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

Monday, February 18, 2019

Is Neo-Fisherism Nuts?

According to my friend and former colleague Steve Williamson, inflation is low in Japan because of the Bank of Japan's policy of keeping its policy rate low. Accordingly, if the BOJ wants to hit its 2% inflation target, it should raise its policy rate and keep it persistently higher. This is what I've called the NeoFisherian proposition. It's a provocative idea because it flies in the face of conventional wisdom. But is it correct? Does it serve as a practical guide for monetary policy? My feeling is that the answers to these questions are "no" and "no." In what follows, I explain why.

At some point in their undergraduate career, students of macroeconomics are introduced to the Fisher equation. The Fisher equation usually stated as R = r + π or, in words:

Nominal Rate of Interest = Real Rate of Interest + Expected Rate of Inflation

For simplicity, think of the real rate of interest as the rate of return an investor can achieve by storing goods across time. For squirrels storing nuts, the real rate of interest is negative. For humans planting corn, it is positive. Whatever its value, let's just fix it at some number and assume it remains invariant over time (this is not a critical assumption for the arguments I want to develop below). Then, the Fisher hypothesis is that the nominal rate of interest should move one-for-one with the expected rate of inflation.

How does the Fisher equation hold up in the data? Let's just say that the evidence is mixed. Fisher himself famously rejected it as being empirically relevant. But over long periods of time, and also across countries, various nominal interest rates do appear positively correlated with the measured inflation rate (taken as a proxy for expected inflation).

Well, correlation is one thing, explanation is another. What is the theoretical underpinning of the Fisher equation? One way to view it is as a no-arbitrage-condition. Suppose that planting a bushel of corn yields 1.02 bushels at harvest (2% real rate of interest). Suppose that the nominal price of corn (its price measured in dollars) is expected to rise 10% by harvest time. What rate of return would an investor demand of a bond promising to deliver money at harvest time? The Fisher equation says that the investor should demand a rate of return of at least 12%. The bond would then deliver 12% more dollars that, if spent on corn at harvest, would leave an inflation-adjusted return of 2%. In this case, the investor would be just indifferent between investing in a corn planting venture and the nominal security.

Viewed in this light, the Fisher equation can be interpreted as the interest rate bond-holders demand, given their outlook on inflation. And, indeed, the standard textbook explanation for why nominal interest rates tend to rise with inflation provides a clear causal link, starting with monetary policy in the form of base-money growth rate:

    [1] increase in money supply growth (spent on goods or delivered as tax cuts or transfers);
    [2] causes increase in demand, which causes prices to rise;
    [3] inflation expectations adjust upwards accordingly;
    [4] bondholders demand higher interest rate to compensate for higher expected inflation.

The interpretation above assumes that monetary policy does not target the interest rate on bonds. Instead, it grows the money supply and lets the market determine the nominal interest rate. But note that monetary policy is targeting an interest rate in this explanation. In particular, the nominal interest rate on central bank money (reserves and currency) is set to zero. This policy goes by the acronym ZIRP (zero-interest-rate-policy). Moreover, there is a Fisher equation that holds for money. It looks like this R_m + LP = r + π, or in words:

Nominal interest rate on money + Liquidity premium = Real interest rate + Expected inflation 

If the nominal interest rate on money is zero, then money must be held for its non-pecuniary benefits (liquidity). The liquidity premium on money is in this case equal to the nominal interest rate on an illiquid bond; i.e., LP = R = r + π. 

Now, if the nominal security yielding a positive interest rate in the story above consists of government bonds (denominated in the domestic currency), then the only way to explain the apparent discount on government bonds is by appealing to an explicit government policy that renders these bonds illiquid relative to central bank money. And indeed, we see restrictions like this in place throughout history. For example, convenient low-denomination zero-interest notes (cash)  versus inconvenient large-denomination notes (bonds) trading at discount. Or consider today, where interest-bearing accounts at the U.S. Treasury are deliberately rendered useless for making payments. Or the Fed's apparent aversion to setting up a repo facility for U.S. Treasury debt in order to enforce a ceiling on its interest rate target path (such a facility would serve to reduce the demand for reserves).

Monetary theorists like Neil Wallace have puzzled forever over the phenomenon of why government bonds should trade at a discount (the so-called, coexistence puzzle). Others, like former Minneapolis Fed president Narayana Kocherlakota, have attempted to rationalize policies that render government bonds illiquid; see here. At the end of the day what is true is the following: the nominal interest rate on government bonds is, one way or another, a deliberate policy choice (for governments that issue debt denominated in the money they issue). This goes for default risk as well. There is no reason to default on debt that constitutes a promise to deliver money that one can costlessly produce. If default takes place in such circumstances, it is a policy choice, not an economic necessity.

Alright, what does all this have to do with Neo-Fisherism and the Neo-Fisherian proposition? I hope everything will fall together in due course. In the meantime, let's assume that the Fisher equation is sound theoretically and holds approximately well in the data. Would this support the proposition? It's clearly not enough because the proposition has to do with causality. The conventional view outlined above is also consistent with theory and evidence. Moreover, the conventional view as expressed through [1]-[4] provides a simple, coherent, easy-to-understand story for why we'd expect to see a positive correlation between interest rates and inflation in the data. It may not be correct, but at least it's understandable. Can the Neo-Fisherian proposition be explained in a similarly simple and compelling way? I think it's important for ideas to expressed in clear and simple terms. If policymakers are going to take the proposition seriously, the underlying economic mechanisms will have to be explained in a simple and straightforward manner. It will have to resonate with listeners at some level.

I've only heard of one mechanism that I find semi-plausible: the idea that a higher policy rate increases the interest expense of government debt which, if not met with a tax increase, must be met by an acceleration in money/bond printing (some empirical evidence here). Alternatively, could it be that an increase in the policy rate serves as a type of cost-push shock that propagates itself forward through some adaptive inflation expectations mechanism? I don't know, but it seems worth exploring.

But this is not how Neo-Fisherians explain the mechanism. You can listen to Steve explaining the mechanism in this David Beckworth podcast beginning at the 21 minute mark. There is also this piece published in the St. Louis Fed's Regional Economist: Neo-Fisherism: A Radical Idea or the Most Obvious Solution to the Low-Inflation Problem? Here is how he explains it (boldfont text representing my emphasis):
But, what if we turn this idea on its head, and we think of the causation running from the nominal interest rate targeted by the central bank to inflationThis, basically, is what Neo-Fisherism is all about [...] But how would this work? [...] To simplify, think of a world in which there is perfect certainty and where everyone knows what future inflation will be. Then, the nominal interest rate R can be expressed as R = r + π, where r is the real (inflation-adjusted) rate of interest and π is future inflation. 
Then, suppose that the central bank increases the nominal interest rate R by raising its nominal interest rate target by 1 percent and uses its tools (intervention in financial markets) to sustain this forever. What happens? [...] after a long period of time, the increase in the nominal interest rate will have no effect on r and will be reflected only in a one-for-one increase in the inflation rate, π. In other words, in the long run, the only effect of the nominal interest rate on inflation comes through the Fisher effect; so, if the nominal interest rate went up by 1 percent, so should the inflation rate—in the long run.
First, I wonder  what "tools and interventions" he has in mind. If the tool involves a sustained increase in the money growth rate, then there's nothing new here--this would just be standard Monetarist reasoning consistent with the textbook explanation [1]-[4] above.

But I think he means something else. As I explained above, the "Fisher effect" is a statement about how expected inflation affects the interest rate, not the other way around. The interest rate in Steve's thought experiment is fixed. Therefore, the "Fisher effect" here must relate to the economic force that causes inflation expectations to rise. What is this force? He doesn't say. One is left with the feeling that, well, since the Fisher equation holds in theory (and to Steve, in the data as well), inflation expectations somehow must adjust to make this true. Ergo, raising the interest rate will eventually lead to an increase in inflation. Central bankers need more than this to go on. In any case, I think that the logic is flawed. Let me explain.

How Neo-Fisherism Leads to Bad Monetary Policy Advice 

Let's take the case of Japan. Japan's inflation rate has been close to zero for a long time. Although I do not know why, Japan wants a higher inflation rate. How is to achieve this objective? 
I laid out my case here: The Failure to Inflate Japan. In a nutshell, the argument is this. [1] Peg the policy rate to zero all along the maturity structure of government debt (the BOJ is doing this); [2] Grow the nominal debt more rapidly until the desired inflation occurs (the government is not doing this).

Steve roundly rejects my line of reasoning--which is, of course, fine--except that (my apologies, but) I don't understand what he's saying (see here):
(i) How does the CB keep R=0 "along the yield curve." How could you have a flat yield curve at zero with positive inflation? (ii) If you're eliminating all taxes and the fiscal authority is financing everything by issuing debt, and the CB is trying to sustain R=0, then something has to give. For example, people start anticipating that fiscal authority can't roll over the debt, default premia rise on the government debt, and CB is forced to increase R to generate the CB profits required to keep the government afloat.
To answer (i), I think the BOJ has shown how it can be done. If the market is discounting JGBs, the BOJ can just buy (or threaten to buy) them up at par. To answer (ii), there is no nominal default risk to consider for Japan--at least, there's no economic reason to default: Japan can print the money it's  promising its bond holders. (And if one is worried about the real default implicit in inflation, remember that increasing the inflation rate is exactly the policy goal here.)

Steve has also pointed out that Japan's nominal debt has already grown substantially, so where's the inflation? The answer is that one cannot just look at supply--one must also consider demand. Evidently, the demand for JGBs has been increasing rapidly as well. If the supply had not accommodated this growing demand, Japan may very well have experienced the mother of all deflations (that demand is not observed and has to be inferred from price and quantity is a key weakness in this story).

Alright, so Steve does not like my way of increasing inflation. What does he recommend as an alternative? The BOJ should raise its policy rate, say from 0 to 400bp, and keep it there. There may be a short-run "liquidity effect," but the inflation will eventually come. How do we know? The Fisher equation. Can you elaborate? The Fisher effect will mean that inflation expectations will rise and inflation will follow. Why should inflation expectations rise? Because ... well, rational expectations ... and the Fisher equation. Can you elaborate? (Rinse and repeat.)

In any case, even if one accepts "rational expectations," the argument is not correct. As I explained above, there are really two Fisher equations:

[Fisher 1]: R = r + π
[Fisher 2]: R_m + LP = r + π

where, in case you forgot, r is real interest rate, π is expected inflation, R is nominal rate on illiquid bond, R_m is nominal rate on liquid bond (including reserves) and LP is a liquidity premium.

The interest rate controlled directly by the central bank is R_m. The central bank can easily set R_m = 0 and then monetize all the tax-cuts that are necessary to increase π. As π increases, so will R, in accordance with the Fisher equation. Could it be that Neo-Fisherians are confusing R with R_m? (This seems unlikely as I know that Steve knows the difference.)

What then is the effect of raising R_m? Well, it's complicated. Much depends on the structure of fiscal policy (Ricardian vs. Non-Ricardian); see here. In some models, raising R_m leaves r and π unchanged, which implies that the liquidity premium on government money LP falls. Eliminating the liquidity premium on government money/bonds is the famous Friedman rule prescription (convention version sets R_m = 0 and π = -r, but R_m = r + π for any π > 0 also works). But in other models, increasing R_m puts upward pressure on the real rate of interest, reducing the demand for investment, leading to economic contraction with no change in long-run inflation; see here.

The point of all this is, IF higher inflation is desired (and I am by no means advocating any such policy), THEN why not keep the policy rate low and use "free lunch" fiscal policies as long as inflation remains below target? Why bother experimenting with the Neo-Fisherian prescription of raising the policy rate that's somehow supposed to make people magically expect higher inflation?

Friday, January 11, 2019

When is more competition bad?

Contrary to popular belief, standard economic theory does not provide a theoretical foundation for the notion that "competition is everywhere and always good." It turns out that legislation that promotes competition among producers may improve consumer welfare. Or it may not. As so many things in economics (and in life), it all depends.

I recently came across an interesting paper demonstrating this idea by Ben Lester, Ali Shourideh, Venky Venkateswaran, and Ariel Zetlin-Jones with the title "Screening and Adverse Selection in Frictional Markets," forthcoming in the Journal of Political Economy.The paper is written in the standard trade language. Like any trade language, it's difficult to understand if you're not in the trade! But I thought the idea sufficiently important that I asked Ben to translate the basic results and findings for a lay audience. I'm glad to say he was very happy to oblige.

And so, without further ado, today's guest post by Ben Lester, my colleague at the Philadelphia Fed.
You can follow Ben on Twitter :  @benjamminlester 

Competition in Markets with Asymmetric Information
By Benjamin Lester

In many basic economic models, competition is good – it increases welfare.  As a result, policy makers often introduce reforms that they hope will reduce barriers or “frictions” in order to increase competition.  For example, the Dodd-Frank Act contains regulations aimed at promoting more competition in certain financial markets, such as derivatives and swaps, while the Affordable Care Act contained provisions that were intended to promote competition across health insurance providers.

In a recent paper with Ali Shourideh, Venky Venkateswaran, and Ariel Zetlin-Jones, we re-examine the question of whether more competition is welfare-improving in markets with a particular feature – what economists call “asymmetric information.”  These are markets where one side has information that is relevant for a potential trade, but the other side can’t see it. Classic examples include insurance markets, where an individual knows more about his own health than an insurer; loan markets, where a borrower knows more about her ability to repay than a lender; and financial markets, where the owner of an asset (like a mortgage-backed security) may know more about the value of the underlying assets than a potential buyer.

Unfortunately, understanding the effects of more or less competition in markets with asymmetric information has been constrained by a shortage of appropriate theoretical frameworks.  As Chiappori et al. (2006) put it, there is a “crying need for [a model] devoted to the interaction between imperfect competition and adverse selection.”

What we do
We develop a mathematical model of a market – to fix ideas, let’s call it an insurance market – that has three key ingredients.  The first ingredient is adverse selection: one side of the market (consumers) know more about their health than the other side of the market (insurers).  Second, we allow the two sides of the market to trade sophisticated contracts: as in the real world, insurers can offer consumers a rich set of options to choose from, consisting of different levels of coverage that can be purchased at different prices.  Last, we introduce imperfect competition by assuming that consumers don’t always have access to multiple insurers: in particular, each consumer will get offers from multiple insurers with some probability, but there is also a chance of receiving only one offer.[1]  Hence, our model allows us to capture the case of perfect competition (where all consumers get multiple offers), monopoly (where all consumers get only one offer), and everything in between.

What we find

One of our main results is that increasing competition can actually make people worse off.[2]  To understand why, it’s important to understand the types of contracts that our model predicts will be offered by insurers.  Let’s say that there are two types of consumers: those who are likely to require large medical expenses (“sick” consumers), and those who are not (“healthy” consumers).  Then insurers will often find it optimal to offer two different plans: one that is expensive but provides more coverage, and one that is cheaper but provides less coverage.[3]  Designed correctly, these two options will induce consumers to self-select into the plan intended for them, so that sick consumers will pay a higher price for more coverage and healthy consumers will pay a lower price for less coverage.

An important property of these contracts is that they fully insure sick consumers, but they under-insure healthy consumers.  Ideally, insurers would like to offer healthy patients more coverage, but they can’t: given the lower price, sick consumers would choose this new plan, making it no longer profitable for insurers to offer it.  This theoretical result – that separating the sick from the healthy requires under-insuring healthy consumers – is a fundamental result in markets where asymmetric information is present.  The relevant question for us is: how does the amount of competition determine the extent to which healthy consumers are under-insured? The answer we find is that some competition can induce insurers to provide healthy consumers with more insurance, but too much competition can have the opposite effect. 

The intuition is as follows.  When consumers are more likely to receive multiple offers, insurers respond by making more attractive offers to consumers, as they try to retain market share.  The key question turns out to be: does increasing competition make them sweeten the deal more for sick consumers, or for healthy consumers? On the one hand, as the offer intended for sick consumers gets better, they have less incentive to take the offer intended for healthy consumers – in the parlance of economics, their “incentive constraint” loosens.  Hence, as insurers sweeten the offer intended for sick consumers, they are able to offer healthy consumers more coverage, and welfare rises.[4]  On the other hand, however, as the offer intended for healthy consumers become more attractive, sick consumers are more tempted to take it – their incentive constraint tightens – and the only way to keep the two separate is to reduce the amount of coverage being offered to healthy consumers, causing welfare to decline.

In the paper, we show that the former, positive effect dominates in markets where insurers have a lot of market power, while the latter, negative effect dominates when the market is relatively competitive. Hence, in markets with asymmetric information, welfare is maximized at some interior point, where there is some competition, but not too much!

Other results and future research
In the paper, we also show that increasing transparency has ambiguous effects on welfare.  In particular, we study the effects of a noisy signal about a consumer’s type – in the insurance example, this could be a blood test or information about an individual’s pre-existing conditions.  We show that increasing transparency is typically beneficial when insurers have a lot of market power, but it can be detrimental to welfare in highly competitive environments.

More generally, our model provides a tractable framework to confront a variety of theoretical questions regarding markets that suffer from asymmetric information, and offers a number of insights into existing empirical studies, too.[5]  For example, there is a large literature that tests for the presence of asymmetric information by studying the quantitative relationship between, e.g., the amount of insurance that consumers buy and their tendency to get sick.[6]  However, according to our analysis, insurers find it optimal to offer menus that separate consumers only when markets are sufficiently competitive, and when there is a sufficiently large number of sick consumers in the population.  Otherwise, they find it best to offer a single insurance plan.  This finding implies that, when insurers have sufficient market power, there will be no relationship between the quantity of insurance a consumer buys and his health status.  In other words, one can’t empirically test for asymmetric information without controlling for the market structure.  This is just one of many positive predictions of our model that we plan to test in the data.

Burdett, K., and K. L. Judd (1983) “Equilibrium Price Dispersion,” Econometrica, 51, pages 955–69.
Chiappori, P.-A., B. Jullien, B. Salanié, and F. Salanié (2006) “Asymmetric Information in Insurance: General Testable Implications,” RAND Journal of Economics, 37, pages 783–98.
Chiappori, P.-A., and B. Salanié  (2000) “Testing for Asymmetric Information in Insurance Markets” Journal of Political Economy,  108, pages 56–78.
Lester, B., A. Shourideh, V. Venkateswaran, and A. Zetlin-Jones (2018) “Screening and Adverse Selection in Frictional Markets,” Journal of Political Economy, forthcoming.

[1] We borrow this modeling device from the paper by Burdett and Judd (1983).
[2] At a high level, the idea that reducing frictions can sometimes make people worse off is not unique to our paper; these types of results are known from the theory of the second best. What distinguishes our result is the context in which it arises, and our ability to characterize precisely when and why reducing frictions (or increasing competition) is harmful.
[3] The negative relationship between price and coverage should be familiar to most readers; see, e.g., the metal tiers (platinum, gold, silver, bronze) offered at
[4] Since sick consumers are always fully insured, consumers’ welfare always rises when healthy consumers are offered more insurance.  On a more technical level, all of our statements about welfare are based on a measure of ex ante, utilitarian welfare.
[5] As a technical aside, unlike many models of asymmetric information and screening, we find that an equilibrium always exists in our environment, that the equilibrium is unique, and that the equilibrium does not rely on any assumptions regarding “off-path beliefs.”
[6] See the seminal paper by Chiappori and Selanie (2000).


The views expressed here are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.

Tuesday, December 25, 2018

Racial Diversity in the Supply of U.S. Econ PhDs

This post is motivated by Eshe Nelson's column "The Dismal Cost of Economics' Lack of Racial Diversity." I was especially struck by this data -- out of the 539 economics doctorates awarded to U.S. citizens and permanent residents (by U.S. institutions), only 18 of the recipients were African-American.
I thought it would be of some interest to see what the data looks like more broadly over other groups and over a longer period of time. I thank my research assistant, Andrew Spewak, for gathering this data (from the National Science Foundation). 

Let's start with the raw numbers first. The data is aggregated into 5-year bins beginning in 1965 and up to 2014. The orange bars represent the number of econ PhDs awarded to U.S. citizens and permanent residents (by U.S. institutions) over a given 5-year period. The blue bars represent the total number of doctorates awarded.

Seems like the number of econ PhDs awarded to U.S. citizens is on the decline and that this decline has been partially made up by the number of PhDs awarded to foreign students.

Now, let's stick with citizens for the moment and decompose the data across various "racial" categories. The following figure reports the share of econ doctorates earned by various groups.

The most dramatic pattern is the relative decline of PhDs awarded to Whites and the increasing share of degrees awarded to Asians (there is also a noticeable uptick in the "Other" category which includes groups like Native Americans). Blacks and Hispanics have made some gains since the early years, but have since stabilized to about a 5% share.

I now reproduce the picture above, but this time looking at total PhDs awarded.

The relative decline of Whites here is even more evident, as is the increasing share of Asians. It is interesting to note that while the share of Hispanics has increased noticeably by including foreign (non PR) recipients, the same is not true for Blacks. One possibility here is that English-speaking foreign black students may be more likely to target the U.K. over the U.S. and that French-speaking blacks may be more likely to target French-speaking institutions in France or former French colonies, like Quebec. (It would be interesting to examine these statistics for Canadian universities). 

Finally, let's take a look at how the share of PhDs across groups lines up with their share of the total population. Here is what the data looks like for the period 2010-2014. 

While White citizens are over-represented, Whites as a whole are under-represented (relative to the domestic U.S. population). Blacks are significantly under-represented both as citizens and including foreigners. Asians, on the other hand, are significantly over-represented--both as citizens and especially if one includes foreigners. Only the "Other" category seems to be roughly representative of the population.

To conclude, there are some clear racial imbalances here. I think most people would agree that increasing Black and Hispanic representation in the U.S. economics profession is a good idea (for many of the reasons highlighted in Eshe's column). Future research into this matter should be informed by the fact that not all minority groups have fared in the same way. It would also be interesting to see how these patterns have evolved in other countries. 

Sunday, December 23, 2018

Does the Fed have a symmetric inflation target?

It's well-known that the Fed has been undershooting its inflation 2% target every year since 2012 (ironically, the year it formally adopted a 2% inflation target). This has led some to speculate whether 2% is being viewed more as a ceiling, rather than a target, as it is with the ECB. The Fed, however, continues to insist that not only is 2% a target, it is a symmetric target.  But what does this mean, exactly? And how can we judge whether the Fed has a symmetric inflation target or not?

These questions came to me while listening to Jay Powell's recent press conference following the FOMC's decision to follow through with a widely anticipated rate hike. At the 16:15 mark, reporter Binyamin Appelbaum (NY Times) asked Powell the following question:
BA: You're about to undershoot your inflation target for the seventh straight year and you forecast that you're going to undershoot it for the eighth straight year...Can you help us to understand why people would be advocating restrictive monetary policy at a time of persistent inflation undershoots? 
Here is how Powell responded:
JP: Well, we as a committee do not desire inflation undershoots and you're right -- inflation has continued to surprise to the downside -- not by a lot though -- I think we're very close to 2% and, you know, we do believe it's a symmetric goal for us -- symmetric around 2% -- and that's how we're going to look at it. We're not trying to be under 2% -- we're trying to be symmetrically around 2% -- and, you know, I've never said that I feel like we've achieved that goal yet. The only way to achieve inflation symmetrically around 2% is to have inflation symmetrically around 2% -- and we've been close to that but we haven't gotten there yet and we haven't declared victory on that yet. So, that remains to be accomplished. 
While this answer sounded reasonable on some level, it did not satisfy the very next inquisitor, Jeanna Smialek (Bloomberg):
JS: Just following up on Binya's question...I guess if you haven't achieved 2% and you don't see an overshoot -- which would sort of be implied by a symmetrical target -- what's the point of raising rates at all? 
Powell replied to this by making reference to the strength of the economy -- growth well above trend, unemployment falling, inflation moving up to 2%, and a positive forecast. In this context, the rate hike seemed appropriate. Again, a sensible sounding answer -- but did it answer the question actually posed?

As I reflected on this exchange, I felt something amiss. And then it occurred to me that people might be mixing up the notion of a symmetric inflation target with a price-level target.

In her question above, Jeanna suggested that if the Fed has a symmetrical inflation, then we should be expecting an overshoot of inflation. But the intentional overshooting of inflation is not inflation targeting -- it is price-level targeting. With an inflation target, one should be expecting inflation to return to the target--not beyond the target.

This would have been a fine answer to Jeanna's question, but isn't it inconsistent with the earlier reply to Binyamin? In that response, Powell left us with the impression that the FOMC has failed to achieve its symmetric inflation goal -- that success along this dimension would consist of actually observing inflation vary symmetrically around 2%. I'm not sure this is entirely correct.

To my way of thinking, an inflation target means getting people to expect that inflation will eventually return to target (from below, if inflation is presently undershooting, and from above, if inflation is presently overshooting). A symmetric inflation target simply means that the rate at which inflation is expected to return to target is the same whether inflation is presently above or below target. To put it another way, symmetry implies that the FOMC should feel equally bad about inflation being 50bp above or below target. Along the same line, persistent inflation overshoots and overshoots should be equally tolerated (given appropriate conditions).
Should a successful symmetric inflation targeting regime generate inflation rates that average around target? It's hard to see how it would not in the long run and if the shocks hitting the economy are themselves symmetric (this is not so obviously a given, but let me set it aside for now). Does missing the inflation target from below for roughly a decade imply that the FOMC has failed to implement a symmetric inflation targeting regime? Powell's mea culpa above suggests yes. But again, I am not so sure.

As I said above, the success of an inflation targeting regime should be measured by how well inflation expectations are anchored around target. By this measure, the FOMC has managed, in my mind, a reasonable level of success (2015-16 looks weak). The following diagram plots the PCE inflation rate (blue) against expected inflation (TIPS breakevens) five years (red) and ten years (green) out.

In my view, the fact that realized inflation has persistently remained below target does not necessarily imply the absence of a symmetric inflation target. Let's take a look at the FOMC's official view on the matter, originally made public on January 24, 2012 in its Statement of Longer-Run Goals and Monetary Policy Strategy. Let me quote the relevant passage and highlight the key phrases:
The Committee reaffirms its judgment that inflation at the rate of 2 percent, as measured by the annual change in the price index for personal consumption expenditures, is most consistent over the longer run with the Federal Reserve’s statutory mandate. The Committee would be concerned if inflation were running persistently above or below this objective. Communicating this symmetric inflation goal clearly to the public helps keep longer-term inflation expectations firmly anchored, thereby fostering price stability...
It seems clear enough that the real goal here is to keep longer-term inflation expectations anchored at 2%.  The idea is that if inflation expectations are anchored in this manner, then the actual inflation rate today shouldn't matter that much for longer-term plans (like investment decisions). If inflation turns out to be low, you should be expecting it to rise. If it turns out to be high, you should be expecting it to fall. Nowhere does the statement suggest we should be expecting under or over shooting -- a characteristic we would associate with a price-level target. As for the phenomenon of persistent under or over shoots, the statement makes clear that the Committee would be equally (symmetrically) concerned in either case.

If one accepts my definition of symmetric inflation target then, unfortunately, we do not yet have enough data to judge whether the Fed's inflation target is symmetric. The policy was only formally implemented in 2012. Since then we've only observed a persistent undershoot and the conditions leading to these persistent downward surprises. Would the FOMC be equally tolerant of letting inflation surprise to the upside for several years should economic conditions warrant? It seems that we'll have to wait and see.

Thursday, December 6, 2018

Working More for Less

I had an interested chat with a colleague of mine the other day about the labor market. In the course of conversation, he mentioned that he used to teach a class in labor economics. Naturally, an important lesson included the theory of labor supply. Pretty much the first question asked is how the supply of labor can be expected to change in response to a change in the return to labor (the real wage). 

My colleague said that for years he would preface the theoretical discussion with a poll. He would turn to the class and ask them to imagine themselves employed at some job. Then imagine having your wage doubled for a short period of time. How many of you would work more? (The majority of the class would raise their hands.) How many of you would not change your hours worked? (A minority of hands). How many of you would work less? (A sprinkling of hands). At the end of the polling, he'd start teaching a standard theory of labor supply and using it to interpret the poll results (substitution vs. wealth effects).

My colleague administered this poll for over a decade. The results were always the same. (How satisfying.)
Then, one day, for no apparent reason, he decided to mix it up a little bit. Instead of asking the class to imagine an increase in the wage rate, he asked his students to consider a decrease in their wage rate. He was expecting a symmetrically opposite response. To his shock, a majority of the class responded that they would work more. Only a minority replied that they would work less or not change their hours.

Surely, this was an anomaly? But when he repeated the experiment with another class, he got the same result. He mentioned it to a colleague of his, who then ran the same experiment with his class and he too confirmed the result. What was going on here? If true, then employers can apparently get more labor out their workers by lowering their wages?!
The phenomenon here seems related to the evidence of "income targeting" among some groups of workers; see, for example, the classic study of New York taxi drivers by Camerer, Babock, Lowenstein and Thaler (QJE May, 1997). Evidently, inexperienced taxi drivers tend to work less when the return to working is high, and work more when the return to working is low. This behavior doesn't quite square with the phenomenon reported by my colleague. The effect there appeared to be asymmetric: students reported willing working more at a lower wage, but also reported willing working more at a higher wage. In both cases, however, it seems that the existence of some fixed obligation (e.g., monthly food and rent payments) plus no ready access to credit could explain why workers might be willing to work longer hours when the return to work declines.

I'm not sure if these findings shed any light on the state of the labor market today. But it is interesting to speculate. Conventional supply/demand analysis isn't always the best guide. 

Monday, December 3, 2018

Does the Floor System Discourage Bank Lending?

David Beckworth has a new post up suggesting that the Fed's floor system has discouraged bank lending by making interest-bearing reserves a relatively more attractive investment; see here. I've been hearing this story a lot lately, but I can't say it makes a whole lot of sense to me.

Here's how I think about it. Consider the pre-2008 "corridor" system where the Fed targeted the federal funds rate. The effective federal funds rate (FFR) traded between the upper and lower bounds of the corridor--the upper bound given by the discount rate and the lower bound given by the zero interest-on-reserves (IOR) rate. The Fed achieved its target FFR by managing the supply of reserves through open-market operations involving short-term treasury debt.

Consider a given target interest rate equal to (say) 4%. Since the Fed is financing its asset holdings (USTs yielding 4%) with 0% reserves, it is making a profit on the spread, which it remits to the treasury. Another way of looking at this is that the treasury has saved a 4% interest expense on that part of its debt purchased by the Fed (the treasury would have had to find some additional funds to pay for that interest expense had it not been purchased by the Fed).

Now, suppose that the Fed wants to achieve its target interest rate by paying 4% on reserves. The supply of reserves need not change. The yield on USTs need not change. Bank lending need not change. The only thing that changes is that the Fed now incurs an interest expense of 4% on reserves. The Fed's profit in this case go to zero and the remittances to the treasury are reduced accordingly. From the treasury's perspective, it may as well have sold the treasuries bought by the Fed to the private sector instead.

But the question here is why one would think that moving from a corridor system to a floor system with interest-bearing reserves inherently discourages bank lending. It is true that bank lending is discouraged by raising the IOR rate. But is it not discouraged in exactly the same way by an equivalent increase in the FFR? If I am reading the critics correctly (and I may not be), the complaint seems to be more with where the policy rate is set, as opposed to anything inherent in the operating system. If the complaint is that the IOR has been set too high, I'm willing to agree. But I would have had the same complaint had the FFR been set too high under the old corridor system.

Alright, now let's take a look at some of the data presented by David. Here, I replicate his Panel A depicting the evolution of the composition of bank assets.
David wants to direct our attention to the period after 2008 when the Fed flooded the banking system with reserves and started paying a positive IOR rate. The large rise in the orange line since 2008 was due almost entirely to reserves and not other safe assets. This suggests that banks were motivated to hold interest-bearing reserves instead of private-sector interest-bearing assets (loans). He writes:
Something big happened in 2008 that continues to the present that caused banks to allocate more of their portfolios to cash assets and less to loans. While the financial crisis surely was a part of the initial rebalancing, it is hard to attribute what appears to be 10-year structural change to the crisis alone. Instead, it seems more consistent with the critics view that the floor system itself has fundamentally changed bank portfolios allocation.
I think the diagram above is rather misleading since all it shows is portfolio composition and not the level of bank lending. Here's what the picture looks like when we take the same data and deflate it by the GDP instead of bank assets,

According to this picture, bank lending is close to 50% of GDP, not far off its historical average and considerably higher than in the decade following the S&L crisis (1986-1995). Here's what commercial and industrial loans as a ratio of GDP looks like:
It's no surprise that bank lending contracted during and shortly after the crisis. One could even make the argument that paying positive IOR contributed to the contraction. But as I mentioned above, one could have made the same argument had the FFR been kept at 25bp. Again, this criticism has less to do with the operating system than it does with where the policy rate was set. In any case, note that commercial and industrial loans are presently above their pre-crisis levels (as a ratio of GDP). 

To sum up, I do not believe that a floor system inherently discourages bank lending as some critics appear to be arguing. Now that the Fed is paying IOR, reserves are essentially viewed by banks as an alternative form of interest-bearing government debt. New regulations since the crisis have induced banks to load up on safe government assets. But as the following figure shows, this has not come at the expense of private lending.
Banks are lending about as much as they have over the past 50 years (relative to GDP). Bank lending as a ratio of bank assets may be low, but this is because banks are loaded up on safe assets--not because they've cut back on their lending activity.

Thursday, November 8, 2018

Smart Contracts and Asset Tokenization

Book of Smart Contracts 1959
In his 1959 classic Theory of Value, Gerard Debreu takes a deep dive into general (Walrasian) equilibrium theory. (Yes, I know, but please try to stay awake for at least a few more paragraphs.)

He studies a very stark hypothetical scenario where people are imagined to gather at the beginning of time and formulate trading plans for a given vector of market prices (called out by some mysterious auctioneer). Commodities can take the form of different goods, like apples and oranges. But they can also be made time-contingent and state-contingent. An apple delivered tomorrow is different commodity than an apple delivered today. An orange delivered tomorrow in the event of rain is different commodity than an orange delivered tomorrow in the event of sunshine. And so on.

For any given vector of relative prices (there is no money), individuals offer to sell claims against the commodities they own to acquire claims against the commodities they wish to acquire. A market-clearing price vector is one that makes everyone's desired trades consistent with each other. How this equilibrium price-vector is achieved is not studied--he is mainly concerned with the less interesting, but still important, question of whether any such price vector might even be expected to exist in the first place.

The theory imagines all relevant trading activity to take place once-and-for-all at the beginning of time. Once trading positions are agreed to, all subsequent good and service flows across individuals over time and under different contingencies are dictated by the terms of promises made at the initial auction. Suppose I had earlier acquired the right for the delivery of oranges next month in the event of rain. Suppose it rains next month. Then the delivery of oranges is made by the orange producer who issued the promissory note now in my possession. In short, contracts look very "smart" in the sense that they can be tailored in any way we want and, moreover, they are assumed to be "self-executing." It's almost as if contractual terms have been spelled out mathematically and enforced by self-executing computer code. Indeed, this is essentially what Debreu assumes.

The Debreu model (also associated with Ken Arrow and Lionel MacKenzie) is often viewed as a sort of benchmark of what one might expect if auction markets are "complete" and worked perfectly (no financial market frictions like asymmetric information, limited commitment, limited communications, etc.) There is no role for money as a medium of exchange in such a frictionless world. As such, it should come as no surprise to learn that monetary theory is devoted to studying economies where these frictions play a prominent role. Financial institutions (governance structures in general, including "the government") can to a large extent be understood as collective arrangements that are designed (or have evolved) to mitigate these frictions for the economic benefit of a given set of constituents (either general or special interests, depending on the distribution of political power).

A recurring theme of the "blockchain" movement is how this new record-keeping technology may one day permit us to decentralize all economic activity. No more (government) money. No more banks. No more intermediaries of any sort. This seems to be, at least in part, what "asset tokenization" is about; see, for example, here: How Tokenization Is Putting Real-World Assets on Blockchains. According to this article,
Tokenization is the process of converting rights to an asset into a digital token on a blockchain. 
This sounds fancy, but as the article soon makes clear, it's basically a variation of an old theme,
There are many proposed methods for taking real-world assets and "putting them on a blockchain." The goal is to achieve the security, speed and ease of transfer of Bitcoin, combined with real-world assets. This is a new form of an old concept: "securitization" (turning a set of assets into a security), and in some cases the tokenization is of securitized assets.
Here's how the innovation is supposed to help small investors (source):
Imagine that you have some property — say an apartment. You need cash quickly. The apartment is valued at $150,000 but you just need $10,000. Can you do this quickly without much friction? To my best knowledge, this is next to impossible.
I often use a similar example in my monetary theory classes. How to liquidate a fraction of one's illiquid wealth? One way is to use a bank (say, to open up a credit line secured by your property). But what he means, I think, is that it's basically impossible to issue a personal IOU representing a claim against the property (and ultimately, against the income that is generated by that property). Well, it's possible, but any such security is not likely to be marketable at any reasonable price. The author has stumbled across the concept of an "illiquid" asset. We use institutions called banks to monetize illiquid assets (banks transform illiquid assets into liquid deposit liabilities). But why do we need banks? Why are most assets illiquid? Economic theory answers: because of the frictions associated with asymmetric information and limited commitment (or lack of trust). O.K., but is there any way to get around these frictions without the use of banks? The same article continues:
Enter tokenization. Tokenization is a method that converts rights to an asset into a digital token. Suppose there is a $200,000 apartment. Tokenization can transform this apartment into 200,000 tokens (the number is totally arbitrary, we could have issued 2 million tokens). Thus, each token represents a 0.0005% share of the underlying asset. Finally, we issue the token on some sort of a platform supporting smart contracts, for example on Ethereum, so that the tokens can be freely bought and sold on different exchanges. When you buy one token, you actually buy 0.0005% of the ownership in the asset. Buy 100,000 tokens and you own 50% of the assets. Buy all 200,000 tokens and you are 100% owner of the asset. Obviously, you are not becoming a legal owner of the property. However, because Blockchain is a public ledger that is immutable, it ensures that once you buy tokens, nobody can “erase” your ownership even if it is not registered in a government-run registry. It should be clear now why Blockchain enables this type of services.
Well, no, to be honest it is not at all clear how "blockchain" solves any of the fundamental problems associated with transforming an illiquid asset into a payment instrument.

We have to keep in mind that "blockchain" is nothing more than a consensus-based database management system (where the data is organized and secured in a particular way). Moreover, any useful innovation found in a blockchain-based database management system (recording data as a Merkle tree, for example) could likely be applied in a non-consensus-based database management system. It's one thing to transfer tokens (or information) across accounts in a database. It's quite another thing to exert your own effort to evict the non-compliant tenant of your 0.0005% share of the apartment you own, especially if other owners are not on board.

It may be that technology will one day eliminate financial market "frictions" and permit widespread asset tokenization (including our human capital), all of which will be traded using smart contracts on an Internet-based auction. If or when that day comes, the people of that world can refer to Debreu (1959) as an economic model applicable to that future world.