The Region

Robert E . Hall

In a discipline that celebrates specialization, Robert Hall is a Renaissance man. And economics is far the richer for it. The Stanford ’s extensive publications over four decades— books, blogs, articles and lectures—provide ready evidence of wide-ranging expertise. (The interview below hardly scratches the surface.) As a labor economist, Hall has produced some of the field’s most influential models of labor market dynamics and essential articles on labor supply, demand and wages. A scholar of fiscal policy, he built the intellectual foundation for the 1986 tax reform bill as well as recent tax proposals. His work in financial theory, consumer and corporate incentives, and government policy illuminates regulatory issues currently under debate in Washington. Innovative analysis of stock market valuation by Hall demonstrated the importance of intangible capital. His studies of entrepreneur ial incentives (with his wife, economist Susan Woodward) and antitrust theory are pathbreaking. Hall’s research on trading through electronic markets—“digital dealing” is his term—provided one of the first lucid explanations of the economics of then-new Internet phenomena such as eBay. Hall’s analytical gifts also have generated important insights on monetary theory and optimal monetary systems. He has done invaluable work as well in growth theory, determinants of , spending on health and economic geography. His erudition has range, depth and quality that few can match. And the profession has recognized this with honors including the Richard T. Ely lecture in 2001, presidency of the American Economic Association in 2010, and fellowship in the American Academy of Arts and Sciences, and National Academy of Sciences. Hall’s public profile, however, is largely confined to the sphere of business cycles, in which he also has unquestioned expertise. That narrow “fame” is due to his chairmanship, for over 30 years, of the committee that determines when U.S. recessions officially begin and end. It is a painstaking and largely thankless task. Pundits and policymakers clamor for the committee’s announcements, but inevitably second-guess the decisions made. Committed to the integrity of process and result, Hall has never bent to pressure, manifest ing time consistency that monetary policymakers can only envy.

Photography by Peter Tenzer

JUNE 2010 26

The Region

THOUGHTS ON U.S. excellent speeches that have described MONETARY POLICY this fully, so it shouldn’t be an issue, and I think more or less it’s not anymore. The Region: Perhaps we could start with Fed can either leave the reserves out monetary policy. What is your broad there but make them attractive to banks view of the Fed’s efforts over the past by paying interest on them, or it can few years to stem the crisis using uncon - withdraw them by selling the correspon - ventional monetary policy and strate - ding assets they’re invested in. Selling gies? assets will be timely because those invest - ments will have recovered to their proper Hall: First of all, I believe you should values; the Fed can sell them and use the think of the Fed as simply part of the funds to retire the reserves. federal government when it comes to So, again, there are two branches to the financial side of its interventions. If the exit strategy: There’s paying interest you look at how the federal government on reserves, and there’s reducing responded initially, it was the Treasury reserves back to more normal levels. that was providing the funds. Of course, They’re both completely safe, so it’s a TARP [Troubled Asset Relief Program] nonissue. The Fed itself is just not a dan - was there using the taxpayers’ money There are two branches to the exit ger. It is run by people who know exact - without involvement of the Fed. Also, strategy: There’s paying interest on ly what to do. And we have 100 percent early in the crisis Treasury deposited reserves, and there’s reducing reserves confidence they will do it. It’s not some - hundreds of billions of dollars at the back to more normal levels. They’re both thing I worry about. Fed, which the Fed then used to buy completely safe, so it’s a nonissue. assets. So there the Fed was just an agent FINANCIAL FRICTIONS of the Treasury. It was as if the Treasury The Fed itself is just not a danger. It is took its funds to a broker. run by people who know exactly Region: That’s reassuring, but I believe Eventually, the Treasury was imped - what to do. you do worry about financial frictions… ed from doing that by the federal debt limit. But the debt limit doesn’t apply to zero interest in the United States, so in Hall: I do, I do very much. funds borrowed by the Fed, so it then normal times with positive market started borrowing large amounts from interest rates, banks try to unload Region: Your recent paper on gaps, or banks by issuing reserves. That is what reserves; when they do so, they expand “wedges,” between the cost of and caused all the confusion about thinking the economy. That does not happen returns to borrowing and lending in this was somehow part of conventional when interest rates in the market are business credit markets and homeowner monetary policy. zero because there’s no incentive for loan markets argues that such frictions I would distinguish between conven - banks to unload reserves. They can’t are a major force in business cycles. tional monetary policy which sets the gain by getting something off their bal - Would you elaborate on what you interest rate and this kind of financial ance sheet if what they buy doesn’t yield mean by that and tell us what the policy intervention of buying what appear to any more. And during the crisis, there implications might be? be undervalued private securities. was no differential, nothing to be gained Issuing what appear to be overvalued by unloading reserves. Hall: There’s a picture that would help public securities and trading them for As the differential reestablishes, tell the story. It’s completely compelling. undervalued private securities, at least which the markets think is going to hap - This graph shows what’s happened dur - under some conditions and some mod - pen in the next year or so, then that ing the crisis to the interest rates faced els, is the right thing to do. In my mind, issue comes up. It would be highly by private decision makers: households it doesn’t make a big difference whether expansionary and ultimately - and businesses. There’s been no system - it’s done by the Federal Reserve, the ary if market interest rates began to rise atic decline in those interest rates, espe - Treasury or some other federal agency. above zero and the Fed didn’t do some - cially those that control home building, thing to either reduce the volume of purchases of cars and other consumer Region: And what are your thoughts on reserves or increase the demand for durables, and business investment. So the best course for a Fed exit strategy? reserves. although government interest rates for So the Fed has two tools, and claims like Treasury notes fell quite a bit Hall: That again gets at this confusion. Chairman Bernanke has been very clear during the crisis, the same is not true for Traditionally, reserves at the Fed pay on this point. He’s given a couple of private interest rates.

JUNE 2010 28 The Region

Even though the Fed has driven the interest rate that it controls to zero, it Between those rates is some kind of The mechanism we describe in our hasn’t had that much effect on reducing friction, and what this means is that textbooks about how expansionary pol - borrowing costs to individuals and even though the Fed has driven the icy can take over by lowering interest businesses. ... When private borrowing interest rate that it controls to zero, it rates and cure the recession is just not hasn’t had that much effect on reducing operating, and that seems to be very rates rise and public borrowing rates fall, borrowing costs to individuals and busi - central to the reason that the crisis has the difference between them is the amount nesses. The result is it hasn’t transmitted resulted in an extended period of slack. of friction. ... That’s a potent source of the stimulus to where stimulus is need - trouble. ed, namely, private spending. Region: So to incorporate that in a The government sector—federal, model seems quite important. Most of the undervalued assets that state and local—has been completely the Fed has bought have been mortgage unable to crank up its own purchases of Hall: Yes, and many, many macroecono - related. ... There would be a case for goods; the federal government has stim - mists have turned their attention to that. expanding that type of policy to other ulated [spending] slightly but not I’ve been following the literature and seemingly undervalued instruments. enough to offset the decline that’s been a discussant at many conferences occurred at state and local governments. of other people’s work on this. In fact, the Fed is giving a conference at the end history prior to the current crisis. That’s Region: Yes, I’d like to ask you about that of next week, and I’ll be presenting my an interesting question, but data on his - later. paper on frictions. torical events aren’t always so easy, so that lies ahead. Hall: So to get spending stimulated you Region: Your model is able, I think, to need to provide incentive for private explain a fair amount of the current Region: And the policy implications? decision makers to reverse the adverse business cycle by incorporating those What can and should be done to reduce effects that the crisis has had by deliver - frictions. frictions? ing lower interest rates. So far, that’s just not happened. The only interest rate Hall: I mainly look at, as kind of a Hall: Good question! Well, it does point that has declined by a meaningful thought experiment, how much of a in the direction of focusing on things amount is the conventional mortgage decline in activity occurs when that like lower rates for corporate bonds, rate. But if you look at BAA bonds or kind of a friction develops. When pri - BAA corporate bonds. They appear to auto loans or just across the board— vate borrowing rates rise and public be undervalued private assets, although there are half a dozen rates in this pic - borrowing rates fall, the difference that’s not been one of the types of assets ture—they just haven’t declined. So between them is the amount of friction. that policy has seen as appropriate to there hasn’t been a stimulus to spend - I show that that’s a potent source of buy or to help private organizations to ing. trouble. I haven’t tried to align it with buy. That would be one way to turn.

29 JUNE 2010 The Region

We’ve concentrated on doing that in form that looting took was through the mortgage-related assets. You can see in ownership of a savings and loan; this the picture that it’s had some effect. was a feature of the savings-and-loan Most of the undervalued assets that the crisis of the late 1980s. Fed has bought have been mortgage As a “looter,” you would use the sav - related. It’s been kind of an obsession ings and loan to attract deposits, pay the with trying to solve these problems as deposits as cash to yourself and then they arise in home building, but home declare bankruptcy. Akerlof and Romer building is only part of the story. The described a number of clever ways of collapse in other types of investment doing that to escape the attention of lax spending has been equally large. There regulators, and that’s the type of thing would be a case for expanding that type you see in many settings. of policy to other seemingly underval - One of the big problems encountered ued instruments. recently is that institutions that have That would presumably result in the become very undercapitalized were still same pattern you’ve seen in mortgages. depleting their equity by paying divi - That policy has been successful—differ - dends. The government has had to push entially successful in depressing mort - very, very hard to get these financial gage rates as opposed to bond rates or institutions to stop paying dividends. It’s a danger whenever you have other areas. Dividends are exactly equity depletion. guaranteed financial institutions that With a government guarantee, it’s exact ly have gotten into a very low capital EQUITY DEPLETION what there’s incentive to do—as situation. They’ve suffered asset value described in that paper. declines, they’ve become extremely Region: Let me ask you about a paper On the other hand, it seems we’ve leveraged and they have this very you wrote in December 2008, on equity been much more successful currently depletion, defined as the “withdrawal of than we were in what Akerlof and asymmetric payoff to the owner: If they equity from firms with guaranteed Romer described as far as preventing go under, it’s the government’s problem; debt.” We’re all well aware of govern - the most extreme forms of this conduct. if they recover, it’s the owner’s benefit. ment bailouts, and implicit or explicit It’s a danger whenever you have guar - That asymmetry, which is the so-called anteed financial institutions that have guarantees of financial institutions… moral hazard problem, is just a huge issue. gotten into a very low capital situation. Hall: That paper was actually reprinted They’ve suffered asset value declines, in a book that just came out, Forward- they’ve become extremely leveraged and Region: Your thoughts about what meas - Looking Decision Making [Princeton they have this very asymmetric payoff to ures can be taken to curb this moral University Press, 2010]. It’s the last the owner: If they go under, it’s the gov - hazard? chapter in this book, which is a compi - ernment’s problem; if they recover, it’s lation of the Gorman lectures I gave at the owner’s benefit. That asymmetry, Hall: The most important thing is to be University College London in October which is the so-called moral hazard sure that financial institutions that are 2008. problem, is just a huge issue. guaranteed by the government have And yet, while we have a lot of insti - large amounts of capital so that the dan - Region: You had a wonderfully provoca - tutions in that setting today, we don’t see ger of them spending the taxpayers’ tive statement in it. You declare that many of them doing things that Akerlof money rather than their own money is equity depletion “appears to be an and Romer described, such as paying very small. That’s a principle that’s been unlimited opportunity to steal from the themselves very large dividends. It’s deeply embedded in our regulations for government.” been difficult to get them to cut the div - a long time. Could you tell us what you mean by idends, but they have not paid out very But I pointed out in this chapter the that? Why does equity depletion occur, large dividends or concealed dividends. principle of so-called prompt corrective and how does it constitute an opportu - So it looks like we’ve been somewhat action, which says if capital goes below nity to steal? successful in preventing the worst kind this mandated level, which is typically of stealing, but the asymmetry is still around 8 percent, then something has to Hall: and Paul Romer potentially a big issue. There are way too be done right away before all the wrote a paper published in 1993 in the many bank failures that should not have remaining capital gets depleted. Brookings Papers that described what occurred and especially should not have We just have not been successful at they called “looting.” The particular cost the taxpayers as much as they did. doing that. We have principles of regula -

JUNE 2010 30 The Region

tion that allow the regulators to say that been many examples discussed of this. a bank is well capitalized even though This is all in retrospect. And I cer - the markets know that it’s not. Banks tainly don’t criticize the people who have been declared to be well capitalized were doing it at the time, especially even when the market value of their Chairman Bernanke. But looking for - debt and the market value of their equi - ward now to the next time this happens, ty have declined to very low levels. convertible debt would be a huge step Regulators seem to ignore something forward. If people at the Treasury could that everyone in the market seems to have just pushed a button to convert the know, which is that they’re shaky. There debt, without needing a new law, they seems to be a lack of willingness to pay would have done it in a second. There’s attention to all the signals that a regulator no doubt about that. They just didn’t should pay attention to. All they do is have that power. look at certain accounting records, which So we need to give regulators that don’t reflect what people know. power through some sort of sensible It’s not easy though. There’s been a security design. Regulators could do large amount of discussion of this topic that, and financial institutions wouldn’t among very knowledgeable financial see it as terribly burdensome because the economists. My colleague Darrell market would know that the probability It takes so long to get [federal government] Duffie here at Stanford has been a par - of this kind of thing happening again is spending up that typically the spending ticular leader. There’s a group called the pretty low. And when it does happen rises only after the recovery has occurred, Squam Lake Working Group, of which again, which will be sometime in the and it comes at completely the wrong he’s a member, that has been advocating next century, that button would be there time. ... But the other fact is that there’s ideas like, as a backstop, having long- to press, and we wouldn’t have the chaos been a small increase in federal govern - term debt be convertible to equity. That that we had in September of 2008. is what happens in a bankruptcy, but ment purchases, but it’s been more than under this strategy it would happen GOVERNMENT SPENDING offset by declines in state and local without a bankruptcy. It would happen AND GDP government purchases. automatically with certain contingen - cies and would solve the problem in a Region: You mentioned earlier the diffi - GDP. There was a vigorous debate, very nice way. It would potentially culty of stimulating the economy, and around here anyhow, on this multiplier increase the borrowing cost, but it I’d like to discuss your work on govern - question. would properly get the incentives right. ment multipliers. The federal govern - The discussion has shifted now A lot of people look to the example of ment’s stimulus package has been a because the premise was that we would Citibank. Citibank’s long-term debt has topic of heated debate among econo - be able to raise government purchases. been selling at a considerable discount, mists, in terms of how much stimulus But, in fact, government purchases have which is a sign that the market knows it’s truly provided and whether more is not increased. that there’s an issue. So instead of doing needed. In a recent paper, you analyze In part that’s because it’s very difficult what we have done, which is give guar - basically what happens to GDP when and time-consuming to actually get the antees of short-term debt with govern - government purchases goods and serv - government to buy more stuff. This has ment investments, the alternative that ices. been a critique of fiscal policy as long as the Squam Lake people are thinking of, Would you give us your rough esti - I’ve been an economist, this notion that and I’ve been thinking of too, is to some - mate of the size of the multiplier in the it takes so long to get spending up that how convert Citibank’s long-term debt current era of very low interest rates, typically the spending rises only after into equity, which is the same thing that and share your sense of the impact of the recovery has occurred, and it comes the market is in effect doing. That would the current stimulus package? at completely the wrong time. eliminate the danger then that the bank couldn’t meet its obligations, in a way Hall: The first thing to say, just looking Region: We searched in vain for “shovel- that is less burdensome to the taxpayer. at the big picture, is that when the idea ready projects.” In retrospect, what we did was to save of a stimulus through federal purchases the economy from a tremendous train program came up in the current crisis, Hall: Yes, “shovel-ready” turned out not wreck. But we didn’t do it in a way that the thinking was, “That’s feasible. We to be. But the other fact is that there’s was as cheap for the taxpayer as it could can increase purchases.” And then the been a small increase in federal govern - have been. And, of course, there have question was how much would it raise ment purchases, but it’s been more than

31 JUNE 2010 The Region

offset by declines in state and local gov - tant. Others have found that to be true. another to get your idea on it! Forget ernment purchases. So I think that the people who looked what’s-his-name’s face! The stimulus bill recognized that that at the evidence of what the multiplier is was a danger. We have had these in normal times and said it’s maybe 0.8 Region: And I think it can be argued that tremendously pinched state and local or 1.0 (which I would agree with) kind that helped pave the way toward the governments. A lot of them have just of missed the point. There was a lot of, I 1986 Tax Reform Act. had no choice when their tax revenue think, inappropriate criticism. declined but to reduce spending. Valerie Ramey, in contrast, has Hall: We like to think so. I’ll accept that. In spite of recognizing that potential focused not on the immediate policy when the stimulus program was question but raised the scientific ques - Region: Twenty-five years later you reis - designed, still the net effect of the crisis tion about the long-run multiplier. Her sued the book, updated of course, and and the policy response was for govern - numbers are ones that I respect and agree continue to advocate it as the “most fair, ment purchases to decline, not to rise. with. They’re more in the 0.9 range. efficient, simple and workable plan on But by very small amounts. Basically, But on the issue of multipliers during the table.” nothing happened to government pur - periods of zero interest rates, because we Given its clear merits and strong chases. And that was in an environment didn’t have any changes in government advocates, why do you think it’s gained in which everybody—and certainly purchases during this one time when relatively little traction in the United Congress was enthusiastic about it—was we’ve reached the zero interest point, we States? willing to go for a program with higher don’t have any good empirical evidence. purchases. But no matter how hard they What we need is a time when interest Hall: One important thing to under - tried to turn the knob, it just wouldn’t go rates are zero and there’s a big increase stand is that contrary to some people’s very far. in government purchases. That just hasn’t impressions, it’s not gone very far in the happened. rest of the world either. Region: So ARRA [American Recovery So we have no way to know through and Reinvestment Act of 2009] was for pure practice; we have to use models. Region: Not in central and eastern naught? The models are very clear that it makes Europe? Mexico, perhaps? a big difference when we’re at the zero Hall: First of all, you have to take it interest rate limit. The normal configu - Hall: Yes, but if you look at their overall apart, as I do in that paper, and ask how ration is that you get this fiscal expan - tax structure, it’s not what we have in much of it went directly into govern - sion—the government buys more, but mind. Their rates are high because ment purchases, which is fairly small, or that triggers sort of an automatic they’ve adopted income tax systems that would stimulate state and local purchas - response from monetary policy to lean work like a flat tax, but they’re on top of es, which was also fairly small. against it. If you shut that down by hav - a very high value-added tax. So the A lot of it was providing income sup - ing interest rates stay at zero, you’ll get a combination doesn’t achieve the low plements, and there you get into the bigger effect. That’s what this literature rates that we were hoping for. question of whether the people receiv - says and it’s quite a big difference. In the U.S., there’s been a lot of back - ing the supplements increased their sliding. It looks like there’s going to be spending or not. That’s a whole other TAX POLICY more and more. The state of California, issue; I’m not commenting on that issue. for example, has a couple of times added That’s a very difficult question to Region: Of course, this raises the issue of surcharges for very high incomes. There answer. taxes, of needing to pay for deficit spend - seems to be a belief that it’s a great idea, To go on to the other part of your ing. And I notice the Time magazine that we can get all the revenue we need question, had there been an increase in cover above your desk about the flat tax. by taxing high incomes, without regard government purchases that was success - to the problems that those tax rates cre - fully achieved, how much would that Hall: From long ago! ate, especially in the longer run. That’s have increased GDP? The answer I got one of the things we talk about in our was around a factor of 1.7, which is at Region: Yes, exactly. Your work with book. There’s more to the logic of low the high end of the range of what most Alvin Rabushka on the flat tax was a marginal tax rates than just the question economists were talking about. huge sensation in the early 1980s, as of who pays the tax. I only reached that by thinking very represented by making the cover of But another factor I would emphasize carefully and reading a lot of recent Time . is that since 1981 when we first promot - commentary on this question of the ed that plan, there’s been a dramatic implications of having a zero fed funds Hall: That’s right. It’s one thing to get widening of the income distribution in rate. That turns out to be very impor - your face on the cover of Time; it’s quite the U.S. That means that the idea of the

JUNE 2010 32 The Region

a value-added tax or other consumption Hall: That’s right. tax in a way that’s progressive. There were two economists on Region: So you’ve been studying that President Bush’s Advisory Panel on process carefully, looking at job search Federal Tax Reform in 2005, Jim dynamics, wage stickiness, wage bar - Poterba and Ed Lazear, who really gaining, productivity, other factors. understood that. They pushed pretty You’ve developed a model that explains hard; that was one of the designs that labor market fluctuations without would make sense for how to do a con - assuming what you consider to be unre - sumption tax, even though it wouldn’t alistically high labor supply elasticity. be a flat tax. The origin of our initial flat tax effort Hall: I think “explain” might be a little was Rabushka coming to me in 1980 bit of an overstatement. I’m not sure and saying, “I know what the people how many of my colleagues would agree want. The people want a flat tax, but I with the word “explain.” [Laughter] I don’t know quite what that is.” And I think “accounting for” might be right. said, “I know what it is because I’ve been thinking about it since I was a graduate Region: Fair enough. What factors have Since 1981 when we first promoted that student.” But, of course, for me, it was a you found most successful in account - [flat tax] plan, there’s been a dramatic consumption tax—an efficient, simple, ing for job-finding rates? And what are widening of the income distribution in fair consumption tax. The flatness wasn’t the key drivers of labor market volatility? the U.S. That means that the idea of the so important but, of course, the flat tax name, which Rabushka contributed, Hall: The important feature that con - poor paying the same tax rate just seems was very important politically. trols the job-finding rate is the incen - less viable than it was when the income tives to employers to create jobs. At any distribution was tighter. Region: Marketing is important. given time, if the incentives are not very strong—it could happen for many dif - poor paying the same tax rate just seems Hall: Yes, but now the idea of tax flatness ferent reasons—then employers will do less viable than it was when the income is understandably not as popular. relatively little to try to recruit workers. distribution was tighter. Job seekers will then have trouble find - The division between a small num - DYNAMICS OF LABOR MARKETS ing jobs, will see themselves at the end ber of winners in the modern economy, of a long line of people waiting for the mostly businessmen and lawyers, as Region: You’ve also done a great deal of job. opposed to most other people, has research on labor markets. In 1982, you Interestingly, the number of people grown significantly. documented the “importance of long- who find jobs each month is more or In “others,” I include doctors, by the term jobs” in the United States. I’m not less a constant. Of course, this changes, way. One of the amazing things that sure that’s still the case. but it’s a pretty good starting point for doesn’t get much attention these days is understanding labor market dynamics the widening division between doctors Hall: It’s still the case. That paper’s been that the number of people who find jobs and lawyers. It used to be that doctors replicated quite a few times. It’s almost a each month is the same in a strong mar - and lawyers competed for the best hous - law of nature. The financial press is con - ket or a weak market. es in Palo Alto. Now they’re all lawyers or stantly telling us how much turnover In a strong market, you have a rela - venture capitalists; they’re not doctors. has increased, how the old days of the tively small number of job seekers, so While there are a lot of good ideas in lifetime job have disappeared. But each one finds it easy. In a good market, flat tax reform, it wouldn’t be remotely there’s no particularly strong evidence it takes the average person about a practical to do it with a single positive of that. There are some interesting month to find another job. In a weak tax rate now. So I play around with sys - changes going on, but nothing that dra - market, there are twice as many people tems that have, say, two brackets. The matic. looking, but each one of them is half as “not-so-flat” tax. But of course that likely to find a job each month; the doesn’t have quite the simple appeal that Region: A 2005 paper of yours argued product of the two—the number look - the “flat tax” did. [Laughter] that job separation was also fairly stable ing for a job and the fraction of them But there’s still a great idea in that book and what was more important was look - who find a job—is the same. which applies to any tax system, which is, ing at the hiring process and job find - So, something like 4 million people it basically figures out how to implement ing. find jobs every month. Even with 10

33 JUNE 2010 The Region

wage. The incentive to create a job is the when it began . Many countries define a difference between what a worker will recession as two quarters in a row of contribute to the business and what the negative GDP growth, and by that stan - worker has to be paid. dard I think the United States would That’s a very simple calculus. But that have entered its recession in, maybe, the seems to vary. In a recession, for various third quarter of 2008. reasons, the profit margin from hiring a worker declines, and that reduces job- Hall: But that gets back to the whole creating efforts, all the things that keep question of, do you include the peak of the labor market moving. And that, in real GDP? We always talk about the date turn, causes it to be difficult for the job of the peak. That helps sort out this tim - seeker to find a job. ing. The peak occurred in the second There’s a great debate going on as to quarter of 2008. However, as you know, just what the factors are that reduce the we declared the peak to be a little earlier additional profit from hiring another than that, December of 2007. worker. For a while, there was the thinking Region: Would you explain what stan - that movements in productivity—pro - dards—I know it’s on the NBER Web Something like 4 million people find ductivity is one of the factors, so if pro - site; it’s very clear there—but could you jobs every month. Even with 10 percent ductivity falls and the wage doesn’t fall elaborate on what standards you use to unemployment, as recently, we’ve still with it, then that reduces the profit mar - determine turning points in business seen the same thing. A very large number gin. But that idea has not worked in the cycles? last three recessions because they were of people looking, very low job-finding periods when productivity was rising, Hall: Actually, it’s not that clear, because rate for each individual, but the product— not falling. So the old productivity story these things are always up in the air. the number of jobs filled—is roughly a has not worked for the last 30 years. [Laughter] There’s a certain amount of constant. It’s a very important fact about But each of us has our own set of ambiguity in what we put on the Web the labor market. ideas. To tie it to what we were talking site. We haven’t resolved some impor - about before, financial frictions have tant questions about how this process percent unemployment, as recently, the same effect. Increasing financial should work. we’ve still seen the same thing. A very frictions reduces the desirability of large number of people looking, very adding workers. That’s especially true if Region: Why really do we need a com - low job-finding rate for each individual, there’s anything about the mittee, a dating committee, rather than but the product—the number of jobs relationship that has an investment relying on a rule of some sort, like two filled—is roughly a constant. It’s a very character. If a worker has to be trained quarters of negative GDP growth? I think important fact about the labor market. and becomes highly productive labor you’ve been on the committee since it Think about a slack market from an in time, then this question of what the began… employer’s point of view. They see there cost of funds is becomes important. A are all kinds of highly qualified people rise in the cost of funds will result in a Hall: I’m the only chairman the commit - out there they can hire easily, so they decline in employment, and that’s tee has ever had, for 32 years. don’t need to do a lot of recruiting— something a lot of people are looking at people are pounding on the door. right now. Region: I didn’t know you’ve chaired it There are many threads to this topic. the entire time! Well then, you’re the Region: And these days they’re census We’re debating actively which ones are right man to ask. Do you think it might takers. most important. be useful for the NBER, in addition to doing what it now does, to also issue Hall: [Laughter] Right! So that’s the first RECESSIONS AND RECESSION something closer to a real-time indica - thing to think about: job creation incen - DATING tor or signal of recessions—that could tives. be revised for false positives or nega - If you ask, how did we get into a situ - Region: People are wondering when will, tives, along the lines that Òscar Jordà ation where job creation incentives have or did, the current recession end, but I’d has recommended? declined? It’s that there’s been a decline like to ask how you and the NBER in the profitability of hiring a worker [National Bureau of Economic Hall: I think we feel that doing some - without a corresponding decline in the Research] committee you lead decided thing like that, and in any sense making

JUNE 2010 34 The Region

it official, would somewhat cloud things pened in ’01. By the alternative measure because there would be enough type 1 of total output, real gross national or type 2 errors [false positives or nega - income, the 2001 recession is quite tives]. We’re very happy to see that type apparent. of research be done; we don’t claim any To me, it’s not an issue because that’s monopoly on this point, and it’s been just looking at GDP. If we look at very instructive. employment, as I did in a 2007 Actually, long ago, in the 1980s, we Brookings paper on the “Modern sponsored a project that informally, Recession,”—by “modern recession” I unofficially put out a recession proba - mean one in which productivity rises… bility index that Jim Stock and Mark Watson prepared. It didn’t work very Region: And monetary policy is under well in the 1991 recession, so they control. stopped doing it after that. And it didn’t work for fairly typical Hall: Monetary policy is stable, exactly. reasons. That was the first recession that But here I think the key point is about wasn’t accompanied by a decline in pro - productivity. With rising productivity in ductivity, so it looked somewhat differ - a recession, you can see a relatively mild ent. So their historical relationships Historical relationships weren’t as stable movement in GDP, and there’s a long weren’t as stable as they hoped. as ... hoped. That’s one of the main rea - period of GDP growth at the same time That’s one of the main reasons why sons why automatic rules haven’t worked. that employment is falling. automatic rules haven’t worked. People ... It’s a classification problem that ... has a When people talk about the jobless have done research on the machine recovery, it’s just another term for pro - shifting structure, and dealing with the approach for years. In fact, when I was a ductivity growth. That’s complicated the graduate student and took a computer shifting structure is the issue. We try very process. The complication in the 2001 science course, my project was to write hard to achieve historical continuity. recession is that productivity rose software that would automate this. So enough to offset employment declines, it’s not a new idea. But it’s never worked The complication in the 2001 recession so we have a very pronounced, obvious very well. is that productivity rose enough to offset recession in employment and what’s employment declines, so we have a very hardly a recession at all in GDP. Region: It would have missed the 1981 pronounced, obvious recession in employ - I’m perfectly satisfied that’s a reces - recession if we’d used the two negative ment and what’s hardly a recession at all sion because I want to balance the two. GDP quarters rule. To the extent you look just at GDP, in GDP. ... That’s material today because, though, it would be hard to call that a of course, we’re seeing the same thing. Hall: You mean 1980. recession. That’s material today because, of Region: Right, 1980. don’t think anyone else does either— course, we’re seeing the same thing. that we know when there’s a recession. GDP reached a very pronounced trough Hall: 1981 was no problem. The 1980 In all the data we look at, certainly in in the summer of 2009. It’s been pretty recession was just one quarter. And peo - the period when we’ve had reliable consistently rising—with one little hic - ple have said that the 1980 recession was data, which is since World War II, cup recently—since then. So on that actually just sort of a prelude to the ’81 there’s never been an episode that’s standard, we say the trough was in sum - recession. We say no, but it’s been said. somewhere halfway between a reces - mer of ’09 or maybe the fall of ’09. sion and a nonrecession. Every reces - But employment is still declining. We Region: It seems it’s more of an art than sion has been clear. And they all see still have not seen a growth month. a science then. unemployment shoot up and typically Everyone is presuming that we will in see GDP decline. March—but that’ll be the first. You can Hall: It’s a classification problem that the We do face issues though. With the plausibly make the trough of GDP be in world seems to want an answer to, but it most recent revisions of GDP, the 2001 June of ’09, but the trough of employ - has a shifting structure, and dealing recession essentially doesn’t exist. It was ment is probably going to be March of with the shifting structure is the issue. a flattening, but as emphasized on our 2010. That’s a long time. We try very hard to achieve historical Web site, there are issues of depth, dura - Not as bad as ’01, when the situation continuity. tion and dispersion, but there was nei - was even worse. The trough in employ - We don’t doubt for a second—and I ther depth nor duration in what hap - ment didn’t occur until 2003.

35 JUNE 2010 The Region

STOCK MARKET VALUATION 1990. eCapital turned positive in 1990. Hall: Flora Hewlett, married to the So during that period, there was some Hewlett of Hewlett-Packard, was my Region: Let me ask about the stock mar - undervaluation. It was very clear the father’s secretary when he was a ket. Roughly a decade ago, you did a lot stock market later decided it was an Stanford professor. If only he’d bought of work on eCapital, eMarkets and stock undervaluation because if you made a one share! market valuation. Your 2001 Richard Ely stock market investment in 1980 and lecture was an example of that. And you held it to 1999, you had a very large Region: You’ve also devoted some time suggested that investors did seem to be excess return in the 20-year period. So I to studying antitrust economics, and fairly estimating the market’s value if think there are still some mysteries. looking at potential for monopoly pric - intangible capital was taken into In spite of the fact that the valuation ing in upstream supplier markets. account. Is that accurate? that we see in the market right now What is your view of the argument seems to be in a reasonable range, the that intellectual property, copyright laws Hall: Well, I talked about some individ - returns since 1999 have been way below and patents inhibit rather than encour - ual cases where I thought you could tell any benchmarks, which suggests that age innovation? the story. There was a discussion of eBay there was some overvaluation then. in the Ely lecture. On the other hand, if Hall: First of all, I think that that’s only you look at the results in my AER INTELLECTUAL PROPERTY been directed at patents. I don’t think [American Economic Review ] paper, it there’s any feature of copyright law. It observes that intangible capital by that Region: You’ve thought and written a protects the expression. There’s an infi - measure was deeply negative in the mid- great deal, in both technical and lay nite space of melodies that composers ’70s to about 1980. Now, positive publications, about the economics of can compose and once they do, it doesn’t eCapital makes a lot of sense, but nega - computers and software, as well as ven - inhibit other composers from compos - tive eCapital is a little hard to swallow. ture capital and entrepreneurs. That ing other songs because there’s this infi - So I’d be careful. seems natural given that you were born nite space. Every expression is complete - There was something weighing down in Palo Alto and have worked here for a ly unique, so when it comes to expres - the stock market from basically 1974 to long time. sion, I don’t think there’s any real issue. I

More About Robert Hall

Current Positions Honors and Awards Robert and Carole McNeil Joint Senior Fellow, , Fellow, American Academy of Arts and Sciences, Econometric Society and Professor, Department of Economics, ; joined and Society of Labor Economists faculty in 1978 Hall of Fame, Money magazine, with co-author Alvin Rabushka for their book The Flat Tax Previous Positions Massachusetts Institute of Technology, Professor, 1970–78 Publications University of California, Berkeley, Assistant Professor, 1967–70 Author of Forward-Looking Decision Making: Dynamic Programming Models Applied to Health, Risk, Employment, and Financial Stability Professional Affiliations (Princeton University Press, 2010), Digital Dealing: How eMarkets Are Transforming the Economy (Norton, 2002), The Flat Tax (with Alvin President, American Economic Association; President-elect, 2009; Rabushka, Hoover Institution Press, 2d ed., 1995) and The Rational Vice President, 2005; Ely Lecturer, 2001 Consumer: Theory and Evidence (MIT Press, 1990), among other books. Director, Research Program on Economic Fluctuations and Growth, Widely published in academic journals, with research focused on National Bureau of Economic Research, since 1977; Chairman, employment, technology, competition and economic policy Committee on Business Cycle Dating Education Member, National Academy of Sciences, since 2004 Member, Advisory Committee, Congressional Budget Office, since 1993 Massachusetts Institute of Technology, Ph.D. in economics, 1967 Member, Oversight Panel for Economics, National Science Foundation, University of California, Berkeley, B.A. in economics, 1964 1989; Advisory Panel for Economics, 1970–72

JUNE 2010 36 The Region

caused any problems. On the other would be recapitalized well before they hand, the patent system has generated became dangerously close to collapse. some very substantial rewards to some We watched in frustration as the regula - true innovations. tors failed to take that action, even You know, it’s all in the details. I don’t though they had promised they would. accept a broad condemnation of the Second, did macroeconomists fail to patent system. I don’t join any of these understand that financial collapse people who say there shouldn’t be any would result in deep recession? Not at business process patents or there all. A complete analysis of that exact shouldn’t be software patents. Some issue appears in an extremely well- good ideas are implemented in software. known and respected chapter in the What is a good idea, and what every - Handbook of in 1999, one stands by, I think, is the notion that written by , Mark Gertler patents shouldn’t last forever. The idea and Simon Gilchrist. Depletion of the of a finite patent life, which is currently capital of financial institutions raises around 20 years, does seem to be an financial frictions to levels that distinct - important part of the design. ly impede economic activity. In particu - The result of that is that the great lar, credit-dependent spending on plant, Patents shouldn’t last forever. The idea of majority of innovations ultimately ben - equipment, inventories, housing and a finite patent life, which is currently efit workers in the form of higher wages consumer durables collapses. That around 20 years, does seem to be an rather than any permanent stream of chapter is an excellent guide to the important part of the design. monopoly profits going to owners. If depth of the current recession. The result of that is that the great that weren’t true, you’d see a huge amount of innovation value capitalized Region: Thank you for a great conversa - majority of innovations ultimately benefit in the stock market, but you don’t, and tion. workers in the form of higher wages rather that’s proof. Consistent productivity —Douglas Clement than any permanent stream of monopoly growth and corresponding real wage March 16, 2010 profits going to owners. growth is demonstration then that the benefits ultimately of innovation are think almost everyone believes in a going to workers. So it’s a great thing. pretty powerful IP rights regime for expression. THE STATE OF ECONOMICS When it comes to the things that patents protect, then the patent regime Region: The past few years seem to have has to do the things that the patent brought about a crisis of confidence in regime claims to do. The patent has to the economics profession, with critics be original; it has to be an innovation. suggesting that macroeconomics has And there the standard of obviousness failed in some fundamental way. It’s a comes in. topic addressed by [Minneapolis Fed If what’s happening is that people are President] Narayana Kocherlakota in somehow able to figure out what the our Annual Report this year. Do you obvious next logical step is and some - agree that the macro profession failed how get a patent on that and then collect the nation during the financial crisis? royalties from that patent even though it doesn’t really make any contribution, Hall: I don’t. There are two parts to the then there’s something wrong with the issue. First, did macroeconomists fail to patent regime. But I don’t think there’s understand that a highly levered finan - any very good evidence that that’s actu - cial system based in large part on real- ally what’s happened. estate debt was vulnerable to a decline People make fun of a lot of the in real-estate ? No way. Many of us patents that the patent office issues, but pointed out the danger of thinly capital - they don’t matter. There’s only a much ized banks. We had enthusiastically smaller set of patents that have ever backed the idea of prompt corrective attempted to be enforced and have action in bank regulation, so that banks

37 JUNE 2010