Nov 20, 2017

Nick Bloom on Economic Uncertainty and the Productivity Slowdown

Productivity has seen a steady decline over the last few decades, and a lack of innovation paired with ineffective management practices may be to blame.
David Beckworth Senior Research Fellow , Nicholas Bloom

Hosted by David Beckworth of the Mercatus Center, Macro Musings is a new podcast which pulls back the curtain on the important macroeconomic issues of the past, present, and future.

Nicholas Bloom is a professor of economics at Stanford University and is the co-director of the Productivity, Innovation and Entrepreneurship program at the National Bureau of Economic Research. Nick joins Macro Musings to discuss his work on the causes and effects of economic uncertainty as well as how to measure uncertainty in an economy. David and Nick also discuss why productivity has slowed down in recent decades and why Nick is not especially optimistic that productivity will really improve anytime soon.

Read the full episode transcript:

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to macromusings@mercatus.gmu.edu.

David Beckworth: Nick, welcome to the show.

Nick Bloom: Hi Dave.

Beckworth: Oh, glad to have you on. We have read your work, and we'd love to get into it in a moment. Before we do that though, with all my guests, I like to ask them how did you get into economics?

Bloom: I don't really have the most direct story here. I was actually intending to go work in investment banking, or certainly make a lot of money. I did well as an undergrad. I got first in Cambridge, which meant I was enabled to get funding to do a master's in Oxford.

Bloom: At the end of that, I actually applied to a bunch of investment banking jobs, consulting, even an IT help desk, which is insane when I look back on it. And the interview that was by far and away the best was actually with the Institute of Fiscal Studies.

Bloom: The IFS is a research think tank in London. It was run by Richard Blundell and Andrew Dilnot, and it's part of UCL. It was just a fantastic interview. They talked about economics, it was fun. It was super interesting.

Bloom: In fact, back in those days, that was 1996, the pay gap versus finance wasn't that big. Whereas now, it would be huge. Anyway, back then, it wasn't a big pay cut. And I thought, "Great." I went to the IFS, and then once I was there, I started doing a PhD part time at UCL-advised by John [inaudible], and just kind of fell into economics. From that point onwards, I just loved it and I stuck with it.

Beckworth: Oh, very interesting. You're now at Stanford University, and you are pretty prolific. I was going through your work in preparing for the show. You have a lot of articles, lot of research done. One of the areas that you're well-known for, is this area of uncertainty, economic uncertainty.

Beckworth: You've created an uncertainty index, a policy uncertainty index. We'll get to that in a minute, but let's jump into this conversation about uncertainty. Let's first maybe define what is uncertainty. How would you define it, and how should we think about in our conversation today?

What is Uncertainty and How Does it Apply Today?

Bloom: It's a great question. There's two ways to think about it. One is the layman's definition, or if you read Wall Street Journal or the Economist. I talk to people all the time. Uncertainty is not knowing the future. You might say, "Hey, look. If I flip a coin, I'm going to be uncertain about what happens."

Bloom: There's also a more formal economics definition, which is slightly different. Frank Knight, the famous Chicago Economist, actually defined uncertainty is what we'll think of might be call it as Knightian Uncertainty.

Bloom: He would say flipping a coin is risk. You have a known probability distribution. You have 50% heads, 50% tails. Whereas, uncertainty is something of which you really have no sense of the probability distribution at all.

Bloom: An example might be the number of coins ever produced by mankind. I really have no idea how to assign a probability to that. Formally, that's uncertainty. Some people called that Knightian Uncertainty. Others kind of call it ambiguity.

Bloom: It's a bit of a mess, but certainly the [inaudible] concept is not knowing each one comes whether you know the distribution or whether you're even uncertain about the distribution is a more fine-tuned definitional difference.

Beckworth: Okay, but for our purposes of discussion today, uncertainty, it's kind of a mix of those two ideas. It's maybe a little more unclear. Is that right?

Bloom: Yes. Economists like to pick hairs over these two concepts, but in practice, I think reality is somewhere in between. You may look at options markets and think, "Well, look. The financial market has options traded all the time," which means people are calculating distributions about various states of the future and trading on it.

Bloom: So, in that sense, you think everyone has a pretty good idea of the distribution of future outcomes. You may then say, "Well, let's think of something that is a period of Knightian Uncertainty." Because I remember when I was starting my PhD, I was thinking a lot about, well, halfway through my PhD, anyway, 9/11.

Bloom: So 9/11, the planes hit the twin towers, markets closed down. We have no idea what's going on. Initially, actually Al-Qaeda took no responsibility for it. Nobody actually claimed the attacks for a while, and there's massive confusion.

Bloom: So, that you think, was a period of pure Knightian Uncertainty. We've never seen anything like that before. We have no idea what might happen. But at the same time, while American financial markets were closed down. European option exchanges were still trading.

Bloom: And you could buy options on the S&P500. So, it tells you that even in periods of seemingly immense Knightian Uncertainty, there are still people prepared to put probabilities on events. So, I think most people would actually put probabilities on most events, and so pure Knightian Uncertainty is actually a pretty rare occurrence.

Beckworth: Well, this has become a hot topic. Obviously, the past decade we had the Great Recession. You've done a lot of research on this. The Federal Reserve has talked about it. The IMF has talked about it. You mention it in your paper.

Beckworth: What got you into this area of uncertainty? Was it the 9/11 episode, or was there anything else that really prompted you along this path?

Bloom: I was the classic grad student. I was burrowing away on extending a paper. This was at 2006 paper in the Review of Economics Studies by Russell Cooper and John Haltiwanger estimating adjustment costs.

Bloom: It's a great paper, and it's well-cited, and I was trying to extend it from capital to capital and labor, which, ultimately, is not that interesting an extension. I was using some mathematical tricks to do it.

Bloom: I was so excited by this, and I start talking to anyone about adjustment costs, and I could see their eyelids starting to droop, and they're struggling to stay awake.

Beckworth: Nice.

Bloom: At some point, I thought, "Who cares?" Ultimately, my extension is not going to change the world, but ugly enough, while I was fiddling around with estimating labor and capital adjustment costs, at some point, I had this kind of eureka moment that if uncertainty was time varying, suddenly these things mattered a lot.

Bloom: I remember going to a conference, then the MBBR Summer Institute presenting it's paper, and about a month before I presented, I managed to figure out and do some basic simulations. And what happened if uncertainty changed over time.

Bloom: And suddenly figured out the look, actually, it will be very powerful. I hadn't read Ben Bernanke's 83 paper, it was kind of doing very much the same thing. So, I was operating in a sense in parallel, but later than him. But I realize that time variation, uncertainty generates powerful effects.

Bloom: And then, in that presentation, Russ Cooper asked me, "Does uncertainty change over time?" And again, I was just finishing up my PhD at the time and I kind of thought, I don't know, I've never looked. And then I was kind of great question. So, from that point onwards, which is the kind of late 90s, I started to look at, "Hey, how can we measure uncertainty and does it vary over time?"

Bloom: And at that point, the best measure of uncertainty was implied volatility on the S&P 500, what's called the VIX, is that kind of a measure of how volatile stock markets will be over the next month. And when I looked at that index, it was incredible. It just, every time there's a recession or a nasty event, it spiked up.

Bloom: So, going back to, for example, the assassination of Jeff Pay. So, back then we didn't have the VIX. I just looked at realized stock market volatility, stock market surge, and were bouncing all over the place. The Cuban missile crisis, OPEC one, OPEC two, Gulf war one. Basically every time there's a nasty event, uncertainty surged upwards.

Bloom: And from that point onwards, I began to focus heavily on... I look recessions and bad events seem denying any carry bad news. That's kind of the first moment or negative PFP or demand shots. They also seem to be associated with big spikes and uncertainty. And maybe it's the second that's causing some of the drops in activity.

Beckworth: And fortunately for you, the Great Recession came along, and really spike interest in this. Unfortunate for humanity, but fortunate for you. The Great Eecession really gave this a research agenda, a new lease on life.

Beckworth: And you had a 2009 and kind of metric article, I believe that looked at some of these issues and in subsequent spate of articles following that, let's go to some of these measures of uncertainty. You mentioned one already, one of the macro measures you, you listed in your papers, volatility of stocks.

Beckworth: You mentioned that the VIX, but there's some other ones. Tell us about bond markets, exchange rates and GDP. Can those also be used to get a sense of uncertainty?

Uncertainty in Bond Markets, Exchange Rates, and GDP

Bloom: Yes. There's a range of different measures and you can break them up into kind of micro or macro and forward looking and realized. So, at least talk at the macro level. Realized is things that actually see there's more volatility in recession.

Bloom: So, for example, you look at GDP growth rates, mostly GDP is growing upwards by two, 3% a year. But suddenly when recessions happen, it plummets downwards. And so, that generates, if you run Gotch models, that generates a big spike in volatility.

Bloom: If you look at industrial production, it depends on bounce up and down much on recession, you look at exchange rates, they tend to bounce up and down when recessions bond prices, stock market prices. So, all of those realized measures of volatility spike up in recession.

Bloom: You can also look at forward looking measures. For example, implied volatility in the stock market, the VIX index, which is being called to theorem index, which is backed out from kind of put and cool options and it looks at expected stock ball that goes up.

Bloom: If you can look at forecast the disagreement. So, if you use things like the separate professional forecasts is from the Philadelphia Federal Reserve Bank. They poll about 50 forecasters. Each I can't remember. Is it each month or each quarter, but whatever it is. You see there's much more disagreement in recessions.

Bloom: So, a normal periods, like now it's kind of, most people are predicting two and a half percent growth next year in 2008, 2009 it was wild disagreement. So those are macro measures.

Bloom: Another one I should say I've worked on recent is even newspapers. I've worked under the economic policy uncertainty index where we just do ultimated scrape of newspapers of the number of articles that appear to mention the words economic or economy uncertain or uncertainty.

Bloom: And one of the basic sets of policy worth like Congress regulation of White House and those index, the frequency of those articles also surges massively in recession.

Beckworth: And that data is available online. Your uncertainty index, correct?

Bloom: Yes. The economic and public policy uncertainty index is available online. We have a website, www.policyuncertainity.com partly that website is just, I should step back a minute. You are in partly right?

Bloom: I was extremely lucky for my PhD research in the sense of the climbing of the great recession. I unfortunately had not bought a house and had some, I'd worked in McKinsey for bit, so I had some savings and for my pension in the stock market and that plummeted.

Bloom: They're all in actually Sterling had moved to the U.S. But all my savings when they the London Stock Exchange and that fell even more because the pound crashed on the financial side. It was a disaster.

Bloom: But on the my career human capital side, it was great because what happened, it was the great recession was obviously a huge recession, but it was accompanied by an unprecedented run up in every measure of uncertainty.

Bloom: And particularly off to the Great Recession. 2009, 2010 I was asked a lot about policy uncertainty, the impact of top and the pretty radical policies that were going on. And so I started working with Steve Davis and Scott Baker to try and measure it.

Bloom: And we started very small, a scraping a lot of newspaper articles because it seemed like really the only way to come up with a measure. And I'd say initially it was a very unambitious project and as we continued to work in that we got more and more interest and it was not high tech, it's super low tech, but it just appeared to kind of a pitch design guys in the sense of it's actually very hard to measure uncertainty.

Bloom: This isn't a perfect measure, but it seems to be not if it was about as good as the other measures out there. So, we put this thing out and we got a lot of pick up from financial sector from the fed, Bloomberg carries the data now. So, it's kind of in the last two, three years it's really taken off. It's just another measure of uncertainty.

Beckworth: And people are using it in their research, right? They're using it in their models and their statistical analysis. Correct?

Bloom: Yes. It's been used. So, the basic issue is you want to measure uncertainty. You want to say you're the fed or research or even a hedge fund. Early on, a lot of the people actually were using it were hedge funds because they wanted to measure the credit spread.

Bloom: So the difference in the interest rate on AAA versus say BBB bonds. So, how much do you pay for the risk of having holding a BBB bond, how much extra interest rate and traditionally they'd use the VIX.

Bloom: So, the VIX is a measure of implied volatility on the stock market. Historically it worked very well. But, post 2009, I think in 2010 the VIX and stock market volatility in general has been trending down and down.

Bloom: And in fact about three weeks ago, so in early October, 2017, the VIX hit an all-time low, nine. So, stock markets are incredibly quiescent and therefore all the measures of uncertainty based on stock market indicators are incredibly low right now.

Bloom: People don't think the uncertainty is low. I mean we have Trump and Brexit and North Korea and all kinds of things. And our policy uncertainty index is actually moderately high.

Bloom: It's not really high, but it's certainly not low. And credit spreads are also about medium. So, what's happened is a lot of hedge funds talked to them have said they'd been using policy uncertainty is another measure of-

Beckworth: Interesting.

Bloom: ... to kind of get a market risk. And it's been, so I really don't want to claim it's a perfect measure. It's based in newspapers, [inaudible] there's all kinds of problems with it.

Beckworth: We'll get to those in a minute. Some of the endogeneity issues. But I'm just curious off the cuff, I'm sitting here listening to you. It strikes me that measure would be useful and trying to maybe explain some of the term premium movements on say 10 year treasury yields, and the central banks had been active in doing large scale asset purchases and they believe that they're having some effect on term premiums.

Beckworth: But I was looking on your webpage, you have a global policy uncertainty index and I imagine something like that, it’d be interesting to see a steady kind of connecting that to the term premiums on long-term treasury bonds. Has anyone done anything like that?

Bloom: You are exactly right. So, struggling to remember who, geez. Her name is [inaudible] but I was at the IMF about three weeks ago and somebody sent me a paper looking at, so an IMF paper looking at term premiums and in fact swaptions which is a kind of complicated measure.

Bloom: It's basically the risk and the term premium and that turns out to be incredibly correlated with our policy uncertainty index, as does in fact the longer end of the implied volatility curve.

Bloom: So just to explain it's a bit more detail. So, the VIX index is an index put out by the Chicago board of options exchange, which is they put out a number everyday day, which is the implied volatility on the next 30 days.

Bloom: The S&P 500 index, when is high it means that the market thinks the index is going to be very volatile and this low, they think it's going to be pretty flat and that thing is very low right now. But there's also a VIX curve, Vol curve which is those from 30 days all the way up to 10 years.

Bloom: And I've got hold of that data from some contacts and Goldman Sachs, you can do it yourself, but looking at that nature it turns out policy and uncertainly is much more correlated with the long end of the curve.

Bloom: So, exactly as you say, a lot of the big policy uncertainty issues right now. For example, Trump's effect on free trade, on tax reform and social security, more generally on free markets stability, the political system, war with North Korea, the breakup of the European Union.

Bloom: Not all of these issues are huge issues in the long run, but they're not very likely to be resolved 30 days from now. So, you're precisely at long run focus measures of uncertainty, unlike the long end of the yield curve or the long end of the Vol curve are actually much more correlated with policy uncertainty. The near term measures.

Beckworth: Yeah, that makes complete sense. And it ties into something I've looked at recently and I've been dabbling around with and that is the safe asset shortage problem.

Beckworth: Getting a little ahead of myself here in an interview, but part of that discussion is if you look at long-term government debt yields across the world and countries that we consider safe assets, safe harbor such as Germany, the United States, United Kingdom their long-term treasury yields have been going down persistently.

Beckworth: Above and beyond that would have been implied from the QE programs, and one of the stories is there's increased risk appetite. Excuse me, risk aversion and people are clamoring for safe assets like treasuries and this uncertainty index would be a great way to flush out that story. And from what you're telling me, it does flush it up very nicely.

Bloom: Yes, exactly. As you know I'm an avid listening to your podcast so I've heard you discussed this before. Yeah, very much. I think if the policy uncertainty index, it has no precise time horizon because it's effectively how much there seems to be discussion of policy uncertainty in the newspapers.

Bloom: You can imagine most newspaper paper articles aren't really the newspapers are looking at, should have in mind that they're the biggest pen in the U.S. So the Wall Street Journal but also the New York Times, The USA Today, Chicago Tribune, Washington Post, etc.

Bloom: So, they are highbrow but they are not financial focus and they are talking about mainstream policy issues like healthcare reform and healthcare reform or social security reform is a big long run issue. The federal government is effectively bankrupt. They were looking 20 years out and something has to be done but we know that none of that's going to happen in the next 30 day.

Beckworth: All right. There's some other measures, but let's move on for the sake of time because I want to get to your productivity slow down discussion as well. Let's move to why uncertainty might be important.

Beckworth: So, you list two channels. I'm drawing from your Journal of Economic Perspectives paper in 2014, it's called *Fluctuations in Uncertainty.* We'll provide a link to that and some of those other articles on the SoundCloud webpage for the podcast.

Beckworth: But in that article you present two channels, two transmission mechanisms through which uncertainty could have real effects on the economy. And I take it independent and beyond and above kind of other forces at work. So, the first one I want to turn to is kind of a risk aversion risk premium there. So tell us how uncertainty would work through that to effect the real economy.

Risk Aversion and Risk Premia

Bloom: So, great question. Why uncertainty math is, in order for uncertainty to math it in models of behavior, what we need is curvature. So, when things are linear, you have what's called certainty equivalents. You don't mind whether you're uncertain or not.

Bloom: And as soon as you have curvature, uncertainty begins to play a role. And there are two places economists typically include curvature and they think about the economy. One is on the household side, which is consumers have concave utility.

Bloom: Each dollar they get, they value less than the last one. And therefore, like security, they hate risks. This is a very old concept in economics. It goes back at least to the Nobel prize winner, James Tobin who was at Yale, Haynes talk about it.

Bloom: The idea here is if risks goes up, if you're more on certain consumers will save more and consume less and that will generate, that drop in consumption at least in an open economy would generate a drop in output.

Beckworth: You also have an uncertainty story tied to risk aversion through businesses as well, right? You mentioned if uncertainty goes up, risk premiums go up, firms want to hoard more cash less likely to invest in capital spending. Is that another angle there?

Bloom: So, that's a more finessed story, but that was actually, looks like it was an important issue around the great recession, which is imagine, basic models.

Bloom: We don't think firms are risk averse. We think firms are owned by diffuse shareholders and they basically, but in practice of course firms are controlled in large part by their managers. This is a corporate governance issue and managers of course certainly don't want to load up on masses of risks.

Bloom: And the last thing they want to do is discover the firm bankrupt. So, there's plenty of evidence and I have a I paper is RG Lynn, Ivana Firo looking at their school. The finance and certainty multiply that when uncertainty goes up, firms hoard cash.

Bloom: And again, this isn't a new finding, it's gone back to decades in the financial literature, but it seems to take on particular relevance now, if you're uncertain, I cut back in investment, I cut back on paying out dividends and I stash cash because I don't want to be left high and dry.

Bloom: So, that's another way for how uncertainty can slow down the economy. It does increase risk premium, but it also leads firms holding risk beam and constantly to stash cash.

Bloom: So, you've got multiple negative effects of risks. Consumers spend less, the risk premium goes up to firms invest less. And also they stash cash. The only kind of thing to push back slightly is go to the capital enclosed economies.

Bloom: This is becomes more complicated because imagine consumers say, consume less. They save more. What we know that in a closed economy, savings equals investment. So, it's less clear how this affects growth in a big economy like the U.S.

Bloom: It's very clear that risk is very damaging for small firm, for mall economies. So, there's a great paper by group in the AAR a few years ago that showing that in small open economies there's South American countries, when risk goes up the domestic economy tanks.

Bloom: The story is basically the money flies out of the country that they look at, for example, Ecuador. And if you're in Ecuador savings definitely doesn't equal investment. Interest goes up. You can imagine, if I'm Ecuadorian, the last thing I want to do, because there's an unstable government is save at home.

Bloom: So, all the money flies out of the country in the kind of free collapses. In the U.S., the risk goes up. Actually there's often was full the fight quality that people save more. But that savings can end up in investments.

Bloom: So, that risk aversion channel, it's very powerful and negative in smaller open economies, in logic closed economies which are more like the U.S. To the whole of Europe it's less obvious actually whether it's negative.

Beckworth: Ben I read that in your articles thinking about it. But one point you do make in the article, it does become very prominent and severe when you hit the zero lower bound. Right? Because at that point you've got a rigidity that it's preventing that market from clearing.

Beckworth: I mean, the story you're telling is you increase savings and interest rates should dropped at some point to where the market clears. But if you get stuck in the zero lower bound, if there's some kind of price rigidity, then all bets are off.

Bloom: That's exactly right. So, another way you can have powerful effects are from big increases and uncertainties with the zero lower bound. And there's a couple of papers on this and it is, I'm blanking on the exact coworker's names, but they show precisely that when uncertainty goes up, you'd like interest rates to drop.

Bloom: So, that savings equals investment. They're jammed at zero. That doesn't happen and you get a nasty recession. And another way is actually prices don't adjust. So, there's a paper by Basu and Bundick in the AAR, the argues, when on uncertainty goes up, savers, consumers want to consume less and save more and prices drop.

Bloom: But if you have sticky prices as you do in these new Keynesian models, again, markets don't clear and you can get drops in output. So, uncertainty through risk aversion can be a powerful negative demand shock. But you need to go beyond the kind of classic linear rewrite business psychomotor and include some of the rigidities we see more recent models, that the Keynesian models or the ZOB. And you can then get very powerful negative drops down in that.

Beckworth: Yeah, I was just reading this, I also was thinking along the lines of what John Cochrane and Roger Farmer have recently advocated. So, John Cochrane, I remember he had a paper in the journal of finance when he was president where he kind of, I don't know if he introduced, but maybe he reintroduced this idea that discount rate theory of the business cycle.

Beckworth: That everything's about the change in the discount rate and the look at the present value of future resources and you're discounting it. And his story is, business cycles are driven by sudden changes in discount rates. But as I look at it, it's not all that different than the certainty shocks.

Beckworth: In my mind anyhow, there's some connection there. I also, we've had him on the show. We also had Roger Farmer on this show and he's a Keynesian, but he is very rigorous with his work. He has what's called a belief function in his model where he's trying to more carefully captures idea of animal spirits.

Beckworth: But both of these, they seem related and they seem related to your idea of uncertainty, right? If there's a sudden uncertainty, shock is going to affect the discount rate is going to affect animal spirits. So, I think there's some overlap. Am I reaching here? Do you agree?

Bloom: No. I very much agree. Interestingly enough, there's a long been a discussion in the literature over the price of risk and the quantity of risk. So in fact, I was discussing it with my graduate students here. But, I always focus on the quantity of risk.

Bloom: The amount of uncertainty going up in our recession. And I think it's almost certainly the case that happens. In fact, I missed earlier discussing, there's also lots of micro evidence, so you look at firm volatility and individual volatility.

Bloom: The other story is the price of risk, which is the risk premium goes up and recessions. And there's also lots of evidence for that. And that's more the John Cochran line. And I I'm pretty sure, I mean, we've seen the data, both happens, both more risk and it's more expensive as relating to animal spirits and Roger Obama's work.

Bloom: I like his work very much and yes, I would see this is a way to meld these together. For example, Martin Schneider and the mics department here at Stanford as work in ambiguity aversion. So, people don't like risks and pessimistic and of course if risk goes up. They become much more pessimistic.

Bloom: And so the kind of intersection of this work with behavioral, generate some pretty powerful effects. If you dislike risks, if you're pessimistic, if you have kind of animal spirit beliefs, you can easily find that upticks and uncertainty generate pretty strong negative effects on investment demand on consumption and therefore drops outcome.

Beckworth: Okay, let's move on to the other transmission mechanism. The other channel. So, we just got done discussing the risk aversion risk premia channel. You also mentioned the real option theory channel. So, explain that to our listeners. How does uncertainly work through it?

The Real Option Theory Channel

Bloom: Yes, this is what I initially focused on in my PhD. I had a motto which is it goes back, I'll come back to this later because I hadn't entirely read the literature.

Bloom: You generated a bit of a heart attack from it. Some point in my job market, but there is a literature Dixit Pindyck, the best known for this literature showing the real options, effective uncertainties.

Bloom: So, the idea here is if it's expensive to buy a piece of equipment and then sell it, or it's expensive to hire or work or train them up and then fire when your uncertainty tend to force.

Bloom: So, on the consumer side, you can think of an equivalent version of you're uncertain as to whether you're going to keep your job. The last thing you do is go out and buy a new car or buy a new refrigerator. This is cool.

Bloom: The real option because you think of before you've gone out and bought a farmer has gone out and bought a new piece of equipment or hire the new work, it has the option to take that powering or investment decision. Once it's done, it loses the option.

Bloom: And as you know from finance, options are more valuable when uncertainty is higher. Therefore, when uncertainty is high, it's more valuable to weight. And therefore, lots of firms wait, you got to go drop in activity.

Bloom: So, the real options model is very simple. It says, when uncertainty goes up firms are more cautious. They pause hiring, they pause investment, consumers are more cautious. They pause, particularly spending on consumer durables and that causes a recession.

Beckworth: Yeah. In a minute, we're going to talk about the productivity slowdown, but I'll just have to ask right here and now when I read that part of your paper, I instantly thought of the productivity slow down. Right?

Beckworth: If firms aren't doing as much capital spending, investment spending is declining, that has to play at some degree into a potential productivity. Slow down. Yes.

Bloom: Yes and no. So, I well before I go any further, I should, if you who attribution to the person, actually when I did, I came up to give a job pool at Stanford and I was talking to Bob Moore, and Bob Moore, I was explaining my job market paper and Bob Moore said, "Hey bro, we've heard this all before."

Bloom: That was a paper by Ben Bernanke that point I panicked and Bob says, but go on. Go on. And obviously I got the job but as soon as I left Bob's office, I basically as soon as I had a break I ran it downloaded Ben Bernanke.

Beckworth: That's not what you want to hear in a job talk presentation.

Bloom: No, it's terrifying. It's been done for the worst, but it turned out it was right. It was totally right. In fact, away, it was very nice for me. So Ben Bernanke, his PhD, the key paper out of his PhD, and I believe his job market paper was on exactly this.

Bloom: He published it in 1983 and a TJ and it talks about how an increase in uncertainty can lead through a temporary dropping out but then a rebound and he had a very nice stylus model in a sense I was doing this quantitatively and bringing data.

Bloom: So, then to come to your question, both my paper and Ben Bernanke's paper show the same style as fact, which is a rise in uncertainty, generates us a pause in activity when uncertainty drops back down again, activity resumes.

Bloom: You can kind of think of it like a stream, the dam, the stream and surely you stop the stream but all the water building up behind the dam and once you remove the dam all flood back down.

Bloom: So, in terms of linking up to the product, you can just slow down and my paper and other work like this, Bernanke much other work, you can have a temporary, quite negative effect of uncertainty in productivity. In order to have a long effect, it needs to be uncertainties.

Bloom: Iron actually probably is rising over time. So, my own view is it probably had quite damaging effects in 2008, 2009, 2010, it would have quashed a lot of R&D projects and investment by firms, but I don't think it can play a role in driving the secular decade by decades and in productivity.

Empirical Importance of Uncertainty

Beckworth: Well, let's move on then to empirical estimates of the importance of this issue. So, you've mentioned some studies that have gone out and I'm going to recall one from your journal of economic perspective paper and this was published in 2014 so maybe things have changed, but you mentioned you estimate roughly like a 9% GDP loss.

Beckworth: That was both like a 3% actual decline in real GDP and then kind of GDP not growing on its trend path. And if I'm correct, you attributed about 3% of that 9% loss and GDP to uncertainty. Is that right?

Bloom: Yeah, those are the numbers in the paper. They have pretty big standard errors around them. The uncertainty dropped. I think it was the things on the run, on the good and the bad side in that number. Obviously the bad side, it comes with pretty big standard errors.

Bloom: I got that off both from a VAR and using a model as you know, with everything in macro everything's endogenous. Everything is driven by retail. So, much as a kid in an impress goes first moment shots or endogenous so my uncertainty shop, so is the ZOB.

Bloom: So is everything policies, board looking, it's a nightmare to attribute this so they should be taken with a lot of a big standard error around them. But those are the best number I could come up with from triangulating both from day to a model.

Bloom: The other thing to point out is that it's a temporary phenomenon. So, uncertainty, going back to the earlier analogy generated a shot drop, but then once uncertainty subsides, that goes and you get a quick to the rebound. So, uncertainty tends to be very damaging in the kind of year of the recession.

Bloom: But I think the full often uncertainty is one of the reasons typically as Milton Friedman described them, recessions have these V-shape so he called it the guitar string theory. As you know, the harder you pull it down, the foster it snapped back. And that's partly because uncertainty generates big drop would also stores up the rebound.

Beckworth: Well, let's maybe step back and ask maybe for a more fundamental question about uncertainty and you brought it up earlier, this question of endogeneity. So, probably many people would say, yes, we believe in uncertainty, but the uncertainty is just a byproduct of bad policy if financial crash something else.

Beckworth: But my takeaway from reading your articles is that it becomes a force of its own. That kind of maybe it starts out because it's something else, but it can grow into its own damaging force that needs to be taken seriously. So, do you see it that way? It maybe it starts out indulgence sleep. It becomes kind of an exogenous independent force of its own.

Bloom: Yes, I see two roles. One is as an impulse, so if you look back of recessions going back over the last 30 years, they're normally driven by some nasty event. So, a war on oil price shock in the case of 2008, 2009, the housing market collapse and well 2001 was 9/11 in large part.

Bloom: So, these events turn out typically to be bad news. They also, which is a negative personal shock. They also increase uncertainty. So, 08, 09 Lehman's collapse. The housing market goes into meltdown. It's clearly bad news, but there's also dramatically more uncertainty.

Bloom: And that's the sense in which I actually think the shocks that hit the U.S. economy are multiple moments. They're not some kind of laboratory clean car PFP shocks, they comprise several elements. One of them is uncertainty.

Bloom: Then the second element is [inaudible] as you got through, I think there's also amplification and propagation mechanism, which is the response of policymakers often can make things worse.

Bloom: So, John Taylor's, argued a lot in this case, Lucas [inaudible] and Petra Varanasi, sorry, at Chicago, have papers arguing that, when times get bad policy makers thrash around and that creates more uncertainty and you can see it in the radical political outcomes you've had both in the U.S.

Bloom: And much of Europe. Recessions have generated more polarizing and more uncertain politics, which I think has increased problem slow down growth burden.

Beckworth: All right, Nick, what about your AER paper where you discuss uncertainty, not just in a kind of a shock sense, but you look at the trend and you've a paper that looks at the growing trend of policy and uncertainty since the 1960s. Can you summarize that for us? What's going on?

Growing Trends of Policy Uncertainty

Bloom: Yes. So, I should be clear, an AER papers and proceedings papers. That's a short summary paper and it's Scott Baker, Steve Davidson, two political scientists, Jonathan Rodden and Canise Princeton and we look at policy uncertainty.

Bloom: We measure policy uncertainty going right back to the beginning of the last century. And we see this U-shape. So, policy uncertainty was high around 1900, it was bad up until the great depression seemed to stabilize in full after world war two and then has been rising ever since. And this maps very closely actually to the U shape and political polarization.

Beckworth: So, there's plenty of evidence showing that politics was much more consensual after world war two, Democrats and Republicans weren't so far apart. Well, as you've seen particularly recently, they're completely non-overlapping and that appears to be one of the factors behind rising policy uncertainty.

Beckworth: So, the Democrats get in they swing one way, the Republicans get in they swing the other. And of course for businesses it makes it very hard to predict what's going on. And that was a great example.

Bloom: I mean, Trump has a pretty radical agenda. It's not clear what's going to be passed and what isn't. Congress is Republican, but it's marginal. And there are certainly a number of Republicans that don't agree with Trump and go to Europe.

Bloom: The same thing in a Europe, for example, my homeland Brexit, It is ever been an uncertainty bomb. That's Brexit. They'll unknown. I mean, nobody has any idea what's happening with Brexit.

Bloom: The politicians it's like some kind of gunfight force. They were shooting each other and falling apart. So, that's very much a sense in which politics become more polarized and I think that's won over the last 20, 30 years. In the U.S. That you're going back to world war two and that seems to increased political uncertainty.

Beckworth: Well, I guess there's something reassuring from that. We've been here before. Maybe there's hope that we can see a decline at some point in the future and political polarization.

Beckworth: Hopefully, it won't require a war, but something might hopefully bring us back down. Well, let's move on to another area you've done a lot of research in, you look at productivity, innovation management issues. So, why don't you share with us some of your findings in particular to the big question that's at hand right now. It's very topical and that is why has productivity growth slowed down so much?

The Productivity Growth Slowdown

Bloom: Yeah, so I have a paper, with John Jones, Mike Webb and John Van Renin probably our ideas getting harder to find, and this for me is kind of a bit of an intellectual journey and I've change my views about 180 degrees over the last five years.

Bloom: So, this paper argues there is a productivity slowdown is a very a good both in that position. It basically argues by looking at the data going back to the 1930s, particularly the World War II. But what we're seeing in the U.S. is a productivity growth has been slowing down.

Bloom: It's not that surprising, but it's not news to many people, but it's actually particularly convincing if you look back over many decades, its productivity growth was about 3% post world war two, it fell to kind of a couple of percent in the 80s, 90s.

Bloom: It's now down to about 1%. So, it's fallen to a third of its long run, third of its postwar value at the same time, R&D expenditure has been doubling roughly every two decades.

Bloom: So, there's a frightening trend in this which is with spending evermore money on research and development, but actually creating fewer and fewer ideas certainly as measured by the growth rate of GDP.

Beckworth: Well, that's very sobering. So, it's getting more expensive to produce new ideas, new innovations?

Bloom: And I live on Stanford campus. I, spend a lot of time talking to scientists and engineers. My dad is actually a scientist the hustle was, it was on a second biotech startup.

Bloom: None of them are remotely surprised by this. So, the story of it is you can think of, there's kind of two areas of history, well maybe three. So, the first year of history is fascinating long and it goes from the Romans almost to a 1,700 AD, where we had almost no growth.

Bloom: So, the estimates of growth is like 0.1% a year, just pitiful, basically zero. And then we have the industrial revolution, which is a huge epoch changing event, but actually growth only jumps up to 1%, which now would seem in a horribly low. But back then was a tenfold acceleration and growth.

Bloom: Actually, from 1800 when the industrial relations started off in the UK has been roughly increasing up to about two, 3% around World War II. And that's an era of what I recall standing on the shoulders.

Bloom: So, that goes back to Newton's quote, which is if I've seen further than you, it's because I've stood on the shoulders of giants, the idea that you can build off other people's ideas and research to be more productive.

Bloom: And if you think of someone like Newton, he lived in an era where he had to work by candlelight. There was no electricity, there was no penicillin, there was no heating, there's no cars. He had a short day to work and it was freezing. He got sick, he died young. But from the 1800s to then about World War II onwards, there's more and more inventions coming on and it's made feature inventors more productive.

Bloom: So, we saw an acceleration in productivity growth around World War II. The second era seemed to, well I guess the third report seemed to set in, which is the apple tree model whereby ideas are getting harder and harder to come by.

Bloom: And here the view is by World War II, we had a large number of elite research, active universities. We had many industrial labs and they're just pouring huge billions of dollars into research and development. And it's actually stomach start to come more and more crowded.

Bloom: And it's becoming harder and harder to find new ideas. And we're very much in that era now. No, that doesn't mean it can't change. It doesn't mean 20 years from now, suddenly we make a revolutionary breakthrough in productivity growth accelerate.

Bloom: I think it's a reasonably safe bet. Looking at recent history where say productivity growth is going to remain reasonably poor as in one, maybe we're lucky 2% a year for the next 10 or 20 years.

Beckworth: Nick, that's not very inspiring, not very hopeful.

Bloom: Yeah. Maybe you have to renew your expectations. So, the one pushback I want to say is our expectations in some ways it just got out of line, right? We used to expecting three, 4% growth, because that's what I knew history has. But if you took an average over the last 100,000, 10,000 years, you'd be ecstatic with one to 2% growth.

Bloom: So, we are actually heading back towards the growth rates we're experiencing around the industrial revolution. A and point B is, it's not even clear. I mean, this could change 20 years from now.

Bloom: In Stanford, I see all kinds of amazing inventions. Things like the driverless car 9, 8 technology, genetic engineering, some of these things you could have revolutionary breakthroughs.

Bloom: The saying is people are over optimistic about change in the next five years and under optimistic in the next 20 years. I just think rather than looking at science fiction, we should look at history and recent history and the picture is not very optimistic on growth rates for the next 20 years.

Bloom: But if you want to look at it optimistically, just compare it to the more distant past and on those grounds doesn't look so bad.

Beckworth: Well, a couple of points on that. I'd like you to respond to first. I wonder if the declining rate of innovation or the more expensive costs of innovation can be a product of our increasingly risk averse society.

Beckworth: So, I've had Tyler Cowen on the show and I kind of latched onto this idea that we, his argument in his book, *The Complacent Class*. As we've become more wealthy, more fluent, we're more risk averse in all areas of our lives. We're comfortable with what we have.

Beckworth: And so, from everything from how we raise our kids to who we let into our neighborhood, to regulations protecting labor, I mean, one example he gives and I've used repeatedly on the podcast is it'd be very difficult for us to build the Hoover dam again because of all the regulations and laws we're not as dynamic.

Beckworth: And there's a lot of other research going on at this speaks to this as well. But I guess I wonder if is the increasing cost of innovation a byproduct of our increasingly risk averse culture that's manifested in these many different ways?

Bloom: Is a very interesting question. So, living out in Silicon Valley, you often get the pushback for the not more risk of us look at the huge, flood of venture capital, which is risk loving investment that's pouring out.

Bloom: On the other hand, you do hear lots of stories. So, one good anecdote in favor of Paella Cohen and I listened to that podcast as your podcast with Paella is think of the way that pharmaceutical innovation has moved from big pharma to biotech.

Bloom: So, my father talks a lot about this stuff 20, 30 years ago. You used to have a lot of research done in big pharmaceutical firms. They were pouring out amazing inventions with massive R&D programs. But what's happened is there's something stultified by the fact they have huge labs. There's a bit of a nine to five safety culture.

Bloom: He says if you visit these pharma zones, the lights are not on, on the weekends, there's hardly anyone in them. Whereas instead, if you go to biotech firms, these small startup that has spun out of universities there, you'll see people working 100 hour weeks taking huge risks, really sweating out because they own the equity.

Bloom: And in fact, the whole business model of pharmaceuticals has moved from, big pharma firms doing R&D in house to now just having a war chest of cash they used to spend on successful biotech startups and therefore effectively they're paying indirectly for the research going into biotechs. And that's one way to get round, the regulatory death and the safety-

Beckworth: Interesting.

Bloom: ... death happening in pharmaceutical firms.

Beckworth: Okay. My other point is one that Alex Tabarrok has raised. He had a Ted talk man, it's probably three, four years ago now, but he talked about the potential for global growth coming from China and India.

Beckworth: We've already seen some of that. But his argument is as more and more Chinese become part of the middle class as their income levels go up. And same thing in India, at some point they're going to really want, their demand for things that we take for granted today.

Beckworth: We'll source, there'll be a huge increased demand for cures for cancer, for AIDS, maybe driverless cars, who knows? But his argument is once you get this huge market over there and you get people who are no longer subsistent living, they're actually a part of the middle class.

Beckworth: You're going to have ideas, the more and more people you have, the greater the likelihood you're going to have an Einstein among that crowd. So, are you hopeful that innovation maybe or research other parts of the world might emerge and make up for the decline? The costliness in the U.S.?

International Innovation’s Role in Fixing the Slowdown

Bloom: Yes, and I think that's already been incredibly helpful. So, there are two trends going on. One is ideas are getting harder to find and in fact Moore's law is a great example of this Moore's law, which is the number of silicon chips or this number of transitions on the silicon chip to double roughly every other year.

Bloom: That's actually held pretty constant since go to Moore made that prediction in the 60s. But the amount of scientists involved in that R&D has got 25 fall. So, it's becoming harder and harder for scientist to make this innovation, then you think, hey, hang on a minute, but Moore's law has kept constant as, am I still making those breakthroughs?

Bloom: Why is that? And the reason is the market's grow ever bigger, do you have I'm entail. Now I know that if I come up with the next generation chip, I'm not only selling to wealthy people in America, I'm selling to most Americans, Europeans and the large massive market in Asia, some of South America, some of Africa, so thankfully grow, there's two forces moving in opposite directions. One is each innovation is getting harder to make.

Bloom: The opposite one is that for a given innovation, we can now sell to a ever-growing market and those two are roughly offsetting each other. The first seems to be slightly dominant and that's why growth is slowly slowing down.

Bloom: But it'd be slowing down a lot faster if it weren't for the growth of the rest of the world. So, yes, absolutely. That's helping, that's preventing a cataclysmic drop in growth is making meaning instead of we having a gradual trend down.

Beckworth: Yeah. But I take your point and Robert Gordon type points that we are still in a world that's very similar to one 30, 40 years ago. We've had a lot of innovations, but we still drive cars, we still live in homes.

Beckworth: Transportation is still relatively the same. What would a radically changed world look like? Maybe Hyperloops and maybe cheaper transportation across the globe in an instant.

Beckworth: So, we are a long ways from maybe this next stage, whatever it might be. So, I think it's a fair critique that you raise along with Robert Gordon. I just want to remain optimistic here and be hopeful.

Bloom: And I don't want to say what pessimism, I gave a very high profile talk out here three years ago, claiming productivity growth is not slowing down. It's on the web. It's embarrassing there in the sense.

Bloom: And changed my mind just by looking at the data. I mean, I'm a data guy and the data shows are stolen. On the other hand, 2% growth isn't bad. The reason, by the way, I think there's such an angst over slowing growth is that we actually not only have slowing growth, but far worse, we have rising inequality.

Bloom: Slowing growth if inequality wasn't rising, would not be such a problem. And nobody really minds that much. 3% versus 2%. The reason it's been so horrible and we've seen the rise in protest politics and I think breaking apart society is inequality is going off at the same time. And actually I think that's the much bigger evil than slowing growth rate.

Beckworth: Okay. One last question. So, you've written on management practices as they relate to productivity. So, share with us your findings and your research on that topic.

Management Practices and Productivity

Bloom: Sure so, long ago I worked at Mackenzie, the management consultant in London and I assume management being a hugely variable and very important. And I kind of got frustrated though when I came back to academia, the economist poo-pooed the whole idea of management.

Bloom: I used to joke that I'd give a seminar and people would hear my English accent and they'd give me a 20 IQ point bump up estimation. But then that see the M word, the management word in the title and deduct 25 points are and a half to two minutes method.

Bloom: So, I should also to say management, it's not like it's a modern topic. Francis Walker who was the founder of the American Economics Association around the 1817, 1880 census. And in fact it was the second president of MIT hence the Walker Memorial.

Bloom: He had his book a paper in the first volume of the quarterly journal economics. So, oldest economics done logging that management drives differences in business performance. It somehow kind of got forgotten about in the 40s, 50s and 60s.

Bloom: And I came back to it with a number of other people. John [inaudible] and Rockway Alyssa in particular, just measuring management practices across firms, using surveys and census measures and showing it's highly correlated with some performance.

Bloom: So, it's not rocket science. It's kind of painful, tedious day to work, but that we just see in the data massive variations in management practices and they're strongly related to performance and coming back to your earlier points, better managed firms tend to be in competitive free markets without much regulation.

Bloom: When they run by professionals and terrible firms or in government or family owned with heavily regulated areas with no competition.

Beckworth: Very interesting. Well, that leaves us with a glimmer of hope that we can get better management and higher productivity growth. Our guest today has been Nicholas Bloom. Nick, thank you so much for joining the show.

Bloom: Great. Thank you very much.

Support Mercatus

Your support allows us to continue bridging the gap between academic ideas and real-world policy solutions.Donate