Are bubble theories fools gold? Or successful alchemy?
There’s been a lot of buzz in the blogosphere recently about attempts by Brad DeLong and Paul Krugman to develop models of asset bubbles. I can’t blame people for trying; we’d all like to understand why we keep getting these crazy price peaks in tech stocks, oil, housing, etc. But as you may have already noticed if you read my EMH post, I am somewhat skeptical. Today I’d like to look at this question from several different angles, starting with a question:
What’s the point of bubble theories?
This is not a rhetorical question; I am actually pretty ignorant of bubble research, and am hoping a reader can fill me in. My initial thought is that the point of bubble theory is to explain why bubbles happen. And doesn’t that mean you would have to be able to identify bubbles in real time? And doesn’t that mean your theory would have to be able to provide useful investment advice? I’m thinking of something along the following lines:
Imagine a model that predicts the optional allocation between index stock funds and Treasury bonds, based on two factors; years to retirement and level of stock market relative to what your model suggests is some sort of “fundamental value.” The percentage allocated to stocks should be positively related to years to retirement and negatively related to the ratio of stock prices and fundamental values.
Is that what bubble theories should produce? And if not, are there any practical implications, or are they simply “just so stories?” I vaguely recall reading that there are “rational bubble” theories, and/or bubble theories that don’t imply it is easy to beat the market. But surely there must be some practical implications; otherwise what’s the point?
For the rest of this post I am going to assume that the point of bubble theories is to find deviations from EMH that have at least some useful implications for financial investment decisions. (If I am wrong, stop reading now and don’t waste any more time.)
Here’s my next problem with bubble theory; the discussion in the blogosphere is not related to rocket science, it is not related to the sort of sophisticated math used to explain options pricing, rather Krugman suggests that what is being discussed are theories about psychology:
Brad DeLong offers a neat little model of speculative fluctuations in asset prices, based on the idea that investors gradually switch strategies based on what seems to work for other people: if people buying stocks seem to be doing well, more people move into stocks, driving up prices and making stocks look even more attractive…
But aren’t sophisticated mathematical models the only area where academics might even conceivably have a comparative advantage? I would think that psychological theories are where you would expect savvy Wall Street investors to have a comparative advantage over academics. I would be intimidated by the thought of modeling bubbles; I would think someone like Warren Buffett would be far better than me at intuiting market irrationality. Yet even Buffett lost about 30% in the market last year. How the h*** am I going to see something that he missed? And not just Buffett, but 1000s of other sophisticated investors who also lost a fortune.
When I first started working on this post I was trying to think of the right metaphor for a hopeless search; the search for the Holy Grail, the fountain of youth, Presbyter John, the alchemist’s formula for gold, etc. Then I realized that the alchemy metaphor had an interesting double implication. To see why, let’s imagine Krugman and DeLong succeed in their quest; what next? It seems to me that there are two possibilities:
1. They keep it secret.
2. They publicize it.
Here is my prediction; they will never get the recognition they deserve for succeeding, even if they are successful. Why? Because the theory will only be useful if it is kept secret, but in that case the public will never know that they succeeded.
Rorty famously argued that “truth is what my peers let me get away with.” I know most people don’t like that sort of post-modernism; but can we at least agree that “what is considered truth is what your peers let you get away with.” In other words Krugman and Delong will have to convince their peers that their bubble theory is correct. But if they do, if it becomes generally accepted that a particular asset price model is superior to the EMH, why wouldn’t the advantage of that model go away? We know Wall Street types feverishly look for any tiny advantage that would allow them to earn above average returns. It seems to me that any generally accepted theory of asset prices would immediately become incorporated into asset prices. But in that case it would no longer predict. And then their idea would be discredited, after all, it’s easy to fit a bubble model to past data. If it doesn’t predict future data, nobody will care. It will be like those market “anomalies” that go away after being publicized. And this is what is especially interesting to me—the bubble model would become discredited even if they were in a sense “right.” (By “right,” I mean that had they kept the idea secret, the model would have continued to perform well.)
On the other hand if they keep the model secret, they can become (somewhat) rich. But not for a very long time, as not even the anti-EMH types think the market is so inefficient that it is easy to quickly gain riches. If it was merely a choice between academic glory and riches, then we would expect any serious academic to go for the glory (unless they already have a Nobel prize.) But that’s not the choice; there is no strategy that leads to academic glory.
By now you may have noticed that this is roughly the quandary that faced medieval alchemists. We now scoff at their foolishness, as they clearly did not succeed. But I’d like to offer a contrarian view. How do we know they didn’t succeed? A formula to turn lead into gold has zero value if it became public. Patent protection was weak, and presumably the lead into gold transformation industry would have been a competitive constant cost endeaver. So even if some obscure alchemist had succeeded, we never would have heard about it. Have you ever wondered how the Bank of Switzerland ended up with so much gold? Just asking . . .
Now let’s turn to the motivation for the search—real world asset price bubbles. Arnold Kling discusses one famous stock bubble, and ends with this observation:
Incidentally, one thing I got from reading Galbraith’s book on financial euphoria is that you can fit the late 1920’s into this model. Holding companies bought stocks, and the holding companies were bought by other holding companies, and so on. You get the spectacular leverage, and on the way up people think that the assets of the holding companies are pretty low risk. Then when people get a little nervous…
Let’s consider the innocuous phrase “then when people got a little nervous.” Why might people have gotten a bit nervous in late 1929? Maybe because after the best two year period in American macroeconomic history (from an investor perspective), the stock market was suddenly flooded with a torrent of bad news.
In the two years prior to the October crash the U.S. economy grew rapidly with slight deflation. This was highly desirable as gold was undervalued after WWI and there was a long term deflation risk that needed to be carefully managed. There were also low marginal tax rates, Federal spending was 3% of GDP, we had budget surpluses, trade surpluses, pro-business Republicans controlled the Federal government, unions were weak, unemployment was low and corporate profitability was rising fast as new technologies were transforming American industry. Stocks would not have been overpriced if the Depression had been avoided.
Then everything went wrong:
In late October the Attorney General announced a crackdown on mergers—which is always bad news for stocks.
The Republican Party in Congress tore itself apart in a bitter fight over the Smoot-Hawley tariff. Hoover’s political ineptitude became painfully obvious. This was the headline story around the time of the crash.
The political situation in Europe (which had been improving), suddenly worsened dramatically with the death of German statesmen Stresemann and the unexpected gains made by German “nationalists” in a referendum on the war debts issue.
And most importantly, the world’s major central banks suddenly and simultaneously adopted highly contractionary monetary policies. This led to a rapid rise in the world’s gold reserve ratio after October 1929; a highly deflationary policy stance.
What is particularly interesting is that 3 of the 4 factors that would cause the Great Contraction (tight money, Smoot-Hawley, and international discord preventing policy coordination) all reared their ugly head in the period from late September to late October 1929. (The last factor, banking crises, wouldn’t begin for another year.)
So here’s my question: Should we encourage young economists to do careful study of historical events like the crash of 1929, looking for rational explanations of what initially appeared to be irrational events, or should we encourage them to think up neat new psychological theories of stock bubbles?
George Bittlingmayer came up with the theory that the change in merger policy contributed to the crash, but the rest of it I developed on my own. (And you already know about my theory of the crash of 2008.) If an obscure professor at Bentley can come up with a plausible rational model of the 1929 bubble, then imagine what we could expect from professors teaching at places like Princeton and Berkeley. Arnold Kling gave some advice to DeLong and Krugman. If I’m not being too presumptuous, I’d advise them to give up trying to turn lead into gold, and start working on the laborious process of trying to extract nuggets of rational behavior from seemingly barren mountains of perplexing economic data. There’s less instant gratification, and there are no short cuts to wealth, but the satisfaction it produces is deeper and more long-lasting.
Tags:
20. May 2009 at 15:27
Well Fama and French do have something of a bubble theory – though they deny that is what it is. But they do show that “value” stocks (stocks with a low ratio of price/book value) have (had) a pretty strong tendency to do better than “growth” stocks(the pretty and popular and already high priced ones). Their work has been around for some time but it still seems to be working ok. They deny that this is irrationality – they claim it is a risk factor – but offer no evidence for the claim. Psychology would seem to play a role – especially for those silly investors who have the odd habit of buying stocks one at a time.
20. May 2009 at 15:29
Scott, i just realized that i’ve never even commented on your NGDP idea. I suggest you get COMPLETELY away from the blame premise no matter how indirect. I see value in having an NGDP futures market, as well as Real GDP. In fact, I believe it would be beneficial to have any and all types of futures contracts because the market is a voting mechanism that would provide important POLITICAL cover for decision makers. I think we might differ in whether the Fed should actually be trading these markets. I believe the main value would come from observing the market’s OPINION and not from being able to influence what the opinion is. Things become too circular and i think we get into a very dangerous game (but i do have to think more about this). I believe we can also benefit from having all kinds of contracts form, the key is to have marketmakers. Schiller’s housing contracts didn’t get as popular as i thought they would, but housign isn’t as commoditized, so they’re not a great hedge. Intrade has all kinds of contracts, but it’s 100,000’s of magnitudes away from how deep the market needs to be. In fact Intrade does have all kidns of GDP and employment bets. To be able to trust markets, the “voting” public/open interest needs to be sufficiently high. therein lies the key. how do we get there?
20. May 2009 at 16:00
“The market can stay irrational longer than you can stay in the black.”
20. May 2009 at 16:04
Suppose you had a theory of bubbles, that did not let you identify a bubble until after the fact (too late to make profits), but did tell you that bubbles were more likely to occur under certain conditions. And that a change in (say) regulations would change those conditions and make the occurrence of bubbles less likely. That would seem to be a useful theory, even though you can’t profit from it.
20. May 2009 at 16:33
Suppose you had a very profitable theory of bubbles, used it spectacularly well without employees, and then retired, and were too lazy/selfish/scared to tell anyone about it? “Scared” meaning: just in case I spend more of my money in retirement much faster than I thought I would, I would like to have the ability to rebuild it in short order, hence I won’t become a part time professor at a college near you….
20. May 2009 at 16:58
JimP, I haven’t studied the F&F model, but I have read about many such models. Here’s my question: You are a young money manager trying to get established on Wall Street. Why not simply invest on the basis of several of these bubble models? Then you should do better than average. But few money managers beat the dartboard. Why? I don’t find any answer I have heard to be convincing. Maybe markets are just slightly irrational, not enough to make up for the expense ratio. Is that the answer?
Alex, Quite a few of your questions are answered in the NGDP post. I discuss circularity, liquidity, etc. I do favor subsidies to insure adequate liquidity.
Coray, My post addressed 40 year retirement accounts (401Ks) That’s where most of my money is. 40 years is a long time. Is the market irrational that long, or is the investor just not understanding what drives the market?
Nick, I see your point, but I just can’t visualize how it could work. You go to Congress or the Fed with an anti-bubble plan. They want evidence. So you show them that bubbles are more likely to occur under some environment (laissez-faire, low interest rates, whatever). But how do you know? You’d have to produce examples of bubbles. But how does one know that a bubble is a bubble? It can’t just be an up and down in price–asset markets are supposed to be volatile. What is too volatile? Doesn’t there have to be some sort of predictable mean reversion? And since things like P/E ratios are observable in real time, wouldn’t that have investment implications? Again, I am not an expert. But I really have trouble visualizing a useful bubble theory that is only useful for public policy, not investment advice. But you have zeroed in on the key assumption that my whole post relies on, I don’t know if I am right in making that assumption.
happyjuggler0, Yes, that is actually the whole point of my “alchemist” analogy. I probably didn’t make it clear enough, but I agree there might be people out there who did figure it all out, and never told anyone. (Of course I was joking about lead to gold, but a market alchemist is possible.)
20. May 2009 at 17:36
Scott wrote:
My initial thought is that the point of bubble theory is to explain why bubbles happen. And doesn’t that mean you would have to be able to identify bubbles in real time? And doesn’t that mean your theory would have to be able to provide useful investment advice?…
Is that what bubble theories should produce? And if not, are there any practical implications, or are they simply “just so stories?” I vaguely recall reading that there are “rational bubble” theories, and/or bubble theories that don’t imply it is easy to beat the market. But surely there must be some practical implications; otherwise what’s the point?
Scott, as you know, my preferred theory of business cycles (and I admit it’s not perfect) is the Mises theory that central banks push down interest rates and set in motion a speculative boom.
But let’s not argue about the Austrian theory here. Let me give an analogy to show that I think your deductions quoted above are overlooking something.
Suppose there is a serial killer who goes up to cartons of milk in grocery stores and injects an undetectable poison with a syringe that leaves a tiny hole, and then the serial killer puts a tiny drop of glue on the hole. So after he is done, it is very difficult to detect which carton has been infiltrated.
All of a sudden people start dropping dead of a mysterious sickness. Doctors figure out after the fact (i.e. during an autopsy) what it was.
Stores review their security video and see the guy in action. But he always has on a hat and glasses, so the detectives can’t figure out who it is. Naturally the police send out warnings to all the stores, but it’s really hard to watch every dairy section in the state–the guy drives all over the place and sometimes he waits for weeks before striking again.
Because of the rarity of the poisonings and the importance of milk, it doesn’t make sense to stop selling milk or to totally revamp the packaging. (Or maybe they switch to selling milk in glass containers, and the killer switches to poisoning OJ cartons.) So in “real time” nobody alters his behavior that much. Some people stop drinking milk of course, but after the initial scare people go back to their old ways, even though they know they aren’t in the original “equilibrium.”
In this scenario, it certainly is useful for people to study why this keeps happening, because if they could stop the guy, then no more poisonings. (This is analogous to the Austrian recommendation to return money and banking to the market.) This is so, even though during the intervention period (when the killer / Bernanke are on the loose), individual consumers might not find it profitable to completely shield themselves from the risks.
20. May 2009 at 17:41
Disregard bubble theory (or not), how does an economist from Waltham (I actually lived in Waltham a couple of years not too long ago, not that that means anything…) who thinks he has a wonderful explanation for the “crash of 2008(ish)”, which he thought of ahead of time, or in real time, and which almost no one else has thought of, actually believe in the Efficient Market Hoax (EMH)?
Isn’t that cognitive dissonance? Anyone, e.g. a clued in hedge fund manager, who was market oriented who thought the way that you do could have made a large bet that we had too tight money, and that this would cause certain obvious (to the hedge fund manager anyway, if not the Waltham professor) market dislocations that weren’t discounted by the market. And made a bundle in the process….
20. May 2009 at 18:58
Scott-
And doesn’t that mean your theory would have to be able to provide useful investment advice?
In a system with multi-equilibriums, you can have a bubbles, but be unable to reliably profit from them. From 2004-2007 there was a significant risk of a currency run. Economists worried about China and the Gulf States deciding to stop buying bonds and treasuries. Hundreds of tons of gold were flowing into the GLD ETF. Monetary inflation was probably over 8% a year, while CD’s only paid 4-5%.
The person on the street did think in terms of “currency runs”. But they did know that the dollar was depreciating, and the best strategy financially was to get out of dollars, ASAP. The dollar was turning into a Latin American currency, with equities and housing playing the role of a store of value.
In a system of low monetary inflation and no risk of a currency run, the equilibrium price for stocks is to be priced on expected cash flows. Calculate the expected dividend yield over twenty years, factoring in inflation/dividend growth. Then compare to the rates of BAA bonds. If expected cash flow from stocks is less than that of bonds, stocks are a poor buy.
But when money supply is growing faster than the interest on bonds, then all bets are off. The equilibrium is for there to be a currency run.
So I think stocks and housing basically priced in the possibility of a currency run (or if not an actual run, priced in the possibility of the dollar going the way of the Peso). It then only took a little bit of deflation to dispel the risk of a currency run and make prices come crashing down.
Yet even Buffett lost about 30% in the market last year. How the h*** am I going to see something that he missed? And not just Buffett, but 1000s of other sophisticated investors who also lost a fortune.
I do think it’s possible for a savvy person to beat the market. When dividends yields are much lower than the yields on BAA bonds (say under 3% ), do not buy stocks. Buy bonds. When dividend yields are higher, invest more in stocks. And when you buy stocks, avoid the “trap” sectors that never pay dividends (tech, airlines, etc ). Buy a dividend ETF that diversifies across sectors (make sure it is not to heavily invested in any one sector, like energy or financial). Also, pay attention to monetary inflation. If measures of monetary inflation seem to be at a sustained high level (NGDP, MZN, total credit creation, gold prices, farmland prices, are all growing >5% a year), than increase investment in equities and gold. But at the hint of a bank run or deflation, pull back fast.
The trouble with the above is that it is a recipe for beating the market by two or three percentage points a year. But it guarantees that you will never blow out the market. The average case is better, but the best case is worst.
Most of the famous investors chose a strategy that was poorer than the above strategy in the average case, but works great in the best case. Their strategy is to try and time the bubble (sometimes accidentally). It works awesome in the best case, but it is a bad strategy in the average case.
20. May 2009 at 20:53
Isn’t that cognitive dissonance? Anyone, e.g. a clued in hedge fund manager, who was market oriented who thought the way that you do could have made a large bet that we had too tight money
That is in fact EXACTLY what Sumner claims DID happen. The market received unexpected information of tight money/low NGDP growth and so accordingly went on a selling spree.
20. May 2009 at 21:00
I think you misjudge one of the key functions of the financial system which is not to avoid bubbles but to blow them. The greatest profits are to be made in skimming money from fools and bubbles offer the ideal prospect. It is the why behind ponzi schemes. It is just most of the time the ability of finance to persuade others to bite is limited, the burden of defectors can be heavy, and all too easily you can end up the fool.
Yes, a successful theory can be self-defeating. Yet it is always tempting to believe what you want to believe. Anyone looking at incomes and housing prices or equivalently debt from 2004 on knew we had a bubble. It was apparent they simply could not be afforded. Even knowing this doesn’t tell you how big the bubble will get, when it will burst, where the wreckage will end up, or how bad it will be afterwards. Some will avoid them, some will hope to get out before they burst, some will play them for what they are worth, some will misjudge them, and some will be suckered into them. Bubbles exist, but they are difficult to predict and difficult to profit from them.
20. May 2009 at 21:26
Let me see. In mid 2007, the price of oil peaked. Was that a bubble? Depends on which side of the trade you are on. For producers, it turned out to be a bubble. For consumers and domestic importers it turned out to be a constraint.
Wait, they are both right! It was an inventory glut somewhere in the middle or beginning of the supply chain. The monetary authorities cannot pop it directly, but their best bet is to create a glut of money in the supply chain at the same level, thus matching the flow of money to the flow of oil.
The only way for the monetary bank to do this is via oil futures, becoming an expert on oil issues, and, well, becoming an oil distributor.
P.S. Scott, I have become a fan. Your blog got me across a difficult theoretical barrier.
21. May 2009 at 04:45
I’ve been working a little on my own version of a bubble theory: there are research results in behavioural economics which can inform these ideas.
(here are a couple of my VoxEU papers which outline the idea: http://www.voxeu.org/index.php?q=node/3548 and http://www.voxeu.org/index.php?q=node/2814)
My meta-idea is roughly in line with Nick’s: if we can come up with a good model of how this happens, and publicise it, it will be less likely to happen in future. Imagine you were the first to come up with the idea that central bank monetary interventions could help stabilise inflation (or NGDP) – your insight might help you make a bit of money, but more importantly it could help the central bank to achieve that stability. This might appear to make the theory self-defeating, because once it’s discovered it is no longer needed – but actually that’s not the case. Once it’s discovered and publicised, it is no longer a private source of profit, but it is still of public value.
Anyway – the distinction here is not between rational and irrational – it’s about making the definition of rationality more subtle. If the EMH really worked perfectly, then the markets in 1929 would have already discounted a certain probability of the changes in US politics, the change in the German mood and the change in central bank policy. Of course the realisation of those events would still have had some impact, but probably not caused a crash. In fact, the likelihood of those events was probably not taken into account at all by most investors. No doubt some people did consider them, but sufficiently few that they didn’t have enough capital or enough time to bring the market back to the correct EMH-given level.
My theory is that investors’ behaviour is rational in the short term, but that they are fairly short-sighted due to cognitive limits and the lack of perfect information. More short-sighted in a fast-changing environment, and more long-sighted (or more “rational” in the conventional sense) when a period of stability gives them more capacity to work out the probable future and carry out arbitrage. The EMH does, in effect, work in the long run, but the rational thing to do in the long run is to quote Keynes on the long run…
I think that this short-sightedness can be systematically described, can give us more insight into asset bubbles, and can lead to policy recommendations; this is what I’m trying to do and probably similar to DeLong and Krugman’s aims too.
21. May 2009 at 04:58
Every economy with fiat money is living with a constant bubble. Here you have a reference to a paper on bubble by Jovanovic
http://www.econ.nyu.edu/user/jovanovi/bubbles4.pdf
21. May 2009 at 05:16
Bob, I don’t quite see how the analogy applies to my main point. Perhaps it applies to my less important point. My main point is what’s the point of bubble theories? What are they supposed to tell us. The theories don’t say “we don’t know who is behind the glasses” they claim to know who it is. If DeLong and Krugman study the market and announce they have no explanation for market movements, no one would consider that a successful outcome. My question is; what would a successful bubble theory look like? Would it be an investment guide? Would it still work after being discovered? That sort of thing. In your analogy we know what a successful investigation would look like–you’d identify the bad guy.
I’m not saying people shouldn’t look into bubbles, they should, and they did. But at this late date what more are we likely to find? They’ve been studied for years. I have a hard time believing some explanation will pop up that will have practical implications. But I guess hope springs eternal.
Happyjuggler0, You misunderstood my argument. I don’t claim to be able to predict the 2008 or the 1929 crashes. I claim to (partly) explain them. They occurred precisely when the bad news hit the markets, leaving no unexploited profit opportunities. I claimed to identify the bad news.
Devin, I don’t see how your explanation responds to my point. You told a story of what might have happened. I could tell a hundred different stories, all equally plausible. What is the purpose of these stories? Is it to help us invest better? That’s what I am asking? If it is an investment guide, then won’t it be worthless immediately after it is known to the public?
On your second point, I don’t agree that successful investors time bubbles. Buffett doesn’t, and he’s pretty successful. If Buffett can’t figure out the right time to sell, how can you or I? Maybe your strategy will work, but I find it hard to believe that lots of rich, sophisticated investors would be in the market when you say they shouldn’t be, if it was clear that you were right. I am asking you to step back from your view about whether you are right or wrong, and think about the earlier problem of how would we even know if a bubble theory was right? How do we evaluate it? If it’s right, won’t it stop working right after it is generally accepted?
TGGP, Thanks, that’s similar to my response.
Lord, I don’t think the purpose of financial markets is to skim money from fools. A person who simply bought whatever stock he drooled on when reading the stock pages, would have done better than all of those sophisticated investors, banks, and hedge funds that invested in Madoff. If you want to say all Madoff investors were fools, that’s fine, but that’s also circular reasoning. On the other hand if you define “fools” as people with IQs below 80, or some other objective criteria, the fact is that their stock investments do just as well as the Mensa club members. In other words they (we) are all just monkeys throwing darts.
Mattyoung, I also addressed this in another comment section. But the short answer is that I see commodity prices as being set in world markets. Countries like China are the big consumers of cement, iron, coal, etc, not the US. Growth in developing countries drove up prices, and when the growth slowed sharply commodity prices plunged. US monetary policy only matters to the extent it affects world growth. Don’t “follow the money,” rather follow the growth. Speculation is overrated. Futures contracts are merely people making bets. Only speculation in the form of actual demand for commodities affects prices, and that sort of speculation has less impact than most people believed. The commodity price bubble even affected “goods” that can’t be hoarded profitably—like shipping rates.
21. May 2009 at 06:18
“Stocks would not have been overpriced if the Depression had been avoided.”
Scott, I have to say I think you’re flat wrong here. Granted, absent the Depression stocks wouldn’t have declined by 80% in 2 1/2 years, but its highly unlikely that corporate earnings would have grown at the same rate as in the previous years. Economic growth in the U.S. has shown remarkable long-term stability over the last 150 years and there’s a strong tendency for reversion to the mean over periods as long as 30 to 40 years. This long-term stability is naturally reflected in corporate earnings. There was a great bull market in the the early 20th century that peaked in 1909 and then stock prices went nowhere for the next 15 years. This is not, of course, an original observation. I need hardly mention the experience of the last 15 years.
Whether stock investors were being “rational” in 1929 is another question. As Richard Posner has recently stated in another context, I don’t think people are any smarter or stupider at one period rather than another. I think in 1929 they had less historical data to go on. One thing I am curious about though, and I would be grateful if you could point me to some literature on the subject. Whether or not there was a bubble in 1928-1929, a lot of people, including members of the Federal Reserve, thought there was. Why did they think that? I understand the kind of data people look at today, like PE ratios, but what did they look at then? I think this is important, because if bubbles imply irrationality, it has to be viewed in the context of the knowledge and beliefs people possessed at the time.
21. May 2009 at 07:14
I like just so stories.
But I really can’t comment on this post because I stopped reading when you told me to.
Oh well.
21. May 2009 at 08:40
I haven’t read Leigh’s papers yet but my “model” goes something like this: each agent, intermediary adopts a diversification strategy that provides an expected return with diversifiable risk. Then an event happens which makes this model no longer tenable which changes behavior shifts to an interdependant-risk multiple outcome (e.g., game theory)where your outcomes now depend on what others do. The risk adverse become hoaders and the prportion of risk adverse grows.
The events that make the diversifiable risk strategy unteneble are the usual culprits: recognition of moral hazard, asymetric information/gaming, adverse selection; together with any change that is perceived to change the payoff.
21. May 2009 at 09:15
Buffet doesn’t time the market in the sense of going all in at the beginning of a bubble and going all out at the top of one, but he certainly times the market in the sense of buying when companies are cheap and not doing so when they are expensive. He has a sense of value independent of price that allows him to make better than average investments, not that he doesn’t make mistakes. He is an owner rather than a speculator with a long term perspective with growth plans rather than trading plans. An ownership premium of 40% makes trading at his scale extremely expensive and ordinarily highly unprofitable.
Most of Madoff’s investors were fools; having money is no more measure of sophistication than what one pays for artwork makes for judgment of artistic merit. Many were gullible. Many flattered themselves into believing themselves sophisticated. Selling Madoff’s funds didn’t make them sophisticated, only partial to the fees they could collect. Not all of them were fools. Some had doubts but probably felt they could get enough out to benefit. Yet they were not the most foolish or it would not have survived as long as it did.
There are two methods of profiting from bubbles, for investors it is timing them, and for dealers it is trading them. Timing is always difficult, but trading is always profitable. It is just necessary to avoid investing the profits into the bubble.
An investment book begins with a tale of casino with two rooms. The red room is crowded and exciting with large sums being won and lost but the house always gets their cut. The green room is sparse and quiet and actually rewards investors for playing but it is boring. Most people prefer the excitement of the red room to the dull rewards of the green room.
21. May 2009 at 10:11
“Imagine a model that predicts the optional allocation between index stock funds and Treasury bonds, based on two factors; years to retirement and level of stock market relative to what your model suggests is some sort of “fundamental value.” The percentage allocated to stocks should be positively related to years to retirement and negatively related to the ratio of stock prices and fundamental values.”
This is excellent investment advice. Do you follow it?
Have you seen the Shleifer and Summers EMH paper “The noise trader approach to finance”?
21. May 2009 at 10:56
Scott, you said, “And most importantly, the world’s major central banks suddenly and simultaneously adopted highly contractionary monetary policies. This led to a rapid rise in the world’s gold reserve ratio after October 1929; a highly deflationary policy stance.”
Did this contractionary monetary policy occur before the week of the crash? What sources of data are you using?
21. May 2009 at 16:10
gnat – interesting idea. Sorry if you’re disappointed upon reading my papers: they don’t actually explain the model as such, except for some hints. They are more about policy responses to an environment where asset bubbles are mainly driven by “irrational” individual behaviour (as opposed to herd effects).
Of course herd effects (in which category I’d classify your model – is that fair?) do actually matter – at least when the individual behaviour remains uncorrected. I am developing a more detailed model which will incorporate those to some extent.
21. May 2009 at 17:07
There are lots of interesting comments, but I’m not sure any are addressing the question of whether bubble theories could be accepted by the public. It still seems to me after all the comments that a bubble theory could never be confirmed. If it were accepted, it would immediately stop working. I’m guessing other people don’t see it that way.
Leigh, I looked at your link and was interested in this paragraph:
“But we know from recent experience that this is not always an accurate assumption. Lord Turner, writing the review of financial regulation for the UK’s Financial Services Authority, identified “self-reinforcing irrational exuberance followed by confidence collapse”[1]. Martin Wolf has said that global finance “is prone to wild swings of mood, from euphoria to panic”[2]. Robert Shiller has written on animal spirits[3]. And even arch-neoclassicist Alan Greenspan pointed out the phenomenon of irrational exuberance twelve years before Lord Turner picked up on it.”
This isn’t so much evidence as it is appeal to authority. These are impressive men, but are they right. Both Shiller and Greenspan argued for irrational exuberance in 1996, before the stock bubble. So it seems to me that they were wrong. Prices soared for years after they told people to sell. (I’m so glad I didn’t listen to their advice.) Previously when I have made that claim people have written me back insisting they were right. But this is what puzzles me—how do we test whether a bubble theory is right? Why do we view them as authorities, if their predictions were wrong? Yes, they have also made some good predictions, but so have I. I made a lot of money in Hong Kong. Does that make me an expert? No, it makes me lucky. And I have also made several bad calls.
I don’t know of any reliable way to identify bubbles. I do know ways of identifying bubbles in past data. I know Auguest 1929, 1937, 1987, and 2008 were great times to sell stocks. And I am ingenious enough to create models that “statistically” prove it was a great time to sell stocks on those dates. Is anyone impressed that I can create such models?
So I start from the premise that we lack scientific knowledge about bubbles, certainly we lack enough knowledge to be even considering public policies to address a problem that we don’t even know is a problem.
I don’t see why you assume investors were ignoring the risk or problems in 1929. Perhaps stocks would have been even higher had investors thought there was no risk. Indeed I am pretty sure stocks would have been far, far higher. We’ve all read the numbers about stock returns from 1926-2000, something like 11% a year vs 3-4% for T-bonds (I don’t recall the exact figures.) The only reasons stocks were able to earn such high returns was high perceived risk. Indeed in retrospect it’s amazing stocks weren’t even higher in 1929, as that gap in return seems way to big to explain by the extra risk of stocks. The real puzzle is not high stock prices in 1929 (they weren’t that high) it was the low prices in 1932, which unlike 1929, look much too low in retrospect.
If I knew nothing about the stock market, but knew the macro history of the US, I would have expected a very strong stock market in the 1920s, followed by a collapse in late 1929. My hunch is that they were pricing in a risk of socialism by 1932, a problem that thankfully was avoided.
If I wanted to attack the EMH, I’d use 1987, there was a huge stock movement up and down with no plausible explanation (that I can find) BTW, I seem to recall there was a merger crackdown in that crash as well.
I would also point out that your theory seems different from Nick’s. He talks about regulation. You talk about publicizing the theory so that people behave more rationally in the future. I actually prefer your approach. Not because I am convinced it will work, but because it might work and I can’t see any way it could do harm.
Sorry if that support sounds luke warm, but I am not a fan of bubble models that I have read. I just don’t think millions of people can suddenly become irrational at the same time. Rather I think they all make common errors, because their brains are wired in similar ways, and they are making the same sort of judgment calls in a world of incredible uncertainty.
Alex, A constant bubble? What does that mean? Are prices always too high? I thought bubbles were unsustainable high prices. Has economics even produced a generally accepted definition of bubbles?
21. May 2009 at 17:53
Phil P, Here is a footnote from a paper I wrote:
During the Great Depression Irving Fisher was widely ridiculed for having called the stock market undervalued in October 1929. And studies by DeLong and Schleifer (1991) and Rappoport and White (1993) provide indirect evidence of bubble-like behavior. But McGrattan and Prescott (2004) suggest that “Irving Fisher was right” about stocks being undervalued in 1929 if one accounts for the value of intangible corporate assets.
I strongly believe that stocks were fairly valued in 1929.
If I had a choice between putting stocks or government bonds into a 401k in 1929, and holding for 40 years, I’d pick stocks, even in retrospect. And of course they didn’t expect the Great Depression. The fact that the US economy has been so stable is often misunderstood. Ex post we know that, but ex ante investors didn’t know if we’d turn out like the US, Japan, Russia, Argentina, etc. I think this partly explains the high returns on stocks. It partly reflects the fact that we have selected the most successful economy in world history to study. If you study a company like GE or Microsoft, you will also find a rate of return that seems to violate the EMH.
In 1929 the US economy looked like GE, by 1932 we looked like GM. In the long run the 1929 investors were right and the 1932 investors were wrong, we are the GE of countries. But I can understand either view, given the uncertainty at the time. The 1932 Argentina investors were probably right, as Argentina ended up more like GM.
I actually think investors had pretty good data back then, I think the main problem was that they didn’t know what would come next in terms of macro history. I don’t know why the Fed decided it needed to pop the bubble. It tried and failed over and over again, with tighter and tighter money (for those who think tight money makes stocks go down) and still stocks rose ever higher. Finally they tightened money so much we went into a deep depression. It was the Depression that made stocks crash, and it was perhaps the worst policy mistake in American history (or at least right up there with slavery.) I say this because I am convinced the Great Depression caused WWII. Hitler was a nobody until the German economy went into depression.
Bill, That means you think bubble theories are not supposed to be useful. Here is my question, what do bubble theory proponents think their models will do? What do they see as the payoff? And how do we tell if the models are successful?
Lord, These “fools” aren’t just rich individuals, but also highly sophisticated institutions. I suppose if you define people who lose money in Ponzi schemes as fools, then your theory is a tautology. Or are you claiming something more than a tautology? Do you predict that people with high IQs would be more successful stock pickers than the less intelligent? Or is there some other testable implication? Just curious.
gnat, There are lots of bubble theories, I just don’t know how to evaluate them. Sorry I don’t have much more to offer, but am running late tonight.
123, I haven’t read that paper. I generally find that any economic concept in economics and finance that is interesting can be summed up in a couple paragraphs, at least the intuition behind it. I do occasionally read summaries of bubble research, and am still waiting for that moment when I say “aha”. Maybe my brain is wired in the wrong way, but most of the anti-EMH that I read does nothing for me. It seems like data mining. What’s the gist of their argument?
BTW, does anyone recall a paper that argued almost none of the major daily stock movements since WWII were associated with “news”. When I studied the 1930s, it seemed like most of the big movements were obviously linked to major news stories. Maybe they looked in the wrong place. I want to reread that paper, but can’t recall who wrote it.
No I don’t follow my own investment advice. But I do buy index funds where possible. Unfortunately my major investment is my retirement account, and I am limited on what I can buy. Interestingly my discretionary portion of the 401k did better (so far) than the company portion, which is even more limited. But that’s probably luck–I certainly am not a great investor.
travis, Monetary shocks are extremely hard to identify. It is the same problem as with 2008. All we can say is that policy seemed to get much tighter after October 1929. My hunch is that this change was partly anticipated by the markets, but I certainly cannot explain the timing of the late October crash. Just that a lot of bad news hit the markets starting in roughly late September. Remember that monetary indicators like interest rates are very unreliable. In retrospect the UK decision to raise rates in late September was a turning point, but I think the markets only gradually caught up to that fact, and when they did, the Wicksellian unobservable natural rate started falling far below the policy rates. I use the gold reserve ratio as an indicator, which started rising fast after October. But it is only reported monthly.
21. May 2009 at 20:17
How does one find nuggets of rational behavior without a psychology theory, by the way?
22. May 2009 at 03:19
Leigh
I did subsequently read your papers and references and I think the ideas are complementay. One question is how expectations are formed and how they change. What are the signals for different investor groups. The second question is how do we model the change irrespective of what caused it. There are numerous influences that explain changed expectations. It not be promising to capture them all. I do think that modeling the change from a stable EMH to one of multiple outcomes (a la Nash) is promising.
22. May 2009 at 03:41
Mattyoung, When economists use the term ‘rational expectations’ they are actually referring to ‘consistent expectations’ or behavior that is consistent with the model. The person may be clinically “crazy” but an economist views them as rational as long as their behavior is consistent with expected utility maximizing models. So you look for where people react in the appropriate way (according to economic theory) to new information.
22. May 2009 at 04:28
Scott,
“Alex, A constant bubble? What does that mean? Are prices always too high? I thought bubbles were unsustainable high prices. Has economics even produced a generally accepted definition of bubbles?”
A bubble is not necessarily unsustainable. A bubble is when the value of an asset is greater than the intrinsic value of the asset (present discounted value of the dividends of the asset). So we value the asset not because of the dividends but because we can sell it in the future. What is the intrinsic value of fiat money? Zero. Money is an asset that pays no dividends we just accept money because we think that tomorrow somebody else will accept money, and that person accepts it because she thinks that somebody else will accept it and so on. Look at the paper by Jovanovic and then look at the paper by Sargent that it is cited there you will see that they are basically the same model.
Alex.
22. May 2009 at 05:09
“A bubble is not necessarily unsustainable.”
“Bandwagon” (Title of book by my former colleague Jeff Rohlfs) is externalities on the demand side in which your buying a product increases its value to me. It includes increasing returns due to “network effects” and “complementary effects”. Since multiple outcomes are possible, there can be sustainable and unsustainable bandwagons. In the dotcom era PSINet–a $10B global backbone company–was liquadated overnite. On the other hand, valuations can be pure unsustainable bubbles. The IBM CFO told the story of selling his dog for $100K in exchange for 2 $50K cats to indicate the unrealality of valuations of Internet transactions. “Internet” companies were valued much higher than the “old economy” based largely on hype.
22. May 2009 at 05:34
I do not think the point of creating a bubble theory is to make money for individual investors.
I think it is to get a commonly agreed upon understanding of how the economy works in this area so that government policies can be designed to minimize the costs of bubbles.
I actually think bubbles are quite useful in that they can cause large amounts of investment is new technologies (railroads, the internet). The benefits of the investment often go to society at large rather than the investors.
Bubbles can also clearly result in major costs (our current situation).
Understanding how they occur – rational expectations says they don’t – may allow us to maximize their benefits and minimize their costs.
22. May 2009 at 05:55
For fiat money to be a bubble, then people would want to hold it because it is increasing in value–deflation.
I think the entire stream of research that ignores that money serves as a medium of exchange and that a medium of exchange is valuable is wrongheaded.
It is true that a medium of exchange doesn’t make sense in a representative agent model. That is a problem with that sort of model.
It is true that modeling money like interest or dividend bearing assets is a stretch. That makes those models of doubtful use in analyzing money.
In my view, the “problem” is that some economists have methodological views that require formalism for science. And when something is hard to deal with in a formal fashion, then they grasp at straws.
22. May 2009 at 06:16
Scott,
Thanks for explaining you psychological theory. Yes, people do react is a way that is consistent with their model of how things work.
“(tight money, Smoot-Hawley, and international discord preventing policy coordination) all reared their ugly head in the period from late September to late October 1929. ”
There was a fairly rapid change in long or medium term outlook causing a contradiction between how investors thought things were working and how they really were working. So the economist is left with the question of why? It is not enough to say the coincidence is interesting.
22. May 2009 at 06:43
Scott, thanks for the info.
Here’s the paper about news and stock movements:
http://www.economics.harvard.edu/faculty/cutler/files/What%20Moves%20Stock%20Prices.pdf
“What Moves Stock Prices” by Cutler et al.
22. May 2009 at 07:42
“I haven’t read that paper. What’s the gist of their argument? ”
they say: This paper reviews an alternative to the efficient markets approach that we andothers have recently pursued. Our approach rests on two assumptions. First, someinvestors are not fully rational and their demand for risky assets is affected by theirbeliefs or sentiments that are not fully justified by fundamental news. Second,arbitrage””defined as trading by fully rational investors not subject to such sentiment””is risky and therefore limited. The two assumptions together imply that changes ininvestor sentiment are not fully countered by arbitrageurs and so affect securityreturns. We argue that this approach to financial markets is in many ways superior tothe efficient markets paradigm.
Our case for the noise trader approach is threefold. First, theoretical models withlimited arbitrage are both tractable and more plausible than models with perfectarbitrage. The efficient markets hypothesis obtains only as an extreme case of perfectriskless arbitrage that is unlikely to apply in practice. Second, the investorsentiment/limited arbitrage approach yields a more accurate description of financialmarkets than the efficient markets paradigm. The approach not only explains theavailable anomalies, but also readily explains broad features of financial markets suchas trading volume and actual investment strategies. Third, and most importantly, thisapproach yields new and testable implications about asset prices, some of which havebeen proved to be consistent with the data. It is absolutely not true that introducing adegree of irrationality of some investors into models of financial markets “eliminates all discipline and can explain anything.”
22. May 2009 at 08:53
Bill,
That is not true. A bubble does not require the price of the asset to grow over time. A bubble forms when the value of a asset is greater than the intrinsic value of the asset (or at least that is the definition I think most economists would agree on). Fiat money has no intrinsic value hence it is a bubble. Look at the similarities:
* We accept dollars in exchange for goods worth more than what he paper the dollars are printed because we think that somebody will accept them in the future.
* We buy a house today at a high price because we think that somebody will buy it in the future.
In both cases the future trade partner accepts the dollars or the house because he/she thinks just like us.
Usually monetary models suffer from indeterminacy of the price level (i.e. there are multiple equilibrium). Jumping from one equilibrium with high value of money to one with low value of money can be thought as a bubble bursting. This is usually what happens when there is a run against a currency.
Alex.
22. May 2009 at 09:29
I doubt most bubbles could be predicted but that is unnecessarily stringent. A theory might still be useful even if it only made predictions after one burst. One might not be able to avoid them but might still be able to address them after the fact. I don’t even see bubbles as something to be avoided at all costs. They may well play an important part in creative destruction. Credit bubbles are a different matter though. They are both relatively easy to see and prevent by requiring lending be limited to retrospective rather than prospective views and based on income as much as asset values. All losses aren’t preventable but systemic losses are. Such a bubble could continue to grow but only through equity rather than debt. Debt makes bubbles far worse because of its systemic nature results in far larger bubbles, its rigidity makes it difficult to deal with, and its liquidation can lead to deflation.
22. May 2009 at 10:13
Alex, Money does have value as a medium of exchange. It’s sort of like motor oil, it makes the other parts work better. It let’s you save time, you don’t have to barter goods.
gnat, “Hype” is kind of vague. Surely investors knew these stocks were highly speculative. Perhaps some dishonest reports were put out, but I think the main problem was that even insiders overestimated how fast the internet was likely to grow. I’m not saying there was no irrationality, but I don’t know how to measure how much. My hunch is that even lots of well-informed people within the tech industry lost money, people who did know what was going on.
dk, If a theory cannot predict well enough to make money, why would policymakers trust the theory when making policy?
Bubbles don’t “clearly result in major costs.” In this blog I have argued that our current recession is due to tight money, not any bursting of bubbles.
Bill, I agree, but would even go further. Suppose the nominal rate is 5%, and the base is $800b. Then there must be a flow of services of at least $40 billion per year, as that is the opportunity cost of holding all of those reserves. The capitalized value of $40 billion per year at a 5% rate is $800 billion. So cash has significant real value to society.
Mattyoung, If I understand your comment you are asking what caused those shocks. That is a complex question that involves political science, etc. There are limits to what economists can do in this area. I focus on issues where the view of causation is useful, primarily monetary policy. If we can show that tight money caused the Depression, it has very useful implications for current monetary policymakers. Of course the obvious retort is that Bernanke studied the Depression, and look where we are now! But I suppose things could be even worse.
Travis, Thanks. The paper is in my view 180 degrees off course. They look at a bunch of market movements, then try to “explain” them with a regression model, and find that they can only explain about 20%. But why is that a surprise? First of all the regression model doesn’t really show which parts of the data are news, but more importantly stocks aren’t going to move on shifts in industrial production, they are going to move on expected future shifts in industrial production. This means they will move on news of shocks that might be expected to cause future changes in industrial production. And of course industrial production is far from the only factor that moves stock prices. and I could go on and on. They are looking at the interwar period and don’t even include the international gold market in their model.
The bottom line is that just because one can’t explain why stock prices changed, does not mean the stock price movement was irrational. After looking at that paper again, I feel much better about my defense of the EMH.
I can’t believe that three distinguished professors think that the EMH predicts that you ought to be able to explain market movements by scanning news headlines. In my Depression study I found that the important shocks were not always in the headlines (although some were, as in the CPS paper.)
Failure to explain =/= irrational
123, Thanks for the summary. They said:
“First, theoretical models with limited arbitrage are both tractable and more plausible than models with perfectarbitrage. The efficient markets hypothesis obtains only as an extreme case of perfectriskless arbitrage that is unlikely to apply in practice. Second, the investorsentiment/limited arbitrage approach yields a more accurate description of financialmarkets than the efficient markets paradigm. The approach not only explains theavailable anomalies, but also readily explains broad features of financial markets suchas trading volume and actual investment strategies. Third, and most importantly, thisapproach yields new and testable implications about asset prices, some of which havebeen proved to be consistent with the data. It is absolutely not true that introducing adegree of irrationality of some investors into models of financial markets “eliminates all discipline and can explain anything.”
I put no weight on the first point at all. It’s like saying perfect comptition doesn’t exist in the real world. Yes, but it’s still a very useful model.
Regarding the second point, I’d like to know more about the “available anomalies.” I had heard that many anomalies went away after the research was published. What does that tell you? Does this provide a strategy that will allow us to beat the market, with a high degree of confidence? If not, why not? What does the trading volume refer to? All I know is that trading volume goes up during times of uncertainty, exactly what you would predict from the EMH. So what anomaly are they referring to there?
The third point says that the model has testable implications that are consistent with the data. I wonder if they used the same data set to test the model, as they used to develop the model? Just asking.
Alex#2, I don’t know what you mean by “intrinsic value.” Most people define value in terms of consumer willingness to pay (the diamond-water paradox.) If people normally hold lots of cash despite the high opportunity cost in terms of foregone interest, I’d say it has great value to them.
Lord, You are making a very different argument from most other bubble experts. People like Shiller definitely do think they can be predicted to some extent, or perhaps I should say identified before they burst. So many bubble researchers think public policy can limit the size of the bubble before bursting. Your argument seems more reasonable to me, but I’d like to see the specifics of what should be done afterward. You know my view of monetary policy and its role.
22. May 2009 at 11:22
Scott,
It is true. But money only has value because somebody else will take the money from you and that somebody will take it because somebody else will take it from him and so on. If some day all of a sudden we decide that we don’t like dollars as a medium of exchange any more then the dollars are worthless (some peolpe argue that the government issuing the currency will always accept their own currency as a way to pay taxes so this pins down the value of money in the end, see the Fiscal Theory of the Price Level). By intrinsic value I mean the consumption value of the good, in the case of an apple it would be the marginal utility of an apple divided by the marginal utility of income. If it is an asset, like a claim to an apple today and one apple tomorrow then it would be the same as before but in the present value form, i.e. the value of one apple today plus the discounted value of an apple tomorrow. Fiat Money has no intrinsic value yet it has market, commodity money has intrinsic value which puts a lower bound on the market value of the good. Maybe if I can’t convience you that money is a bubble good I can convience you that bubble goods are money. Bubble goods are bought not to be consumed (or at least some of them are). Bubble goods are bought (on the margin) just to be sold. Money is bought (with goods and services) not to be consumed but to be sold (in exchange of goods and services). Just go back to the introduction of money into an OLG economy. You will see that there is always an equilibrium in which the value of money is zero and (under certain parameter values) there is at least one equilibrium in which money has positive value (the value of money is the inverse of the price level) and in fact there are usually many other equilibriums in which the price level does many different things. Please check this references
A Model of Commodity Money (Sargent& Wallace)
http://minneapolisfed.org/research/sr/sr85.pdf
(there is a publised version but it is restricted)
Bubbles in Prices of Exhaustible Resources (Jovanovic)
http://www.econ.nyu.edu/user/jovanovi/bubbles4.pdf
Of course I’m only looking at “rational” bubbles (i.e. bubbles with rational forward looking agents) and I’m sure that there are other models of irrational bubbles but this is what I know of.
Alex.
22. May 2009 at 11:27
Scott, the main idea of the paper is limits to arbitrage (including limits of ability to profit from identification of bubbles). The whole paper is well worth the read, and it resembles more a long blog post.
The available anomaly that they explain is the closed-end fund discount, this anomaly has not yet disappeared.
Their model is not based on a data set.
22. May 2009 at 11:43
Scott, you said, “In my Depression study I found that the important shocks were not always in the headlines (although some were, as in the CPS paper.)”
Is this study in your manuscript or have you published parts of it? If so, could you direct me to the relevant papers? I am very interested in your analysis.
Regarding the CPS paper, I was actually surprised they were able to explain as much as they did given the variables involved.
I am confused by Table 2 of CPS. How does column one of Table 2 differ from Table 1? How did they get different R^2? In Table 1, adding additional lags didn’t add to R^2, but in Table 2, R^2 increased with more lags. I assume they are using the same adjusted R^2. Puzzling.
22. May 2009 at 12:42
Scott, I think that a ‘true’ bubble theory wouldn’t necessarily be discredited even if it were rapidly priced into asset prices. Suppose that the bubble theory said that P/E ratios should never rise above 20 and offered proof. Once integrated by the markets, p/e ratios would never rise above 20. The bubble theory wouldn’t be discredited because the necessary antecedent conditions would never exist after publication.
22. May 2009 at 15:58
Alex, You said:
“If some day all of a sudden we decide that we don’t like dollars as a medium of exchange any more then the dollars are worthless”
Isn’t this also true of impressionistic paintings? What if we suddenly decide we want more “realistic” paintings on our walls?. And what if we decide we don’t like toast any more, are toasters then worthless? I still think money has real value as a medium of exchange. So much so that if the government collapsed, I think it is possible people would keep using fiat money. I suppose it would be hard to test this. Money is also a weird “network good” isn’t it?
I have never been convinced by the Sargent/Wallace approach to money.
123, Interestingly, I’ve bought closed-end funds. And I once profited when Harvard bought one up and opened it up. So arbitrage isn’t impossible–indeed if (like Harvard) you have 30 billion it’s easy to spot these inefficiencies and profit from them. But I also seem to recall there was a principle-agent problem. The closed end fund fought hard against Harvard. I do agree that closed end funds seem to violate the EMH. Indeed it seems to me that managed funds in general violate the EMH. Isn’t this one area where the anti-EMH and pro-EMH positions have identical implications? You can use the existence of managed funds to show that people aren’t rational. The implication of that observation is that people should buy indexed funds. But that is also the implication of the EMH itself.
If you recall from my earlier EMH post, I actually don’t have any sort of religious faith that the EMH is right. What I have is a visceral distrust of people who tell me that investors are fools and all you have to do to make a lot of money is buy when P/Es are low and sell when high, or some other similar scheme. So I sort of play the devil’s advocate, trying to convince people to be less sure of their own views of the market than they seem to be.
I’ll try to look at the paper sometime, but right now I am about to cut back on the blog, as I am way behind in other work.
I don’t see how a model can not be based on a data set. it seems to me that all macro/finance models are based on data sets. How could it be otherwise? Maybe we are defining terms differently, but if you build an anti-EMH model, you can’t help having observed certain anomalies before building the model, and you naturally try to build the model to explain those anomalies. I’m not saying there is anything wrong with that, but it simply means that the models can’t truly be tested until you can come up with a new data set.
And it definitely means you can’t use classical statistical “confidence intervals.”
You recall hearing about one in a quadrillion events occurring in financial markets (1987, 2008, etc)? That’s what happens when you put models to work in different data sets–they don’t predict very well, because they were constructed through data mining, not economic theory.
travis, Sorry, I spent about 10 minutes on the table and couldn’t figure it out either. Could you email me at Bentley and let me know what kind of project you are working on (and if you are an academic.) I’ll definitely send you something useful, but it would help to know the project (what years, etc.)
travis, You might be right about the 20 P/E ratio theory. it would depend on the specifics of the theory. You might have a bubble theory that said buy when X occurs, but that no longer workers because X occurs less often and only when prices are low for very good reasons. But your point is a good one. It also raises a question in a different direction. There has been a lot of finance research since 1980, and I think the stock market seems crazier than ever. So it seems like investors are just brushing of the Shillers of the world. I don’t know if that’s good or bad, but it sure is interesting. BTW, Shiller’s ideas seem very plausible to me from one perspective. From another perspective they seem to make it too easy to determine when to buy and sell. So I am still an agnostic.
22. May 2009 at 21:51
Scott,
“Isn’t this also true of impressionistic paintings? What if we suddenly decide we want more “realistic” paintings on our walls?. ”
* Look at the paper by Jovanovic. There he is trying to show that Vintage wines behave like bubble goods.
“And what if we decide we don’t like toast any more, are toasters then worthless?”
If preferences change then prices will be affected. I can affect the value of currency in a model without affecting preferences but adjusting expectations about future values of the currency.
“I still think money has real value as a medium of exchange. So much so that if the government collapsed, I think it is possible people would keep using fiat money.”
True, but that value depends on the belief that fiat money will be used and accepted in the future, i.e. it will have value in the future. Just like the high price of an asset today depends on the belief that the asset price will be high tomorrow.
“Money is also a weird “network good” isn’t it?”
What is your definition of an asset bubble? “Network” goods are exactly the goods that are prone to be subjected to a bubble. To have a bubble we need to the value of the good to be determined not by its “consumption” value but by the value of what we think we can sell it to others. People paid millions for a house in CA not because they enojyed living there but because they thought they would be able to sell to somebody else just like I accept dollars not because I like to look at the faces of American Presidents but because I think I will be able to give the dollars to somebody else in the future in exchange for something else. The same thing can happen to paintings, wines and big stones in the pacific.
Alex.
23. May 2009 at 05:07
Alex, If that is your definition of bubble, that’s fine, but I don’t think it is the definition that most people use. There is a big difference (in my view) between sudden changes in the net present value of returns from common stock, and a change in the perception of how others will value the network good in question. Most people define bubble as market irrationality. If the people you mention want to come up with another definition, that’s fine. BTW, I don’t think bubble theory helps us understand monetary economics, because as a practical matter people do not expect others to suddenly lose interest in using Federal Reserve notes as money. So I don’t see any practical applications to this view.
23. May 2009 at 07:10
Scott,
I don’t think you are right. There is a definition of a bubble which is independent of the rationality or irrationality of agents. That bubbles are associated with market irrationality is another issue. I would say that most people will argue that bubbles are the result of irrational behavior, other people argue that bubbles can happen even with rational agents, while others will argue that what we think are bubbles are not bubbles.
I don’t think you are right about the relationship between monetary economics and bubbles either. First of all money is an asset and as such it is subject to the same principles as all other assets. Second, it might be the case that the Federal Reserve notes (dollars) have not exhibited busts, but there are many other currency which have been subject to runs (sometimes succesfull and sometimes unsuccesfully). Here is a references, you might not like the quality of the research or think that it is not important. But there is little I can do about that,
Are currency crises self-fulfilling? The case of Argentina
http://bura.brunel.ac.uk/bitstream/2438/927/1/02-26.pdf
Are the perceptions so different? I sell pesos and buy dollars because eveybody else is and the currency is depreciating. I buy Google because everybody else is and the stock is skyrocketing. I buy a house in CA because everybody else is and the housing prices are soaring.
I understand that my views on monetary economics were shaped by different events than you. You probably think about the US in the 30’s when you think about monetary policy examples. I think about the 80’s in Argentina. You think about how to avoid small deflations I think of avoiding hyperinflations which means you don’t really have to worry about sudden jumps in the value of the currency.
Alex.
23. May 2009 at 10:13
Scott-
Devin, I don’t see how your explanation responds to my point. You told a story of what might have happened. I could tell a hundred different stories, all equally plausible. What is the purpose of these stories?
The point is that the risk of a currency run was priced into asset prices. Ordinarily, stocks should trade at a price where the expected future cash flow from dividends is equivalent to the cash flow you’d receive from investing in bonds of equivalent risk (something like BAA bonds). If prices are too high and dividend yields too low, smart investors will sell their stocks and buy bonds.
But if there is a risk of currency run, or if the rate of monetary inflation is greater than the rate of interest, then there is no floor on dividend yields. In a currency run, bonds would become worthless, dollars would be worthless, and only real assets like stocks or housing would maintain their value.
In 2007, dividend yields were ridiculously low. I think part of the reason for this is that stock prices were reflecting the risk of a currency run/hyperinflation scenario. When the risk of a currency run dropped off dramatically, stock prices plummeted, and dividend yields returned to more reasonable levels.
If it is an investment guide, then won’t it be worthless immediately after it is known to the public?
If everyone had my theory of bubbles there would very quickly be a currency run and gold would spontaneously remonetize.
If policy makers had my theory of bubbles, the wouldn’t let monetary inflation grow at freaking 8% a year. They would keep monetary inflation well under the interest rate for CD’s, preferably at zero percent. At zero percent monetary inflation I think bubbles would mostly disappear. It would become obvious that stocks should be priced based on dividend yields.
On your second point, I don’t agree that successful investors time bubbles. Buffett doesn’t, and he’s pretty successful. If Buffett can’t figure out the right time to sell, how can you or I?
Let me clarify. The well known, famous investors got rich by using strategies which are suboptimal in the average case but great in the best case. One example of this is timing bubbles. Another example is a 100% equities strategy where you pick stocks. Another example is being first to a popularize a certain strategy, and then having other people follow your lead.
Obviously, by definition, the average investor cannot time bubbles. If the average investor was a genius, bubbles would only exist in a situations where there is no Nash equilibrium price.
BTW, Shiller’s ideas seem very plausible to me from one perspective. From another perspective they seem to make it too easy to determine when to buy and sell.
Shiller’s theory is only partly correct. First, it’s not earnings that matters, but cash returned to the shareholder – ie dividends and stock buybacks (in practice, stock buybacks usually just cancel out dilution, so dividends are the thing to watch). Second, you have to compare the cash flow to alternative investments. If government subsidies drive down bond yields, dividend yields will go down too. Third, it’s not just current cash flow, but all future cash flow. So if monetary inflation increases, stocks will trade at a lower dividend yield, because investors know that the dividend will increase overtime. At a rate of 5% monetary inflation, with BAA bonds trading at 8% interest rate, stocks should trade at around 4% dividend yield. If monetary inflation drops to 0%, stocks should trade at 8% dividend yield. If monetary inflation is 10%, and bonds trade at 8% yield, then, well, there is no Nash equilibrium price for stocks. Hello bubble!
23. May 2009 at 10:16
Alex-
Money is the bubble that never pops. This is true of fiat money or commodity money. Most demand for gold is monetary demand. The floor price of gold is irrelevant when the floor price would be about 95% lower than its current price.
23. May 2009 at 12:25
Alex, Money isn’t just another asset. It’s the only asset with a fixed nominal price. That means the only way for the value of money to change is for all other prices to change. If not for that attribute, we wouldn’t even have a field called “monetary economics” it would be too unimportant. I suppose you could say it is like other assets in the sense there there is a supply and demand for money.
You don’t need bubble theories to explain Argentina’s hyperinflation in the 1980s, the simple supply and demand theory will do fine.
Devin, The purpose of the post wasn’t to examine bubble theories, but to consider what bubble theories are for; what’s the purpose of developing a bubble theory? My hunch was that the only plausible purpose was prediction, and I see no evidence they predict. But if they did, I still probably wouldn’t know about it, because anyone smart enough to think up a bubble theory that predicts, isn’t stupid enough to tell the rest of us, and destroy the profit opportunity it provides.
23. May 2009 at 15:26
Scott: “.. Money isn’t just another asset. It’s the only asset with a fixed nominal price. That means the only way for the value of money to change is for all other prices to change. If not for that attribute, we wouldn’t even have a field called “monetary economics” it would be too unimportant. I suppose you could say it is like other assets in the sense there there is a supply and demand for money.”
Off topic, but that, in a nutshell, is precisely where you and I disagree. Let me Fisk you, if I may 😉
“Money isn’t just another asset” Agreed!
“It’s the only asset with a fixed nominal price.” Yes, that’s ONE of the ways it differs from all the other assets, and it’s important. But what you should ALSO have said is: “It’s the only one that is a medium of exchange”.
“If not for that attribute, we wouldn’t even have a field called “monetary economics” it would be too unimportant.” No. Even if money were not the medium of account, but were the sole medium of exchange, it would still be very special, and very important, and we would still need a field of monetary economics.
“I suppose you could say it is like other assets in the sense there there is a supply and demand for money.” No, strictly speaking, there is no identifiable demand and supply of money. Because there is no unique market for money. Every other good, in a monetary exchange economy, has a market in which it is exchanged for money. And we can speak of the demand and supply of that good in that single market. But money has no market, It appears on one side of every market. If there are n goods, plus money, there are n different definitions of the demand for money.
Sorry. I couldn’t resist. But that short passage illustrated so perfectly our differences. You are a medium of account guy; I’m a medium of exchange guy. Slowly, but surely, I will convince you!
23. May 2009 at 15:43
Another relevant Vox article (not by me this time) was published today: http://www.voxeu.org/index.php?q=node/3598
It’s by Gareth Campbell of QUB and looks at the British railway stock bubble of the 1840s, and concludes that…it wasn’t a bubble after all. Probably similar to Scott’s conclusions about the 1920s?
Incidentally footnote 2 contains two definitions (popular and “economic” definitions) of “bubble”.
23. May 2009 at 16:11
I agree with Nick.
Nick, would you agree with the following paragraph: The medium of exchange is a good that stores value in between transactions, thus allowing indirect exchange. All demand for a medium of exchange is demand to hoard the good. Sometimes people only wish to hoard for a short time and they immediately trade their pay for food. But people also wish to hoard for long time periods, perhaps as a buffer in case of emergencies or job loss, or perhaps in anticipation of a future increase in consumption.
Also, I’d be interested in hearing your answer to a question I just posed to Scott. In the American economy, which actual good represents the medium of exchange? Dollar bills and ledger entries at the Federal Reserve ( MO)? Bank notes (M1)? All zero maturity deposits(MZM)? All zero maturity deposits plus treasuries?
23. May 2009 at 17:44
Good questions, but I’ll have to pick up these tomorrow.
24. May 2009 at 02:23
Devin: yes, I think I would roughly agree with your paragraph. I might put it a little differently.
Suppose there are n goods (including the good that may or may not be used as money).
In a barter economy, there are (n/2)(n-1) markets, so any of the n goods can be swapped against any of the (n-1) other goods. (The “n/2” appears in the formula because if there is a market in which apples can be swapped for bananas, it’s also a market in which bananas can be swapped for apples).
In a monetary exchange economy, with a single medium of exchange, there are n-1 markets. One good (money) can be swapped in every market. The other n-1 goods can only be swapped in one market.
So in a monetary exchange economy, we can identify the excess demand for each of the n-1 non-money goods, because it is only swapped in one market. But there are n-1 different excess demands for money.
If the relative price of apples and bananas is right, but the price of money is wrong, what happens? In a barter economy, people can still swap apples and bananas, so apple producers who want bananas, and banana producers who want apples, can still trade. But in a monetary exchange economy, the apple and banana markets get screwed up as well. The apple producers won’t buy bananas until the banana producers buy their apples, and the banana producers won’t buy apples until the apple producers buy their bananas. Gridlock in all markets, even though all relative prices are right, because the price of money is wrong.
What’s the closest measure of medium of exchange in the US? Dunno. M1 I think. It’s because commercial banks can directly cancel offsetting obligations in a central clearing house for cheques, and only have to hand over reserves for the difference.
You can’t buy apples with treasuries. You first sell the treasuries for money, then sell the money for apples. Treasuries are like apples. They are not the medium of exchange.
24. May 2009 at 05:36
Nick, I don’t have any strong presumption that the medium of exchange role is unimportant, rather I have a strong presumption that the medium of account role is important. That comes from observing how changes in the value of money cause all sorts of relative price changes that destabilize economic activity. In my study of the Depression I found that easy money raises real stock and commodity prices, and also reduces real wages. These relative price changes seemed to explain much of the concurrent fluctuations in output. So it’s more an Occam’s razor sort of thing.
I’m not used to thinking in terms of models with complete short run price rigidity, but for what its worth here’s what I would say about money supply and demand:
The Ms and Md curves are on a graph with nominal base money on the horizontal axis, and the value of money (1/P) on the vertical. Under a fiat regime the supply of money is totally inelastic, and is controlled by the central bank. The demand curve is a rectangular hyperbola. When the supply of money shifts, the price level moves proportionately in the long run. If prices are rigid the nominal interest rate changes, causing an equal shift in money demand and demand in the short run. I suppose in an economy without interest rates you would simply have a shortage of money if the supply fell.
If I am not mistaken, in micro we assume the price of apples is not measured in nominal terms, but in terms of all other goods, or the real price of apples is on the vertical axis. The big difference is that in the money market, the price of money is fixed in the short run.
I am a bit puzzled by the apples for bananas example. There are two possibilities. One is that the relative price of apples and bananas change during the monetary disequilibrium. In that case, barter doesn’t solve the problem, as we are assuming nominal price rigidity in terms of money, and that implies relative price rigidity of apples and bananas. So barter is impossible.
I imagine you had in mind a situation where no relative prices change, other than the equilibrium relative price of money (or course the market relative price of money is unchanged, due to nominal rigidity.) Suppose there is 5% too little money. There is still enough money to buy lots of apples and bananas, just not to completely equilibrate the markets. Those markets have excess supply exactly equal to the 5% shortfall of money (but we also have to convert stocks to flows with some velocity concept). And that is equally true under barter or a money medium of exchange. Under barter, the medium of account is just another good with a price that happens to lead to excess demand, putting all other markets in excess supply. So I don’t see the difference. But there may well be a difference, as I am not used to working with this sort of model.
Leigh, Thanks, I agree with Campbell. I was interested in Campbell’s assertion that some economists favor having the Fed try to pop bubbles. They did try that in 1929, and it “worked.” I agree with Bernanke that it is hard to know when market prices are at the wrong level. I would add that if the Fed thinks it knows this, I’d rather see them invest taxpayer money in the market, and use the profits to pay down the national debt. That intervention would help stabilize the market in question, without screwing up monetary policy, a la 1929.
Devin, Off the top of my head I’d say there are two ways of thinking about media of exchange. In a sense, the base is the only real medium of exchange, as you almost always transfer base money in any transaction. (If you give shares of common stock to your neighbor for his boat, I’d consider that barter, as the nominal price of stock isn’t fixed.) However, this transfer of base money can be delayed with checks, credit cards, debit cards, etc. So I suppose those are in a broader sense media of exchange. It’s not my area of expertise. I also think these other media could be viewed as instruments that reduce the demand for base money, that economize on the need to hold cash.
Nick#2, I see you also had a somewhat vague answer to Devin, but I would have thought you’d have a really strong view, as you argue the medium of exchange is important. As for the medium of account, it is now clearly the base, but under the gold standard there were arguably two media of account–gold and cash, with a fixed rate of exchange. So that made things pretty complicated. And don’t even ask me about bi-metallism.
24. May 2009 at 14:56
Nick and Scott-
What’s the closest measure of medium of exchange in the US? Dunno. M1 I think. It’s because commercial banks can directly cancel offsetting obligations in a central clearing house for cheques, and only have to hand over reserves for the difference. You can’t buy apples with treasuries. You first sell the treasuries for money, then sell the money for apples. Treasuries are like apples. They are not the medium of exchange.
First, I’d note that figuring out an answer to this question is a perquisite to understanding modern monetary economics. I mean, you kind of have to be able to define money, before you can analyze it! The modern economy is far too complex to understand inductively. The only hope of understanding it is too come up with a simplified model of a monetary economy, deductively understand how it works, and then map the result to the real thing. But doing so requires understanding which actual good represents “money” – ie the medium of exchange.
Second, I’d remind you both that the medium of indirect exchange is not just the good that people use to conduct transactions. It’s the good that provides no direct utility, but is used to store value in between transactions. So if people stored dollars in their vaults, but out of convenience always converted their dollars to dimes to actually perform transactions, dollars would be considered part of the medium of exchange.
If we say that base money or M1 is the medium of exchange, we have to deal with a number of oddities.
#1 There is on the order of $30-$40 trillion in dollar denominated debt world wide. Yet both BASE and M1 are around $1.6 trillion. This is an extraordinarily high ratio of debt to actual quantity of the medium of exchange. A country like China simply purchased dollars directly, instead of hoarding T-bills, it could corner the market on dollars! It would own the world. Practically this wouldn’t work, but it gives you an idea of how out of whack the ratio is.
#2 The ratio of Treasury debt to actual money is close 7 to 1. Not only that, USG is bleeding red. Not only that, it has been bleeding red almost non-stop for four decades, and it shows no prospect of ever actually starting to pay down the debt. Despite this, almost every investor and financial institution considers Treasuries completely riskless with zero default risk. Again, puzzling.
#3 The typical American individual or company has very small hoards of the medium of exchange. $1.6 trillion comes out to about $5,000 per person, compared to a per capita income of $35k. That strikes me as a very low ratio. That means the average company and person has about one month of savings total in the form of the medium of exchange. Of course, the average person or company does not perceive it this way. The person also holds savings accounts, CD’s, and treasuries, all which are considered liquid, zero risk paper.
These puzzles all stem from one cause: most people studying monetary economics have not fully realized the implications of a fiat currency.
A dollar is a piece of paper backed by the full faith and credit of USG. USG has a functioning printing press, and it can print any number of them it wants. Any bank account that is FDIC insured is also essentially a piece of paper backed by the full faith and credit of the USG. So are treasuries.
The trouble is that there is no equivalent in a gold standard economy of a CD or Bond issued by the currency creator. Thus economists have modeled Treasuries as debt, when really, they should be considered just another form of the medium of exchange, like dimes.
Under a gold standard, if you want to hoard the medium of exchange for a long time, you put gold in a vault. But under a fiat standard, if you want to hoard the medium of exchange for more than one month, you are always better off holding treasuries or CD’s. Thus Treasuries, CD’s, Savings accounts, anything backed by the full faith and credit of USG should be considered part of the medium of exchange.
Another way to look at it is to imagine USG just creates a bunch of T-bills and hands them out to people. The people who receive the T-bills will have a permanent improvement to their balance sheet. As a result, they will be able to spend more. This spending will increase NGDP and inflation. So again, from a monetary standpoint, T-bills matter.
Imagine an economy was under a 100% gold standard, and gold was the medium of exchange. Many people would hold gold in vaults for long term storage. Now imagine someone invented a magic “gold incubator”. You would put your gold in it, seal it tight, and when you opened it months later, the amount of gold would have magically increased by a couple percent. But if you opened it to early, the gold would get corroded and you would lose it.
In this economy, you would see a split. Everyone who was holding gold for short term use would continue to hold gold. But everyone holding for the long term would put their gold in gold incubators. Now, would the gold put in the incubators suddenly not be called part of the money supply? Would it be considered not the medium of exchange? I argue that it still is every bit the medium of exchange as before, its simply taking a slightly different form to maximize yield.
T-Bills are like the gold incubators. Except, because it is a fiat currency, USG never actually has to print the dollars to put inside the incubator. It simply has to make a creditable promise and back the bills with its full faith and credit. And since there is a lot of demand for long term savings, most people just roll over their T-bills and USG never actually has to print the dollars.
So in summary, in a fiat currency, the medium of exchange consists of the actual currency, and all goods that the currency issuer has guaranteed to be redeemable in the currency. In the U.S. that means M2 plus CD’s plus T-Bills. It probably also includes money market funds, now that the Fed has introduced the AMLF and MMIFF. The difference between T-bills and dollars is about as important as the difference between dimes and dollar bills. It’s a difference of convenience, not of monetary economics.
24. May 2009 at 16:09
Devin: I can define “medium of exchange” theoretically. But defining it empirically (which particular empirical measure fits that theoretical definition best) is not always so easy. Is “GDP” always a good empirical measure of “output”? No. But it’s one we often work with, faut de mieux. Does “employment” as defined by US bureau of Statistics (is that your equivalent of Statistics Canada?) mean precisely what you want it to mean theoretically? Probably not. Ditto M1. But we still talk about output, employment, and the medium of exchange.
“Second, I’d remind you both that the medium of indirect exchange is not just the good that people use to conduct transactions. It’s the good that provides no direct utility, but is used to store value in between transactions.”
I disagree. Sometimes the good that is used as a medium of exchange does give utility. Cigarettes, cows. It happens.
“So if people stored dollars in their vaults, but out of convenience always converted their dollars to dimes to actually perform transactions, dollars would be considered part of the medium of exchange.”
I disagree. I would then say that dimes are the medium of exchange, and dollars a store of value. It’s why I would exclude broader monetary aggregates from medium of exchange. If you can’t write cheques on your savings account, or use a debit card on it, I wouldn’t include it.
“A country like China simply purchased dollars directly, instead of hoarding T-bills, it could corner the market on dollars! It would own the world. Practically this wouldn’t work, but it gives you an idea of how out of whack the ratio is.”
But the reason it wouldn’t work is that dollars are a medium of exchange, and Tbills aren’t. If China tried to swap (sell) all its Tbills for dollars, other people would not be indifferent to swapping their dollars for Tbills. Because they need those dollars as a medium of exchange. Tbills wouldn’t do the job. If China tried it, the price of Tbills against dollars would drop without limit (unless the Fed printed more dollars, which it would).
I do like your “gold incubator” thought experiment. It’s rather like paying interest on currency, except you get no interest unless you can prove you have kept the currency in your possession for one month. It’s like buying a non-tradeable GIC, rather than a one month Tbill.
24. May 2009 at 17:10
But defining it empirically (which particular empirical measure fits that theoretical definition best) is not always so easy.
Right. But you at least need to have a definition that’s close, and you need to know exactly how the various aggregates map to the original model.
BTW, does it strike you as weird that it is so hard to define the money supply? I mean, the dollar is a fiat currency. The Federal mints go to great lengths to prevent counterfeiting. Yet it cannot actually keep track of what the money supply actually is.
I disagree. I would then say that dimes are the medium of exchange, and dollars a store of value.
No, both dimes and dollars would be the medium of exchange. Dimes would simply be the denomination used for transactions.
And I don’t get why suddenly you are making a distinction between the medium of exchange and the store of value. Before you seemed to agree with me that they were the same thing. In order for a good to be used for indirect exchange, it must store value in the time period between transaction A and transaction B. If I temporarily convert the good to a different form for storage, that hasn’t changed the nature of the good, as long as the convertibility is guaranteed.
To make the issue even more clear, imagine a country under a gold standard. Most people hold gold bars in their vaults, but they use small coins for transactions. Free coinage is completely allowed. Which is the medium of exchange? Again, I’d argue both. In general, if there is near perfect convertibility between object A and object B at a fixed ratio, then from a monetary standpoint those goods are the same thing.
But the reason it wouldn’t work is that dollars are a medium of exchange, and Tbills aren’t
The Fed would always print more dollars to prevent the price of the T-Bill from dropping below $1. There is guaranteed convertibility, so T-bills should be considered part of the medium of exchange. It would be like China only buying up gold coins but not gold bars in my previous example. People would simply coin more gold pieces.
25. May 2009 at 03:38
Devin:
“BTW, does it strike you as weird that it is so hard to define the money supply? I mean, the dollar is a fiat currency. The Federal mints go to great lengths to prevent counterfeiting. Yet it cannot actually keep track of what the money supply actually is.”
Yes, it is a bit weird. Was it Leland Yeager’s analogy? A coat keep you warm and dry, but a promise to pay one coat does not function as a coat; it does not keep you warm and dry. A dollar bill functions as a medium of exchange, but a promise to pay a dollar can (sometimes) also act as a medium of exchange (chequable demand deposits at banks, for example).
It’s as if someone allowed counterfeit CD’s, provided they were clearly marked as counterfeit, and did not have the same label as the genuine article, even if they worked just as well in playing music.
This is the view that leads to calls for 100% reserves (you are allowed to make one cassette copy of each CD you own, provided you do not play the original).
If the central bank were profit maximising, this would never be allowed. But they aren’t, and have decided to allow copying, provided copies are clearly marked as such, and in a different format. I think this is socially efficient, even if it’s not profit-maximising.
“And I don’t get why suddenly you are making a distinction between the medium of exchange and the store of value. Before you seemed to agree with me that they were the same thing. In order for a good to be used for indirect exchange, it must store value in the time period between transaction A and transaction B.”
Being a store of value is a necessary but not a sufficient condition for a good’s being a medium of exchange. All media of exchange are stores of value, but not all stores of value are media of exchange (most aren’t).
25. May 2009 at 08:20
Nick-
I don’t want to get an argument over semantics. If you define medium of exchange to only mean the denomination that is actually used in transactions, then you need a new word to describe “a medium of exchange and all stores of value linked by fixed and creditable convertibility ratios”. Frankly, I thought the phrase “medium of exchange” already meant that. But whatever that is, is the subject of monetary policy.
Let me make the question a bit more applied: Let’s say that the Fed decides that in order to get out of the recession, it wants NGDP for the next year to be 10% greater than what it would otherwise be. It wants to trigger this increase by a one time, supply side increase in the money supply. The Federal Funds rate is already at 1%, and for various reasons the Fed intends to keep it there. The Fed actually wants to print the money. At a first approximation, this would require a 10% increase in the money supply. What aggregate should the Fed use? Approximately how much money should the Fed print? How should the money get injected?
But they aren’t, and have decided to allow copying, provided copies are clearly marked as such, and in a different format.
Yeah, basically, any institution that can create accounts backed by FDIC insurance essentially has a limited license to print money.
25. May 2009 at 10:44
Devin: since what the Fed actually controls is the monetary base, I would start with a 10% increase in the base, under the assumption, reasonable as a first approximation, that all the other monetary aggregates would rise by about the same 10% also. But you can’t do that and expect to be able to hold the Fed funds rate fixed at the same time. (And my guess is that it would rise, not fall, BTW.)
If I wanted to get better than that approximation, I would need a full-blown macro model. Or else adopt Scott’s plan.
How to inject the new money? In today’s circumstances, I would suggest buying the stock index. S&P500 say.
25. May 2009 at 12:33
Nick-
Would your answer change at all if I said USG in general? So disregarding the actual laws that bind the Fed right now, if the government in general wanted to cause a 10% increase to NGDP, would you still suggest targeting the base?
25. May 2009 at 16:40
Devin, It seems to me that your essay shows the dangers of deduction:
1. You said if China got all of our base they “would own the world.” Actually they would own some scraps of paper, that pay zero interest. The US would just print up more scraps to meet the US demand for currency.
2. No, people aren’t always better off holding T-bonds rather than hoarding cash. Most cash is hoarding because it is more convenient than T-bonds (especially more anonymous.)
The question of what is money is partly a question of convenience. But virtually all macro models that I know of define “money” as an asset with a fixed nominal value. Thus inflation is, by definition, a depreciation in the value of money. Among assets with a fixed nominal variable, you pick the one that is most convenient, or most useful in terms of explaining macro shocks. In my view base money currently serves that role, whereas gold did prior to 1968. I like base money because the Fed has a monopoly on the supply, and thus you merely have to model the demand side of the market. Bonds don’t have a fixed nominal value—so it’s possible for both the price of goods and the value of bonds to rise at the same time. With money that is not possible.
Nick, I mostly agree with your responses to Devin.