 Good morning everyone to the second day of the ECB forum on investment and growth in advanced economies Yes, there's contributions have been extremely lively interesting. There were debates until late in the evening And so today we'll continue in the same vein, but we'll shift to business cycles growth and macroeconomic policy Before I give the floor to even mash I would like to quickly do a couple of housekeeping and remind you of certain things So in terms of media all sessions will be live webcast therefore on the records and can be reported upon life and as yesterday any Conversations here off on the sides will be off the record At the risk of repeating myself Please take the opportunity to go and see the posters on displays on your way to the coffee break more back I've just seen the president speaking to to various students and we have quite a time limit on the voting It's today at lunchtime. So please make sure you cast them. I Also wanted to add something related to donations We were all moved by what happened and the president announced that we will make a financial donation to the causes in Portugal we've selected to the Solidarity Concert Runtos Porto's that was held last night It actually already raised over a million and the funds will be handed over to now That's merely cordias portuguesas who are actively working for the victims In the affected areas and the second one the second donation will be for the leader portuguesa de bomberos Which is the Portuguese national fire brigade? And you need to know that 90% of the of the firefighters in Portugal are voluntary So we'd like to join up efforts. We will send you an email to your iPads We will send you emails to your email addresses and therefore we hope that we can contribute and match What others are doing and therefore show with our financial support and do whatever we can and now the floor is for you much Who will chair the second session on business cycles growth and macroeconomic policy and as yesterday's after the presentation of the papers? and the discussions you will have the possibility to ask questions and With that without further ado over to you Eve. So good morning everyone on this beautiful rainy day after such a long period of drought and I Have the difficult part to maintain the very high bar that was put yesterday by the discussions And I'm convinced that we will have as interesting discussion around today with the quality of the speakers on this panel and In order to ensure it. Let me also say that the interaction with the public has been extremely positive yesterday and I am also determined to give the floor and the word to the public again today That's the reason why I would also Tell you that I will be very strict on the timing that is allocated to our speakers and We have today I Don't look at anyone in particular We have today a very distinguished parallel We will discuss business cycles growth in interaction with macroeconomic policy and also in particular monetary policy For that purpose, we have asked for two papers the first one which will be presented by Bob Hall is on sources and mechanism of stagnation and impaired growth and The it will be afterwards it will be discussed Exactly the same procedure as yesterday, and I would say same procedure as last year and The second paper will be gross and complementarity between structure reforms and macroeconomic policy and Both papers have in common that we I think will address the issue Are we in a prolonged recession are we in a kind of the is still secular stagnation a word to be used but in any case how do we see the interaction With gross and low innovation and productivity with hysteresis effects saving investment imbalances we also I think we'll discuss the depth overhang how it still interacts and the Reduction in the working-age populations in our advanced economies and then finally I think one important element is also the inequality and political uncertainties that To some extent have a little bit diminished in Europe, but they'll still per se pervasive at the global level So I don't want to make a long introduction. Let me immediately start with first Paper that will be presented by Bob Hall Let me tell you that he is the senior fellow at Stanford University's Hoover Institution and He was also a member of the National Presidential Advisory Committee on productivity in the US his paper Will be discussed by Goethe Eggertson who is professor at the Brown University and Who has worked at research departments in the IMF from 22 From 2002 to 2004 and afterwards for a long stint also at the Fed New York Until 2012 the main focus of Goethe's work is the analysis of monetary and fiscal policy over the business cycle So please Bob, could you make the start? Okay, that's a great pleasure to be here I Only two weeks ago. I was at the conference of the central bank of Portugal there couldn't be a bigger difference in one respect the in the directions for that Event it says dress code colon. There is no dress code Now I couldn't help noticing the contrast between What I thought was quite a downbeat message from chairman Bernanke in the opening dinner And then a much more optimistic view about Europe from President Draghi Yesterday sufficient to elevate the euro by several percentage points. I don't think I've ever been president when a market ever moved before The but I think I'm in what you'll see here today is more on the Bernanke side Ben was Talking mainly about the Particularly bad things that are happening at the bottom of the earnings distribution in the US I'm going to talk about something which is I Think an important part of that story, which is even the average worker in the US and other economies Has not done very well not so much from the Crisis although that had an immediate Powerful short run effect, but I'm going to talk more about the longer run effects that really came into play in the US and other countries around the year 2000 One of the one of the themes of the work that I've done recently which is cited in this paper is that the Although the US had a bad recession Starting at the end of 2008 It was only a typically bad recession. It was about as bad as the 1982 recession And the US economy bounced back from it Real about it at about a normal speed The real problem in the US was that it was superimposed on Longer-run adverse developments, especially with respect to earnings, which is what I'm going to talk about today And those adverse developments began well before the crisis Yeah, and in the in the paper that by the way just came out as a national bureau working paper I and three other authors developed that theme. I hope convincingly if not convincingly at least a great length And so I invite you to take a look at that which is now available In this effectively it's published form okay, so I'm going to talk about stagnation Stagnation is thought to be yesterday's issue. You know Larry summer has made a big deal out of this Now almost four years ago And I think generally sort of stopped talking about it because both in Europe and the US we've seen a movement In the US's case back to full employment and in some European countries as well but at least a lot of progress in Europe and so talk of stagnation has somewhat died down with fairly good by recent standards growth But but in fact We know that the crisis Worsened the stagnation one of the things I'll show you is How much of that stagnation was already underway before the crisis in many countries? So so I think it's still appropriate to talk about stagnation And stagnation is not just a factor that affects people at the bottom of the income distribution I'm going to talk about national averages, which are Something that can be documented from national income accounts But but you should think of that as one of the moments of a distribution of earnings And in the US at least the widening of the distribution which has continued Makes this the fact that it's mean the growth of the mean has dramatically slowed down as amounts to stagnation as far as I'm concerned So the the measure that I'm going to Talk about today and focus on and decompose Is real labor earnings per member of the population? It's not a number that Gets a lot of attention It's a very comprehensive number in the sense that by earnings, I mean Compensation so it includes fringe benefits It's the national income accounts are one of the best ways to understand how much of earnings today in a modern economy comes In kind from employers and not directly in cash It's not nearly as bad a number as German Bernanke mentioned that he said quoted this number Which I think is somewhat misleading that there has been no increase in mean earnings or median earnings Since 1980 that's when the numbers we're talking about here. There was quite substantial growth In the time prior to recently in fact it started growing again I think it's a very useful number to look at because In most advanced countries most people do not have very much non-labor earnings so so sort of the average person is Consuming based on earnings alone So and that's important because of the fact that in a number of countries as I'll show you the share of National income going to earnings has declined. That's particularly true in the US But it's also true in some other countries not all countries So so I'm going to look at a number which is a little bit unfamiliar But I think actually Is the single number that that captures the kind of things that I want to talk about and as soon as Larry Summers launched the discussion of stagnation We realize well, how do you measure stagnation? And I think an answer Real earnings per member of the population notice that it's not particularly not real earnings per worker because a lot of the Changes that have occurred in the US and other economies is the fraction of The population that's at work. So I want to sweep that into this calculation Numbers based on just looking at workers don't convey Certainly in the case of the US don't convey some of the really important things that have happened with respect to labor supply Okay, so I talked about that So here's the here's the results of these calculations now these numbers come from the harmonized accounting system of the OECD By the way, this is my first venture into It's throughout my career Absolutely been a governing principle of my career that I study only the US economy on the grounds that the US economy is Sufficiently complicated to occupy at least a career but for this group it seemed to me that Having developed this accounting system and sort of thought through these issues with respect to the US Which I've done in a number of earlier papers including one. I just mentioned This is just transplating those calculations, but if you start asking me details About any country but the US You'll embarrass me. So I refer you to the OECD for better or worse for The documentation of these numbers so so the bottom line of this analysis in terms of Macro economics is that macro economics is way more complicated and hard than you think and the reason for that conclusion is that the The economies the six economies that I'm looking at which are identified here are all advanced economies They all have access to the same technology. They all have economies. They're organized along the same principles of a mixture of Government activity and but mostly relying on the production side on private activity They were subject to a worldwide financial shock in 2008 Everything ought to be the same all these economies by everything we teach in terms of how there are random shocks and pinching and Go through an impulse response function that the orthodoxy of modern macro when you feed the same shocks into the same model With the same impulse response functions, you get exactly the same results, right? But that's not the way things work as you can see from this picture so these this is the outcome variable that I'm interested in and It's divided evenly between Pre-crisis and post-crisis seven years pre-crisis seven years post-crisis And you can see the effect of the crisis almost everywhere as a decline in 2008 2009 But the pre-crisis development one of the things that I didn't realize is that Germany was stagnated pre-crisis and the crisis energized Germany and German earnings have risen much faster post-crisis than pre-crisis Who to thunk? So so that's kind of a surprise Maybe a little less surprising is that the performance of Britain Which was very good pre-crisis Was at least average so that by the end of the period the growth over the 14 year total period 15 year total period was Was at the top The US is not so great The notice that the US had the biggest shock effect and That's because as you'll see the the rise in unemployment in the US was higher than other countries and this shows the importance of keeping track of The fact that when there's fewer people working average earnings per member of the population will decline It's particularly important for the US where unemployment is a particularly volatile mean reverting variable Okay, so this is what we're going to try to untangle To untangle that I'm going to do Something that comes very naturally to me because Robert M. Solo was my thesis advisor at MIT And he had recently this tells you how long ago. I was in graduate school He'd recently come up with this completely ingenious idea, which is which is really dominated macro measurement ever since of log linearizing the production function And then observing that you could measure the elasticities as factor shares. So that gave a completely transparent way of thinking about how a set of macro variables interact and Give rise now in his case the outcome variable was just output So he just log linearized the production function directly I'm going to do something a little different because it's going to be talking about real earnings rather than output But philosophically, this is what? Solo taught us and so things like this are almost always identified with solo as a solo decomposition So the first thing is that Real earnings the outcome variable that I'm interested in is by definition the labor share times some measure of real output Now you can get into the details, but it turns out not in terms of these log changes that we're going to focus on here It's not out We don't need to measure it doesn't really matter exactly how you measure real output However, you measure it you want this to hold as an identity So you're going to have a corresponding way of measuring the share a lot of these things I write down or there's no behavior here These are all identities And the behavior is just reflected in in the decomposition So we'll we'll get to Economics at a certain point right now. We're just talking about identities Okay, so then real output is the output per unit of labor input times the volume of labor input So this is going to be a labor productivity Exercise of course, it's going to be driven by total factor or multi-factor productivity But I'm going to state it in a form which is often used in growth theory, which is to state things in terms of Per unit of labor input So so that's something to keep in mind Now now here's one thing you have to read the appendix of this paper to Understand this but output per unit of labor input depends on multi-factor productivity with a with a power and you can look in them You look on the last page of the paper, which is in your iPad to see that And and also a function of the capital output ratio So this is going to enable me to talk about the effect of capital accumulation Or and or the slowdown in capital accumulation as a determinative factor But normalizing by the level of output and that's quite instructive. I think that's a that's a useful thing to do And then finally we can break down the volume of labor input and this is where the most original parts of this work occur, not that it's breathtakingly original Into hours per worker workers per member of the labor force which Captures unemployment members of the labor force per person of working age, which is labor force participation And finally the people of working age is a fraction of the total population Which is measures the dependency ratio and that's an important factor and it varies across the countries we're looking at Okay, so so this gives rise to a seven-way breakdown. So there are seven different things to think about in understanding The the issues of stagnation first I mentioned for the labor share second productivity measured Solo's way that is a multi-factor or I don't know why this is but in the US It's called total factor productivity TFP, but in Europe. It's called multi-factor The capital output ratio I just talked about hours per worker The employment rate which is just one minus the unemployment rate to keep it in the same sign as a measure of positive measure of the amount of work that people are doing The ratio of the labor force the working age population and the ratio of the working age population of the population of all ages The last by the way over this period is actually not very important It's a it's a pretty steady trend. We're not going to spend a lot of time on them But the others are all we'll get to the economics as soon as we start talking about each of these Okay, then just to summarize in a table The findings that were also shown in that in that picture, which is the growth rates of real compensation Remember the population now averaged over the the seven years Pre-crisis and the seven years post-crisis. You can just again see the heterogeneity in responses and basically the six countries Arranged in alphabetical order Fall naturally into groups don't ask me why but the alphabet helps here so France and Germany resemble each other and having Low or even slightly negative in Germany's case performance pre-crisis and Definitely leading the pack in terms of post-crisis performance measured by this measure Then the second to Italy and Spain have better pre-crisis performance But post-crisis these appear to be countries at least on a timing basis where there was a tremendous swing toward negative Performance measured by by this measure and then finally the two so-called Anglo-Saxon countries Britain and the US Had pretty good pre-crisis performance are very good in the case of the UK and somewhat negative Post-crisis performance So again, you can see these closely similar economies Get very different results, you know, and I Can I this breakdown? I think it's going to help people think about this But believe me, I don't have all the answers here by any means Obviously, I know much more about the US so my discussion of these will will tilt toward the US Okay, so labor share This is now suddenly the hot topic and quantitative macro in the US and it should be should be in in Europe as well Germany is the standout in which the shares labor share has actually been slowly rising throughout this period just completely completely smooth behavior in all the other five countries, especially Spain You see a plunge in the labor share Well again, the heterogeneity is an interesting question So quite a few papers have been written Several of which I've been called a referee called upon the referee by the way, so I've acquired I haven't written on this subject, but I've been paying a lot of attention to to the evidence And it seems that the topic that we talked about yesterday which is They the Of the various things that might affect the labor share, but you know first of all solo taught us That we probably ought to think about the labor share It's kind of a constant. You know, there's come under Cobb Douglas Lots of things that you would think would affect the labor share don't affect it that's one of the first things to understand because With you know the elasticity of Cobb Douglas technology, then you get an exact offsetting of things that would normally change the share So you can't think about well did wages rise relative to the cost of capital? That shouldn't affect the labor share at all under Cobb Douglas and Cobb Douglas seems to be a pretty good approximation the factor that that we I think we need to focus on is the role of Changes in the price marginal cost ratio the price marginal cost ratio in the industrial organization economics is a a Is a measure of market power, but it's market power very generally described for example in monopolistic competition You have market power even though we think that that there's zero Profit available to the marginal entrant So it's not as if there were barriers to entry or conspiracy, but still there is market power so That's an idea that that has been pushed pretty hard by quite a number of authors in the in the US with respect to this decline in the Labor share, which is one of the three big factors. We're going to see labor share productivity and and at least in the case of the US participation But of course it's heterogeneous. Why should it be so much bigger in Spain? as Spain isn't the headquarters of the high marginal cost Price cost ratio companies like Google Marginal cost at Google is essentially zero So they have infinite marginal cost price marginal cost ratio But Spain is the country where you see the biggest change Now that I think is something that we ought to be looking at then you see these cluster of countries and they see Germany Just sailing along with slightly rising I Think probably the productivity Data are are well known to everybody That is productivity growth has been terrible in all these countries We see it as kind of a crisis in the US There's a lot of discussion of why earnings growth has been so low in the US Say how can it be? We have a very tight labor market. We have extraordinarily It's extraordinarily difficult for an employer to find a worker in the US today that by the measure that measure of Duration to finding a new worker the US labor market is tighter than it's ever been The number of people who have just entered unemployment as a fraction of labor force in the US since we kept records since 1948 is lower than it's ever been. We have the tightest labor market. You can imagine not that far after this terrible recession But but it's not driving earnings growth And the reason it's not writing driving earnings growth is the productivity growth is low It's left out of many accounts But you can you can see even though it's the best over this whole period. It's been very low Since in the post crisis period But we can't find any evidence that it's the result of the crisis that we looked into in this paper that I mentioned before And then other countries the UK, which is a strong performer by most measures by by productivity growth is terrible So this is this is in some ways the central problem of modern capitalism is that we just can't get any productivity growth anymore What lies in the future is is not something I'm going to get into but This is just shocking relative to the earlier record, especially in the US The capital output ratio there's there has been a Lot of discussion of the shortfall of investment, but the shortfall of output is so much higher than the shortfall of investment that The decline in output relative to the capital stock, which is a slow-moving state variable has actually raised the the crisis Increased the capital output ratio Saturated the economies in capital because of the decline in output and that decline comes from all the other factors The adverse factors that I've talked about so So you can see that there's no country where the capital output ratio plunged because of lack of investment even though investment Was relatively weak for the different reasons we learned about yesterday Okay, weekly hours per worker is not not a big factor. It's much more volatile in the US as this shows It's much more likely that workers will work short hours in the US than other economies, but Between beginning and end. This is actually not a big factor. So I won't dwell on it The employment rate of the labor force Is how unemployment gets into the story and it's and it's really important And it's often a neglected if we look only at per worker then we leave out this critical point And you can see that in the high unemployment Country of the whole group, which is Spain It's a very meaningful notice then all these pictures I'm showing you the volt vertical axis is the same and these are all additive factors I didn't mention that before but you can see that this and productivity In terms of short run fluctuations Keeping track of unemployment, which is something that is not done in most analyses, believe it or not We talk about the importance of unemployment. We ignore it Do these calculations and labor force participation? Only the US has had this collapse of labor force participation Which is the result which is now the topic of a very large literature But it's very specific to the US I won't go into that now for a while time. I mentioned that's nothing else get that so conclusions Unitary theories of stagnation are unhelpful Don't buy Hansen don't buy I won't mention other names Because it's no unitary theory works You got to pay attention to what actually is happening separately in each of these six countries And I guess so another factor is that it seems like the zero lower bound is not such a big deal several of these countries have escaped the zero lower bound, but they haven't escaped stagnation and So So each each country has its own story And I was asked to talk about policy but anticipating I wrote out of time I'll just say one thing and that is that financial fragility Was a big contributor to at least the short run effects of the crisis. We do need to work very hard Developing more robust financial systems will avoid the short run problems but we'll still have to live with a variety of longer run problems and there the the heterogeneity of results and these Economies extend me even though I knew it was gonna get something like this, but the amount of it is just quite striking. Thanks Thank you Bob and thank you for respecting your time as well. I will now hand over immediately to Goiti and To discuss this very focused labor market focused analysis by Bob well, thank you very much for inviting me here to discuss this paper that I Found very interesting. So This is how I'm gonna organize my discussion first I'm gonna talk about what I consider as sort of a Stylized facts that that Hall presented in in his framework. They're gonna be two of them so in macro just realized recently that Stylized fact is a term invented by macroeconomists in particular Caldor and that's because we don't really have fact We have just tendencies. There's always an exception Germany in this case and then I'm gonna talk about the core of The paper which is the what I call Hall's decomposition Which I think is a very useful way of looking at the data and namely Decomposing the increase in real income into seven different factors And I'm gonna present it slightly differently than than than Bob did in a way. I hope is useful. So Then of course, I'm gonna talk about interpretation and here I'm gonna separate paths with Hall and suggest Bob and suggest that unitary theories are gonna be very helpful to understand what is going on Let's mention Hansen, but we can mention all the names and and then emphasize some disagreements So the two stylized facts that that Hall emphasized and this is the picture he Showed us. I'm gonna show in a second. The fact number one is that the Great Recession marked the fall and income Pair capital and you know And this year is the or the income per member of the population. It has been falling Or growing slower ever since the crisis and I'm putting a line here in the crisis This is just a figure that Bob showed us and I guess one of the things that Bob emphasized here was that The performance of these countries are very different which he suggested might be Means that no unitary theory is useful to think about this. I suppose I'm gonna My theme is gonna be somewhat counter to that in fact I'm gonna suggest that perhaps it's useful to think of all of these Countries has having been subject to a crisis shock or a fall in the natural rate of interest the key difference is that Germany was subject in a to a differential way to the other countries that were part of the European monetary union So for some reason here the Okay, yeah, yeah, so Germany was differentially affected than these other countries that are part of the monetary union And then United Kingdom United States had had their own monetary policies and a key difference here That in the reaction of these economies can be seen by The way in which they adjusted labor inputs appears to be different and I'm gonna emphasize that fact and so there seemed to be some difference in labor market institutions that May be useful to think about this figure and the differential effects but bottom line stylized fact number one there seems to have been this market decline in real income per member of the population that is Really marked by by the crisis Stylized fact number two and again Germany has the odd man out That a large part of this can be explained in the in the fall in the sheer of income that goes to labor This is the labor sheer and we can see here particularly Spain She's a market decline in labor sheer but we see this across all countries except Germany Now one thing that is a little bit of a puzzle to me relative to the what we saw Yesterday was that there we saw some evidence that labor share had also been falling in in Germany But perhaps that was the sample period was longer, but that's something certainly that would be worth looking further into Okay, and here you see that this fall is sort of Starts to Become more severe right around the crisis So how can we account for this drop in income per member of the population? And the key thing in this in this paper is to propose a simple and very useful accounting framework Which is sort of akin to growth accounting, but has much more details in decomposing the labor input and allow for time-varying Factor shears. So let me just briefly give you a little bit of a feel for what what it is doing here So the key Innovation here relative to the typical growth accounting is to go into a lot more detail in the volume of labor input What I'm calling calling LT here in particular. So, you know, where does LT come from? Well, it comes from the population Big NT, right? But then there's only a fraction of that population that is of working age Which is HT over and T and a fraction of that that are member of the labor force. So that's the And then a labor force participation then those that participate in the labor force not all of them are employed So that's this term and then finally Those who are in fact employed there may be changes over time in the number of hours. They work per week So this is a very useful way to think about it to decompose the labor input into these different for different margins and that's precisely what Bob does and then Then he also takes account of the share of income that goes to labor and if we take a cop dog Less production function that would be given by alpha here finally the volume of capital You can also take a countdown and finally TFP, which I guess we should multi-factor productivity here in Europe, but I'm used to call it okay, so That's just a simple production function and giving a sense of where what the kind of decomposition is that he's doing and then you know at the at the at the core of it though is is looking at the income That goes to labor or let's say log of L alpha with the share that goes to labor income over YT over the population and one of the very nice thing about what and I'm gonna this is just to Preview to you how I'm gonna present his decomposition So let's say that this is the income per growth member of the population in a certain period What you can do then and what he does and it was perhaps not as apparent because he was Showing you all the countries but for each component at a time Let's think of a single country that grows by this blue bar. You can decompose that into Growth in the working age to total population labor force participation employment rate Now this of course could also be negative hours per worker Capital output and finally TFP and all of the the point is all of these bar up bar should sum up to the big one Right, so it's a very useful way of doing this decomposition So let's take a look at how this looks for these different countries that that Bob was looking at Okay, so this here shows United States and the red here shows you the fall and income from 2007 and 2014 while the blue shows you the increase in income from 2000 to 2007 and then it is decomposed in these different factors below. So what we can see here is that the Large part of this what I would emphasize I guess is that a large part of this can be really explained why it has not been increasing as much in 2007 to 2014 is this decline in labor force participation So that's the green bar the very negative green bar. No, I mean the red bar The other thing is that you see a fall in TFP relative to the blue, right? And finally you see that the labor share has declined here in in the case of the US So I would say to I guess three main factors Volume of labor as measured by labor force participation Declining labor share and then a fall in TFP is the story in the US in Spain you also see that there's a big decline in the volume of labor But there is not showing up in labor force participation But instead the employment rate and to some extent ours workers But again a second element to that story is the fall in the labor share total factor Productivity here is not doing a whole lot Nor is is capital output In Italy you get this again the same story that the volume of labor is falling and here It is the combination of employment rate and ours worked and then finally the labor share and again here you don't see much in The TFP is playing a lot smaller role And you can see here the scale is bigger the fall is much bigger in Italy France you see somewhat similar picture of the fall. There is a smaller in real income over that latter period the grad is Is minus one But again here you see the volume of labor is falling and it's the employment rate and ours per worker that is falling and then again The labor share so you see do see commonalities here falling in labor inputs and Labor share the key difference which measure of labor input is falling. Okay Odd man out here though is the United Kingdom there what it seems to be entirely driving it see appears to be TFP that the fifth odd man out is Germany where you don't even see any fall in income whatsoever. So let me just Move on from there. So, okay. So what have you learned? I think there's a bunch of stuff We have learned here that is interesting first. What is not this story aging here is not this story, right? Contribution of working age to total population not important to explain the drop in income per population But I mean that is one narrative that has been popular for people to explain the stagnation That's not playing a role here. Another thing that is interesting here is that lack of investment is not this story Contribution is very small and that comes directly out of this decomposition. That's interesting And there are some that have argued that financial friction was holding back investment and that is driving that this is suggesting to us That is not that important. Now. Finally. It's hard to tell any coherent story with total factor productivity It's all over the map in these countries Okay, so what is this story? Well, the story I would suggest is that in most countries, Frank, Italy, Spain There is a drop in labor input via employment rate and ours worked in US this shows up as we had dropped in labor participation The quick key question then is what is does the drop in labor input? Why does it show up differently in the US versus the European countries? Bob has suggested that the in this paper he referred to that the the fallen Labor force participation is unrelated to the crisis I'm very skeptical of that and if if it did in the US. Why didn't do it in these other countries? I think a more reasonable interpretation is that There was a demand shock labor input fell across all these economies But they did so in a different way because they have different labor market institutions But I think that's a key question to ask here. Why did the show up and differently? What is the story? Well, there is all these countries except for Germany show a market drop in labor share violating sort of a classic calder stylized fact and I guess one key question is what is driving the drop in labor share and I think there was a very interesting discussion here yesterday I have had thought of this as being just due to an increase in monopoly power in the US But it looks like this is a phenomenon that is also taking place in Europe And it's much harder to tell that story in Europe apparently at least according to the talks we heard yesterday So then the question is what is explaining it if not? Increase in monopoly power or it cannot be the only story if that has if that development has not been happening in Europe Okay, so peculiar number one that the TFP rather than change in labor input is trying the UK slowdown and income growth Drop in labor share still there peculiar number two Germany well, what's new so not well on that But I think that perhaps is because they were not affected as strongly by the financial struck shock for for whatever reason So the bottom line here is that nowhere is lack of capital formation a big deal nowhere is aging a big deal Labor share falling across the board in Spain Italy France the labor the labor income stagnation also show But subs in labor input while in the UK the productivity seems to be the main culprit And Germany no stagnation Let's now talk a little bit in the two minutes. I've left a little bit about interpretations One thing I think I find a little odd about and he didn't really Bob didn't speak too much about it is that This notion that Insufficient demand would have to show up only we have one one measure of labor input namely the employment rate I see no reason why you wouldn't expect a low demand showing it for example in participation or our worst work and and so I think that's So in and that has been my theme here perhaps late for rationing is simply happening We have a different mechanism across the country's depending on the labor market institution I also see no reason why demand as Josh cannot show him show up in total factor productivity In fact, there's one of the papers here of the young Economist that is modeling exactly that channel where it works through innovations and You know, that's the hysteresis effect and even Bernanke suggested another hysteresis mechanism in his talk if I understood him right Which is essentially the financial crisis got people out of work and they became addicted to opioids So, you know, that's a that's a pretty classic hysteresis effect where you permanently scar the labor force force with Unemployment so bottom line. I don't see anything here in the result that suggests that a slowdown is not driven by demand and Now does that matter well it matters because If you say that that is not driving it you sort of give Monctry follow policy a free pass which I think would be a mistake is especially here in Europe and also it sort of Ignores the obvious thing that these countries have in common. So Which is that so what would you have expected in a demand driven recession? Well, you would have expected that Interest rate would be low inflation below target and then an output gap I suggest to you that we have seen all three things in the countries in question And there's nothing that says that that output gap needed to show up in NXT So I'm out of time a very interesting decomposition my interpretation is somewhat different I don't think we know exactly what the source of this is But I think this paper gives a very valuable tool tool via this simple decomposition. Thank you Thank you Goethe. I think now it would be fair to ask Bob whether he wants to have a short reaction Not beyond five minutes because we also get five minutes maximum maximum to react to the statement of Gaudi and his Questions, okay. Well, first of all, thank you for a Very insightful summary of the paper. So I Which I have no fault I think on this question of I think he accused me of demoting demand policy and I realized I Didn't I didn't get into in this discussion, but I realized That's right It does seem to point in the direction that Demand effects and therefore demand policy, which is something that's normally within our grasp and still is with fiscal policy if we're not completely unable to exercise fiscal policy I'm inclined in that direction. I recognize that that the evidence here is not Dispositive One of the things you mentioned is that there's no reason in principle why Demand wouldn't enter the determination of labor force participation, but that's something I have studied very carefully in the US And there I think the evidence is is almost completely persuasive The decline in participation began before the crisis was was well underway by 2008 They you if you look in the participation numbers There's a there's a little bit of a sign of the recession itself But then so unemployment Rows rapidly and then decline pretty rapidly participation just continued downward There's the the fact that participation continued downward at the time that the labor market was tightening and has now reached an unprecedented degree of tightness I Think pretty much knocks out for the US Now I certainly agree with what you said that labor market institutions Are going to be part of the answer to this and there's a very very large literature on that point which which I think would support that But nonetheless, I think you know given given first of all the importance of Factor total total factor productivity and very little evidence Supporting the demand theory. There are recent papers that suggest that support the demand theory But they don't to me they don't survive just staring at the at the time series evidence There's not historically been say in the US not much of a correlation in particular if you look at the timing of TFP measures in the US they go exactly the wrong way for theory that they're demand induced It was exactly when demand seemed to be falling most rapidly when when we had the most satisfactory TFP growth and then when the economy began to recover TFP growth fell so the timing just doesn't fit together I Am afraid that we're going to have to face up to the fact that of the volatility of Output or real earnings in this case Demand is not that big a deal If you look at say two to three year changes They're dominated by TFP not by demand That's a point that I've been pushing for a long time. I wrote a extremely badly received Jackson Hole paper in 2005 that pushed that point very hard, but Very distinguished Harvard economists got up and said everything is wrong in this paper and they sat down Carried the day So So I recognize that I probably haven't prevailed on that point, but it would this is particularly in the US where the US economy Traditionally has bounced back on the demand side quite rapidly unemployment in the US unlike Continental Europe is a very strongly mean reverting variable There's there's a well-known paper by Blanchard and Summers that made that point long ago, and it's it's exactly true today So so at least in the US Even even when imprisoned by the zero lower bound Unemployment melts away in a way that is quite convincing that what's happening is that the economy is returning to normal in terms of demand supply relationships and Zero zero to support any theory of hysteresis The recent Brookings paper looks very Very hard at hysteresis and comes up with nothing. So so you know is a great idea, but You know it doesn't seem to sharpen the data I Think I reckon this will be part of our discussion certainly quite interesting to see whether the demand components were not responsible for the great recession and what would then be also the appropriate monetary policy in response But I think also the disentanglement of cyclical and structural factors have a role to play let me now turn to our second paper and this paper is called Gross and the complementarity between structural reforms and macroeconomic policy it will be introduced by Philly Baguio who is professor at the Collège de France and Also at London School of Economics He is a fellow of the econometric society and of the American Academy of Arts and Sciences Philip has received numerous prices as well and his introduction of 25 minutes Will be followed by a discussion that will be Chad Jones who has Done important research on long-run economic growth and he is a professor at the Graduate School of Business at Stanford University and also he has been a research associate of the National Bureau of Economic Research and He also has received Distinguished rewards in the past so let's start with Philip and You want me to do 25 in a go or 20 and then five or whatever. Okay. We'll see No, okay I get going I will cut the oxygen after 20 minutes. Sorry. Oh, yes, 25. Okay. You have it in front of you So, thanks very much So this is joint work with any scar will be we see her from the IS and Emanuel Fari and I think I have to make the disclaimer that none of what I will say are the views of BIS as such Okay, particularly the title. No, I don't know So of course there is this big issue why is Europe less resilient than the US Why is it that following the recent big financial crisis, you know, Europe had more sluggish growth than the US so there's been various you know explanations Put forward for example, my friend Jean-Piz any fairy who is not here today, but was on the program You know emphasize the fact that the sequence the policy sequencing in Europe was suboptimal compared to the one in the US following the crisis Other people have said well more fundamentally There are two things that you have in the US that you don't that you don't have in some European countries starting with my own Is a lack of is in fact particularly is the failure to implement structural reforms and also people have blamed at the European level a lack of reactivity in macroeconomic policy compared to compared to US so in 2014 Mario Draghi pointed to the Complementarity between proactive monetary policy on the one hand and structural reforms on labor and product markets on the other hand He said there is that much I can do But on the other hand Countries have to do their homework and they have to do structural reform in particular my own country has to decide to do serious structural reforms so Here I will focus on product market competition and we argue that there is complementarity in inducing more growth between More proactive monetary policy a monetary policy that react to the business cycle more and a more competitive environment Okay, so that's that's what we are trying to do here So there is first a model which I will not inflige Which I will not impose on you but the first part of a paper is a formal model in In our model you have an economy where firms can make growth enhancing investment But they are subject to liquidity shocks to survive liquidity shocks They need to reinvest in their project at the interim period But but anticipating that they might face liquidity shocks and therefore have to reinvest They they may want to hold some liquidities To start with to to have to do less Reinvestment exposed you see I mean so so they may have to sacrifice because they may want to make sure that they have Enough to reinvest at the interim period So what you do is that when you know you face liquidity shock you tend to hold liquidity to not invest all your liquidities in the first Place because you want to make sure that if you have reinvestment needs you With your profits plus the the whole debt liquidity you have enough to reinvest and continue the project. Okay so a main a main prediction of this is that of course a More counter cyclical interest rate particular lower interest rate in recessions Would reduce the amount liquidity firm of liquidity firms need to hold to with our liquidity shocks. Okay, and so The predictions that come out of that model is first that the more likely liquidity shock the more growth enhancing it is to have more counter cyclical interest rate, but But particularly when you have high competition because when you have low competition Firms make low rents make high rents. So when they have high rents, they can always Will use these rents to reinvest But when you have high competition There is less rents that you can use at the interim period to reinvest and therefore you need to hold more liquidity And that's where a more counter cyclical interest rate would help firms. You see what I mean? That's the that's where the interaction between Interest rate and competition Competition plays, okay So So the to test this we do the following we will use a difference in difference approach Comparing pre versus post omt industry growth in a sample of euro area countries. We relate the change in growth To the unexpected change in tenure government bond yields between pre omt and post omt the announcement of omt lowered The the yields the bond yields and and stimulate growth. You see I mean, that's the idea So the results to summarize is the unexpected drop in government yields following omt Had a positive effect on growth in sectors with higher Pre-existing indebtedness of course is sectors that are indebted or are more likely to face liquidity shocks that are affected by the by the decrease in bond yield induced by the omt, but all the more But all the but only in countries with low market product market regulation where you have high competition So we see this interaction between product market competition and the omt policy Okay, so it's twice a tribute to draggy first the idea of the complementarity and the omt. So it's well So and pre-existing levels of indebtedness as as a drag on growth the more so in countries with high With low product market regulation, which means with high degree of competition. I look I watch the time so Why look at before I Get desigrated Why look at before versus? After omt announcement because we are interested more generally in the effect of monetary policy on growth But of course these raises always an endogeneity if you you may are you capturing the effect of cutting interest rate in bad times on growth? Or are you capturing the fact that it's easier for fast-growing country to have a more counter-cyclical policy? So how do we address this endogeneity issue is we look at omt as the ECB response to the Europe Europe and sovereign debt crisis and we look at the unexpected at the effect of the unexpected change in government Don't use following the announcement of omt and is the effect of that thing on growth that we are looking at okay So some historical context over 2011 2012 Some major euro area countries face severe spikes in government bond yields raising prospect of sovereign default Then the omt program was a commitment by the european central bank to buy government debt under some strict conditionality Omt was targeted at relatively short maturity bonds Yet its announcement was followed my massive change as we will see in long-term bond yields beyond and above what has been expected And what we are using is the unexpected part you see them The the forecast error the effect of the forecast error in the young long-term young Bond yield so we measure the unexpected component in bond yields by the forecast error Defined as the difference between the realized bond yield in a quarter Q of your Y and The forecasted bond yield for that quarter at the end of the previous year And we average over quarters for any given year and what we do is to compute the the change in the forecast error In the average forecast error over quarters between the period 2011 2012 and the period 2013 2014 pre omt post omt. Okay, so we look at how empty change the forecast error Okay in long-term bond yield Okay, so what you can see here I know if you see very clearly, but you have the blue bars and the red bars Okay, so the red the red line is the forecast error in the policy rate in the three months interest rate Which is the the policy instrument, but what's interesting is the blue bars Those are the forecast error in the tenure yields and what you can see is that for Spain and Italy you see before the omt You see you have the long-term rate were above the forecasted rate. You saw I mean you were Things were worse than what you expected But but then after the omt is the other way around the long-term rate. I will our way below the forecast Levels you see you have a decrease. You see the forecast the real goes below the forecast. Okay, so there is a Drop in the only you see the in this In this forecasted error you can see that in Spain and Italy and those are countries where you have high level of debts you have a highly indebted and and if you take if you take for example other countries like You see here, of course the rays you see the the the fact that you have above forecast long-term yields before was due to the European sovereign crisis and then for the dotted line is the omt and then the omt Changes things then after that it goes the other way around you have the realization goes below forecast Okay, so there is a dramatic effect of the announced omt But if you look at other countries for France and Germany, which were not in so dramatic that problem if you don't you don't see The same effect you see Okay, we should look at the US and UK you don't see the same effect So it's particularly important in countries where indebtedness was high. Okay, so So now I can look at what unemployment Spain On employment was above forecast You know when you have the sovereign debt crisis before the omt, but after the omt unemployment goes below forecast Okay, and same in Italy if you look now at growth you see What you see now the red line the red line is the the GDP growth the error in GDP the red line is Is the is the growth The the the error in growth forecast and what you see is that prior to the omt the true growth Would be with would would go be you know more and more below the forecast, but you see that the move Okay, and that's that's true for Spain and and Italy, but particularly for Spain okay, so So now what I want to do is to look I know that the omt has an effect on the long-term bond yields I saw that this has effect on growth on employment, but I want to know how the the effect How this interacts with product market regulation? Okay, so what I will do there is that I look at product market regulation You can see I have nine minutes left So you see between France of course we have the highest product regulation in France, you know Visit France. It's a very lovely country. There is a lot to do there But Emmanuel but with Emmanuel I'm very optimistic. Okay. I should not say that that's not the view of the BIS. Okay, so the Okay, so and what's very interesting also is to look at the right-hand side because the right-hand side You have the various components of product market regulation and the very interesting one is the is the is the is the green one you see that the but you see barrier to The in particular you see the barrier to trade and investment are particularly high in in The barrier to trade and investment are particularly high in Spain Belgium and France and Let's so in Germany Portugal Austria and Italy and we see that this plays a big role later on the buyer The barrier to trade and investment. Okay, which are good major product market competition. So I get back to those later So remember Spain Belgium and France high barrier to trade and investment compared to the other countries Okay, so now what are the I have eight minutes left? Okay, so I look at the time So what what is my main regression? My main regression is to agree is to is to regress the you see the the the growth you see the the forecast error on growth post in 12 to 14 on on its on the forecast error In in 10 12 for countries and sectors S Dino sector C country. Okay, so it's cross country cross sector The beta the beta I have a Yeah, of course the pointer the beta D zero that's the coefficient on the indebtedness and I expect that beta D zero Should be should be negative. That means existing that reduces investment in growth. I mean, it's clear But what's interesting is the is the interaction with the change in the forecast error Following the omt and in fact, I expect that beta C zero should be positive. That means that the more in depth And you are the more The the the low the lowering of the your own long-term bond yield following omt should boost your investment and growth you know you see what I mean and That's the first equation what we do in the second equation is to add the interaction with product market regulation So you have the So you see the term you add the interaction term with product market regulation the beta D one should be positive It means that that reduces investment and growth less so when you have high monopoly runs less So when you have low competition because when you have low competition you can use your interim profits to refinance yourself and investment You see what I mean, so but on the other hand, of course when you have more product market competition the you see the when you have Then the beta C1 tells you that a decrease in bond yield will boost your investment and growth But less so when you have more product market regulation Okay, that's why the beta PMR is product market regulation is the and in the inverse of product market competition Okay, so now let me show you I have five minutes left this result So what you see if you look at the second row first you see the second row is the beta Is the beta is the interaction between sectoral indebtedness and unexpected up in yield and you see that the You see the more indebted you are the more growth will be be value-added growth or labor market productivity growth Capital productivity growth the more those guys benefit from a drop in yield following omt But now you go to the third row and what the third row shows you is the is the in the triple interaction with when you add product market regulation and you see that so what the third row tells you is that when you Have a drop in yield it boosts investment and growth when you are more indebted But less so in you if your sector has higher product market regression Let's so if there is more monopoly power in your sector and that's the complementarity between the monetary policy and and and Product market competition. Okay, so I want to I skip I go fast there There what you can look is you can look at the transmission channel. You can say well Why why is the lower bond yield help firms because it will reduce their interest payment in particular So here what we do is to do exactly the same exercise, but instead of regressing The growth we regress the interest payment in rest of regressing growth or investment We regress the interest payment and what we see in fact is that You see the second row there will tell you that You see the lowering of the yield following omt will reduce interest payment all the more in more indebted sectors That's makes sense But the third row tells you that we that will do this less so when you have higher monopoly rents The lower yield will reduce interest payment But to a lesser extent if you have more monopoly power, okay So conclusions the main results we looked at the effect of unexpected drop in long-term government bonds following the announcement of omt We found that heavily indebted sectors benefited disproportionately from the unexpected drop in long-term government bond yield following omt But only in countries with low index of product market regulation In other words with countries with higher degree of product market competition So what are the next steps in this research is to look not only at product market, but look at labor market We want to redo to take as an interactor product labor market regulation You know that in France now we are engaged in a in a big labor market reform that will be key to save the world, okay And So that's the thing that that's one thing we want to do but the second thing here I use sectoral data But in fact what we want to do is to you firm level data and the whole problem is too much firms and to have Firm level measure of credit constraint and for that you need to more to to match firms and banks So shoulder of Gabrielle shoulder of has done that for a subset of firms in the US We have comprehensive firm level data and we just got the permission to match You know in France you are very good But you need to get permission from various institutions and each time they change the director So when you get the green light from one they change the director of the other one You see I mean and so you need periods where you have two remaining directors and And So it takes for very long. Well, I would have shown you the firm level results, but I couldn't because that's what I am aiming at Okay, so So what I will do later, but I think that I do later So you tell me you are the boss should I talk about the other thing because after since Chad we talk about something else Wonder what I was about to say was one response to Chad But I think you should be I think it's relevant for central bankers is to explain to you that we don't know how to measure TFP growth and because we don't and we don't know how to measure aggregate inflation And I think it's I guess it's relevant to this assembly and I would like to have five minutes to explain why But do you want me to do it now or to do it after Chad? Let's first let's have Chad first so that you can rest a bit. Yeah, because excellent. Thank you Okay, but I want to thank Philly The high amount of discipline of his presentation and respecting the time so You have a very difficult to roll now to discuss All the messages that we have received wonderful. Well, it's certainly a pleasure to to be here and participate in the session The paper that Philippe just presented is a is a classic Aggie on paper It's it's ambitious. It's provocative and it ingeniously address addresses a fundamental issue of importance to the ECB and many other economies As I considered my remarks, I appreciated the fact that I'm surely the least knowledgeable person in this esteemed group Concerning monetary policy and omt and the effects of omt. And so that left me with a conundrum I realized they had a comparative disadvantage in commenting on Philippe's presentation So with apologies to Philippe and his co-author I took the second alternative that the organizers offered me and that is to provide my own Perspective on the topic of our of our session growth and advanced economies and in particular I want to think about this question of How do we understand what accounts for the slow productivity growth that we've seen in the United States in Europe in the last 15 years? So that'll be the thrust of my remarks I want to begin by just reviewing the basic facts very quickly. You're familiar with them and we've seen some of them today My remarks are gonna basically say what is what is the recent growth literature? How does that help inform us about the sources of the productivity slowdown? So I want to review the theory briefly and then discuss the evidence from from the recent literature And then conclude with prospects for the future Okay, so this first graph is is one that I'm sure is familiar to you But still I thought it was helpful to look at the evidence I'll start with the US and I'll show you the the European TFP growth on the next slide first we have Private business sector total factor productivity growth and you can see You know from 1990 to 2003 growth was Fairly strong at 1.2 percent and then it fell by 40 percent in the second half of this period To 0.7 percent. So that's a sort of a useful summary of the productivity slowdown the slowdown and growth in the United States a Question that's commonly been raised in the literature is to what extent can mis-measurement account for this so You know the the rise of Google and Facebook and the the free stuff that we all we all receive Does surely there's some mis-measurement there and can that mis-measurement explain Some of the slowdown. I think the answer is yes and Philipp will comment to some extent on the extent to which mis-measurement can Can help on the other hand we know mis-measurement is not the whole story. So there's been recent work by Chad Severson John Fernald and other people arguing that mis-measurement is not the whole story and my version of that I guess I found interesting to look at the manufacturing sector where we think things are relatively well measured and where you know at least arguably a lot of the Facebook Google Innovations aren't necessarily impinging strongly on this sector and the striking thing there as you see the productivity slowdown is much much more dramatic As we saw yesterday. This was the rapidly growing sector Historically so one point six percent growth in the first half of this period in the second half productivity basically stagnated completely So a growth rate of zero point two percent. So really dramatic productivity slowdown. So How do we understand this fact is going to be the first question then extending the analysis to Europe? This shows you total factor productivity in the US again and then France Germany Italy in Spain these numbers are from the pen world tables latest database and You know, you just see that the slowdown is even more striking in Europe. So Germany shows slower productivity growth in the US since 2000 but basically France Italy and Spain all have productivity levels that are essentially lower than they were in 2000 France may be two percent lower Spain may be five percent lower and Italy may be you know more than ten percent lower and so It's not just enough to explain why is productivity growth slowed But at least in many of these economies productivity growth has been negative for a long period of time and that that provides a particular challenge. I think To understand Okay, growth theory offers Two basic ways to understand the determinants of TFP The first are these idea growth models that Philippe pioneered along with Peter Howard and Paul Romer And if we want to understand slowing productivity or maybe a decline in productivity I think the decline in productivity the idea-based models are less helpful for but certainly for slowing productivity We might think about it as coming potentially from two sources. First. It could be harder to find new ideas Maybe there's something with the production function for new ideas that says it's getting harder. This is something that Bob Gordon's emphasized Second it could be where we're looking less intensively So maybe we're just not looking as hard for new ideas as we used to be so the inputs could be could could be not performing as well So idea-based growth models is one sort of way to get a handle on TFP The other theory we have is misallocation And this is something that's been appreciated More in the last decade than ever before and I think it's one of the important contributions of the growth literature in the last decade It's associated with Restucia and Rodgerson and Shane clean out and the basic idea which is very clever is that misallocation the misallocation of resources of capital and labor or other inputs misallocation at the micro level When you aggregate up Shows up as low TFP at the macro level so TFP is about how are you using your capital and labor in the aggregate? You've got a certain amount of capital You've got a certain amount of labor well if that stuff is misallocated at the micro level that means the economy is less efficient And it shows up as aggregate TFP and so misallocation. I think provides a particularly Useful way to get a handle on declines in TFP because rising misallocation Would reduce the level of TFP and so that that's a that's a particularly useful Insight for possibly explaining some of the the weak performance that we've seen in productivity So those are the two the two theories that I want to explore and I'll take them in turn So first the evidence on ideas So so this slide draws on some recent work I've done with Nick Bloom John Van Rien and Michael Webb And the basic idea this project is to apply the solo growth accounting methodology To the idea production function Okay, and and what we find kind of everywhere we look is that ideas and in particular I guess the exponential growth that results from the discovery of new ideas Are getting harder to find exponential growth is getting harder to achieve Okay, and the way to the way to start to understand this is with the equation in the middle of this slide Virtually all idea-based growth models rely on an equation like this where economic growth Driven by the discovery of new ideas is on the left hand side It's produced by the discovery of new ideas and on the right hand side. We see that weirdest growth come from well It comes from research effort Multiplied by the productivity of those researchers Okay, so I want to use this kind of equation as an accounting framework to understand where our growth is coming from and whether What's happening to idea TFP Essentially, it's a solo exercise we can measure growth rates reasonably well We can measure research reasonably well and so we recover TFP in the idea production function as the residual from applying this equation Okay, and we've done this in a bunch of different places I did this long ago for the aggregate economy for the US economy and there you know that growth rates You know over long periods of time are relatively stable if anything they've been declining recently What about research effort research effort is rising enormously in the US economy? So if you go back to the 1930s the US economy today has roughly 20 times more research effort than it did in the 1930s Well, if you combine that with stable or declining growth rates and enormous rises in research effort You see that you know by definition in some sense idea TFP has to be falling dramatically and we find that you know since the 1930s It's roughly 20 or 25 times harder to generate the exponential growth through research, you know today than it was in the 1930s What we do in this new paper is apply this methodology Everywhere we can at the micro level to see what's going on with sort of you know detailed innovations So we look at Moore's law. We look at agricultural productivity You know the production of wheat cotton soybeans and corn we look at various metal medical innovations So pharmaceutical innovation Mortality associated with cancer or heart disease And we look at firm level data from Compusat So a very diverse range of sources that involve you know different data measurement issues And things along those lines and the robust finding we find everywhere We look is that you know idea TFP is falling quite dramatically So my favorite example of this comes from Moore's law. So Moore's law as you know is this statement that The density of computer chips has been doubling every two years for much of the last half century And since the doubling time is constant that corresponds to a constant exponential growth rate So a growth rate of like 35% per year. So really rapid growth. Well, how is that growth been achieved? How has that stable growth been achieved? Not surprisingly It's been achieved by throwing more and more resources at the problem So the number of researchers looking for ideas related to Moore's law has grown tremendously In fact, you know one way to summarize this is to say it's essentially 75 times harder today To generate that doubling of computer chip density than it was in the 1970s So so to conclude this slide, I mean basically what we find is everywhere we look ideas are getting harder to find The interesting thing though with respect to today's conversation is we find that's always been true That was true in the 1970s and the 1980s and the 1950s and the 1960s It's always been getting harder and harder to find ideas And so we've been throwing more resources at the problem to generate the exponential growth We don't find any significant evidence I would say that something changed in 2000 that it's even harder today than it was in the 1990s And so I actually think there's a lot to this story that ideas are getting harder to find But at least when we've looked at the evidence we haven't found that this is a source of the slowdown in productivity since 2000 Okay One other thing I want to draw your attention to on this equation Well, I've got the the slide up is that the number of researchers Rising is the key to offsetting declining idea TFP to get productivity growth in the economy And so I think of this as the red queen model of economic growth You have to run faster and faster to maintain the same constant growth rate And if we ever slow down our running That would slow down growth. And so that's what I want to turn to next what what's happening on the input side Let me skip this slide and go here. This shows you research employment R&D employment and select economies And one thing to appreciate is that, you know, the productivity growth in the US or in Portugal depends on ideas created everywhere So you want to look at the worldwide inputs not just, you know, researchers in Portugal or researchers in the United States It's I've got a collection of economies here You could see the interesting thing is the input growth research input growth has slowed down So in the EU is is the smallest slowdown say from 3.7 to 3.1 The US has a slowdown of about a third from 3.2 to 2.1 percent And then intriguingly Japan has a remarkable slowdown that you know goes back to the 90s in their research effort And so This at least is suggestive evidence to me that there may be something associated with slowing research effort In terms of explaining the slowdown in productivity and in fact one of the young economists Martin de Ritter Suggest that one reason for this after the financial crisis is the interaction between the financial crisis and financial constraints and research Investment so that could be part of the story Okay, let me turn out the evidence on misallocation. I Mentioned before that you know this misallocation literature I think it's one of the highlights of the growth literature in the last decade The insight is relatively straightforward. It's that the efficient allocation of resources Requires equating the value of the marginal products of inputs across firms And so you can use dispersion in the marginal product of inputs as a measure of misallocation And so Shane clean now and a number of follow-up papers have explored this The the recent paper by bills clean on and run looks at how this has changed over time in the US economy and The way I've plotted it here is they're asking Relative to full efficiency to what extent does misallocation Reduce efficiency and the surprising stunning fact that they that they document is in the green line Which is if you just apply the Shane clean or methodology to the US economy over time and ask what happens It looks like misallocation has increased dramatically The allocation was two-thirds efficient in the late 1970s and it's maybe one-third efficient by the end of this sample so an enormous increase in misallocation, which is Probably too large to be believed. I know when when when these authors saw this finding they were very puzzled They looked hard at it and everyone had worried about measurement error And so the point of this paper is to develop a correction for measurement error And they argue that when you correct for measurement error, it looks instead like the top lines It looks like the US is 80% efficient and that's sort of stable over time And so which of these lines is correct I think is is still an open question and a very important issue So to some extent increasing misallocation could be happening in the US, but but but maybe not What about in European economies? so there's a nice paper by Gita Gopinath Lucas Kerobonis and Sebnem Calimli Azcan and Carolina Viego Sanchez that look at Spain in particular and then they also have some evidence for Italy and in some other EU countries and Here I'm plotting the the standard deviation of The log-marginal revenue product to capital and marginal product of labor and you could see an increase in dispersion Significantly large be a 30% increase in dispersion and these authors claim that that could lower TFP by 7 to 12 percent in Spain and By 5 to 15 percent in Italy And so this sort of rising misallocation could be an important contributor to slow productivity in Europe This has been expanded upon by a bunch of papers. Let me not go through those in detail Let me just spend the last minute talking about prospects for the future On this slide, I guess I would highlight the middle bullet point I would say that the most optimistic reason I can find to think productivity growth might not Be so bad in the future is China and India each of these economies by themselves are as large as the US Europe in Japan And so you know as these economies develop how many future Edison's and Einstein's are they going to produce who? Historically haven't been close enough to the frontier to realize their potential But increasingly you're going to contribute to the idea frontier and then on misallocation I would say there are many open questions here There's some evidence in these papers that I didn't get a chance to discuss that Misallocation across sectors associated with the financial crisis sort of that too much construction leading up to the crisis Might be a factor of misallocation here What happens you if you apply this mismeasurement correction to the Gopinath et al. Results I think is an open question and then finally what are the fundamental economic forces that are behind this rising misallocation That's something the growth literature is working on Enormously right now, and I hope to be able to report to you some good answers in the future. Let me stop So in this very unconventional panel Since we discussed complementarity, I will hand back to Philip to have the presenters privilege of a few minutes to make some complimentary remarks So can I get back my slides? I can Show my slides. Okay. Okay. Okay. Excellent. Thank you. Oops. We're getting there Okay, so Thanks Chad. Oh, there's a really insightful analysis and so I Wanted to But to do so that way we launched the debate and I tell you the debate on that. There are several There is one thing I wanted to mention Very much in relation to what Chad has discussed for which might be of interest for you know this audience is We tried to make the argument with my colleagues Antonin Bergeau, Timo Bopart, Pete Clignot again, same Clignot and you will Lee That TFP growth is typically mismeasured and why is it mismeasured because you see what we know is the monetary values of objects So we take this glass between yesterday and today. We know the monetary value if the monetary value has increased the price We know it's pure inflation because the glass has not changed. Suppose I change the glass marginally Okay, I could and the monetary value has gone up. I could more or less There's entangle what's inflation and what's the real improvement in the glass But what happens when you have new goods replacing all goods then then we don't know what is inflation and what is not So what statistical institute do typically is they use something they call imputation in in Europe We call that extrapolation because we like you use different words from the Americans But it's exactly the same the same thing we do we miss in the same way with different words Okay, so what we do is that for any product category? We use the goods that have not been replaced by new goods that have not been subject to great to calculate the inflation rate On those goods and then we extrapolate we say that should apply to the whole product category and The question is how much growth are you missing and you are missing a lot the more creative destruction you have the more missing growth You are the more TFP growth you are missing and this is a suggestive feature here I represent on the very horizontal access I rank us sectors by the extent of the creative destruction here I measure it by job creative destruction, but you could also measure it by firm churning and On the vertical access you have the correlation between TFP growth and the flow of patterns And you see that the more creative destruction there is in a sector the lower the correlation between major TFP growth And the number of patterns I mean the less innovation is reflected in growth Okay, so that was the that what started us and in fact what we did is that we were able to compute In fact we and of course because the more creative destruction you have the more you rely on Imputation and we screw up when we rely on infusion because typically new goods are better than all goods So when you do the imputation you overestimate inflation and you underestimate productivity growth And you see that's going to be quite dramatic We find if I we have a method for the whole US economy and we find you see that the you see you go from the Measured growth measure TFP growth in the US over this period to true TFP growth And you see that in fact you are adding every year more or less 0.6 0.7 percentage point of TFP growth that we are missing you see I mean so it means that in the US between one fourth and one third of TFP growth is simply missed by the calculation, okay, and I've done we've done the same thing for France and And we had a lower period and France also you have those are the measures for France and so you hear you have the US France comparison and But in France is dramatic because in France in fact half of the major TFP growth is missed What the TFP growth is double of what you measure it in France But because Francis has no TFP growth is not because and what's interesting that this number in spite of what we call dynamic Dynamism has not gone down over time But the missed growth has become a bigger part of total growth because total TFP growth has gone down So of course now when you go for inflation targeting monetary policy You have to recalculate because if we know that we have even less inflation than we say That how can that not be of interest to the crowd here in this room? It should be So that's why the first two remarks on on on Shad, which are very interesting on we had this discussion yesterday You find that ideas are harder to find in given sectors. The question is we may keep finding new sectors Where again we can have so maybe we may have decreasing returns to ideas in any sector But maybe we keep finding new sectors those are the question we are wondering yesterday And I wanted to have your your feedback and on the misalocation I think that I was we were wondering and that's the discussion I have with Ufuk axi-git and with Gilbert set at the Bank of France It could be that the you know the lowering of bond yield You know the the is the quantity easing of the easier monetary policy, which was had many good aspects to it Maybe also allowed more inefficient firms to remain you could look at the effects of those on firm dynamics And it could be that the monetary policy itself had it is good to lower interest rate in recession Maybe there is a downside to do it too long is because you may have an effect on firm dynamics You may allow less efficient innovators to remain and prevent entry by more efficient innovators So what would be very interesting and that would be a way to to match your part and my part of this We'd be to look at how monetary policy if you maintain low interest rate too long You you affect negatively the firm dynamics and that could be Feeding back into the misalocation and that I think would be something this is sorry Thank you Thank you very much This session Started slightly behind time so I don't want you to prevent to participate in this discussion my attempt will be to channel all the energy from Philip into the audience for this discussion and I Would therefore ask that we take 15 minutes of the coffee break and Maybe that we would then have 20 minutes discussion So I will call for first round of Questions On this side, please do as always present yourself for the convenience of those who don't see you For Kavila and German consulate of economic experts Coming back mostly to the first two talks. I mean Robert Hall and go dickets and Pointed out there was something unusual in Germany or the heterogeneity was particularly pronounced and since we in our console Keep looking at Germany every year and write a report for the public and for the government I just wanted to offer some additional Comments on that first of all if you look at Robert Hall's Chart the change in Germany happened actually in 2005 not in 2007 That's when he kind of starts rising the measure you looked at in other measures, too So, you know in terms of thinking about what's driving it when would are you talk about what happened before what happened in the period before and That was also quite different period to some of the other countries you looked at for the euro area for example, you know In Germany, we had a huge unification boom real estate boom Which we're we took many years to get out of the side effects So just at the time when Spain was embarking on this real estate boom We were the sick man of Europe according to the economist If you look at 2003 2004 then The things that changed at the time in terms of the policy perspective was massive tax reductions So this is the only major tax reduction we had in a long time labor market reform and also a very Focus of the unions on job security versus versus wage and Then if you look at the period afterwards There's been a change in the structural rate of unemployment We've really created a low wage sector which massively expanded employment and so from the supply side Created demand you know change the demand conditions in the country and you have to look at it as a two-sector economy At the same time we have a high wage export sector Which didn't grow that much in employment, but certainly grew and then There's some element of luck you know We were hit by the recession in terms of output very strongly in terms of employment not much because they kept on the employment and We had an element of luck if you think of you know, there's not just the countries you looked at But the demand for that export sector came back from Asia from China and from California That's where all the Porsches are driving around and So I think summing up Robert Hall country to a gaudy emphasized that you really have to look at the supply side and I think that You know if you look at the heterogeneity in Europe I think you have an additional evidence for that So I think most of what happened in Germany was more driven by the supply side Factors and that I think that also changes the policy recommendations your good gift in terms of productivity You have to think of composition effects which to which are part of it For example, you know the of course productivity stopped growing very much as we increase the size of the low-wage service sector Right because there you don't see much productivity growth. That's only a part of it but maybe an interesting element and just my final point in terms of In coming in quality we're still a country with a huge amount of redistribution. So even after those reforms And so in fact income inequality in spite of after tax and transfers didn't increase since 2005, okay? Thank you folk. I take some more questions. Please be more concise to take as many questions as possible Michael Berger homeboat University I wanted to pile in also on the German thing Germany has got different institutions and what happened in Germany is really Exceptional Bob Hall shows that and it's interesting to know why first thing is they increased labor force participation from 60% In the 1980s to just under 80% and this has been done by increasing part-time work So looking at earnings per person is gonna pick that up in a very interesting way And it's kind of like work-sharing because total hours in Germany haven't changed much since the 1990s They've grown a little bit since the Since 2010, but they've actually been pretty stable the second thing is that wages have not grown since the mid 2000s and third dispersion of wages has gone up So what Fulker was saying about the low wage sector is exactly right the question is why did Germans take up those low-paying wages? And the answer is they had labor market reforms that basically forced them to by cutting unemployment benefits Duration and replacement rates. So that's the the key thing is understanding this It's not really a miracle. It's just basic neoclassical economics working the right way Thank you on the first paper If the issue is stagnation, why don't you use GDP per head as an index? Okay, if you want to measure bargaining power, perhaps the labor share is better than real earnings So if you measure GDP per head, definitely there you will see that investment played played a huge role in In the ranking Now all the discussions I fully agree that the demand had a role to play in all this Especially if you look at GDP per head I'm not sure that as far as the labor share is concerned We should ignore investment. I mean it comes from yesterday's paper, especially the second paper That the labor share fall implied the operational Surplus got up, but it was not invested. So perhaps a virtual Question would would be what would have happened if the operational surplus had been investment in the labor share Thank you Thank you. I will now ask the first paper discussions and present of To answer these first three questions Would you start Bob? Well, okay, I think that the the two remarks about Germany are helpful and It just illustrates the point that I think to understand a lot of the findings In this kind of a decomposition you have to know a lot about what was going on in the country so You know the ideal version of this would have a one expert from each of the six countries rather than Then one person trying to cover six countries who only knows about one country, which was in fact what happened So I welcome that kind of explanation I'm certainly aware that that Germany had important changes and in labor market institutions On this question of GDP per person First of all I would be be should be clear that I would think of GDP is on the income side not the product side So I don't think that I don't think there's any special question about the role of investment And again as I explained at the beginning The the the focus on this is thinking about earnings as the basis for the well-being of Most ordinary workers Obviously GDP per person is also is a very useful thing to know for example since GDP is basically the tax base If you wanted to know what's the scope for redistribution you'd want to know about GDP per person I totally agree with that but if you had to pick one thing to try to deal with a stagnation issue which was what I wanted to do then I think Compensation per person this is a more natural singles thing to do, but I'm certainly not against measuring GDP for sure I'm not sure I have that much to add, but I guess I think of I think of it as Having been subject I guess to different idiosyncratic shock and I think it is useful to think about a common mantra I mean Germany needed different things than Spain and Italy and that explains the big divergence And I think you know it's important to think in terms of the Monterey Union to make sense of that It's harder if you take a purely neoclassical approach like the production function approach that that Bob Paul had I guess one puts back on on the Rigidity, I mean the this notion that the labor market in the US has never been as strong. I guess that does sort of Beg the question why is the inflation not higher? No, no because there's no basis in in theory for the notion that Market tightness Affects the rate of change of a wage or a price. I mean that's an idea that seeped into economics But has just remarkably little basis. There's no there's no micro theory that tells you that rates of change depend on gaps that's just something that got into Thinking in the post-war period, but it's lacked any meaningful microeconomic foundation or put it differently It's not the case when markets are very strong that The typical seller says oh the answer to my problem with strong markets is to raise price You know it may we have a theory of pricing a micro theory of pricing that doesn't doesn't incorporate that at all It's just something that got pasted into Macro there's a famous in the 1970 Origin of sort of modern micro macro There's a famous passage. I quoted many times saying it's time to forget p dot is equal to lambda times D minus s and have a real theory and those of us who believe in that real theory of don't Anyway, I have a new paper with Tom sergeant that elaborates this attack on the Phillips curve I'll let you take a look at it Okay, I think this invites for more questions Over here So I have a question to chat Jones Presentation which actually I found very interesting. So this idea of the idea production function where you need to throw more and more people In in order to generate the same amount of growth. I was just wondering whether that would not imply that The wages or the marginal productivity of an additional workers actually Decreasing and whether that would not imply that the wages in that research sector should Decrease over time. I see there are two different forces here Growth operating on a larger base may partly compensate for that But what what's the implication of this theory for wages in that sector? Okay, thank you and then a child's good Much of the analysis Charles good one Much of the analysis about European differences relates to real exchange rates being Very favorable for Germany very bad for Italy and Spain, but there was no mention of real exchange rates from Second question is that the speakers all seem to assume That TFP is an exogenous factor driving wages and earnings How about reverse causality? When labor is cheap and plentiful and profitability is high Why bother to try and get the last ounce of Output from your workers Can't it be that the problem is that low wages are driving low productivity rather than the reverse Thank you, and then I still have philly plane at the end the issue of the productivity in the US their pure economic analysis has shown that partly the issue is firms in the US outwardly shifting productivity to overseas locations so to for example contract manufacturing in in the Far East or Tax planning in terms of a location of intellectual property elsewhere Means that maybe and it can be quite big according to pure economic analysis That the rate of productivity growth in the US is understated and this is mechanically at the economic incentive to do so It's pretty strong given the US tax system Thank you Sure sure absolutely So this first question was it was a great question if we're if we're throwing more and more resources at the idea Production function doesn't that imply that the marginal product of those resources is going down So wouldn't we see wages going down and is there a problem there? I think there are two answers to that question. The first is at the economy wide level The the general equilibrium force is basically population growth So where are we getting the resources economy wide to throw into this production function part of it Is that the the population of the economy is growing the growth in the population? Provides the demand as well And so the growth in population provides the supply the growth in the population provides the demand And so in general equilibrium models the wage of researchers grows along with the wage of workers And there's no there's no problem. They are everything's in line when you look at the micro evidence I think new forces get can be involved. So for example with Moore's law We're clearly throwing more and more resources at Moore's law at a much greater rate than we are in others in other parts of the economy. I think that has to be explained by some, you know General-purpose technology features of information technology, and so that would provide the demand side there, you know Across the sectors researchers have to have wages that are equated. It's the same misallocation point And so I think everything hangs together when you look at the general equilibrium model Okay, so real exchange rates You know as a as a student of solos I have to respect the the framework that is If there's a production function you can log linearize the production function and and you can do interesting things Something like real exchange rates aren't left out of the story But they're a source of movements of those variables, but they don't shift the relationship That's that's the whole idea of Productivity measurement So I don't think it's appropriate to say that that framework leaves out real exchange rates Of course, there's a there's a whole separate question whether the whether trade is properly incorporated That was the outsourcing question the last question And I Guess all we can say is that that Doing national income accounting with outsourcing is a huge challenge You can see that in state data historical state data of the US by far the most productive state in the United States in 1980 was Puerto Rico And that's because Puerto Rico was the location of highly productive Plants That have been moved there because of a tax advantage But and therefore they had all the income which then translates into productivity booked in Puerto Rico But but it was in it was it just meant you had to leave Puerto Rico out of your state-by-state analysis As far as TFP, it's very important to understand. This is a measurement exercise not a causation exercise If just because we measure TFP doesn't mean we're taking it to be exogenous Now of course when you start talking about What a what accounts for wage growth and then you're beginning to get into that territory, but the exercise itself is just measurement so Then the question is How accurate is the measurement? But there's no there's no taking a log linearization of the production function does not take a stand on on the Exogenous and there's no kind of metrics in this work It's all just log linearization, but that gives it its power It's a limited tool, but I think it's a pretty powerful tool, but you just have you have to be careful in interpreting it Okay, thank you. I will have to close the session. Unfortunately. I have two last short questions They will be able to be asked, but I don't know if we have time to give answers So we have two minutes 35 seconds first question. Thank you. Just good to have a question for Philip one of the surprising facts in the graph that Bob showed is a Friends in France is very stable resilient doesn't grow very fast, but it grows very stately very little volatility Still we've seen no reforms. How can you put the two things together? I've been puzzling over that question for decades So I was wondering the countries in your sample how specific they are in terms of the of the financial market performance in these countries you and since your channel always works for financial markets How much of these coefficients would actually depend on the status of the would be country specific and depending on the status of the financial markets and How that would affect actually also looking at the interaction between financial market reform sleep public market Okay, Philip you have one and a half minute And so it's typically sectors that are more prone So it's countries that are more indebtedness, but also we look into sectors sectors that are more prone to be Financially constrained or liquid Actually to be shocked that are not the most affected by the by the combination of the interest rate And and of all the more when they have high competition You see when there is high competition in the country So that's that's what we so it's a character We look at that characteristic of the country But you can look at characteristic of the sectors of using a kind of Rajan Zingales say are the sec is the sector more prone to be liquidity construction and those is for those that that it's more that the thing is more dramatic So in fact with a firm level what we've been I said we have firm level measure of credit constraint that's what we get through the Bank of France and France, you know, we were we were mediocre in in the down and the up core and We never go down much, but we slowly decline income And that's the dramatic part as well is because it took time to realize that things the boat was sinking when the boat sings very slowly You see don't realize and at some moment you have to that moment you have to react because the euro in a sense is an anaesthetic You said I mean before nine, you know in a 83 you had to react and something like that did not happen since then Okay, thank you very much. I think this concludes the session Thanks a lot Present us discussant coffee break