The Collision of Economics and Ecology

all right everyone maybe I like to get started now awesome wine in your hands right not too many glasses of wine it seems right so it's a very pleasure to have Lord may gave the keynote speech today you have the details there we have the tweet address whatever is called here if you want to tweet about it and ask questions I will share the session it's a it's a true honor to have Bob main here who we see he is a professor at Oxford and imperial college he was the past president of the Royal Society he was the UK chief scientific adviser to the government and he is currently a member of the Economic Affairs Committee in the House of Lords and most importantly of course is a chairman of my systematic risk center here at the LSE now I have to have to remember this works good I'm going to talk the talks roughly divided into three bits the first bit I'm going to talk partly about how I got started developing interests branching out of my earlier life which started as a chemical engineer as an undergraduate ended up with a PhD in theoretical physics spent a few years at Harvard the division of Engineering and Applied Physics went back to Sydney is a physicist went to Princeton as a biologist and now I seem to be transmogrified to a certain extent into a finance person and I'm going to say a little bit about that and some of the parallels with systems in ecology then secondly I'm going to talk about applying some of the things one has done in ecological systems to get some insights into the dynamics of model financial systems and the third section I will focus on some of the regulatory implications of all of this yeah I've got various thanks to people that I will touch on as I go along but first and most really heartfelt and important is George Sugihara if it hadn't been for George I wouldn't be here at the moment talking to you but George got involved in an interesting study in 2006 the US National Academy of Sciences and the Federal Reserve Bank of New York in a very precent way long before anything happened observed that the increasingly elaborate financial instruments that were being invented all with a focus on individual institution just sometimes individuals maximizing profit with minimum risk nobody was thinking about possible implications for the system as a whole and the US National Academy of Sciences and the Federal Reserve Bank of New York put together a study group to look into this issue and in doing so they looked for people who might really read across from other disciplines infectious diseases propagation of infectious diseases ecosystems and George was a natural George having been spent five or six years managing to which a bank securities USA before going back to fisheries complex multi-species fisheries George was a good person to get involved and George got me involved and indeed we wrote a little paper about it published in nature which Gillian Tett has many of you'll be familiar with us from Financial Times and Gillian Tett wrote a nice essay about it called the birds and the bees and the big banks but she wrote that after a few years after we actually published the paper because the paper was also published before everything sort of went belly-up one of the things George did while he was on this committee was asked how are the planks in the you in the USA he connected up and never and that prompted a study and this is a diagrammatic very impressionistic result of the study which was built on monitoring some my memory is right the first digit seven but I forget where there was seventy thousand or seven hundred thousand email transfers of cash the way we stand send money direct order in America and it involved this network some nine and a half thousand banks interest it's a very interestingly biased to a few big banks sixty-five percent sixty five of the bank's account for much than half 65% of the transactions if my memory is correct however just simply knowing what the system is and who is how many banks each how many connections each Bank has is really not enough to say anything much about the distribution the one of the first caveats is you really need to know not just a degree distribution how many banks are connected to a large number helming to a small number you need to know the full structure of who interacts with whom and more important even than that you need to know whether the very active bank if you go back to what prompted this is a publication with genetic Gupta and Roy Anderson in to do do with the early days of hiv/aids on the west coast in the United States you need to know – one thing to know how many people have a lot of partners and how many people have a few partners but what is really important for the transmission of that infection is do the very active people differentially mix with other very active people or are they as it were predatory among single people in large numbers or just insured are the interactions associative in which as it were shy people associated people and very active people with very active people are they at the opposite extreme disassociative inland are very active people mainly interact with not very active people are they just simply proportionate to the number of people that you do interact with and the answer is quite consequential as you can see from that if they are did the way to maximize the spread of infection is to have disassociative connections associative connections rise faster but before there were drugs against dying of AIDS they tended to end up with a smaller completed epidemic and rather distressingly had turned out by and large the things were more disassociative that's some just some sense of the unexpected and but interesting complexities around this look few years earlier – before all this thing that turned the into world an ecologist and epidemiologist Princeton I had got interested in ecology this was the late 60s long ago after the audience probably weren't born but it was a time when consciousness was raised in various directions and in Australia a Society for social responsibility in science was formed led by the co-author of what was internationally the top-selling ecology takes Charles birch and I was drawn into that and I thought I should really learn what I'm being socially responsible about and I started reading in ecology books and I came across and remember this is an interesting time ecology is a young subject the word itself is a not much more than a century of them and the world's all the psychological society with which were celebrating this meeting the celebrated its centenary last year but it was a tight it's early days very sensibly were just describing things but round the 60s there was a rising movement of trying to get a better grip on fundamental questions and one of the tentative things that was put forward by evil and hutchinson a really major figure in ecology Britain who was based at Yale and had many his students for our major figures in the discipline now but his most notable student was chap called Robert MacArthur at Princeton but they had come to the conclusion mistaking basically some of them it was an elaboration of an idea due to Elton they got mentioned earlier today namely the complicated ecological systems with lots of species and lots of interactions were by virtue of the complexity more robust to the students and I one of the books I read was a book by Ken what called air ecology and resource management and he set out this conventional wisdom and then he said he found it unconvincing because his experience and I felt it I thought like my experiences and more complicated something is the more things go wrong and so I that evening after reading it I decided I would look at something others just as a toy model it's the sort of thing I like doing and I said suppose I have a community of M species in banks in species each one of which in this ecosystem if isolated with no interactions with the others had intraspecific mechanisms that if it was impact didn't know it moved away from its average equilibrium point would return if it became extra abundant it would tend to fall back if it became low birthrates and shrank it would tend to bounce back suppose in short I had a matrix of these species with minus one down the diagonal and say each of these species would recover from disturbance in unit time and zeros everywhere else and then I sprinkled in interactions at random plus or minus so A plus plus would be cooperative minus minus competing and twice as probable at random plus minus at be predator-prey and each of these interaction each species would have on average M of these interactions and they would be of strengths alpha and I proved an interesting extension of a theorem actual arising in particle physics by Jean Vigna which said which asked questions about what would be the stability of the thing and he proved it for anti-symmetric matrices but in what I was able to show was that if the number of species in was very big than the number of if the number of species times the average connectin that is to say if total number of connections per species which was little m over n times the square on the average strength exceeded unity the normalization time for recovering then the system would tend to collapse I have great delight in fact I'm sure I've got it on and one other trend one of the power points this is now called the Mae fitnah Theory this what this said is real systems real ecosystems are not randomly assembled but what it did do is essentially redefine the agenda of research in this area which is still continuing because what you've got to ask is in a complicate in an ecosystem there is a pressure towards getting more species and more complexity as more Nicias arise as the thing gets more complicated and then someone will appear to fill the niche and so you have an effective trade-off between evolutionary events driving more species and lots more complexity and the instabilities these generate mitigating against it and I think question that we're still pursuing but it's not a completely solved problem by any means is how do complicated ecological systems have special non-random interaction structures within them that tend to reconcile added complexity with not being too unstable and that that is a quest still in progress but there's good reason to believe it's not a foolish quest because there are this very interesting paper by Jennifer Dunham for the South Bay Institute and Doug Irwin suggesting that the Burgess Shale the thing that factored so much attention a tentative reconstruction of the communities they're reveal predator-prey ratios which are very similar to most modern ecosystems but this is still at work in progress so that was the background motivation now I'm going to sketch some read across to ecological things once drawn into this enterprise that George was involved with 2008 things went pear-shaped and I think first my first reaction my first reaction always is a very narcissistic lecture but we'll get less so as we go on but when I get involved I have a very short attention span as a scientist and I hop around a lot but when I get into something new I tend to try and think about what I would what I think would be an interesting thing to do before I learn too much about and what I did in this case as I said let me see how this theorem about too much complexity works here the nodes are going to be banks but the network is going to be a bit trickier because the nodes are more complicated and node in an ecosystem as a species and the things coming in are things though preying on it competing with it and those things going out are benefits to it and they're just very simple got arrows that coming in or they're going out but for banks it seemed to me not having learned much about it because I didn't want to be circumscribed by learning what I ought to think I'd I thought there's four sort of things going on in notes there's money coming in from outside deposits and money coming in from within sister and this money going out to the outside in mortgages and stuff and there's money going out in bank loans and then having thought about that and done some calculations which I'll show you but before I got stuck into that I asked Nikolas vo one of my friends in my pink was here earlier but it's gone what did he think of it he said that's absolute that's just ridiculous but here's a paper just been published by people at the Bank of England that might set you on the right track track and I'm delighted to say this was the model that they were using which has as you will see the liabilities must see but you so that wiped out the entire net worth then this map will fail and this bit here wasn't covered by the followers so that kind of shock would be attenuated in each phase but it gave you a sense of what was going on and I then asked what is this going to look like as things propagate and this pretty quickly but this is discovered black skulls oh not like skulls sorry the 1933 legislation that said you can't be a high street bank and a gambling back glass-steagall that's right in the bank stuff so which what proportion are casino in High Street and suppose a single bank comes down to bring their other banks what help be they in turn will now fail the fragility of the system was that you wanted the sort of balance on High Street and Casino from this simple and artificial model at this point we had reared that it was now clear that there were things wrong with the banking system and I could got to know Mervyn King before he was actually governor of the Bank of England it was while I was chief scientific advisor and the ESS the economic Social Sciences Research Council very wise these days and it was about to get a new director and committee was put together with three civil servants and myself as chief scientist and a token civil suit Oken social scientist and the three civil servants wanted one candidate and I and the token social scientists wanted another candidate and we prevailed and I thought this is none usually wise and skillful social scientist it was more than King while he was deputy government so at this point in this narrative I called Mervyn and I said I've been doing some things and I have what I think are some interesting ideas why don't we bring George Sugihara across from the state and a couple of other of my students who are doing interesting things on networks and just have a bit of a chat and so on and from that we did that and from that came these things I'm going talk about them them work with Andrew all day than others one of the first things that I did in elaborating this was go beyond the simplicity of the first model and start looking at exactly how you might do a slightly more realistic model of the system with different kinds of assets but more importantly and this was I got the things I was doing now I've done a lot of them flex the glass-steagall II thing I just did there's an analytic approximation but at this point I wanted to do some big computational things and my PhD thesis beckons way back in Sydney was the first big computation on the first mainframe in Australia and it left me with an abiding desire as many of my acquaintances know to deal with computers through the medium of graduate students so I got a graduate student in marilla of path II there was a very very evil person now he was a graduate student of one of my colleagues Angela MacLean that he's now senior lecturer Imperial College not many years later and I bought and it was it was nearly really said what we need to do is characterize the way these things propagate is not just automatic debts being propagated it's people getting scared it's a lack of confidence and he said let's try and include that however crudely and his idea and this is his idea is to first of all characterized the confidence of the banking system the financial system as a whole by calculating a sort of solvency health from ratio of the value of all the assets you hold at a point in time as a ratio to the number of an to assets and a liquidity health which was the fraction of the interbank loans that have not been tampered with and that's a measure that things see we take that as a measure of the confidence in the banking system or in our model network banking system with specified kind of over dispersed network and then we do an analogous definition of the health of an individual bank and then the analogy there of the sort of asset health is the value of the bank's assets against their initial value and the bank's liquidity health is the fraction of the loans that the bank can settle immediately either through liquidity because it's got adequate liquidity or should we were selling short-term assets and then having calculated the health of each individual bank in the model and the overall system health if you've got two interacting banks I&J and the product of their health is less than one minus the cost systems whole confidence thing then both banks should turn all their long loans into short as opportunity arises and you can see if if the confidence is high that'll never happen and as the confidence gets very low that will easily happen and more dramatically if they get less than one minus C squared which is always going to be smaller than one minus C then both banks call in all the loans and then we model that and look and see if the sorts of things that happen and without going into sort of endless detail this is the sort of thing we find suppose we have a system of banks this is I my memory is right it's a system of some two hundred banks there's little banks and big banks all the little banks are the same size all the big banks are 25 times bigger and there are eight of them out of 200 which is not as bad as a statistic I've seen the United States where 1.6 percent of all US banks hold 65% of all assets so that would have been more extreme than our 8 and now the banks when I say that the big banks the a big banks are bigger they have 25 times more loans going in and out and their individual asset classes are 25 times as big and I now let one of these banks chosen at random could be bigger company little much more likely to beautiful and that's the blue and the that's the blue and the red line together and then I say this has reserves such that one bank fails as liquidity goes down below point eight eight you only you've got in reserve only eight percent of your total assets how long will it be before more banks go down and you see if you do this without this confidence effect but just the mechanical thing I showed earlier you'll have to get the very low levels of capital reserve before another Bank goes down but once another Bank goes down it really will start to top conversely once you include these confidence effect once the first Bank goes down it doesn't take much for the whole system to come down and the next slide or now I'm going to just shuffle these slides briefly the next slide does a big simulation of this and what this is which is looking in more detail if the index bank you choose to fail is a small one of the small index banks on Aria 92 then following large it doesn't do much damage the conversely book site is conversely if you hit one of the big in big banks it may be show up well on wave that figures drawn people bring the holes essentially the whole system down now I'm going to go back to the two I just skipped having done all this we were to try and test some of these ideas we were both astonished but the degree to which there's very little things in the way of experimental data tested against thisis kind of thinking but there's a chap called Shem at Princeton who really is rather keen on this sort of stuff and he accumulated the data from Morgan Stanley first quarter of 1962 the second quarter of 2011 and then he sort of analysis asking what's the quarterly change in assets and he found that he had collaborate he had corroborated the tentative opinion projective learning more about death than then about equity and so we went back to our model not having known anything about this but taking the same model and running it to see what it would have said and Morgan Stanley would have done and it is remarkably in agreement and that we found rather reassuring come to the last bit now what are the implications of all of this I want to read this off with what I think is the most interesting single piece of data rate relating to the past events that I'm aware of what this is Andy Hall Dame did it go back to 1880 what's the ratio of all UK bank assets to GDP it's just sort of normalize out information and it trickles along with five ten twenty percent fluctuations at point five five percent for a century up to nineteen seventy I'm amazed that there's so little attention has been given to this I'm not aware so I I may say modestly I know we have anybody else who's even asked what happened around 1970 but changed things and I can tell you that around 1970 is a more or less around the time when I was doing my period during the PhD thesis on the bloody computer and it was just beginning to give the computational ability to handle the way people invented ways of taking the same acids splicing and dicing him selling him again taking more rent from them nobody in their right mind thinks that the real assets in the UK banking system had increased from 0.5 to 8 but nobody was asking the question I think that really is a very distressingly revealing fact on the other hand it does it was my wife who found this my wife is reading and Robert Skidelsky biography and john maynard keynes and turns out my pure coincidence i knew i had a very interesting bloke sitting beside me on the economics affairs in the lords and very nice and he often shattered i knew his name was robert but i didn't sort of relate to his a second name which is get Skidelsky and judith showed me this is a quote from page 23 of the condensed version of the three-volume thing which rather summed up my own rather unkind view which i published somewhere in which I said to conventional mode of discourse in most of the economic discussions I'm familiar with would be more familiar to Socrates Athens than to a post enlightenment physicist as people have ideas they're clever people they have ideas and then they talk about the ideas this is an unkind way of saying the same thing now I'm just going to digress for a moment this is from a paper that I wrote with some other people and Joe Stiglitz he used to be our neighbor in Princeton and it's weaving up to saying some things about what that's the models that arrive morals that arise this is what these things are sorry that's a sort of rather unkind and people who can read quickly that's a rather unkind view of some of the valuation of these dodgy instruments and underlying insofar as there's a theoretical justification for the way these things were priced and I'm drawing now from a very interesting paper that I'll cite explicitly on the next PowerPoint but insofar as there is a pricing mechanism for these things it largely relies on these things from the mythological land scampia global equilibrium and there was a very interesting paper and sitting just here three seats back on this side and ears catchy ollie these three people's when they were still in tree when he was still in tree years did a really beautiful paper deliberately simplified model of derivatives and looking at that in a competitive situation and indeed it's a it's not only is it an elegant paper but it's a really mathematically beautiful paper it draws upon something even but the tough the hardest thing I ever encountered in theoretical physics with a thing called the I scene theorem and this is this calculation is that being tougher than that but it what it does do it suggests that you once you get these things too complicated they go belly-up so a drink I come to the concluding bit of all this which is what morals emerge from it well they're all fairly familiar to you and so won't again spend much time on that rather leaving a bit more for questions but these the fact that these these things are not so they're just over the counter they're not clear what they are trading for and more than that if you would as this thing says if you were to translate them it's been estimated by a very to various degrees of the horror that if you estimate it how much how many pages you'd have to construct in a convention consents prospectus that many of these instruments that are being traded it's sort of mind-boggling so one of the things you could do is try to make derivatives more transparent have them treated over exchanges ensure that they're not written by government-issued entities don't allow people to get a guarantee from the government that I matter how stupid they are they don't have to worry the taxpayer will pay for it and to end the priority they receive in bankruptcy that thirdly fear and of course ensuring we produce the hidden subsidy that is received from the public and the second reform B would make information on prices volumes and exposures available to regulators and the public another recommendation is all stuff from Andy all day but their's comes partly from the stability and complexity modeling's large capital reserves are definitely something we should look at but more importantly big banks should hold relatively larger capital reserves we've seen that model and we've seen that in things playing out realistically the big banks have bitterly resisted that using ludicrous argument that they have bigger reserves by and large it will go up to the crash they are having smaller reserves in relation to capital than the middle banks and they argument there is they wanted all the money working and when they were told and argued with so you should have bigger reserves they said well our reserves are bigger than the small bang furthermore the reserves tended to be run down in that boom that euphoria is the alleged earn lost I've lost that slide that I'm so fond of but as that thing was going from its century-old 0.5 up to 8 people were running down reserves big banks and little banks and clearly the the moral emerges that in boom times banks should hold bigger capital reserves and in bust times they should hold and worse now when the government gives you money cost-of-living percent to take a mortgage inside and they're hoarding things because they feel better doing that it's in fact in busts times and to get things going again nation they could indeed I want to end by observing a really couple of really good things that I don't have with me for distribution but anyone who wants to get all them it could be in touch with me the first one is a superb essay by Benjamin Franklin at Harvard not Benjamin Franklin Benjamin tree the good Friedman he wrote in a couple of years ago now in the spring issue of the American Academy of Arts and Sciences have a big conference each year and he wrote a superb essay on the financial system and one of the things he did and I'm not quite sure what this means because I only know the words but he said the cost of running the US banking system shorty men are mainly salaries I suspect the cost of running the US banking system 30 years ago was equivalent to 10 percent of all profits earned in America halfway to the present 15 years ago the cost was 20 percent of doubled and just before the crash it was up to 33 percent and he concludes his article by saying the time has come to ask how much it is costing this financial system whose basic purpose should be to allocate capital efficiently in a free market system now I think it's a very wise and not particularly energetically acted upon recommendation and I've written two books I'm just going to mention in conclusion and the last slide I'll put up oh sorry I've got a couple more slides this was just showing we how leverage went went crazy mr. turn goddamn I've forgotten to turn one of my slides over you missed a few things and you missed some of the reorganization of banking but nothing really interesting because I sit them all in anyhow and this is a set of bantam along here the volume of the balloon is the magnitude of the thing I'll so I'll leave that there's a some reference of people interested in but the two other papers one I've only just become aware of it was in one of the questions that's often posed well first of all last year for about a year we had a seminar series it organized by David vines whom some of you will know colonists at Oxford and one of his colleagues which was already entered to the notion of should there not be a duty of care for bankers the way there is for the medical profession or the legal profession so that when you behaved really badly you suffer quite draconian consequences and he's just produced that this book has just been published by Oxford and I thought it was really quite an interesting book but it didn't pursue the question that I kept trying to agitate forward in this seminar and finally gave up which is why as it is didn't some of these people go to jail the hidden darlin sacks after all who was aggressively marketing to clients instruments that he they were taking a massive short positional you would have thought and there's an excellent in the January issue of the new review of books some of you'll be familiar with there was a superb article by Judge Jed Rakoff RAK oh if it entitled the financial crisis why no top executives why have no top executives being prosecuted and so I leave you on those rather unkind comments well my time and the all things fine series and I trusted along here I find it I don't find it and released implausible because the instruments that drove that increase my view you know you now have the computers that enable you to do fancy things then you can take a given asset and if you're clever enough but even in this I just believe that so in some senses become like a price so you pay the price but you we were having conversation earlier today you will remember that the put on the special session presentation from the head of the Bank of Scotland who would explain motion that quite ridiculous and you will recall I did ask him sort of did he have no conscience essentially and was he's grieved and such I was really quite rude to him that was there was a lot wetter lively a sessions of the meeting and I I can see how people who the people most the people who did this probably didn't feel they were bad people I just feel a bit of bad luck I've been doing something trying to look after their clients but it I find it hard to believe that the major driver in this was the rent you could take from the activities you could generate that weren't available to you earlier the thing that David mines and others organized had people who had been involved in the lawyers thing and there again if you remember Lloyd's here's this bunch of people who get money for doing essentially nothing except lending their name and then all of a sudden they discover they've agreed by then they get their name that they're putting themselves at risk they're having to pay for something that went pear-shaped and they're very indignant that I want to do it and I just find that peculiar I guess I brought up as a Scottish presbyterian hit that may just be and but I think you can't rely on intrinsic virtue what we need are stricter rules that we have and better oversight that we had it was certainly transparent the fact that some of these markets were so low paid don't know if you read one of them best essays on this that and many of you must have mm mm I forget the exact date I think this 2002 what he just taken over an insurance company that had lots of derivatives he took one look at them and got rid of them all and in his letter is explained while he got rid of them all and he said he be hated them by say these these are financial instruments of the voice was there and from some Bobby furthermore with a bit of track record of managing money I think you had a group of people riding on there no regulation that said if this goes bust you're playing through and because in since it couldn't be because much of the money was high street money right back to class speed this graph doesn't seem to predict that kind of great event or frequency of event at all so I'm wondering how much explanation there is in that really whether you're at 30 percent or at 600 percent you might be one of the first things I noticed and the the answer that was given was yes it was a bad time but the assets held by bank bank assets for sure and that's so you don't see much order is conditional and having some big banks it's good to require them preserves but it was included with the benefit of having a few big mouths yes well that would be another way out of it but I thought if I were to write a paper that we were to write a paper and suggested something so under such credibility as we have I can see the point that small banks often serve different purposes from big banks and the big banks can do things the System Restore is a more efficient way towards achieving because and that's Christian we I'm surprised basically I think I'm going to repeat myself a bit because I I think the way to fix it is to try and make sure that you keep a closer eye on you keep a closer eye on capital reserves and you keep that closer eye in relation to the size and activity level thing and you do not I've sort of blanket guarantee that you can do foolish things and the government will step in and compensating for and ideally if you if you could you can begin with an interesting example they probably banking set up by Building Society football teams and all sorts of social things and then started I'd that became chairman man thought everybody else has done so in some sense some oversight of the banking system as a whole to make sure that the people who were in charge of the bank are people who actually knew something about banking that's not a good person not a bad person but he's an evolutionary biologist who writes good textbooks and I guess he thought he was doing the right thing for the bank because everybody else was waiting in some sense but that's not something one would have I wouldn't have the guts to write that down as a recommendation because you know if you did on the other hand have regulations on how much you could live abuse that probably would have Reserve more regulation yeah regulations the next quarter nice you

Leave a Reply

Your email address will not be published. Required fields are marked *