 Mae'n cael ei wneud, mae ydych chi'n gweithio'r cwm pan cryschol wedi gwybod hwnnw, ac mae gennych nid o'r gweithio'n gwybod y ffaith ac oedd yn 2008 o'r cymdeithas yn cael ei gweithio. Ond nid yw'n gweithio'n ei wneud o'r David Vine. Mae'n cael ei ffwrdd o'r pryd, ac mae'n gwneud agnodau a'n ei cofnod o'i wneud o'r ei wneud o'r cwm. So David, rydyn ni'n gweithio'n ei wneud o'u'r ffwrdd? Bench bilang, dwi'n fawr i'n gwahanol, rydyn ni'n d greni phos gan ymgyrchol? Rhaid i ddweud. Mae hyn yn ystafell yn ddi hereb yr ysbylio, ac mae geisio'n gweffredig ynogi. Rydyn ni'n ffawr cymdeithio ar hyn i'n gweithio. Ond rwyf amser sy'n gwybod i'r dda ffae ffordd o'r ddwy Emhrae ac mewn addysgau. Roedd ni i'r mynd yma sy'n ddwy'r mach chipedd syddtwch yn gwybod. Fel cyfnod y benshmark model oedd yna mor ddau'r ddweudio'n sgwrs a'r ddweudio'n gwybod a'r gwybod yn beirio, ac rydw i'n ddim yn ei ddweudio'n ddau'r bod yw ddweudio'n ddweudio'n ddweudio'n ddweudio'n ddweudio'n ddweudio'n ddweudio'n ddweudio. Rhywbeth y gallwn wedi gweld yn gweithio'n ddweudio. Olivia Blonshard famously said that the state of macroeconomics is good about three months before the world collapsed. As a result of this mismatch between what I believed and what was going on around me during the global financial crisis, I set about with a group of colleagues to have a project which will lead to a double issue of the Oxford review of economic policy on rebuilding macroeconomic theory. And the project that we've set ourselves is, imagine asking to build an approach to macroeconomics that you can teach to finally your undergraduates or first year graduate students. With the kind of clarity that happened after the Keynesian Revolution when people were teaching Keynesian economics and people believed they had something useful. It's clear that as we went through the GFC what we were teaching in that first core course in macro to graduate students just didn't bear any relation to the kinds of questions we'd needed to think about. In this paper I'm going to talk very quickly about having been through paradigm shifts before. I'm then going to describe this model that I felt so comfortable with until the world collapsed around me and my model. And then in the later parts of my talk I'm going to talk about in a rather conventional way about what to do next. And it's conventional in the sense that this is what you would want to teach people to understand macroeconomics. It's much less useful than the kind of work that John has been engaged in which is what you might call real policy modelling for people actually giving advice about policy. And what I'm going to say is in a sense much less fundamental than what David's going to say which attacks I think in a very interesting way the fundamental premises of the whole modelling framework that we use. Let's just describe the 30s because they give you an idea of what a paradigm shift really is. When Keynes was confronted in the Macmillan committee with Montagu Norman in 1930, Montagu Norman said, I can't understand what all the fuss is about, industry should be able to adjust itself. And Keynes in 1930 didn't have the equipment to describe why macroeconomic policy was necessary to do the stuff that adjustment of wages in an individual uncompetitive industry couldn't do. By 1936 he'd written the general theory with nominal wage rigility and what that provoked was a need to understand the consumption function, the multiplier and liquidity preference in order to be absolutely sure why the interest rate wouldn't spectacularly well adjust to ensure that the economy kept on working without unemployment problems. But more than that, as Hicks and others showed, this analysis in the general theory introduced general equilibrium theory. That's to say that what happens in one market affects another. That when there's unemployment in the labour market people get less income and buy less goods and that causes unemployment to get worse. And watch that going backwards and forwards between different markets. And I think of that as a real paradigm shift. That's to say new content including the multiplier and new method doing general equilibrium theory. In the 70s and 80s the problem was not mass unemployment but what you might call mass inflation. And the results of that for economic thinking were very mixed and the battles of that time still live on. People like me adopted an evolutionary approach. We had introduced the Philips curve. We persuaded ourselves that it was vertical in the long run so that that meant that in the end if you wanted higher employment in the economy it needed to be supply side policy. We thought of the need in the end to put a nominal anchor on the economy in the way that the ISLM system had never done. Which became after 15 years of battles monetary targets, the exchange rate mechanism in the end inflation targeting. And we gradually understood how to connect this up with the analysis of economic growth which we'd all been doing. I think of this as evolutionary and it was certainly a new way of thinking about policy but it evolved out of what we'd done before. And it led directly to the New Keynesian model that I was so falsely comfortable with in what I described. But there were rock throwers in the American Midwest whose approach was much more revolutionary. They had two ambitions first coming from the Lucas critique which argued that if people behaved in one way and you changed policies of course people would change their behaviour and to understand this you really needed to do all the things written on that slide. That's to say micro found your models, have optimising agents, have them forward looking and constrict all this analysis to consistency with rational expectations. And that dogma as a way of thinking about decision making which I think David would rightly ridicule from the world he lives in has been enormously hegemonic in the subject ever since. And you'll see that and it'll form part of this discussion in this panel. The second thing these Midwesterners wished to do was to show that Keynesian economics was a bunch of nonsense that markets would clear themselves. We'd go back to the Martialian economics that Maynard Keynes had escaped from in 1930. And that view although there are still many who live well east of Boston, well west of Boston, but nevertheless their dominance has disappeared. But we're left with this insistence on consistent optimising micro founded rationality as a way of doing macro even though we are New Keynesian. Let me just describe to you what we are as New Keynesian. It's an economy which those of you who are economists will know has a consumption function which is forward looking called an oiler equation, a forward looking investment function. It has forward looking and consistent inflationary processes although consistent and forward looking there is the kind of rigidity that means that policy is necessary. And inflation can result when aggregate demand is different from aggregate supply and you have a tailor rule and you have a central bank following something like a tailor rule to discipline this economy. And we know very well how to think about what that economy does. This shows in a way which you might not be able to read enough to make much sense of the effect of a permanent productivity shock in this economy. To the top left up there you have output falling that causes investment to collapse down top line, the capital stock gradually falls. The interest rate has to be cut to ensure that the demand for output which has gone down because of investment falling doesn't fall too much and that fall in the interest rate will counteract the disincentive to invest caused by the loss of productivity. Wages will gradually fall but the crucial point to describe in that story is that you end up back in a straightforwardly sensible equilibrium that looks just like you were before except that the economy is less productive. But the crucial point is how strong an attractor in technical language this equilibrium is. Similarly if you do an inflation shock with this model inflation the bottom right hand corner goes up so the central bank will raise the interest rate, output will be depressed, that will cause both consumption and investment to come down. Notice that the capital stock will first fall and then rise back up again in the top right hand corner. This operation of the monetary policy committee will have profound effects right through the production system of this economy. But again crucially this economy at the end of these slides goes right back to where it was at the beginning and that's what I'd somehow persuaded myself during the great moderation the economy was really like. Why is this a world that made me increase, I've lost the construction of that sentence. Let me start it again and just say because I believe this I felt increasingly foolish as I watched the global financial crisis evolve around me. Look at what happens to pretend actual output in this economy which is the US which was growing at that steady rate before 2007 and look at actual in the red line at the bottom. Now the purpose of showing you those slides previously was to show you that when you disturbed this economy that we thought we understood you go right back to where you started. And that doesn't go right back to where it started 15 years no 10 years later. Now you could say and there's been stupid attempts to rescue this way of thinking by saying that the global financial crisis was a productivity shock. And that's why we're lower and some of the discussion about the secular stagnation on the first night had that aspect. But there's much more interesting important stuff to say which are there put Turner put very clearly. This is a problem that demand 15 years later is still below supply. Now that's just not consistent with the model I simulated with those pictures a minute ago. And why is the model stupid in being confronted by something like this. There's no risk premium in this model. And secondly, the powerfulness of rational expectations looking forward is what gives this the attractor property of going back to where it started. Everyone understood stands. It's like bringing up your children. You teach your children to believe that the world's going to be all right. And when they're confident and well brought up, they do things which are consistent with the world being all right. And it turns out to be all right and that's self fulfilling. That is not what's happened over the last 10 years. But secondly, this experience of the last 10 years is not consistent with that very interesting paper which was presented on the first day. About the effects of autonomous demand expansion on the economy. And I flicked through these slides very fast. But the important point is from these that that lecture, these simply photographs of the slides of that lecture, you increase demand permanently in this economy. And 10 years later, the blue line is well above the red line that's zero. And you remember in my model that I simulated when you shock demand, it just goes back remorselessly to where it started. Look what happens to inflation. That's meant to rise in my model in order to make the central bank raise the interest rate and pull the economy back to where it just doesn't happen. Look why it doesn't happen because when there's a stimulus to demand, people accumulate more capital and that enables the economy to produce more to sell. And satisfy the extra demand. Look at that blue line just keeps on rising that stock of capital. And finally people work harder to employment top right hand corner is higher. None of this is consistent with the model that I described to you. Furthermore at this INET conference, we've had three other issues that are very important and not in the model. Inequality, which we all care about, but it matters in macro. This top 1% grabbing all the income, which has been so much a part of our discussions reduces the willingness to spend. Secondly, a dear Turner talked about the declining cost of capital and that reduces investment. And if consumption and investment are both low, that means that fiscal policy will be necessary when the economy is in a depressed state. And none of this is consistent with the previous understanding. Now, can we do something that's helpful? I've got more things to say on these slides, but I want to focus in on a very concise, simple blunt answer to this question provided my colleagues, Wendy Cullen and David Soskus, which gives you an answer that's consistent with reorganising your thought in a way that's helpful. Supposing you take this model that I've been happily simulating for you and you do five things to it. Supposing you say that there's a zero bound to interest rates, what am I trying to do? I'm subjecting this imaginary model to a very big negative shock like in the global financial crisis in 2008. And I say suppose interest rates reach their zero bound so you can stop being able to control this economy with monetary policy. Suppose for reasons which Paul Krugman many others are puzzled by, inflation doesn't start deflating when there's lots of unemployment, second point. Suppose that there are what we can technically describe as strategic complementarity amongst investors, which are easy to understand in a climate of uncertainty that we know about. I won't invest because you won't invest because I won't invest. Fourthly suppose that in such a world there's very little technical progress, an explanation of the very poor productivity that we're experiencing. Why? Because there's very low investment. And finally and crucially suppose there is not this happy sunshine in the rational expectations way that leads us all to believe that one day everything will be a right again like my young children and getting on with being alright again. Those five things are sufficient to turn that model that I described to you and simulated into a model which can stagnate in the kind of long run Keynesian unemployment equilibrium that that picture there tends to suggest we're living in. And it seems to me that a task of those of us who teach macroeconomics to undergraduates and to graduate students is to say well when the times are good it looks like those lovely pictures. But when the times there can be bad times when it looks like that not so lovely picture. And that leads me to my conclusion which clicking back through many more suggestions of details which I don't want to go through to say what do we learn from this exercise. I think that there is a valuable search underway by people like me who are teachers. That's to say to find a simple straightforward model which can when things are good describe a world in which things are good. But which is also capable of clearly explaining to us the list of things which taken together can make the world continuously not good in the way that I've described. And having done this that will enable with that ambition we will we will do two things. First of all we'll see off this hegemony of rational optimizing forward looking consistency as a requirement for everything that we think about. But secondly and I wanted to end this way we'll make room for the sort of stuff that John Muir Bower does. That's to say we will make it no longer intellectually respectable intellectually unrespectable to be not using these assumptions and to be saying valuable things about policy. There's like the 16th century after the reformation we have to recognize there's no longer a true church in macroeconomics. Thanks very much. Right the future of macroeconomics why central bank models failed and how to repair them. So I'm going to skip through some of the critiques and go straight to this one. The new Keynesian dynamic stochastic generic equilibrium model. Well my view is that would have been nice because I don't think the theory was new. In fact the theory was outdated outdated by the asymmetric information revolution of Stiglitz and Aquilof. We've both been at the conference. It wasn't Keynesian. Was it ignored the coordination failures some of which David just talked about especially between the real economy and finance which David didn't talk about. It wasn't dynamic enough because in the real world lags real world lags are very different from the way these models describe. They're much slower actually. Hardness stochastic stochastic means to do with distributions. So we've got two kinds of distributions we have cross section distributions heterogeneity agents differ in all kinds of ways and we have time series distributions probability distributions over time events happen with uncertain outcomes. So both of these the true uncertainty in the real world was ignored essentially in these models and hardly general equilibrium missing most of the system feedbacks. Some of which I'll talk about. Now we know from David Henry's work that rational expectations and intersempel optimization break down very seriously when there are structural breaks in the economy and radical uncertainty. Of course the economy is subject to many structural breaks. Think of globalization technology. Think of the uncertainty we face about artificial intelligence and robots and so on. So let me give you a brief outline. I'll start with the representative agent model and why it's so wrong. Then talk about information liquidity constraints income uncertainty. Explain why the revolution in credit market architecture which be the big focus of this conference actually the lunchtime post lunchtime session today for example and why debt matters. We had a wonderful session on debt yesterday then move on to consumption because that's 60 or more percent of total GDP. It's really crucial to understanding how the economy operates and it's the key weakness actually of the New Canes in DSG model. And then talk about what we can learn by empirically modeling consumption together with the other things that households do namely the portfolio decisions. Right. So the lack of a representative agent is should be pretty obvious in real business cycle models. There's a model of a representative degree of unemployment. Every household has the same level of unemployment which is ridiculous. We know unemployment risk varies hugely by occupation education. We know in the real world there are credit constraints mortgage defaults negative equity very hugely across households. How can we address this? We can no longer work with a representative agent that represents the average of the economy. We have to think about stochastic aggregation working with means and other distribution parameters. And doing so we can still make progress. Now one very important insight from this very famous paper by how talk in 1956 is that what happens in the micro level need have no relationship in terms of functional forms with what happens at the macro level. So with his assumptions you've got no substitution in the technology at the micro level but a high degree of substitution at the micro level through the extensive margin. We've done work using these ideas on UK mortgage defaults in which we can track negative equity using aggregate data and it works really, really well. Okay income uncertainty liquidity constraints and buffer stock saving that's what was missing in the New Keynesian DSG model. Angus Deaton we've got the Nobel Prize laid the micro foundations for this approach. With liquidity constraints and serious uncertainty many households indulge in buffer stock saving and they have shorter time horizons and the more uncertainty they face the shorter the time horizon. And in his book he summarises the evidence against the simple permanent model. Chris Carroll has done a lot of work in the same area. Now in Angus Deaton's 92 book, the Clarendon Lectures at Oxford, he makes the point that we really need to integrate, we need to go beyond the simple buffer stock model and integrate the treatment of illiquid assets like pensions, stocks, bonds and houses. Into the model of the household. And very recently, 20 years on, but very belatedly the professionals come round to thinking in actually implementing his suggestions. So Kaplan, Violante and the other papers here that I've mentioned are doing a really good job in looking at the micro of what's going on there. So I mentioned that credit market architecture has gone through huge changes in the web version of the paper. I'll talk about the US, which is really interesting, but just for the UK we had absolutely radical changes compared to the 1970s. So we abandoned exchange controls in 79, the banks invaded the mortgage market, the building society responded, they were given new liberties. A second way of innovation happened through the centralised mortgage lenders that invaded the market. Then we had a mortgage crisis and a credit crunch. And then in the mid 90s, a new drive for liberalisation, particularly in the bites of that market increased securitisation, a new breed of centralised lenders who weren't working through the high street. And we've tracked a measure of what happened there. Since 2008, because we had a very serious credit crunch and partial relaxation after that. So taking these shifts in credit market architecture into account is really crucial to understand what's going on. Now, we all know why debt matters. Irving Fisher's debt deflation theory summarised here, you understand this very well. You'll know about Adair Turner's work and Schulrich and Taylor and Mien and Sufi and so on. In fact, back in 1990, Anthony Murphy and I wrote a paper about the unsustainable credit boom that had happened in the late 80s. And we argued that it was going to have very negative consequences, which indeed it did. So when I said that the New Canes and DSG model lacked general equilibrium, I mean these are the kind of feedback loops that were completely missing in that model. So in the US, the mortgage and housing crisis had feedbacks through lower demand for housing, big impact on the construction sector, big impact on consumption. And then the negative feedback on the finance side, bad loans fed into the ability of banks to advance credit, credit spreads wide and so on. So you can see that without serious public intervention, the great recession would have been far, far worse. So I'm particularly interested in understanding the consumption linkage in that story. So we need a more general consumption function and an income forecasting equation to capture the consumption channel in that financial accelerator. So a really important aspect of this is we want to take full account of the structure of balance sheets. So it's not enough just to look at net worth. Net worth is liquid assets minus debt plus illiquid financial assets plus housing wealth, that's net worth. I think it's really important that we disaggregate the balance sheet of households when we try to understand what happens to the household sector. And moreover, that we take into account the shifts in credit constraints that have occurred over time. Now my criticism of the New Canes and DSG model applies to a degree to a non-DSG models of the Federal Reserve, of the Bank of England model and the LBR model, and indeed the new ECB non-DSG models which are being developed. Because in those models they impose a net worth constraint. So the only way in which asset prices and credit and so on debt enter the consumption function is through net worth, and that is incredibly restrictive. Now a key aspect of this is to understand differences between countries as well as differences over time. So if you have a very restrictive, a conservative credit market that suggests that aggregate consumption is going to fall when house prices rise. And that's because future first time buyers have to save more for their deposit. And if people don't have access to home equity withdrawal, then they can't spend the increased collateral that they have. Whereas liberal mortgage markets like the UK or the US imply the opposite. You have a low down payment constraint, you have to save very much to enter the housing market. And the access to home equity means there's a big response in the aggregate to an increase in housing collateral. So this is a picture of what happened to the mortgage credit conditions index in the UK. You see the big liberalisation from 1980 onwards, the early 90s credit crunch and the much more serious credit crunch in the recent past. Now we've implemented these ideas in a model of UK consumption that incorporates these shifts in credit and disaggregates the balance sheet. And what can tell very good stories about what happened both in terms of the long term and in terms of the recent crisis in the UK. So the red line is the consumption to income ratio. You can see the fluctuations in it. The blue line is the contribution of the credit conditions index directly to explaining those variations. And the pink line is the interaction of housing wealth and credit conditions. So that says that when credit conditions are loosened, the housing collateral effect is much more powerful than when credit conditions are tight. But given this liberalisation and this increase in house prices, there's a big offset which is the dotted declining line which shows the contribution of liquid assets minus debt. So as people build up debt, so the burden of debt, the permanent burden of debt on consumption becomes more and more serious. That's the payback effect of liberalisation. And so when in the crisis you have both the credit crunch and the collapse of house prices, you get a very sharp fall in the consumption to income ratio. But households can't leverage fast enough. They're stuck with high levels of debt and therefore they have to cut consumption. The model has other things in it. It's got the stock market, it's got income expectations, it's got interest rates, all of which are quite important. But let me skip over that. So some of the insights, well, one learns a great deal about why it is that the consumption of income ratios trended up over time both in the US and the UK. And of course what happened during the build up to the crisis and during the crisis. Another really important insight is into these shifting correlations between debt and economic growth. Ikeen talked about this the other day. He made the point that there's a positive relationship between credit growth and a negative relation to the level of debt with economic growth. And in my approach you can reconcile this because when you have shifts in credit supply that explains why it is at certain periods there's a positive correlation between debt and consumption. But credit crunches obviously arise when that changes. Our models also take uncertainty quite seriously and we can proxy this pretty well with the change in the unemployment rate. Moreover, in the UK where we have a floating rate mortgage market there's a big cash flow effect in the aggregate for changes in nominal interest rates when nominal interest rates go up, the cash flows of people in debt fall. So we need actually not just a consumption equation but we need to model the combination of consumption and balance sheets. And we need to do that by estimating a system in which we can extract this ambiguous thing that's really quite hard to observe directly namely credit conditions. So we do that in a latent variable approach and I've done some work with people at the Bundesbank and the Bank of France and the results are really interesting. I think I've used 15 minutes if I... You've had 15 minutes but you can have a minute more. So with this approach one can tell some new economic stories. One is that money transmission in Germany for example is very different from the US or the UK. The role of demography is really interesting. Demography matters a great deal for the composition of portfolios and then feeding into portfolio composition it then feeds slowly into consumption. So one can work out long term implications of the big shifts in demography that are taking place. Now the part of the story is that the role of debt is very important so the empirical relationships that say that high debt of things being equal cause low consumption cause some doubt on central bank policy as has been discussed several times in the conference. You know if you have appeared a very low interest rate for a long time in a courage household to build up more and more debt in the long run that high debt is going to constrain consumption in the future and it's not a good long run policy. So I'm not claiming... I'm talking about macro time series work here. I'm not claiming that micro evidence is unimportant. Micro evidence is really important but macro shifts in credit conditions and asset prices have macro effects that can't be ignored. So finally to return to the New Canesian model. Defenders of the model claim that it's flexible and that we need that approach to tell economic stories built on micro foundations and to incorporate the expectations. Well I disagree. The older equation that David talked about is a straight jacket and it's strongly rejected by the data. It's one of the most rejected equations in macroeconomics. The claim macro foundations are sand. There is no representative agent. The information economics revolution implies heterogeneity and short horizons in the face of liquidity constraints and serious income uncertainty. The key economic stories about finance and the relationship between finance and the real economy can't be told in the New Canesian DSG model. But these new insights into household behaviour do help. Let me stop there. So my economics training ended in 1969 at King's College Cambridge where I mostly learned about Canes and didn't get to hear about New Canesianism. After which I became a social scientist and a psychoanalyst and I'm going to talk about something slightly different. So why I think we're concerned about macroeconomics is because we start from the problem that we didn't see the problems coming. And then as I understand it we have quite a bit of difficulty understanding the way out of the problems at the moment or indeed whether we really know where we're going. So it seems to me the future of macroeconomics is to re-establish a credible reputation because its raison d'etre really is to support macroeconomic policy and to try and be reasonably convincing. And I think the kind of populist staff and all that is quite important because actually we do need models that we at least believe in and central bankers can believe in if we're going to avoid sort of very unproven ways of thinking about these things. So I think there are three ways in which macroeconomics might re-establish a credible reputation. One is clearly by incorporating some secure advances made in other areas of economic thinking such as information economics, game theory etc. Secondly I think there's a huge scope for new methods of analysis and data handling. If Doin Farmer had been here he would have talked about that but from physics and computer science and so on. But what I want to do is raise a much more fundamental issue which I think lies at the heart of economic thinking. And this is basically that human economic agents simply cannot know the economic facts of the world or coordinate on them except through their human interpretive and perceptive capacities. And this of course is necessary based on their brain architecture and psychology and located in specific social environments. So the point I'm trying to make here is you get a lot of talking economics about things like fundamentals, about information and so on. But how does that get into a human head? And that's really what I'm trying to talk about and that social science and psychology are interested in the fact that as you all know really if you try to consider what do you think is going to happen to the economy in the next 10 years, you actually have to interpret data using the capacities you have. And in modern social and brain science facts are simply not available for action except by embodied and socially influenced perception and memory. And I'll talk a bit about that but first it means that the first thing that I think people have to do is actually study much more directly. I think what John just was talking about was a move in this direction, but study much more directly how do economic actors actually behave in different parts of the economy. So the alchemy of finance by George Soros was a kind of autobiographical account which in my opinion gives a far more accurate idea of what goes on in finance than most if not nearly all economics textbooks. Why Wages Don't Fall in Recession by Truman Buley? You may or may not know a mathematical economist who got fed up with the arguments in macro and went and talked to employers up in the northeast of the United States and asked them during a recession what they actually did. The picture he got was very different from anything in the models. Herbert Simon, he has written a great deal. Well Herbert Simon was a key member, one of the first five people who implemented the Marshall Plan. So he knew a great deal about actual detailed implementation of what happened in the economy and of course he went on to get the Nobel Prize and to make many, many contributions. My own somewhat small contribution is that I did study 52 asset managers and it was quite clear to me from that book that the way people behave in financial markets has very little to do with much of a theory. What I got out of that was that what matters is social interaction, what Soros calls fallibility, so because agents are interpreting the world they obviously interpret the world wrong. And if a group of agents all interpret the world wrong for a long time that can happen and everything will go according to those quote misinterpretations but under radical uncertainty you can't say they're misinterpretations because you don't know what the correct thing is to do. So in my view none of the descriptions of decision making in these studies are consistent with standard models and nor is that the case for many, many other studies which I've mentioned in the full version of this paper. So what we're looking for is a economic macroeconomics which has what one might call ecological validity and bounded rationality to use Simon's term. But the problem here is that Simon's idea has been completely misinterpreted in a great deal of economics. His statement is this from the quarterly journal of economics I think actually 1955 not 1946. The task is to replace the global rationality of economic man with a kind of rational behaviour that is compatible with the access to information and the computational capacities that are actually possessed by organisms including man in the kind of environments in which such organisms exist. So this is not optimisation under constraints. This is two sides to it. On the one hand the possible limits of computational capacity but on the other hand the limits to knowing what the right answers are what it is to do based on the situation in which you actually find yourself in which is radical uncertainty. So in my view the general assumption of global rationality and the assumption of risk rather than uncertainty throughout economics in fact has really avoided the proper study of how economic actors coordinate and that's what the future needs to look at. Now in the work that the team I have and I do at university college we've developed something called conviction narrative theory which comes out of the finance study that I mentioned. And the basic idea here in fact is remarkably similar to the way I understand Keynes from my days at his college. So actors supplement and support reasonable calculation Keynes thought with what he called animal spirits. And so to put aside thoughts of ultimate loss as a healthy man puts aside the expectation of death. If the animal spirits are dimmed and the spontaneous optimism falters leaving us to depend on nothing but a mathematical expectation enterprise will fade and die. Though fears of loss may have a basis no more reasonable than hopes of profit had before. So the key point is neither profit nor loss you can particularly know about. And this is basically what you could consider the problem of action. So you turn the problem around under uncertainty it's not a question of selecting the correct action. It's a question of how do people know that it's safe or right to act at all. Now in conviction narrative theory is a new social psychological theory of decision making which asks exactly that question. How do economic actors manage to act in radical uncertainty and with what consequences for the way they coordinate when their decisions are aggregated. And what we say is that agents individually adopt conviction narratives that could be a firm adopt one narratives they think are accurate and feel a true that are subjectively capable of supporting action. Because cognitively and effectively through that thinking and their feelings they manage both the anticipations of potential gain and loss associated with its uncertain consequences. Now the point about narratives and it's a crucial point is that they can respond much more rapidly than the fundamentals underlying it. And that's the kind of thing we've been seeing in the last few years in spades in a lot of different areas. The world hasn't changed really a huge amount surprisingly in the last 10 years if you will look at actually the details of many actual lives. Of course there have been all sorts of things but the narratives about them change much more quickly and can simply flip like that. So the model of a conviction narrative theory therefore and it's very different if I had time I'd explain it to you than the model of psychology for example with Daniel Kahneman and the standard behavioral economics. Here we have cognitive deliberative processes and emotional processes interacting all the time. And the point here is that cognition is always embodied is always in the body. The brain is constantly engaged in a relationship to the capacity for survival reproduction and so on. And so that any kind of narrative has to feel good before we adopt it. So this leads to a conviction narrative and that then leads to action. So in this theory you see this complicated graph here which tries to bring together the different things that are influencing people when they take action. So for example the fund managers I studied would have a high level narrative which sort of tells them what they're doing and what they're trying to do and in which they believe. This assist them with opportunity identification. So opportunities come along such as fit according to particular narratives and haven't talked about that later perhaps at this time. But these are particularly relevant or particularly sent to them in the particular local social environment in which they're operating. So if you're in a particular fund management house or in a particular part of the world there'll be particular heuristics rules models conventions that you draw on and which give you confidence. That they're the things to do because that is the way things are done in that particular environment. And it particularly important the other aspect here which is trusted sources. So which people if you're in finance which are the people who advise you which are the bits of information do you trust. There's a huge amount of psychology dealing with the fact that what people think they trust makes a huge amount of difference. So you can dress people up to look trustworthy and they say exactly the same as people who don't look trustworthy it's subjective and you'll get a completely different result. Same with presentational components that is the way things are presented. So all this adds up into generating two potential sets of feelings and I can illustrate that if you're listening to me now. So as you're listening to me you're thinking or you are whether you're conscious of it or not you're kind of going yeah that makes sense. No I don't like the sound of that that sounds very dodgy get the idea. And so as you're listening to me or as you're appraising narratives you are thinking yes or you're thinking no you go forward or back. So what happens in the conviction narrative is the overall in order to act in the particular way obviously the approach emotion has to trump the negative one. Now the key point about all this is that I think if you give a lot of thought to it this is probably the way most major decisions are made within an economy but that's not where it ends. The point of this model is it then focuses you on OK so how do people get convinced. What is it that's going on that gets people convinced and one of the aspects there is the narrative circulating in society or the state of things in the world. And because of course unlike in economics in real life people are looking around and you know what's he thinking what's he thinking what's he doing. There's a massive in finance in particular just massive amounts of looking around to see what other people are doing. And so the this is a theory of how individuals make decisions but related to the if you like the ecology of narratives and so on that are around about how things are done. So I won't go into it but we have something called divided state theory which is a theory about the state of mind in which you can be when you're making decisions. And they're kind of two states of mind you can be in one is the one that you would think we'd all love to be in which is that when evidence comes along we evaluate it according to its value as evidence. The other what's called a divided state is we evaluate things purely on the basis of what we already feel. So if you look at this diagram here supposing you already have some sort of narrative idea about what's happening. If you get new information and you're in a integrated state that's a kind of state you'd all want to be in then you will modify your narrative and modify what you what you're going to do. But if you're in a divided state then if you get information which fits makes you doesn't make you feel massively uncomfortable you may pay attention to it which will of course increase your tendency to do what you were previously going to do. This can help to explain bubbles but if you get information that is makes you feel uncomfortable then there's a strong tendency for people not to pay attention. So this is the kind of thing I would say happened leading into the financial crisis. Now we can actually do something with this I'm going to take a couple of extra minutes. We can actually do something with this which we explained in some detail in the paper and in published work by taking this intuition that emotions of approach and avoidance. If you find them in narratives or in the Guardian or in all the over time that these are actually giving you information about if you like the state of confidence or in Cain's term the state of animal spirits in the economy. Now if you look at this particular graph what we've done there is we've taken all the writers news articles between 1996 and 2014. That's the several million. We've used the computer of course to extract from these articles all the articles that mentioned the word liquidity. We've then counted the presence of words that indicate approach and words that indicate avoidance and we've drawn a nice time series which you see there and you see that mentions of the word liquidity when it's going down. It means there's an increasing amount of avoidance or anxiety when it's going up an increasing amount of approach. So you see that leading into the financial crisis the word liquidity is increasingly occurring in context where there's less and less avoidance doubt anxiety those kinds of things. So this particular technique may be usable with complex topic ideas because of course this is exposed right but there is a possibility of using that way things are going with computer methods to identify topics to be of some use in trying to see which kinds of things are going on in the economy. We also have a second method of doing it which I won't elaborate because of time but where we can look at the extent to which narratives are coming together in a consensus sense or they're more dispersed. And you can see from there that actually again leading up to the financial crisis narratives were increasingly there was a consensus about narratives so whereas afterwards there's been much more uncertainty and things are much more dispersed. We can also use this idea in a completely separate way which is to say well movements in this because that was to do with topics movements in in the economy in terms of the amount of approach and avoidance emotions in the whole economy may give you some information about what is happening. So here is for the Canadian economy that's for the UK economy and that's for the US economy and you can see there that you've got RSS that's a relative sentiment shift which is this approach avoidance. You've got GDP and you've got fixed investment and I don't think I've got time to really go through it but you can see if you looked at it that this goes up and down with the kinds of things you would expect it to be associated with. There's no causal implication with those things that are written there but there is as you'll see from this. You can actually see if you look at how RSS has behaved all those numbers that are in bold right show at least two standard deviation moves in that in a particular period from the norm. So the mean value over the period 2003 quarter two to 2007 quarter quarter two. And you see there that for a whole bunch of countries from 2007 Q2 you begin to get this very significant change in the RSS long before long before people realize what was going on into the into the economy. You probably remember that the credit market froze in August 2007 that the layman thing wasn't until off the chart that wasn't until Q3 that most central banks were still forecasting that the economy was going to expand in Q3 2008. There is actually Granger causality between RSS and these measures of GDP in to some extent in these three economies and for those of you like vector auto regression models. There's also clear effective RSS on the US production employment UK production employment and Canadian production and these effects are up to 18 to 20 months out. Thank you very much. Right. I'm going to open up to questions in a minute just to make a couple of two or three points. One is that from where I was sitting as a journalist it seems to me quite obvious that the model in 2008 was about to blow up and that was not because I used any economics to do so it was just because I used basic journalist techniques and two of the things we always learn is never trust anybody at any time about anything. So question everything very deeply always be deeply skeptical about it and ask who is benefiting from this particular paradigm and it was quite clear that the financial markets were using this model to justify their increasingly reckless behavior. The other point was that journalists from their own experience know when people are winging it and there was a real extent to which people were winging it back before the crisis. In my view they'd constructed these beautiful models that didn't really seem to have any great grounding in reality and journalists find out what's going on by talking to people and walking. We walk around talk to people observe things and the models that were there to justify what was going on didn't seem to have any real bearing on what we were seeing with the benefit of our own eyes and just talking to people. So when readers said to me this looks like a big bubble in the housing market in the financial markets in 2006 it was hard to disagree because it obviously was. You need to use your own eyes and a little bit of economic history to actually see that this was just a repeat of what happened before. David made a good point about how these big shifts in economics tend to happen when there's a big shock to the system in the 1930s there was one in the 1970s there was one. I'm not really you'd imagine that there'd be a really big fundamental shift in economics as a result of 2008. Now I'm not quite sure I'm not quite sure why that hasn't happened. The status quo seems to have clung on by its fingertips and I'm not sure there's been that much change in certainly not the way there was in the 1930s or in the 1970s. And to me looking at it now you imagine that the economists would have been run out of town like the old medicine shows were in the Wild West for peddling fake remedies but they haven't. I mean there's still quite well dug in and I canvassed this with a couple of the panellists and one argument was that it's a question of incentives that there's an incentive to actually get in the top economic journal. I'm not I must say I'm not totally still not totally convinced that that must that's the reason I think we might want to explore something along those lines just couple of final points. I mean I loved John's description of the neo Keynesian dynamic stochastic general equilibrium model that is not new it's not Keynesian it's not dynamic enough it's not stochastic enough it's not general equilibrium enough apart from that absolutely fine. And I just wonder whether even with David Vines is five refinements we can ever you know whether that's going to actually sort things out. And I think my final point is this that in economics I think there's always a tendency to think we know more than we actually do and pretend a level of expertise that maybe we just don't have. I think that in a sense central banks are part of the problem because we've handed over an awful lot of power to central banks as these supposed technocratic experts and it wouldn't be very good for these central banks to come along and be asked well what's going to happen to the economy and the honest answer is well we don't really know but they don't actually to a large extent is that the future is incredibly uncertain but we have better than with this enormous power and we've just got to keep our fingers crossed and hope that they have learned some lessons from the crisis otherwise we could be in for something quite bumpy. So that's really my four penneth. Are there any questions specifically for John because he has got to shoot of I'll take some questions for the panel generally but if you can come forward and come to the mic. Yeah there's a question there. Yeah and can you come forward as well and then you would. Richard Hanfield University of Bath. I've just got one for David at this point. In John Maynard Cain's by Minsky he criticises the Hicks Hansen General Equilibrium LSLM as a bastardisation of basically Cain's disequilibrium model and that whole process in fact he's quite critical of the whole of that general equilibrium concept being applied to Cain's and stuff and then goes on to develop his own theories within that. That would be one point. The other one is more about understanding that in fact I think the most macroeconomics don't actually understand banks and that they don't understand banks because I've spent many years working you know I'm actually late into the macroeconomics so it's a is that they actually don't understand the accounting practices of banks and the things that they're managing their liquidity and how that works within banks and they're managing their credit risk and how that affects it and how they have to write things off and so on. So and that falls very much into David's view is that they have to forecast all of this and they have this sentiment of is it going to be good or is it going to be bad and how they reinforce each other with that and I think that there's one thing I have not observed in any variation of model is actually the modelling of the balance sheet, the statement of cash flows or the profit and loss statements within the banking structure of a macroeconomic model because it exists, it's there, that's what they manage on a day-to-day basis. In fact they manage it by the hour. So that's why two pennyworthies might say. John, do you want to take the point about banks? Yeah, just a quick one. In this work that we did on France, we used the NPL ratio of the French banking system as a major ingredient in the credit conditions index. So it turned out that from 1990 onwards what happens to the non-performing loans in French banks is very closely related to their ability to advance credit. So it actually gives you a linkage between what's happening in the banking system and what's happening to the household sector, which is part of understanding what's going on, not the full story. Well banks are heavily constrained by their credit risk profile and how they balance that credit risk against the liquidity in such a sense and the way they have to write things up as well. David, do you want to take the first point? Well, which bit of Keynes is the right one? I was reminded as you asked the question of our discussion about Adam Smith on the first day. There are at least two Adam Smiths and neoclassical economics has had a field day with the wealth of nations whilst ignoring so many of the other insights. I think it's right to say that there are many John Maynard Keynes, that the ISLM clarification of the central idea and the general theory was fundamentally important. Without that, there wouldn't have been a way to teach economics clearly that inspired in a valuable way that generation of people between the end of the Second War and the 1970s, which made an extraordinary difference. But that wasn't all of what's in Keynes. I agree with you about the importance of his study of finance and Minsky's ideas, but I would just add sideways that I've written a book on Keynes arguing that he was fundamentally concerned with globalization and the international world, which is yet having grown up as he did in the empire, having a father and friends that managed the world, writing about how to manage the world, led to his work on Bretton Woods, completely ignored in much of contemporary macroeconomics, but fundamental to an understanding. So there we are, we've got at least three important John Maynard Keynes in play. OK, thank you. Is there a question here? Hi, I'm Hannah, I work at rethinking economics. I just wanted to say thank you to the panel, both about your candidness about mistakes of the past, but it's also great to see solid work being done and creating alternatives to rational economic man. I wondered if I could just garner some of your expert opinions on a couple of futuristic macroeconomic policies, and if you don't like any of the ones I suggest, please share your own. The first one is the establishment of the maximum internal wage ratio. The second one would be the nationalization of broadband and internet access, and third one would be a three day weekend. John, I'm going to ask you to, I know you've pushed for time, so any of those float your boat or not float your boat? The internal wage ratio, I have a lot of sympathy for that. I think the CEO reward model is very seriously broken. There's no real evidence that great bifocation of top to bottom is actually led to any great improvement, anything, isn't there, hasn't led to higher investment, hasn't led to higher productivity, it's just led to higher wages of people at the top. The three day weekend, that's a tough one. The labour market is a bit of a conundrum, and obviously with what may be happening to artificial intelligence and robots, we might have to seriously address this issue. But of course there are many people for whom work is actually a hobby, and imposing a three day weekend on them wouldn't actually stop them from carrying on. I think part of the problem, I mean, it's kind of issues been raised quite a bit in the conference, this type of problem. So it seems to me that what we should be trying to think about is how could we get an answer to that question, or a set of answers to that question, based on a set of assumptions that we could make very transparent, which would allow sensible discussion on the issue. Because there's been a lot said in the conference which could allow you to think that, and I suppose the contribution I would make to it would be to say that I really do think that narratives really matter, and that economists are a bit obsessed too quickly with the idea, you know what the real world is, and a narrative is just a narrative, actually to a very significant extent narratives make the world and can go on making it for a very long time. Now that doesn't mean you can just do what you want, but I mean the most impressive narrative was the, I'll do what it takes of the ECB, which as far as I can see was just a few words, that people believed is very unclear whether he could have carried it out, whether it was really constitutional. But it, and then once you'd got over, once that narrative had been accepted, it then changed things. So I would look at the kind of questions you're raising in that way. What narratives would there be around a three day weekend then, or the nationalisation of broadband? Well nationalisation, I mean you'd have to, you know what I'm trying to say is that there presumably would be arguments for and against, but you could, in the end you'll win that argument by whether or not the narrative catches on. So a lot of the policies we're discussing is really about trying to understand what would be the objections people, you know it will come up to that, you know not just the objections that people are not going to get their profits, but what will be the objections and then how would that be, we work through. Can I take that narrative story further on the three day week, that we were reminded that Keynes who keeps on appearing here there and everywhere said in the economic consequences for our grandchildren that he imagined by about 2017 we'd all be working a 15 hour week and enjoying ourselves in the other many hours of the week. And I think one of the real reasons we're not doing that is what Adair Turner and others have talked about, their desire for positional goods and the consumption that that leads us to undertaking in order to be better off than the people around us. And that's essentially a narrative story. And you say if people are busy doing that then it makes them very hard for them to find a way of supporting enough resources to keep the national health service running and to run the schools and to do other things. And we would be wanting to work less hard if only if we persuaded ourselves that we already had enough road schools and things that we need as well as enough positional consumption to feel better than everybody else. See I would say that's a hypothesis but what we need to be doing is actually establishing it. I think it's quite a lot of ways of arguing that actually people aren't only in a race for positional goods and that there are different ways of motivating and so on and so forth. I mean I think we don't know because once you're within a social system that's operating in a certain way it operates in that way but it can change. There is one piece of evidence which is the performance of the French economy where limits on the working week I think proved pretty unsuccessful in stimulating employment. Hannah what do you think? Well I only work four days a week and I'm pretty happy. It allows me to study in my free time. It allows me to socialise. A lot of my friends are pursuing more flexible working because it has a good impact on our mental health. I think all of these are indicators that a future macroeconomist needs to be taking into account but thank you for your comments and consideration. I think one of the things that I mean Cain's actually said in the economic concerns granture was that productivity would go up about eight fold between the 1930s and which it has done. I mean he was right about that wasn't it? It's just that we haven't actually decided to use it in this way. In some senses there's an argument on your side that we're choosing to use our greater wealth in not satisfying ways. I mean although different countries do actually apportion improvements in productivity and growth in different ways don't they? So you know if you go to the states when the economy grows they tend to work longer hours rather than take more leisure time. The Europeans take more leisure time so there is a choice. I think that's absolutely right. Anybody else? Rob Smith and I'm in Denmark at the moment. I'm fortunate enough to be in quite a heterodox economics department. It's a slightly controversial question and that would be that we've talked now about how some of these macroeconomic models they fail to predict the crisis. But there were some guys that had pretty good models that took into account some of the things that Steve Keane mentioned yesterday about macroeconomic balances or other imbalances that have feedback effect into the economy. And a gentleman by the name Doug Betsomer wrote quite a good summary article of several of the guys who did a good job of predicting the problems that we have now. Some of the contextual problems that led up to but also the stagnation problems that would follow. And I haven't heard any of the names mentioned now that we're looking at the future of macroeconomics. Some of the contemporary ones that I've had the fortune of reading are Wynd Gwydli, Robert Blecher, Randall Ray. And some people, or one lady who is here now has talked far more about the methodological aspects of macroeconomics at Sheila Downer and one of her colleagues, Victoria Schick. And I don't hear any of that being integrated into what's going to follow from a modeling point of view. So there is a textbook out by Mark Levoin, Wynd Gwydli, the late Wynd Gwydli, who have incorporated fundamental uncertainty into a long-term modeling programme, or paradigm you might call it. But I haven't heard any of that. I was wondering why is that? Why haven't we heard about these people who have been working on it? Wynd Gwydli started in the 1970s on it, and then Nicholas Colder, who was in the bank, so just a bit curious. David, you're right, there were people in advance of the crisis who actually said there's something, I think Steve Keen was one of them who is coming in a new era of being. He was another one, so there were a few people. Take that point on, David, that's an interesting point. I wrote my first article working with Wynd Gwydli way back then, and I think that that stuff was driven out of macroeconomics by the Euler equation that John ridiculed. But the balance sheet work that John talks about, the empirical, careful, thorough balance sheet work, is essentially a taking forward of that set of ideas that Godly and others have had. So I wasn't able to talk about everything in exactly 20.0 minutes, and one of the things that I would have talked about is the way in which a study of finance is leading us to want to understand that radical rise in risk premium, which happened during the crisis in 2008. And the person to ridicule as an inappropriate model in not being able to deal with that isn't just what I put on the blackboard, but the very famous textbook by Woodford, which has no financial system and no banks and no nothing in it of that kind. And I think the last 10 years have made a serious push in directions model. Many people are now working on models with financial systems in which there can be shocks that radically propagate through rising risk premium and bankruptcy and leverage cascades that cause the economy to collapse. We couldn't have done that 15 years ago. I think it's a good question actually, but I think it's a question of which is kind of two ways you have to think about it. One is the whole question of how do ideas get taken up and not just in economics but in other places. And really, you know, the relationship between ideas in science and whether they're taken up in policies already a big issue, even without going into sort of controversial ideas like in economics. And then the second problem is in economics itself. I think a problem of evidence. So the fact that economics has not grown up as a discipline which made trying to test ideas against each other as its main way of proceeding, which we know is difficult. But it's, for example, you know, you'll all have seen the film The Big Short. And you know, the guy in The Big Short goes out and has a look at the housing and sees all these for sale signs in all the thing. And this kind of thing is not so impossible. And I think to understand the narratives that are going on to understand what's going on, we need to consider not junking what we've got now but adding to it much more empirical ways of looking at things. And testing whether theories are what I call group think theories divided state theories or whether, you know, they actually work. And clearly some of these models, which my two colleagues on them, some of the, you know, they just don't contain stuff that is realistic. Very quickly because our time's up. Just to say that on your side in that discussion with David, I think we macroeconomics has been a discipline in which algebraic coherence has triumphed over appealing to facts. It's what you teach your graduates to do now, did until the crisis. And I think that way of, as many of my colleagues have said, 15 years ago, young students and their mid-career colleagues spent their time understanding the way the economy works. Until the crisis, people spent their time understanding the way theory could help them tell a precise piece of argument about the economy. I think the effect of that failure in the crisis is forced people to being much more serious in the way you would want. That's a very, very good thing in my view. So on that hopeful note, I'm going to call this session to a hand. John's gone, but I'm sure you'd like to join me in thanking the two remaining panellists, David Vines and David Tuckett. Thank you very much indeed. It's been a good session.