From Senior Thesis to Publication

One of the most challenging and rewarding aspects of my job at Haverford College is advising senior theses. At Haverford, every student in every major writes a senior thesis (or equivalent capstone project). I co-teach the senior thesis course in the fall with two or three colleagues, and advise thesis writers in the spring. In the fall, students work on choosing a topic and research question, conducting a literature review, and developing a research proposal. In the spring, they work one-on-one with an advisor to see the proposal through.

Every completed thesis is an achievement, and we celebrate all of our seniors with a champagne reception on the due date. But today, for the first time, one of my advisee's theses has successfully made its way through the peer review process and is published!

Samantha Wetzel, Haverford College class of 2018, excelled both as an economics student and on the basketball team. We published "The FOMC versus the staff, revisited: When do policymakers add value?" in Economics Letters. (This link should provide temporary free access to the published version; here is the SSRN working paper version.)

The thesis, and this letter based on it, follow up on Romer and Romer's (2008) similarly-titled paper. Here is the abstract:

The Board of Governors staff and the Federal Open Market Committee both publish macroeconomic forecasts. Romer and Romer (2008) show that policymakers' attempts to add information to the staff forecasts are counterproductive. In more recent years, however, policymakers have improved upon staff forecasts. We show that policymakers' value-added is greater when economic conditions are unfavorable or uncertain.

For other undergraduate researchers hoping to write a successful, potentially publishable thesis, I have a few bits of advice based on this advising experience. First, start early if possible. Samantha came up with her thesis topic during junior year, when she read Romer and Romer (2008) in my course on the Federal Reserve.

Second, ask a well-defined question. This may be the hardest part. Your thesis (like Sam's) will probably be much longer than the 2000 word limit of Economics Letters, with a longer literature review and more robustness checks etc., but you should be able to easily explain what you asked, what you found, and how you found it in a few pages. Your thesis is far more likely to be successful if you have a primary hypothesis that you can precisely state. (Even if you think you can state it, write it down to be sure!)

Third, come up with a plan so that you can make progress each week-- use your advisor to help you decide on a timeline for key goals and to hold you accountable. For student athletes like Samantha, this means thinking in advance about when key games and travel will be, and planning accordingly.

Rethinking Macroeconomic Policy

I had the pleasure of attending “Rethinking Macroeconomic Policy IV” at the Peterson Institute for International Economics. I highly recommend viewing the panels and materials online.

The two-day conference left me wondering what it actually means to “rethink” macro. The conference title refers to rethinking macroeconomic policy, not macroeconomic research or analysis, but of course these are related. Adam Posen’s opening remarks expressed dissatisfaction with DSGE models, VARs, and the like, and these sentiments were occasionally echoed in the other panels in the context of the potentially large role of nonlinearities in economic dynamics. Then, in the opening session, Olivier Blanchard talked about whether we need a “revolution” or “evolution” in macroeconomic thought. He leans toward the latter, while his coauthor Larry Summers leans toward the former. But what could either of these look like? How could we replace or transform the existing modes of analysis?

I looked back on the materials from Rethinking Macroeconomic Policy of 2010. Many of the policy challenges discussed at that conference are still among the biggest challenges today. For example, low inflation and low nominal interest rates limit the scope of monetary policy in recessions. In 2010, raising the inflation target and strengthening automatic fiscal stabilizers were both suggested as possible policy solutions meriting further research and discussion. Inflation and nominal rates are still very low seven years later, and higher inflation targets and stronger automatic stabilizers are still discussed, but what I don’t see is a serious proposal for change in the way we evaluate these policy proposals.

Plenty of papers use basically standard macro models and simulations to quantify the costs and benefits of raising the inflation target. Should we care? Should we discard them and rely solely on intuition? I’d say: probably yes, and probably no. Will we (academics and think tankers) ever feel confident enough in these results to make a real policy change? Maybe, but then it might not be up to us.

Ben Bernanke raised probably the most specific and novel policy idea of the conference, a monetary policy framework that would resemble a hybrid of inflation targeting and price level targeting. In normal times, the central bank would have a 2% inflation target. At the zero lower bound, the central bank would allow inflation to rise above the 2% target until inflation over the duration of the ZLB episode averaged 2%. He suggested that this framework would have some of the benefits of a higher inflation target and of price level targeting without some of the associated costs. Inflation would average 2%, so distortions from higher inflation associated with a 4% target would be avoided. The possibly adverse credibility costs of switching to a higher target would also be minimized. The policy would provide the usual benefits of history-dependence associated with price level targeting, without the problems that this poses when there are oil shocks.

It’s an exciting idea, and intuitively really appealing to me. But how should the Fed ever decide whether or not to implement it? Bernanke mentioned that economists at the Board are working on simulations of this policy. I would guess that these simulations involve many of the assumptions and linearizations that rethinking types love to demonize. So again: Should we care? Should we rely solely on intuition and verbal reasoning? What else is there?

Later, Jason Furman presented a paper titled, “Should policymakers care whether inequality is helpful or harmful for growth?” He discussed some examples of evaluating tradeoffs between output and distribution in toy models of tax reform. He begins with the Mankiw and Weinzierl (2006) example of a 10 percent reduction in labor taxes paid for by a lump-sum tax. In a Ramsey model with a representative agent, this policy change would raise output by 1 percent. Replacing the representative agent with agents with the actual 2010 distribution of U.S. incomes, only 46 percent of households would see their after-tax income increase and 41 percent would see their welfare increase. More generally, he claims that “the growth effects of tax changes are about an order of magnitude smaller than the distributional effects of tax changes—and the disparity between the welfare and distribution effects is even larger” (14). He concludes:

“a welfarist analyzing tax policies that entail tradeoffs between efficiency and equity would not be far off in just looking at static distribution tables and ignoring any dynamic effects altogether. This is true for just about any social welfare function that places a greater weight on absolute gains for households at the bottom than at the top. Under such an approach policymaking could still be done under a lexicographic process—so two tax plans with the same distribution would be evaluated on the basis of whichever had higher growth rates…but in this case growth would be the last consideration, not the first” (16).

As Posen then pointed out, Furman’s paper and his discussants largely ignored the discussions of macroeconomic stabilization and business cycles that dominated the previous sessions on monetary and fiscal policy. The panelists acceded that recessions, and hysteresis in unemployment, can exacerbate economic disparities. But the fact that stabilization policy was so disconnected from the initial discussion of inequality and growth shows just how much rethinking still has not occurred.

In 1987, Robert Lucas calculated that the welfare costs of business cycles are minimal. In some sense, we have “rethought” this finding. We know that it is built on assumptions of a representative agent and no hysteresis, among other things. And given the emphasis in the fiscal and monetary policy sessions on avoiding or minimizing business cycle fluctuations, clearly we believe that the costs of business cycle fluctuations are in fact quite large. I doubt many economists would agree with the statement that “the welfare costs of business cycles are minimal.” Yet, the public finance literature, even as presented at a conference on rethinking macroeconomic policy, still evaluates welfare effects of policy using models that totally omit business cycle fluctuations, because, within those models, such fluctuations hardly matter for welfare. If we believe that the models are “wrong” in their implications for the welfare effects of fluctuations, why are we willing to take their implications for the welfare effects of tax policies at face value?

I don’t have a good alternative—but if there is a Rethinking Macroeconomic Policy V, I hope some will be suggested. The fact that the conference speakers are so distinguished is both an upside and a downside. They have the greatest understanding of our current models and policies, and in many cases were central to developing them. They can rethink, because they have already thought, and moreover, they have large influence and loud platforms. But they are also quite invested in the status quo, for all they might criticize it, in a way that may prevent really radical rethinking (if it is really needed, which I’m not yet convinced of). (A more minor personal downside is that I was asked multiple times whether I was an intern.)

If there is a Rethinking Macroeconomic Policy V, I also hope that there will be a session on teaching and training. The real rethinking is going to come from the next generations of economists. How do we help them learn and benefit from the current state of economic knowledge without being constrained by it? This session could also touch on continuing education for current economists. What kinds of skills should we be trying to develop now? What interdisciplinary overtures should we be making?

Consumer Forecast Revisions: Is Information Really so Sticky?

My paper "Consumer Forecast Revisions: Is Information Really so Sticky?" was just accepted for publication in Economics Letters. This is a short paper that I believe makes an important point. 

Sticky information models are one way of modeling imperfect information. In these models, only a fraction (λ) of agents update their information sets each period. If λ is low, information is quite sticky, and that can have important implications for macroeconomic dynamics. There have been several empirical approaches to estimating λ. With micro-level survey data, a non-parametric and time-varying estimate of λ can be obtained by calculating the fraction of respondents who revise their forecasts (say, for inflation) at each survey date. Estimates from the Michigan Survey of Consumers (MSC) imply that consumers update their information about inflation approximately once every 8 months.

Here are two issues that I point out with these estimates:

I show that several issues with estimates of information stickiness based on consumer survey microdata lead to substantial underestimation of the frequency with which consumers update their expectations. The first issue stems from data frequency. The rotating panel of Michigan Survey of Consumer (MSC) respondents take the survey twice with a six-month gap. A consumer may have the same forecast at months t and t+ 6 but different forecasts in between. The second issue is that responses are reported to the nearest integer. A consumer may update her information, but if the update results in a sufficiently small revisions, it will appear that she has not updated her information. 

To quantify how these issues matter, I use data from the New York Fed Survey of Consumer Expectations, which is available monthly and not rounded to the nearest integer. I compute updating frequency with this data. It is very high-- at least 5 revisions in 8 months, as opposed to the 1 revision per 8 months found in previous literature.

Then I transform the data so that it is like the MSC data. First I round the responses to the nearest integer. This makes the updating frequency estimates decrease a little. Then I look at it at the six-month frequency instead of monthly. This makes the updating frequency estimates decrease a lot, and I find similar estimates to the previous literature-- updates about every 8 months.

So low-frequency data, and, to a lesser extent, rounded responses, result in large underestimates of revision frequency (or equivalently, overestimates of information stickiness). And if information is not really so sticky, then sticky information models may not be as good at explaining aggregate dynamics. Other classes of imperfect information models, or sticky information models combined with other classes of models, might be better.

Read the ungated version here. I will post a link to the official version when it is published.

The Low Misery Dilemma

The other day, Tim Duy tweeted:

I expect the FOMC minutes to reveal that some participants were concerned about low inflation while others focused on low unemployment.

— Tim Duy (@TimDuy) August 16, 2017

It took me a moment--and I'd guess I'm not alone--to even recognize how remarkable this is. The New York Times ran an article with the headline "Fed Officials Confront New Reality: Low Inflation and Low Unemployment." Confront, not embrace, not celebrate.

The misery index is the sum of unemployment and inflation. Arthur Okun proposed it in the 1960s as a crude gauge of the economy, based on the fact that high inflation and high unemployment are both miserable (so high values of the index are bad). The misery index was pretty low in the 60s, in the 6% to 8% range, similar to where it has been since around 2014. Now it is around 6%. Great, right?

The NYT article notes that we are in an opposite situation to the stagflation of the 1970s and early 80s, when both high inflation and high unemployment were concerns. The misery index reached a high of 21% in 1980. (The unemployment data is only available since 1948).

Very high inflation and high unemployment are each individually troubling for the social welfare costs they impose (which are more obvious for unemployment). But observed together, they also troubled economists for seeming to run contrary to the Phillips curve-based models of the time. The tradeoff between inflation and unemployment wasn't what economists and policymakers had believed, and their misunderstanding probably contributed to the misery.

Though economic theory has evolved, the basic Phillips curve tradeoff idea is still an important part of central bankers' models. By models, I mean both the formal quantitative models used by their staffs and the way they think about how the world works. General idea: if the economy is above full employment, that should put upward pressure on wages, which should put upward pressure on prices.

So low unemployment combined with low inflation seem like a nice problem to have, but if they are indeed a new reality-- that is, something that will last--then there is something amiss in that chain of logic. Maybe we are not at full employment, because the natural rate of unemployment is a lot lower than we thought, or we are looking at the wrong labor market indicators. Maybe full employment does not put upward pressure on wages, for some reason, or maybe we are looking at the wrong wage measures. For example, San Francisco Fed researchers argue that wage growth measures should be adjusted in light of retiring Baby Boomers. Or maybe the link between wage and price inflation has weakened.

Until policymakers feel confident that they understand why we are experiencing both low inflation and low unemployment, they can't simply embrace the low misery. It is natural that they will worry that they are missing something, and that the consequences of whatever that is could be disastrous. The question is what to do in the meanwhile.

There are two camps for Fed policy. One camp favors a wait-and-see approach: hold rates steady until we actually observe inflation rising above 2%. Maybe even let it stay above 2% for awhile, to make up for the lengthy period of below-2% inflation. The other camp favors raising rates preemptively, just in case we are missing some sign that inflation is about to spiral out of control. This latter possibility strikes me as unlikely, but I'm admittedly oversimplifying the concerns, and also haven't personally experienced high inflation.

Macro in the Econ Major and at Liberal Arts Colleges

Last week, I attended the 13th annual Conference of Macroeconomists from Liberal Arts Colleges, hosted this year by Davidson College. I also attended the conference two years ago at Union College. I can't recommend this conference strongly enough!

The conference is a response to the increasing expectation of high quality research at many liberal arts colleges. Many of us are the only macroeconomist at our college, and can't regularly attend macro seminars, so the conference is a much-needed opportunity to receive feedback on work in progress. (The paper I presented last time just came out in the Journal of Monetary Economics!)

This time, I presented "Inflation Expectations and the Price at the Pump" and discussed Erin Wolcott's paper, "Impact of Foreign Official Purchases of U.S.Treasuries on the Yield Curve."

There was a wide range of interesting work. For example, Gina Pieters presented “Bitcoin Reveals Unofficial Exchange Rates and Detects Capital Controls.” M. Saif Mehkari's work on “Repatriation Taxes” is highly relevant to today's policy discussions. Most of the presenters and attendees were junior faculty members, but three more senior scholars held a panel discussion at dinner. Next year, the conference will be held at Wake Forest.

I also attended a session on "Macro in the Econ Major" led by PJ Glandon. A link to his slides is here. One slide presented the image below, prompting an interesting discussion about whether and how we should tailor what is taught in macro courses to our perception of the students' interests and career goals.

Labor Market Conditions Index Discontinued

A few years ago, I blogged about the Fed's new Labor Market Conditions Index (LMCI). The index attempts to summarize the state of the labor market using a statistical technique that captures the primary common variation from 19 labor market indicators. I was skeptical about the usefulness of the LMCI for a few reasons. And as it turns out, the LMCI is now discontinued as of August 3.

The discontinuation is newsworthy because the LMCI was cited in policy discussions at the Fed, even by Janet Yellen. The index became high-profile enough that I was even interviewed about it on NPR's Marketplace.

One issue that I noted with the index in my blog was the following:

A minor quibble with the index is its inclusion of wages in the list of indicators. This introduces endogeneity that makes it unsuitable for use in Phillips Curve-type estimations of the relationship between labor market conditions and wages or inflation. In other words, we can't attempt to estimate how wages depend on labor market tightness if our measure of labor market tightness already depends on wages by construction.

This corresponds to one reason that is provided for the discontinuation of the index: "including average hourly earnings as an indicator did not provide a meaningful link between labor market conditions and wage growth."

The other reasons provided for discontinuation are that "model estimates turned out to be more sensitive to the detrending procedure than we had expected" and "the measurement of some indicators in recent years has changed in ways that significantly degraded their signal content."

I also noted in my blog post and on NPR that the index is almost perfectly correlated with the unemployment rate, meaning it provides very little additional information about labor market conditions. (Or interpreted differently, meaning that the unemployment rate provides a lot of information about labor market conditions.) The development of the LMCI was part of a worthy effort to develop alternative informative measures of labor market conditions that can help policymakers gauge where we are relative to full employment and predict what is likely to happen to prices and wages. So since resources and attention are limited, I think it is wise that they can be directed toward developing and evaluating other measures. 

Thoughts on Angrist and Pishke's "Undergraduate Econometrics Instruction"

Joshua Angrist and Jörn-Steffen Pischke, coauthors of "Mastering 'Metrics," have just released a new NBER working paper called "Undergraduate Econometrics Instruction: Through Our Classes, Darkly." They argue that pedagogy has not kept pace with trends in economic research in the past few decades:

In the 1960s and 1970s, an empirical economist’s typical mission was to “explain” economic variables like wages or GDP growth. Applied econometrics has since evolved to prioritize the estimation of specific causal effects and empirical policy analysis over general models of outcome determination. Yet econometric instruction remains mostly abstract, focusing on the search for “true models” and technical concerns associated with classical regression assumptions. Questions of research design and causality still take a back seat in the classroom, in spite of having risen to the top of the modern empirical agenda. This essay traces the divergent development of econometric teaching and empirical practice, arguing for a pedagogical paradigm shift.

The "pedagogical paradigm shift" they call for would include three main components:

One is a focus on causal questions and empirical examples, rather than models and math. Another is a revision of the anachronistic classical regression framework, away from explaining economic processes and towards controlled statistical comparisons. The third is an emphasis on modern quasiexperimental tools.  

Since I am relatively new to both teaching and economics-- I didn't major in economics as an undergraduate, and did my Ph.D. from 2010 to 2015-- the first economics course that I designed and taught at Haverford quite naturally adhered to many of Angrist and Pischke's recommendations. The course, which I taught in Fall 2015 and Fall 2016, is called Advanced Macroeconomics, but is essentially an applied econometrics course on empirical macroeconomic policy analysis. The students in the course are typically juniors and seniors who have already taken econometrics.

On the first day of class, we read excerpts from the 1968 paper "Monetary and Fiscal Actions: A Test of Their Relative Importance in Economic Stabilization" by Andersen and Jordan. The authors want to test whether "the response of economic activity to fiscal actions relative to that of monetary actions is (1) greater, (2) more predictable, and (3) faster." They use very simple regression analysis, essentially regressing changes in GNP on changes in measures of monetary and fiscal actions. This type of regression is now called a "St. Louis Equation," since Andersen and Jordan were at the St. Louis Fed. I ask my students to interpret the regression results and evaluate the validity of the authors' conclusions about policy effectiveness. With some prodding, the students come up with some ideas about potential omitted variable bias and data concerns. But they don't think about reverse causality or the idea of a "controlled statistical comparison." I introduce the reverse causality issue, and much of the rest of the course focuses on quasiexperimental tools.

The course has no textbook, but we use "Natural Experiments in Macroeconomics" by Nicola Fuchs-Schundeln and Tarek Hassan as the main reference. The course has four units: consumption, monetary policy, fiscal policy, and growth and distribution. In each unit, I assign natural experiment or quasiexperimental papers as well as other papers that attempt to achieve identification via other means, to varying degrees of success. The reading list was influenced by Christina Romer and David Romer's graduate course on Macroeconomic History at Berkeley, which introduced me to the notion of identification and ignited my interest in macroeconomics.

Angrist and Pischke also argue that "Regression should be taught the way it’s now most often used: as a tool to control for confounding factors" in contrast to "the traditional regression framework in which all regressors are treated equally." In other words, the coefficient of interest is on one of the regressors, while the other regressors serve as "control variables needed to insure that the regression-estimated effect of the variable of interest has a causal interpretation."

This advice on teaching regression resonates with my experience co-teaching the economics senior thesis seminar at Haverford for the past two years. Over the summer, my research assistant Alex Rodrigue read through several years' worth of senior theses in the archives and documented the research question in each thesis. We noticed that many students use research questions of the form "What are the factors that affect Y?" and run a regression of Y on all the variables they can think of, treating all regressors equally and not attempting to investigate any particular causal relationship from one variable X to Y. The more successful theses posit a causal relationship from X to Y driven by specific economic mechanisms, then use regression analysis and other methods to estimate and interpret the effect. The latter type of thesis has more pedagogical benefits, whether or not the student can ultimately achieve convincing identification, because it leads the student to think more seriously about economic mechanisms.

Why are Long-Run Inflation Expectations Falling?

Randal Verbrugge and I have just published a Federal Reserve Bank of Cleveland Economic Commentary called "Digging into the Downward Trend in Consumer Inflation Expectations." The piece focuses on long-run inflation expectations--expectations for the next 5 to 10 years-- from the Michigan Survey of Consumers. These expectations have been trending downward since the summer of 2014, around the same time as oil and gas prices started to decline.  It might seem natural to conclude that falling gas prices are responsible for the decline in long-run inflation expectations. But we suggest that this may not be the whole story.

First of all, gas prices have exhibited two upward surges since 2014, neither of which was associated with a rise in long-run inflation expectations. Second, the correlation between gas prices and inflation expectations (a relationship I explore in much more detail in this working paper) appears too weak to explain the size of the decline. So what else could be going on?

If you look at the histogram in Figure 2, below, you can see the distribution of inflation forecasts that consumers give in three different time periods: an early period, the first half of 2014, and the past year. The shaded gray bars correspond to the early period, the red bars to 2014, and the blue bars to the most recent period. Notice that there is some degree of "response heaping" at multiples of 5%. In another paper, I use this response heaping to help quantify consumers' uncertainty about long-run inflation. The idea is that people who are more uncertain about inflation, or have a less precise estimate of what it should be, tend to report a round number-- this is a well-documented tendency in how people communicate imprecision.

The response heaping has declined over time, corresponding to a fall in my consumer inflation uncertainty index for the longer horizon. As we detail in the Commentary, this fall in uncertainty helps explain the decline in the measured median inflation forecast. This is a remnant of the fact that common round forecasts, 5% and 10%, are higher than common non-round forecasts.

There is also a notable change in the distribution of non-round forecasts over time. The biggest change is that 1% forecasts for long-run inflation are much more common than previously (see how the blue bar is higher than the red and gray bars for 1% inflation).  I think this is an important sign that some consumers (probably those that are more informed about the economy and inflation) are noticing that inflation has been quite low for an extended period, and are starting to incorporate low inflation into their long-run expectations. More consumers expect 1% inflation than 2%.

More Support for a Higher Inflation Target

Ever since the FOMC announcement in 2012 that 2% PCE inflation is consistent with the Fed's price stability mandate, economists have questioned whether the 2% target is optimal. In 2013, for example, Laurence Ball made the case for a 4% target. Two new NBER working papers out this week each approach the topic of the optimal inflation target from different angles. Both, I think, can be interpreted as supportive of a somewhat higher target-- or at least of the idea that moderately higher inflation has greater benefits and smaller costs than conventionally believed.

The first, by Marc Dordal-i-Carreras, Olivier Coibion, Yuriy Gorodnichenko, and Johannes Wieland, is called "Infrequent but Long-Lived Zero-Bound Episodes and the Optimal Rate of Inflation." One benefit of a higher inflation target is to reduce the occurrence of zero lower bound (ZLB) episodes, so understanding the welfare costs of these episodes is important in calculating an optimal inflation target. The authors explain that in standard models with a ZLB, normally-distributed shocks result in short-lived ZLB episodes. This is in contrast with the reality of frequent but long-lived ZLB episodes. They build models that can generate long-lived ZLB episodes and show that welfare costs of ZLB episodes increase steeply with duration; 8 successive quarters at the ZLB is costlier than two separate 4-quarter episodes.

If ZLB episodes are costlier, it makes sense to have a higher inflation target to reduce their frequency. The authors note, however, that the estimate of the optimal target implied by their models are very sensitive to modeling assumptions and calibration:

"We find that depending on our calibration of the average duration and the unconditional frequency of ZLB episodes, the optimal inflation rate can range from 1.5% to 4%. This uncertainty stems ultimately from the paucity of historical experience with ZLB episodes, which makes pinning down these parameters with any degree of confidence very difficult. A key conclusion of the paper is therefore that much humility is called for when making recommendations about the optimal rate of inflation since this fundamental data constraint is unlikely to be relaxed anytime soon."

The second paper, by Emi Nakamura, Jón Steinsson, Patrick Sun, and Daniel Villar, is called "The Elusive Costs of Inflation: Price Dispersion during the U.S. Great Inflation." This paper notes that in standard New Keynesian models with Calvo pricing, one of the main welfare costs of inflation comes from inefficient price dispersion. When inflation is high, prices get further from optimal between price resets. This distorts the allocative role of prices, as relative prices no longer accurately reflect relative costs of production. In a standard New Keynesian model, the implied cost of this reduction in production efficiency is about 10% if you move from 0% inflation to 12% inflation. This is huge-- an order of magnitude greater than the welfare costs of business cycle fluctuations in output. This is why standard models recommend a very low inflation target.

Empirical evidence of inefficient price dispersion is sparse, since there is relatively minimal fluctuation in inflation in the past few decades, when BLS microdata on consumer prices is available. Nakamura et al. undertook the arduous task of extending the BLS microdataset back to 1977, encompassing higher-inflation episodes. Calculating price dispersion within a category of goods can be problematic, because price dispersion may arise from differences in quality or features of the goods. The authors instead look at the absolute size of price changes, explaining, "Intuitively, if inflation leads prices to drift further away from their optimal level, we should see prices adjusting by larger amounts when they adjust. The absolute size of price adjustments should reveal how far away from optimal the adjusting prices had become before they were adjusted. The absolute size of price adjustment should therefore be highly informative about inefficient price dispersion."

They find that the mean absolute size of price changes is fairly constant from 1977 to the present, and conclude that "There is, thus, no evidence that prices deviated more from their optimal level during the Great Inflation period when inflation was running at higher than 10% per year than during the more recent period when inflation has been close to 2% per year. We conclude from this that the main costs of inflation in the New Keynesian model are completely elusive in the data. This implies that the strong conclusions about optimality of low inflation rates reached by researchers using models of this kind need to be reassessed."

Whose Expectations Augment the Phillips Curve?

My first economics journal publication is now available online in Economics Letters. This link provides free access until November 9, 2015. It is a brief piece (hence the letter format) titled "Whose Expectations Augment the Phillips Curve?" The short answer:

"The inflation expectations of high-income, college-educated, male, and working-age people play a larger role in inflation dynamics than do the expectations of other groups of consumers or of professional forecasters."

Update: The permanent link to the paper is here.

False Discoveries and the ROC Curves of Social Science

Diagnostic tests for diseases can suffer from two types of errors. A type I error is a false positive, and a type II error is a false negative. The sensitivity or true positive rate is the probability that a test result will be positive when the disease is actually present. The specificity or true negative rate is the probability that a test result will be negative when the disease is not actually present. Different choices of diagnostic criteria correspond to different combinations of sensitivity and specificity. A more sensitive diagnostic test could reduce false negatives, but might increase the false positive rate. Receiver operating characteristic (ROC) curves are a way to visually present this tradeoff by plotting true positive rates or sensitivity on the y-axis and false positive rates (100%-specificity) on the x-axis.

Source: https://www.medcalc.org/manual/roc-curves.php

As the figure shows, ROC curves are upward sloping-- diagnosing more true positives typically means also increasing the rate of false positives. The curve goes through (0,0) and (100,100), because it is possible to either diagnose nobody as having the disease and get a 0% true positive rate and 0% false positive rate, or to diagnose everyone as having the disease and get a 100% true positive rate and 100% false positive rate. The further an ROC is above the 45 degree line, the better the diagnostic test is, because for any level of false positives, you get a higher level of true positives.

Rafa Irizarry at the Simply Statistics blog makes a really interesting analogy between diagnosing disease and making scientific discoveries. Scientific findings can be true or false, and if we imagine that increasing the rate of important true discoveries also increases the rate of false positive discoveries, we can plot ROC curves for scientific disciplines. Irizarry imagines the ROC curves for biomedical science and physics (see the figure below). Different fields of research vary in the position and shape of the ROC curve--what you can think of as the production possibilities frontier for knowledge in that discipline-- and in the position on the curve.

In Irizarry's opinion, physicists make fewer important discoveries per decade and also fewer false positives per decade than biomedical scientists. Given the slopes of the curves he has drawn, biomedical scientists could make fewer false positives, but at a cost of far fewer important discoveries.

Source: Rafa Irizarry

A particular scientific field could move along its ROC curve by changing the field's standards regarding peer review and replication, changing norms regarding significance testing, etc. More critical review standards for publication would be represented by a shift down and to the left along the ROC curve, reducing the number of false findings that would be published, but also potentially reducing the number of true discoveries being published. A field could shift its ROC curve outward (good) or inward (bad) by changing the "discovery production technology" of the field.

The importance of discoveries is subjective, and we don't really know numbers of  "false positives" in any field of science. Some never go detected. But lately, evidence of fraudulent or otherwise irreplicable findings in political science and psychology point to potentially high false positive rates in the social sciences. A few days ago, Science published an article on "Estimating the Reproducibility of Psychological Science." From the abstract:

We conducted replications of 100 experimental and correlational studies published in three psychology journals using high-powered designs and original materials when available. Replication effects were half the magnitude of original effects, representing a substantial decline. Ninety-seven percent of original studies had statistically significant results. Thirty-six percent of replications had statistically significant results; 47% of original effect sizes were in the 95% confidence interval of the replication effect size; 39% of effects were subjectively rated to have replicated the original result; and if no bias in original results is assumed, combining original and replication results left 68% with statistically significant effects.

As studies of this type hint that the social sciences may be far to the right along an ROC curve, it is interesting to try to visualize the shape of the curve. The physics ROC curve that Irizarry drew is very steep near the origin, so an attempt to reduce false positives further would, in his view, sharply reduce the number of important discoveries. Contrast that to his curve for biomedical science. He indicates that biomedical scientists are on a relatively flat portion of the curve, so reducing the false positive rate would not reduce the number of important discoveries by very much.

What does the shape of the economics ROC curve look like in comparison to those of other sciences, and where along the curve are we? What about macroeconomics in particular? Hypothetically, if we have one study that discovers that the fiscal multiplier is smaller than one, and another study that discovers that the fiscal multiplier is greater than one, then one study is an "important discovery" and one is a false positive. If these were our only two macroeconomic studies, we would be exactly on the 45 degree line with perfect sensitivity but zero specificity.

Macroeconomics Research at Liberal Arts Colleges

I spent the last two days at the 11th annual Workshop on Macroeconomics Research at Liberal Arts Colleges at Union College. The workshop reflects the growing emphasis that liberal arts colleges place on faculty research. There were four two-hour sessions of research presentations--international, banking, information and expectations, and theory--in addition to breakout sessions on pedagogy. I presented my research in the information and expectations session.

I definitely recommend this workshop to other liberal arts macro professors. The end of summer timing was great. I got to think about how to prioritize my research goals before the semester starts and to hear advice on teaching and course planning from a lot of really passionate teachers. It was very encouraging to witness how many liberal arts college professors at all stages of their careers have maintained very active research agendas while also continually improving in their roles as teachers and advisors.

After dinner on the first day of the workshop, there was a panel discussion about publishing with undergraduates. I also attended a pedagogy session on advising undergraduate research. Many of the liberal arts colleges represented at the workshop have some form of a senior thesis requirement. A big part of the discussion was how to balance the emphasis on "product vs. process" for undergraduate research. In other words, how active of a role should a faculty member take in trying to ensure a high-quality final product of a senior thesis project versus ensuring that different learning goals are met. What should those learning goals be? Some possibilities include helping students decide if they want to go to grad school, teach independence, writing skills, econometric techniques, the ability to for an economic argument. And relatedly, how should grades or honors designations reflect the final product and the learning goals that are emphasized?

We also discussed the relative merits of helping students publish their research, either in an undergraduate journal or a professional journal. There was a lot of lack of clarity about how it affects an assistant professor's tenure case if they have very low-ranked publications with undergraduate coauthors, and a general desire for more explicit guidelines about whether that is considered a valuable contribution.

These discussions of research by or with undergraduates left me really curious to hear about others' experiences doing or supervising undergraduate research. I'd be very happy to feature some examples of research with or by undergraduates as guest posts. Send me an email if you're interested.

At least two other conference participants have blogs, and they are definitely worth checking out. Joseph Joyce of Wellesley blogs about international finance at "Capital Ebbs and Flows." Bill Craighead of Wesleyan blogs at "Twenty-Cent Paradigms." Both have recent thoughtful commentary on Greece.

Trading on Leaked Macroeconomic Data

The official release times of U.S. macroeconomic data are big deals in financial markets. A new paper finds evidence of substantial informed trading before the official release time of certain macroeconomic variables, suggesting that information is often leaked. Alexander Kurov, Alessio Sancetta, Georg H. Strasser, and Marketa Halova Wolfe examine high-frequency stock index and Treasury futures markets data around releases of U.S. macroeconomic announcements:

These announcements are quintessential updates to public information on the economy and fundamental inputs to asset pricing. More than a half of the cumulative annual equity risk premium is earned on announcement days (Savor & Wilson, 2013) and the information is almost instantaneously reflected in prices once released (Hu, Pan, & Wang, 2013). To ensure fairness, no market participant should have access to this information until the official release time. Yet, in this paper we find strong evidence of informed trading before several key macroeconomic news announcements....Prices start to move about 30 minutes before the official release time and the price move during this pre-announcement window accounts on average for about a half of the total price adjustment.

They consider the 30 macroeconomic announcements that other authors have shown tend to move markets, and find evidence of:

• Significant pre-announcement price drift for: CB consumer confidence index, existing home sales, GDP preliminary, industrial production, ISM manufacturing index, ISM non-manufacturing index, and pending home sales.
• Some pre-announcement drift for: advance retail sales, consumer price index, GDP advance, housing starts, and initial jobless claims.
• No pre-announcement drift for: ADP employment, durable goods orders, new home sales, non-farm employment, producer price index, and UM consumer sentiment.

The figure below shows mean cumulative average returns in the E-mini S&P 500 Futures market from 60 minutes before the release time to 60 minutes after the release time for the series with significant evidence of pre-announcement drift.

Source: Kurov et al. 2015, Figure A1, panel c. Cumulative average returns in the E-mini S&P 500 Futures market .

Why do prices start to move before release time? It could be that some traders are superior forecasters, making better use of publicly-available information, and waiting until a few minutes before the announcement to make their trades. Alternatively, information might be leaked before the official release. Kurov et al. note that, while the first possibility cannot be ruled out entirely, the leaked information explanation appears highly likely. The authors conducted a phone and email survey of the organizations responsible for the macroeconomic data in their study to find out about data release procedures:

The release procedures fall into one of three categories. The first category involves posting the announcement on the organization’s website at the official release time, so that all market participants can access the information at the same time. The second category involves pre-releasing the information to selected journalists in “lock-up rooms” adding a risk of leakage if the lock-up is imperfectly guarded. The third category, previously not documented in academic literature, involves an unusual pre-release procedure used in three announcements: Instead of being pre-released in lock-up rooms, these announcements are electronically transmitted to journalists who are asked not to share the information with others. These three announcements are among the seven announcements with strong drift.

I wish I had a better sense of who was obtaining the leaked information and how much they were making from it.

Forecasting in Unstable Environments

I recently returned from the International Symposium on Forecasting "Frontiers in Forecasting" conference in Riverside. I presented some of my work on inflation uncertainty in a session devoted to uncertainty and the real economy. A highlight was the talk by Barbara Rossi,  a featured presenter from Universitat Pompeu Fabra, on "Forecasting in Unstable Environments: What Works and What Doesn't." (This post will be a bit more technical than my usual.)

Rossi spoke about instabilities in reduced form models and gave an overview of the evidence on what works and what doesn't in guarding against these instabilities. The basic issue is that the predictive ability of different models and variables changes over time. For example, the term spread was a pretty good predictor of GDP growth until the 1990s, and the credit spread was not. But in the 90s the situation reversed, and the credit spread became a better predictor of GDP growth while the term spread got worse.

Rossi noted that break tests and time varying parameter models, two common ways to protect against instabilities in forecasting relationships, do involve tradeoffs. For example, it is common to test for a break in an empirical relationship, then estimate a model in which the coefficients before and after the break differ. Including a break point reduces the bias of your estimates, but also reduces the precision. The more break points you add, the shorter are the time samples you use to estimate the coefficients. This is similar to what happens if you start adding tons of control variables to a regression when your number of observations is small.

Rossi also discussed rolling window estimation. Choosing the optimal window size is a challenge, with a similar bias/precision trade-off. The standard practice of reporting results from only a single window size is problematic, because the window size may have been selected based on "data snooping" to obtain the most desirable results. In work with Atsushi Inoue, Rossi develops out of sample forecast tests that are robust to window size. Many of the basic tools and tests from macroeconomic forecasting-- Granger casualty tests, forecast comparison tests, and forecast optimally tests-- can be made more robust to instabilities. For details, see Raffaella Giacomini and Rossi's chapter in the Handbook of Research Methods and Applications on Empirical Macroeconomics and references therein.

A bit of practical advice from Rossi was to maintain large-dimensional datasets as a guard against instability. In unstable environments, variables that are not useful now may be useful later, and it is increasingly computationally feasible to store and work with big datasets.

Another Four Percent

When Jeb Bush announced his presidential candidacy on Monday, he made a bold claim. "There's not a reason in the world we can’t grow at 4 percent a year,” he said, “and that will be my goal as president.”

You can pretty much guarantee that whenever a politician claims "there's not a reason in the world," plenty of people will be happy to provide one, and this case is no exception. Those reasons aside, for now, where did this 4 percent target come from? Jordan Weissmann explains that "the figure apparently originated during a conference call several years ago, during which Bush and several other advisers were brainstorming potential economic programs for the George W. Bush Institute...Jeb casually tossed out the idea of 4 percent growth, which everybody loved, even though it was kind of arbitrary." Jeb Bush himself calls 4 percent "a nice round number. It's double the growth that we are growing at." (To which Jon Perr snippily adds, "It's also an even number and the square of two.")

Let's face it, we have a thing for nice, round, kind of arbitrary numbers. The 2 percent inflation target, for example, was not chosen as the precise solution to some optimization problem, but more as a "rough guess [that] acquired force as a focal point." Psychology research shows that people put in extra effort to reach round number goals,  like a batting average of .300 rather than .299. A 4 percent growth target reduces something multidimensional and hard to define--economic success--to a single, salient number. An explicit numerical target provides an easy guide for accountability. This can be very useful, but it can also backfire.

As an analogy, imagine that citizens of some country have a vague, noble goal for their education system, like "improving student learning." They want to encourage school administrators and teachers to pursue this goal and hold them accountable. But with so many dimensions of student learning, it is difficult to gauge effort or success. They could introduce a mandatory, standardized math test for all students, and rate a teacher as "highly successful" if his or her students' scores improve by at least 10% over the course of the year. A nice round number. This would provide a simple, salient way to judge success, and it would certainly change what goes on in the classroom, with obvious upsides and downsides. Many teachers would put in more effort to ensure that students learned math--at least, the math covered on the test--but might neglect literature, art, or gym. Administrators might have incentive to engage in some deceptive accounting practices, finding reasons why a particular student's score should not be counted, why a group of students should switch classrooms. Even outright cheating, though likely rare, is possible, especially if jobs are hinging on the difference between 9.9% improvement and 10%. What is changing one or two answers?

Ceteris paribus, more math skills would bring a variety of benefits, just like more growth would, as the George W. Bush Institute's 4% Growth Project likes to point out. But making 4 percent growth the standard for success could also change policymakers' incentives and behaviors in some perverse ways. Potential policies' ability to boost growth will be overemphasized, and other merits or flaws (e.g. for the environment or the income distribution) underemphasized. The purported goal is sustained 4 percent growth over long time periods, which implies making the kind of long-run-minded reforms that boost both actual and potential GDP--not just running the economy above capacity for as long as possible until the music stops. But realistically, a president would worry more about achieving 4 percent while in office and less about afterwards, encouraging short-termism at best, or more unsavory practices at worst.

Even with all of these caveats, if the idea of a 4 percent solution still sounds appealing, it is worth opening up the discussion to what other 4 percent solutions might be better. Laurence Ball, Brad Delong, and Paul Krugman have made the case for 4 percent inflation target. I see their points but am not fully convinced. But what about 4 percent unemployment? Or 4 percent nominal wage growth? Are they more or less attainable than 4 percent GDP growth, and how would the benefits compare? If we do decide to buy into a 4 percent target, it is worth at least pausing to think about which 4 percent. 

Firm Balance Sheets and Unemployment in the Great Recession

The balance sheets of households and financial firms have received a lot of emphasis in research on the Great Recession. The balance sheets of non-financial firms, in contrast, have received less attention. At first glance, this is perfectly reasonable; households and financial firms had high and rising leverage in the years leading up to the Great Recession, while non-financial firms' leverage remained constant (Figure 1, below).

New research by Xavier Giroud and Holger M. Mueller argues that the flat trendline for non-financial firms' leverage obscures substantial variation across firms, which proves important to understanding employment in the recession. Some firms saw large increases in leverage prior to the recession and others large declines. Using an establishment-level dataset with more than a quarter million observations, Giroud and Mueller find that "firms that tightened their debt capacity in the run-up ('high-leverage firms') exhibit a significantly larger decline in employment in response to household demand shocks than firms that freed up debt capacity ('low-leverage firms')."

Source: Giroud and Mueller (2015)

The authors emphasize that "we do not mean to argue that household balance sheets or those of financial intermediaries are unimportant. On the contrary, our results are consistent with the view that falling house prices lead to a drop in consumer demand by households (Mian, Rao, and Sufi (2013)), with important consequences for employment (Mian and Sufi (2014)). But households do not lay off workers. Firms do. Thus, the extent to which demand shocks by households translate into employment losses depends on how firms respond to these shocks."

Firms' responses to household demand shocks depend largely on their balance sheets. Low-leverage firms were able to increase their borrowing during the recession to avoid reducing employment, while high-leverage firms were financially constrained and could not raise external funds to avoid reducing employment and cutting back investment:

"In fact, all of the job losses associated with falling house prices are concentrated among establishments of high-leverage firms. By contrast, there is no significant association between changes in house prices and changes in employment during the Great Recession among establishments of low-leverage firms."

Politicians or Technocrats: Who Splits the Cake?

In most countries, non-elected central bankers conduct monetary policy, while fiscal policy is chosen by elected representatives. It is not obvious that this arrangement is appropriate. In 1997, Alan Blinder suggested that Americans leave "too many policy decisions in the realm of politics and too few in the realm of technocracy," and that tax policy might be better left to technocrats. The bigger issue these days is whether an independent, non-elected Federal Reserve can truly be "accountable" to the public, and whether Congress should have more control over monetary policy.

The standard theoretical argument for delegating monetary policy to a non-elected bureaucrat is the time inconsistency problem. As Blinder explains, "the pain of fighting inflation (higher unemployment for a while) comes well in advance of the benefits (permanently lower inflation). So shortsighted politicians with their eyes on elections would be tempted to inflate too much." But time inconsistency problems arise in fiscal policy too. Blinder adds, "Myopia is a serious practical problem for democratic governments because politics tends to produce short time horizons -- often extending only until the next election, if not just the next public opinion poll. Politicians asked to weigh short-run costs against long-run benefits may systematically shortchange the future."

So why do we assign some types of policymaking to bureaucrats and some to elected officials? And could we do better? In a two-paper series on "Bureaucrats or Politicians?," Alberto Alesina and Guido Tabellini (2007) study the question of task allocation between bureaucrats and politicians. In their model, neither bureaucrats nor politicians are purely "benevolent;" each have different objective functions depending on how they are held accountable:

Politicians are held accountable, by voters, at election time. Top-level bureaucrats are accountable to their professional peers or to the public at large, for how they have fulfilled the goals of their organization. These different accountability mechanisms induce different incentives. Politicians are motivated by the goal of pleasing voters, and hence winning elections. Top bureaucrats are motivated by "career concerns," that is, they want to fulfill the goals of their organization because this improves their external professional prospects in the public or private sector.

The model implies that, for the purpose of maximizing social welfare, some tasks are better suited for bureaucrats and others for politicians. When the public can only imperfectly monitor effort and talent, elected politicians are preferable for tasks where effort matters more than ability. Bureaucrats are preferable for highly technical tasks, like monetary policy, regulatory policy, and public debt management. This is in line with Blinder's intuition; he argued that extremely technical judgments ought to be left to technocrats and value judgments to legislators, while recognizing that both monetary and fiscal policy involve substantial amounts of both technical and value judgments.

Alesina and Tabellini's model also helps formalize and clarify Blinder's intuition on what he calls "general vs. particular" effects. Blinder writes:

Some public policy decisions have -- or are perceived to have -- mostly general impacts, affecting most citizens in similar ways. Monetary policy, for example...is usually thought of as affecting the whole economy rather than particular groups or industries. Other public policies are more naturally thought of as particularist, conferring benefits and imposing costs on identifiable groups...When the issues are particularist, the visible hand of interest-group politics is likely to be most pernicious -- which would seem to support delegating authority to unelected experts. But these are precisely the issues that require the heaviest doses of value judgments to decide who should win and lose. Such judgments are inherently and appropriately political. It's a genuine dilemma.

Alesina and Tabellini consider a bureaucrat and an elected official each assigned a task of "splitting a cake." Depending on the nature of the cake splitting task, a bureaucrat is usually preferable; specifically, "with risk neutrality and fair bureaucrats, the latter are always strictly preferred ex ante. Risk aversion makes the bureaucrat more or less desirable ex ante depending on how easy it is to impose fair treatment of all voters in his task description." Nonetheless, politicians prefer to cut the cake themselves, because it helps them get re-elected with less effort through an incumbency advantage:

The incumbent’s redistributive policies reveal his preferences, and voters correctly expect these policies to be continued if he is reelected. As they cannot observe what the opponent would do, voters face more uncertainty if voting for the opponent...This asymmetry creates an incumbency advantage: the voters are more willing to reappoint the incumbent even if he is incompetent... The incumbency advantage also reduces equilibrium effort.

An interesting associated implication is that "it is in the interest of politicians to pretend that they are ideologically biased in favor of specific groups or policies, even if in reality they are purely opportunistic. The ideology of politicians is like their brand name: it keeps voters attached to parties and reduces uncertainty about how politicians would act once in office."

According to this theoretical model, we might be better off leaving both monetary and fiscal policy to independent bureaucratic agencies. But fiscal policy is inherently redistributive, and politicians prefer not to delegate redistributive tasks. "This might explain why delegation to independent bureaucrats is very seldom observed in fiscal policy, even if many fiscal policy decisions are technically very demanding."

Both Blinder and Alesina and Tabellini--writing in 1997 and 2007, respectively-- made the distinction that tax policy, unlike monetary policy, is redistributive or "particularist." Since then, that distinction seems much less obvious. Back in 2012, Mark Spitznagel opined in the Wall Street Journal that "The Fed is transferring immense wealth from the middle class to the most affluent, from the least privileged to the most privileged." Boston Fed President Eric Rosengren countered that "The net effect [of recent Fed policy] is substantially weighted towards people that are borrowers not lenders, towards people that are unemployed versus people that are employed." Other Fed officials and academic economists are also paying increasing attention to the redistributive implications of monetary policy.

Monetary policymakers can no longer ignore the distributional effects of monetary policy-- and neither can voters and politicians. Alesina and Tabellini's model predicts that the more that elected politicians recognize the "cake splitting" aspect of monetary policy, the more they will want to redelegate it to themselves. Expect stronger cries for "accountability." However, the redistributive nature of monetary policy, according to the model, probably strengthens the argument for leaving it to independent technocrats. The caveat is that "the result may be reversed if the bureaucrat is unfair and implements a totally arbitrary redistribution." The Fed's role in redistributing resources strengthens its case for independence if and only if it takes equity concerns seriously.

Targeting from Below

Inflation targeting (IT) was widely adopted by central banks in both industrialized and emerging-market countries in the 1990s and 2000s. Typically, the objective for switching to an IT framework has been to reduce and stabilize high and volatile inflation. Studies of IT find that it has been successful in regards to this objective.

But how does IT fare when inflation is instead too low? Michael Ehrmann of the Bank of Canada addresses this question in "Targeting Inflation from Below: How do Inflation Expectations Behave?" He notes that the Bank of Japan adopted IT in an environment of undesirably-low inflation. Likewise, when the Federal Reserve announced a 2% inflation target in 2012, core inflation had been below 2% for some time. "Although designed to lower inflation and inflation expectations," Ehrmann writes, "IT is now charged with the objective to raise them, a challenge that has not yet been studied extensively."

Since inflation targeting is supposed to work by anchoring expectations near the target, Ehrmann studies the inflation expectations of professional forecasters to compare the performance of IT when inflation is persistently low, near target, and persistently high. He uses data from Consensus Economics for Australia, Canada, the euro area, France, Germany, Italy, Japan, the Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the United Kingdom and the United States. He uses three different indicators of the extent to which inflation expectations are anchored: (1) the extent to which expectations depend on lagged inflation; (2) forecaster disagreement; and (3) the extent to which inflation expectations get revised in response to news. On all three counts, he finds that under persistently low inflation, expectations can become disanchored. That is, inflation expectations are more dependent on lagged inflation; forecasters disagree more; and inflation expectations get revised down in response to lower-than-expected inflation. He also finds that when inflation is persistently low, expectations do not get revised upward in response to higher-than-expected inflation.

These findings are important. We tend to worry about inflation expectations becoming disanchored when inflation goes too high above target or stays above target for too long. This is partly why the inflation target gets treated more like a ceiling than a symmetric target. But in the current situation in the U.S. and Europe, it may be that keeping inflation too low is weakening the anchoring of expectations. A temporary burst of above-target inflation seems unlikely to damage the anchor.

Mixed Signals and Monetary Policy Discretion

Two recent Economic Letters from the Federal Reserve Bank of San Francisco highlight the difficulty of making monetary policy decisions when alternative measures of labor market slack and the output gap give mixed signals. In Monetary Policy when the Spyglass is Smudged, Early Elias, Helen Irvin, and Òscar Jordà show that conventional policy rules based on the output gap and on the deviation of the unemployment rate from its natural rate generate wide-ranging policy rate prescriptions. Similarly, in Mixed Signals: Labor Markets and Monetary Policy, Canyon Bosler, Mary Daly, and Fernanda Nechio calculate the policy rate prescribed by a Taylor rule under alternative measures of labor market slack. The figure below illustrates the large divergence in alternative prescribed policy rates since the Great Recession.

Source: Bosler, Daly, and Nechio (2014), Figure 2

Uncertainty about the state of the labor market makes monetary policy more challenging and requires more discretion and judgment on the part of policymakers. What does discretion and judgment look like in practice? I think it should involve reasoning qualitatively to determine if some decisions lead to possible outcomes that are definitively worse than others. For example, here's how I would reason through the decision about whether to raise the policy rate under high uncertainty about the labor market:

Suppose it is May and the Fed is deciding whether to increase the target rate by 25 basis points. Assume inflation is still at or slightly below 2%, and the Fed would like to tighten monetary policy if and only if the "true" state of the labor market x is sufficiently high, say above some threshold X. The Fed does not observe x but has some very noisy signals about it.  They think there is about a fifty-fifty chance that x is above X, so it is not at all obvious whether tightening is appropriate. There are four possible scenarios:

1. The Fed does not increase the target rate, and it turns out that x>X.
2. The Fed does not increase the target rate, and it turns out that x<X.
3. The Fed does increase the target rate, and it turns out that x>X.
4. The Fed does increase the target rate, and it turns out that x>X.

Cases (2) and (3) are great. In case (2), the Fed did not tighten when tightening was not appropriate, and in case (3), the Fed tightened when tightening was appropriate. Cases (1) and (4) are "mistakes." In case (1), the Fed should have tightened but did not, and in case (4), the Fed should not have tightened but did. Which is worse?

If we think just about immediate or short-run impacts, case (1) might mean inflation goes higher than the Fed wants and x goes even higher above X; case (4) might mean unemployment goes higher than the Fed wants and x falls even further below X. Maybe you have an opinion on which of those short-run outcomes is worse, or maybe not. But the bigger difference between the outcomes comes when you think about the Fed's options at its subsequent meeting. In case (1), the Fed could choose how much they want to raise rates to restrain inflation. In case (4), the Fed could keep rates constant or reverse the previous meeting's rate increase.

In case (4), neither option is good. Keeping the target at 25 basis points is too restrictive. Labor market conditions were bad to begin with and keeping policy tight will make them worse. But reversing the rate increase is a non-starter. The markets expect that after the first rate increase, rates will continue on an upward trend, as in previous tightening episodes. Reversing the rate increase would cause financial market turmoil, damage credibility, and require policymakers to admit that they were wrong. Case (1) is much more attractive. I think any concern that inflation could take off and get out of control is unwarranted. In the space between two FOMC meetings, even if inflation were to rise above target, inflation expectations are not likely to rise too far. The Fed could easily restrain expectations at the next meeting by raising rates as aggressively as needed.

So going back to the four possible scenarios, (2) and (3) are good, and (4) is much worse than (1). If the Fed raises rates, scenarios (3) and (4) are about equally likely. If the Fed holds rates constant, (1) and (2) are about equally likely. Thus, holding rates constant under high uncertainty about the state of the labor market is a better option than potentially raising rates too soon.

Guest Post: Estimating Monetary Policy Rules Around The Zero Lower Bound

I hope you enjoy this guest post contributed by Jon Hartley. 

As the Federal Reserve moves closer to normalizing monetary policy and moving toward a federal funds rate “lift-off” date, I’ve created MonetaryPolicyRules.org, a new website that provides up-to-date interactive graphs of popular monetary policy rules.

Since the federal funds rate has hit the zero lower bound, Taylor rules have received a lot of criticism in large part because many Taylor rules have prescribed negative nominal interest rates during and after the global financial crisis. Chicago Fed President (and prominent monetary policy scholar) Charles Evans stated about the Taylor Rule that “The rule completely breaks down during the Great Recession and its aftermath”.

The discretionary versus rules-based monetary policy debate endures, most recently with the Federal Reserve Accountability and Transparency Act being introduced in Congress, followed by a series of dueling Wall Street Journal op-eds by John Taylor and Alan Blinder. Rather than thinking about Taylor rules as a prescription for monetary policy (in a normative economics sense), what has been left out of the discussion is how Taylor rules can accurately be a description of monetary policy regimes (in a positive economics sense) even if the central bank does not explicitly follow a stated rule.

Tim Duy has accurately pointed out in a recent post that using the GDP and inflation forecasts also provided by the FOMC for 2014 through 2017 (and beyond), no traditional monetary policy rule captures the median of the current fed fund rate forecasts (commonly known as the “dot plots” released by the Federal Reserve on a quarterly basis as part of their Delphic forward guidance) which are considerably lower than either the Taylor (1993), Taylor (1999), Mankiw (2001), or Rudebusch (2009) rules would estimate.

What’s also worth noting is that in the early to mid-2000’s the federal funds rate was considerably lower than what any of the above classic monetary policy rules would estimate. This in large part is because all of these rules were estimated using data from the “Great Moderation” of the 1990’s, which was then led by a very different Federal Reserve than we have today (note those rules fit the federal funds effective rate data very accurately during the 1990’s).

Source: Tim Duy

The real question is how can we estimate a monetary policy rule that describes the Bernanke-Yellen Fed, while also addressing the problem of the zero lower bound for nominal interest rates?

One interesting idea that has gained some popularity recently is the idea of measuring a “shadow federal funds rate” (originally hypothesized by Fisher Black in a 1995 paper, published just before his death, which uses fed funds futures rates and an affine term structure model to back out a negative spot rate). The idea nicely estimates the potential effects of quantitative easing on long-term rates  while the federal funds rate is at the zero lower bound (and for that reason I’ve included the Wu-Xia (2014) shadow fed funds rate on the site). With the shadow fed funds rate in hand, one can now estimate a monetary policy rule with a standard OLS regression. One issue with this methodology is how there is a lack of consensus around what to use as input data for a shadow rate which can give you very different results (Khan and Hakkio (2014) observed that the Wu-Xia (2014) shadow fed funds rate looks remarkably different from the rate calculated by Krippner (2014)).

Wu-Xia (2014) and Krippner (2014) Shadow Federal Funds Rates (in %)Source: Khan and Hakkio (2014), Federal Reserve Board of Governors, Krippner (2014), Wu-Xia (2014)

One other solution to the problem of estimating a monetary policy rule at the zero lower bound is an econometric one. Fortunately, we have Tobit regressions in our econometric toolbox (originally developed by James Tobin (1958)) which allow us to estimate Taylor rules while censoring data at the zero lower bound.

In my Taylor rule that is estimated with federal funds rate from the Bernanke-Yellen period, censoring data at the zero lower bound using a Tobit regression, I use both y/y core CPI inflation and unemployment. In another version, I use the Fed’s new Labor Market Conditions Index (LMCI) as a labor market indicator though both yield relatively similar results*. Importantly, these estimates indicate that the Bernanke-Yellen Fed puts a much higher weight on the output/unemployment gap than the Mankiw (2001) rule estimated with data from the Greenspan period.

Tobit Taylor Rule Using Unemployment Rate and Core CPI as Inputs:

Federal Funds Target Rate = max{0, -0.43 + 1.2*(Core CPI y/y %) – 2.6*(Unemployment Rate-5.6)}

Using unemployment and inflation forecast data from the latest Federal Reserve FOMC meeting’s Survey of Economic Projections, I input these data as inputs into the Tobit Rule and Mankiw (2001) Rule. Matching these federal funds rate targets implied by the rules with the median federal funds forecasts provided by both the Federal Reserve “dot plots” and the Survey of Primary Dealers (note: the expected Fed Funds rate path from Survey of Primary Dealers has significantly fallen below the Fed dot plot Fed Funds rate path as noted by Christensen (2014)). Unfortunately, we do not have precise data on which dots belong to which Fed officials (otherwise, we could try to construct a Taylor rule for each Fed president and board member). Compared to the Mankiw (2001) rule, the estimated Tobit rule much better matches the median forecasts provided by the dot plots and the Survey of Primary Dealers.

Federal Reserve Forward Guidance/Survey of Primary Dealers Fed Funds Rate Forecasts versus Tobit Rule and Mankiw (2001) Rule Using Fed Unemployment and CPI Forecasts

Janet Yellen has spoken fondly of the Taylor (1999) rule in the past as she has previously stated in a 2012 speech that “[John] Taylor himself continues to prefer his original rule, which I will refer to as the Taylor (1993) rule. In my view, however, the later variant--which I will refer to as the Taylor (1999) rule--is more consistent with following a balanced approach to promoting our dual mandate.”

It is no surprise that the Tobit rule estimated with more recent data comes much closer to accurately describing the Fed’s forward guidance than the Taylor (1993) rule. However, what is really interesting is that the Tobit Rule is also much closer to describing the Fed’s current forward guidance than the Taylor (1999) rule, which remains far off.

Footnote:

*An important issue which has been addressed recently with the introduction of the Fed’s new Labor Market Conditions Index (LMCI) is how do we measure improvement (or lack thereof) in the labor market? While the U.S. unemployment rate for September was 5.9% (the lowest level since July 2008), the figure fails to capture a number of fractures in the economy remain which are not reflected in the unemployment rate. One item included in the LMCI (but not reflected in the unemployment rate) is the high U-6 unemployment rate (which factors in individuals who are underemployed, working part-time for economic reasons and would rather have full-time jobs) currently at 11.8%. Another is subdued wage growth that is not commensurate with drops in the unemployment rate as history would suggest. The labor force participation rate is at historical lows of 62% (in large part due to the number of retirements (a secular demographic trend) and to some extent due to discouraged workers (a cyclical trend) according to a recent Philly Fed study).

A previous post on this blog astutely points out that correlation of 12-month changes in LMCI with 12-month changes in the unemployment rate is -0.96, suggesting that “the LMCI doesn’t tell you anything that the unemployment rate wouldn’t already tell you”. While the economists who developed the LMCI list on the Fed’s website the correlations of 12-month changes (which tell you the tendency of large 12-month figures moving together), I would argue that this is accurate of long-term labor market trends, while the correlations of monthly changes with the LMCI is a more accurate representation of the extent the measures move together in small short-term labor market movements. Doing so indicates that the monthly level of the LMCI has a -0.82 correlation with the unemployment rate, suggesting that the LMCI is not completely redundant in short-term labor market movements, incorporating some parts of the mixed economic narrative told by dampening wage growth, low labor force participation, and high amount of underemployed part-timers.