred because of the Treasury action or an equivalent action on the part of a Fed QE program.

So, if the point is that financing short-term at low rates entails risk relative to financing long-term at high rates, then I agree. The question, however, is not whether QE exposes the taxpayer to duration risk. Of course it does. The same would be true if the Treasury was to shorten the average duration of its debt on its own. Whether the Fed or Treasury performs this operation is immaterial (from an economic perspective, not necessarily from a political perspective). The fundamental question then is whether government debt is structured optimally to begin with. To the extent that the fiscal authority relies too heavily on long-term debt, then some QE on the part of the Fed could have the public finance benefit I described in my post.

Does the Treasury rely too heavily on long-term debt? I honestly do not know. Those who warn against the Fed shortening duration must implicitly assume that Treasury debt-structure is optimal. Maybe they're right. But as the following figure shows, the spread between long and short rates is usually (though not always) positive.


If this pattern could be relied on to hold in the future, then the question to ask of a QE program is not whether it entails duration risk (it does). Rather, the question is whether the endeavor constitutes a positive net-present-value project.  The data above suggests that it may very well be. This is not the same as claiming there's a "free lunch." The Fed has saved the government over $80 billion a year over the past seven years. That's about $600 billion in cost savings--money that the government would have had to secure through other means to service its debt. It was fluke, the critics will say. An ex ante gamble that pays off ex post is not a justification. Sure. But billions in remittances, year after year, one fluke after another?

Let me briefly address a few other points made by Larry. In the quoted passage above, he warns that the Fed is exposed to suffering a capital loss on its assets should market interest rates rise, that this has happened to private banks in the past with dire consequences, that the result may be lower (possibly negative) remittances to the Treasury.

I think these concerns are overstated. Warning against the prospect of having to reduce remittances is not an economic argument against exploiting a profit opportunity (although, one could make the case on political grounds).


Yes, if the market price of bonds fell, the market value of Fed assets would decline. Should this be a concern? U.S. treasury bonds are nominally risk-free and the Fed normally holds its assets to maturity. Larry's reference to the plight of U.S. thrift institutions is not, in my view, relevant to the issue at hand. Because Fed liabilities are money (essentially equity shares accepted by all as a payment instrument), the Fed cannot go insolvent in the same way a private company can (although politically, the optics would be bad). In any case, the proper way to view the issue is from the perspective of a consolidated Fed/Treasury balance sheet. The issue concerns optimal maturity structure and the impact on the government's consolidated balance sheet as market interest rates change.

It is also Larry's view that the Fed has been carrying "significant" default risk through its holdings of mortgage-backed securities. While I think there are good reasons to let the MBS portion of the Fed's balance sheet run off, I think Larry is once again overstating concerns (even if I grant his point about potential market distortions). The MBS purchased by the Fed are not legacy assets--they constitute senior tranches of securities issued well after the real estate market bottomed out. In short, they're about as safe as securities backed by private assets get. That's not to say that the Fed should be holding such assets. My own view is that the Treasury should have issued bonds against such securities. Doing so would have gone some way to alleviating the worldwide shortage of safe assets.

Larry concludes with a plea for "normalization." Let's go back to a world where bank reserves are scarce. But why? In what world does it make sense to render liquidity scarce? Milton Friedman argued that optimal debt policy entails eliminating all liquidity premia. Almost every economic model I am aware of supports this notion. An appeal to follow historical norms is not an economic argument.
 
Let me sum up. I think there are two types of arguments one can bring to bear on the question of central bank balance sheet policy as a fiscal debt-management device. The first is economic and the second is political.

The political argument is based on the idea that an independent and accountable central bank is a good idea. To protect central bank independence, it would be unwise for a central bank to operate with the expressed purpose of managing the debt--a realm guarded jealously by the fiscal authority. I think this is a good and practical argument against becoming too enthused with large central bank balance sheets. But this is not the argument that Larry is making.

Larry's argument is an economic one. An economic argument against the Fed helping to manage the debt, suggested to me by my colleague Steve Williamson, is that the Fed does not really have the tools to do the job properly. In particular, the Fed cannot issue debt that can circulate widely. The only liability it can issue--reserves--is clearly inferior to short-term treasury debt. In particular, reserves can only be used by banks with reserve accounts, while treasuries can circulate outside the banking system. This is a good point. It basically says "since the Treasury can do the job better in principle, better the Fed not do it at all." But this not the argument Larry is making either.

Larry's argument implicitly assumes (possibly correctly) that the Treasury has structured its debt optimally. If it has, then there is no "free lunch" for the Fed. If this is the case, then I would have to agree. (And it's heartwarming to see Larry thinking so highly of a government agency's ability to run its operations!)

As I reflect back on the original argument I put forth, I think I would now put it in the following way. First, I would not advocate that the Fed explicitly pursue balance sheet policy as a way to help government financing. As I made clear in my original post, the recommendations I was making were predicated on the assumption of the Fed achieving its dual mandate. If in the course of achieving the dual mandate the Fed is saving the Treasury billions of dollars, then what's the rush to reverse operations? The fear of exposing the taxpayer to duration risk may be a reason--if you feel the Treasury already has its debt structured optimally. But the fact of consistently positive and sometimes very large remittances from Fed to Treasury for years (and even decades) might lead one to question that assumption.

Wednesday, February 15, 2017

A public finance case for keeping the Fed's balance sheet large

Former Fed Chair Ben Bernanke recently asked a question concerning the optimal long-run size of the Fed's balance sheet (Should the Fed keep its balance sheet large?). Bernanke comes down on the side of "keeping the balance sheet close to its current size in the long run." While he does not explicitly say how "size" is defined, I think it's clear he means the size of the balance sheet measured relative to the size of the economy (say, as measured by nominal GDP). According to this measure of size, the Fed would have to grow its balance at the rate of nominal GDP growth.

In addition to the reasons reported by Bernanke, I think there's a public finance argument to be made for keeping the Fed's balance sheet large--at least, under certain conditions--like ensuring that the inflation mandate is met. Let me explain.

Let's begin with a picture that most people are familiar with. 
Prior to 2008, the Fed's balance sheet was under one trillion dollars in size. Prior to 2008, it grew roughly at the same rate as the economy. Most of these assets consisted of short-term U.S. treasury securities. Most of these asset acquisitions were financed with zero-interest money (currency in circulation). Since 2008, the Fed's balance sheet has grown to 4.5 trillion dollars. The composition of assets has moved away from short-term government debt to longer-term debt and mortgage-backed securities. Most of these asset acquisitions were financed with low-interest money (reserves).

Is 4.5 trillion a big number? Well, yes. But then, the U.S. is a big economy: the U.S. nominal GDP for 2016 is close to 19 trillion dollars. So in measuring the size of the Fed's balance sheet, it probably makes more sense to measure size as a ratio. The following graph plots the size of the Fed's balance sheet as a ratio of nominal GDP.
Prior to 2008, the size of the Fed's balance relative to the economy averaged about 6%. The balance sheet size peaked in 2014 at just over 25%. Note that by this metric, the Fed's balance sheet has been contracting since 2014.
 
For the record, note that the large expansion in the supply of Fed money was associated with historically low rates of inflation:
Now let's talk about the business of banking. I like to think of a bank as an asset transformer: a bank converts relatively illiquid assets into relatively liquid liabilities. The Fed buys relatively high-yielding (but safe) securities, like U.S. treasury bonds and AAA-rated mortgage-backed securities. It pays for these acquisitions by issuing liabilities (printing money) in the form of low-interest reserves.
The Fed transforms high-interest government debt into low-interest Fed liabilities (money).

The difference between the interest the Fed earns on its assets and the interest it pays on its liabilities is an interest rate spread. Isn't it wonderful to be able to borrow at low rates and invest at high rates? This is precisely what the Fed did with its large scale asset purchase (LSAP) program. Apart from any other effects that this intervention had on the economy, it resulted in huge profits for the Fed. Keep in mind that any profit made by the Fed is remitted to the U.S. Treasury (and thus, ultimately, to the U.S. taxpayer).

So just how much money does the Fed return to the treasury each year? I'm glad you asked, here you go:
In recent years, the Fed has been returning about $80-90 billion per year to the U.S. Treasury. While interest rates were higher in the past, the Fed's balance sheet was much smaller--and so while the profit margin was high, the volume was low. Today the profit margin is smaller, but the balance sheet is much larger. (The distance between the red and blue lines represents the Fed's foregone profit since it started paying interest on reserves in 2008).
 
What sort of rate of return does the Fed make on its portfolio? The following graph plots Fed payments to the Treasury as a ratio of the Fed's assets.
Since the bulk of the Fed's assets are in the form of U.S. government bonds, it should be no surprise to learn that the rate of return has generally followed the path of market interest rates downward. Still, in recent years, the annual rate of return is about 2%. Given that the Fed is presently financing these assets with cash (0%), ON RRP (0.25%) and IOER (0.50%), the profit margin is still significantly positive (though one wonders about the scope for further policy rate hikes if market rates remain low).

In light of this analysis, why are some people calling for the Fed to reduce the size of its balance sheet? Usually the concern is that a large balance sheet portends higher future inflation. But we've been living in a world of lowflation for many years now and we're likely to stay there for the foreseeable future (though central banks should of course remain vigilant!). There is, in fact, some theoretical support for the notion of reducing the Fed's policy rate (subject to the dual mandate and financial stability concerns); see, for example: The Inefficiency of Interest-Bearing National Debt.

Reducing the Fed's balance sheet at this point in time seems like a needless loss for the U.S. taxpayer. Given that the Treasury is marketing a bond, who do you want to hold it? If the debt is held outside the Fed, the government needs some way to pay the 2% carry cost of the debt. The government will in this case have to reduce program spending, increase taxes, or increase the rate of growth of debt-issuance. Alternatively, if the Fed holds the debt, the carry cost is generally much lower. This cost-saving constitutes a net gain for the government. So why not take advantage of it?

***
P.S. I realize there are some who argue that a central bank enables big government. Since the government is too large, we need to end central banking and, in this manner, starve the beast. But this argument amounts to "let's make the government less efficient in terms of financing their operations--that'll force it to get smaller." This line of argument strikes me as naïve--I'm not sure what would prevent the government from simply substituting into different methods of finance. If you want smaller G, then lobby Congress to make G smaller. But given that smaller G, it should still be financed in the most efficient manner possible. And that means following the prescription above.