Vernon L. Smith (2014) - Rethinking Market Experiments in the Shadow of Recessions: The Good and the Sometimes Ugly; Propositions on Recessions

Thank you very much, ladies and gentlemen. I want to thank the organisers. The Countess for inviting me to this wonderful event. And I want to thank you for coming. The last 7 years has been an incredible learning experience for me. My co-author, Steve Gjerstad, and I - as you know we are not macroeconomists or monetary economists; we’re experimentalists. We started following the Great Recession, I think it was in 2006 or ’07. The time stamp in my memory is that I submitted an article to the Wall Street Journal in December 2007. The title of that article was that ‘We have met the enemy and he is us’. The Wall Street Journal changed the title to ‘The Clinton Housing Bubble’. (Laughter) You know, let you and him fight. I thought that Bill Clinton was simply an instrument of very popular policy. And so I mention the Tax Relief Act of 1997 in the article, but they picked up on that. You don’t get a chance to choose your, in the media you don’t get a chance to choose your titles you know. They do it. So we began by studying the Great Recession in some detail. And our reaction was: wow, this is an outlier, this is very unusual. We then went to the depression. And we saw the same thing. We saw the housing origins. We saw a balance sheet recession - and that term emerged early in our work. And then we looked at all of the 12 recessions in between the 1st and the 14th. And we saw no balance sheet recessions but an ordinary cycle, involving the interaction between the housing market and interest rate policy and Federal Reserve policy. I want to begin with the Great Recession and the Depression, showing you a couple of slides. And I want to relate, in particular, 2 rather polar components of the GDP, 2 experiments: housing and consumer non-durables. Here are the components of the GDP leading up to and then the Great Recession. Notice that all of these components are normalised on the 4th quarter of 2007, the beginning of the Great Recession. Here is non-durable consumer goods - we always separate that in our charts, in our book. Durable consumer goods. Plant and equipment. Investment and housing. Notice that housing peaked about 80% above its 4th quarter 2007 level and in the early part of 2006. It fell 1, 2, 3, 4, 5, 6 quarters in a row before the depression began. During the entire post World War 2 period, if you had a decline in new housing, single and multi-family new housing expenditures, in excess of 10%, you were going to have a recession. Then it bottomed out down here. And essentially remained steady until it started up in 2012. Here’s the Great Depression, same sort of thing, same chart. Here new housing expenditures rose 60% from 1922 to ’25, steadied in ’26 and fell in ’27, ’28 and then ’29. And that ushered in the Great Depression. All these other components of GDP growing. All decline coincidentally. Everybody taken by the same surprise. And housing then continued to fall, bottomed out down here in 1933, rose in 1934. I remember 1934 like it was yesterday. I was 7 years old and I lived on a Kansas farm. And in 1934 the bank possessed the farm. Good riddance. I look back upon that now from the perspective today. And I am so happy that no one came and said, you can stay on this farm. You will still owe the same amount of money, but we will lower your payments, at a lower interest rate, and stretch your loan and you can stay here. We would have been in a black hole of negative equity for at least 10 more years. We escaped. Now I want to go back to some early experiments of mine. Very briefly explain demands experiments. Those experiments really quite surprised me and generally were a surprise, I think, in the profession. We hadn’t expected markets to do that well. Here is a recent supply and demand experiment. This is in a high school workshop that we do. We always begin with this as an introduction. And this is a multilateral bid-ask market. And people are ... here is the first trade. And it’s a bid-ask market. Here’s the bids rising, asks falling. Until finally someone accepts the standing bid or ask and you have the first contract. And then an auction for a new unit begins and the bid-ask spread narrows until you have a contract and so on. The line here connecting all of these, that’s the contract line. And as each trade is made the buyer or seller is logging in to an accounting sheet, that person’s profit. The buyer buys for less than the value that he’s been assigned. The seller sells for less than the cost that we have privately assigned them. And they log in. They capture their surplus at the moment of trade. And then, after the experiment is over, we show what no one knew before, and that is the aggregate supply and demand for that market. And this one is very characteristic. The prices tend to converge from below into the equilibrium with a symmetric supply and demand. If you do a survey and ask people questions, one of the things you find out is that buyers always think the sellers are doing better than they are. But, in fact, in these markets very commonly, and most prominently, sellers do not as well as buyers. Buyers are perfectly capable of taking care of themselves in these markets. Here this is probably about the second or third experiment I did way back in the 1950s. The first couple of experiments I did were also symmetric supply and demand. And I thought that’s why it was that I was getting these results. And so I took a very asymmetric case and shot down this notion that it was an accident of that symmetric supply and demand. Notice now the defining features of exchange in these experiments. The buyers and sellers earn the surplus with each trade. Each is role-specialised - you know before you go to the market whether you are a buyer or a seller and you do not switch roles because of the price. All trades are final, no retrading. The expected gain is the utility value only. You do not gain from resale. These markets also show very fast stable equilibrium convergence. And you can do far more complex markets. We’ve done electric power markets with up to 9 nodes. And each node sellers are putting in bids to buy or sell electric power. And then it’s computer-assisted to run the algorithms to optimally distribute the power among the buyers. Those markets also find the equilibrium very quickly. So these are markets for hamburgers, haircuts, airplane trips, hotel space. These are non-durable consumer expenditures, and they are huge. If you subtract government expenditures from GDP this is about 75% of private product - theorem. The instability comes from the other 25%, primarily housing-mortgage markets. This is an important - I want to spend a little time on this chart. This is the price of housing relative to the price of all other goods charted. And I particularly want you to notice - here’s the run up to the Great Recession – that there was a relative price bubble that peaked in 1979, a second one in 1989. And notice that in those bubbles we have a convex price path. The cords are below the function. That’s the common standard bubble. And also it's true of the bubbles typically we observe in the lab. They do not have this concave shape that we have. This is really quite unusual in the Great Recession. One way to model these bubbles in the lab is to assume you have 2 types of investment sentiment. You have fundamentalists. Fundamentalists buy in proportion to the discount from fundamental value and they sell in proportion to the premium. And then the other type of trader is a momentum trader. And they simply buy in proportion to the current rate of change in price. The interactions between those 2 kinds of investors, depending upon the convex mixture of those 2 types of sentiment you can get, basically generate any kind of a bubble or none. But it’s the momentum traders, you see, that can give you this sort of a pattern. And I also want to mention: Notice here that in 2001, right about here, the market has reached this previous all-time peak. interest only loans, subprime loans, the miracle of the negative equity loan. All of this great innovation in mortgage finance had come along. We already had a self-sustaining expectation of rising prices firmly established. That bubble already had a head of steam by 2001, before all of these other products came in. Also I want to mention, this is something we talk about in the book, the paradox of economic policy, monetary policy. Suppose in 2001, Greenspan could see that we were heading for disaster, bound to head for disaster. He intervened to squelch this bubble. He would have lost his job. Imaging hearings in Congress and Greenspan goes and he shows charts. And he says, oh, I headed off a much bigger disaster. And he shows the kind of charts maybe that you’re going to see here. Who is going to believe it? The other thing, finally, I want to mention about this chart: We don’t know why it is that people get caught up in self-sustaining expectations of rising prices. Don’t know the answer to that question. But there is reason to believe that things can trigger the set of circumstances that get you into that situation. And I think a really good candidate in 1997 was the Tax Relief Act passed in the Clinton administration - enormously popular. You would be able to take up to half million dollars in capital gains on a house tax-free. If the holding period had been 2 years, you could do this many times. You could buy another one, hold it 2 years and do it again. Democrats loved it. Republicans loved it. Libertarians loved it. Now you see why I called my article ‘We have met the enemy and he is us'. This bubble is driven by the flow of mortgage funds. You’ll notice there’s a surge here that accounts for the ’79 peak. And a surge in mortgage flow, in mortgage credit here. And then a huge surge here, tripling of the flow of mortgage credit. People are buying. You’re buying something that lasts 50 to 100 years and you’re buying it mostly with other people’s money. By the way, I think the origin of OPM, the other people’s money, that’s Adam Smith. He uses it in the 'Wealth of Nations'. This flow of mortgage funds collapsed and went negative about right in here, about early 2007. The first time in a peace time year that the flow of mortgage funds became negative since 1930. The flow of mortgage funds in the depression from about 1930/31 it went negative; it never turned positive again until 1938. Now we’ll just look briefly at asset bubbles here in the lab. And, by the way, the performance here was far worse than we expected. We kind of inappropriately extended our earlier learning to this asset environment. And this was supposed to be a very transparent environment in which we would not get bubbles. People are given endowments in cash and shares. There’s a well-defined expected dividend at the end of each period, a random draw. And it’s explained to the subjects just before they go into the first period: In the first period your're going to be trading an asset that will subsequently, over 15 periods, enjoy 15 dividend draws. So that fundamental, that dividend holding value of a share, that you’re trading in the first period, is 15 times the average of the expected dividend. Each pair is 24c, that’s $3.60. We tell them. They pay no attention to that. At the end of the first period we say, alright, now there’s 14 periods left, 14 draws. Fundamental dividend holding value is $3.36. Here’s where the trading is, down here, with inexperienced subjects. They come up, they peak out and they crash near the end. Bring that same group back, the same environment - and this is the pattern. Bring back a third time. They’re gradually converging over time in different sessions, sequence of sessions. They’re learning by experience. This is the same chart except there’s deviations from fundamental value. And our model of bubbles predicts that if you endow people with more money, relative to shares, you’ll get a bigger bubble. And that’s because of the momentum trading. And this is the total of 12 experiments, and these means are plotted. Notice all of these tend to be convex paths. Oh, and by the way, we discovered in 2000 a very simple way to get rid of these bubbles. Very simple - couldn’t believe we hadn’t thought of that before. You declare the dividend but you don’t pay it in cash; you escrow it into an account. And it’s paid at the end of the experiment. There’s no flow of new cash into the system. We don’t get a bubble. We did 7 of these, we got one tiny, very tiny miniscule bubble. All of the rest traded at fundamental value. We thought, well, if that’s true, suppose we replicate these experiments and we declare the dividend and pay half of it in cash and the other half at the end. Now we start to get bubbles again, but smaller. These are interesting experiments by Lahav. He has done very long - 200 trading periods. These are crossing markets. Each period you put in a PQ and it’s either a bid to buy or it’s an offer, a PQ offer, to sell. And those are crossed to yield one market clearing price. And notice there’s a group down here: it gives us the bubble, it collapses, and you get another bubble, comes back. These are entirely home-grown. There are no external events precipitating these. These are just entirely ... and then here’s one that trades pretty much at fundamental for the whole period. Let’s look at severe downturns as balance sheet recessions. And, particularly, the Great Recession and Depression as balance sheet recessions. Here is the anatomy of negative equity in the housing market. We have here the value of all single multifamily homes, plotted from 1997. It’s growing not only because prices are growing, but also because you’re pumping out more units into that stock. Here is the growing mortgage debt to finance those. It’s growing not quite in step. The difference is equity. Equity is growing here from around 6 trillion to over 12, peaking out in 2006. And then this market collapses and it collapses against fixed, even growing. The debt continues to grow slightly, gradually going up. And so all of the impact of that decline in value is on equity. So over a 15-year period, from 1997 clear up this continues out into about 2012, you have total equity in the stock of homes, US homes, fell from 6 trillion to 5 ½. It dropped by ½ trillion dollars. No wonder we’re stuck in low growth. That’s an incredible balance sheet shock. Here is the period for the ‘20s and the depression. The main difference here is that debt is growing more closely in step with value. Notice how flat equity is. It doesn’t grow as much during the entire period. Growing up to 1929, and then it collapses. Equity is not restored to its 1929-level until 1940. This is a typical post-war period. We get some crunches, but we don’t have a major decline in equity like we had in the depression and the Great Recession. In none of these cases does the Federal Reserve completely lose control over the housing market through interest rates. You can bring it back. The ordinary business cycle - what is that? We call it the consumer housing cycle. It’s the 12 recessions here in between the first and the last. Here’s a balance sheet recession. Here’s another one. You only get one of these about every 80 years. No wonder you don’t expect it. And you’re not prepared for it. And these recessions here - looks like there are some kind of false positive signals from housing as a leading indicator. But here’s a collapse in housing. A recession doesn’t occur. Here’s the collapse in housing. A recession doesn’t appear. But we had this build-up for the Korean War. The same thing here: We had a build-up due to the Vietnamese war. If there are not large numbers of damaged balance sheets in the economy - if basically balance sheet is in pretty good shape - it is true that government spending will trade off against private spending as one normally in the flow models expects. Always, in every case, when the recession ends housing is coming back. This is almost an exception here, except we’re stuck in low growth. Proposition 3: How do you achieve escape momentum from recession when there’s large numbers of damaged household and bank balance sheets? Well, you don’t do it with monetary policy. As Sir Denis Robertson said in the 1930s: It’s like pushing on a string. You also don’t do it by government deficit spending, and for the same reason. There’s too many low, zero and negative equity balance sheets. World War 2 spending was not an exception. By 1940 we had experienced 10 years of balance sheet repair. Eventually you repair those balance, and government spending will start to behave normally. So with the coming of the war in 1940 and balance sheets repaired, we started to ... that doesn’t mean that the same amount of spending in 1930 would have had that effect at all. Proposition 4: Bankruptcy and default as a balance sheet repair and reboot process. And a couple of prominent examples. Sweden is an example that in the early ‘90s, when they basically faced a balance sheet recession with a collapse in investment and housing market, they put their banks through. The banks went through bankruptcy, they zeroed out the equity. And they got their system going again. Also the FDIC model. The FDIC model is ... 489 banks failed over at the FDIC from 2008 to 2013. They weren’t big news items at all. What you heard about was Bank of America and City Group being bailed out. You didn’t hear about all of these balance sheet repair processes that were going on all over with small and medium-sized banks. Now why is bankruptcy and default important? New capital then directly supports new lending, undiluted by cash claimants whose investments failed. You don’t have to dilute the return on new activity by paying for these historical claimants who you rescued from the consequences of their decisions. In 1900, Henry Ford was producing the model T automobile. By 1920 he had produced over half of the automobiles that had ever been produced in the world. He was growing, he was reinvesting. Imagine that you had required him, that return to be diluted and paid to the carriage makers, the livery stables and the horse breeding farms that were all becoming obsolete. Think of that as the whole problem of what you face when you bail out investors and give them a claim on a future output and activity of the economy. You drag down, and you’re going to slow that down, and you’re going to be stuck. Proposition 5: The political process protects incumbent investors from bankruptcy and default. They’re there, they’re very visible. They make contributions to political parties. They help - in the United States choose who is going to be secretary of treasury. They help choose advisors. They are very visible. We know who they are. You don’t know who all the investors are. They’re invisible. The ones that are going to be investing to give you the recovery, they are not around. You don’t even know who they are. The bail-out champions in the United States were Bank of America and City Group. City Group actually was 3 times before they finally got their balance sheet more or less stable. But both of those companies their shares sell today for 75% of book value, a big discount. Investors - you can’t hide this from investors; they don’t trust the books. Wells Fargo Corporation, which never needed a bail-out, that was pretty evident early. A number of reasons for this: They sell for 175% of book. So the market here is telling you a lot about what’s going on there. Really, prediction is difficult because the US has no experience with massive downturns, bail-out cures, and continued balance sheet damage. Proposition 6: Stock market crashes, unlike housing market crashes, do not bring recessions. We got that fixed. The late 1920s, in the private sector and the banks started to raise margin requirements. In 1933 the New York stock exchange required that of all member brokers, they required their customers to have 50% margin. And it was codified in the SEC Act of 1934. So this put property right limitations on the use of other people’s money. And also these were call loans. If you fell into negative equity you got a margin call, and you put up cash. If you don’t put up the cash they sell you out. You’ve authorised on a margin account for them to do that. So as a result the balance sheets get cleaned up as you go. When the stock market falls, margin debt falls in debt - so there’s no build-up. The Federal Reserve didn’t understand this in 2005. They had a conference in 2005. Is there a housing bubble - yes. The answer was yes, very clearly. Because housing prices are out of line with the prices of other goods, with rents, with median income. But they didn’t expect it to be that serious because the fallout from the dot-com crash of 2001/2 was relatively modest. They didn’t understand the difference between this chart and the one I showed you earlier. This is not rocket science, it’s pretty simple stuff. I’ll let you do your own forecasting here. Is this sustainable? This is the ratio of home price to income. And here is kind of the base line, way back here. The return in the crash never got back to the base line. Various stimulation programmes arrested that. And then we’ve now had a recovery in the price of homes. But income is not moving, median income is stuck. So we’re clear up, we’re half to this level now. Ladies and gentlemen, I’ll let you await the consequences of this. Thank you.

Vernon L. Smith (2014)

Rethinking Market Experiments in the Shadow of Recessions: The Good and the Sometimes Ugly; Propositions on Recessions

Vernon L. Smith (2014)

Rethinking Market Experiments in the Shadow of Recessions: The Good and the Sometimes Ugly; Propositions on Recessions

Abstract

In all economic recessions consumer non-durable expenditures are relatively stable, characteristics that have long applied to certain experimental markets. The supply and demand experiments of the 1960s and 1970s reliably and rapidly converged to the predicted equilibrium. These markets also were for non-durables: all sales were final; items could not be re-traded; subjects each knew their buyer (or seller) role in advance of entering the market.
In eleven of the last fourteen U.S. Recessions, new housing expenditure was a leading indicator of decline. Moreover, economic recovery is always associated with housing. The experimental asset markets of the 1980s and 1990s likewise exhibited unexpected and unpredictable tendencies to bubble and crash. The larger the endowments of cash or the greater the inflow of cash, the bigger is the bubble; housing bubbles are also associated with the inflow of mortgage credit money.
In the macroeconomic accounts, re-tradability, long durability, and credit inflow seem to be the key to instability.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content