Bursting Sumner’s Bubble, the Climax
I added the below in an update to my original post, but I want to make sure y’all see this one:
Thanks to the sleuthing of Dan, we are able to produce the following work of beauty:
(A) “I’m a believer in the EMH and hence skeptical of the idea of bubbles, a least as the term is usually interpreted. But I’m in the minority, the vast majority of people think bubbles exist.” — Scott Sumner, 9/7/2011.
(B) “Next we have to discuss what we mean by ‘bubble.’ Most people mean a sharp rise in prices, followed by a big decline. I agree with most people.” — Scott Sumner, 1/12/2012
(C) “I am constantly amazed that so many highly intelligent economists and finance-types seem incapable of understanding something as simple as an asset price bubble…[My argument] means that if bubbles are big price advances followed by substantial declines, then a world without bubbles would violate the EMH. Which means asset price bubbles do not violate the EMH.” — Scott Sumner, 8/9/2012
I submit that there are only two ways to reconcile (A) through (C):
Door #1: Sumner claims that not just his definition, but the definition of “bubble” used by most people, has changed during the last 11 months.
Door #2: Sumner claims that a “sharp rise in price, followed by a big decline” is quite a different thing from “big price advances followed by substantial declines.”
My work is done here.
More like “Dan’s work is done here”.
za-za-za-ZING!
OK then WordPress’s work is done here.
Well then, uh, Matt Mullenweg’s work is done here.
Yeah sorry, I went the wrong way with my joke. My bad.
Come on, you should have kept it going!
Matt Mullenweg
-> Matt Mullenweg’s parents
-> natural selection
-> “initiator” of natural selection
-> “initiator” of existence
-> God
-> God’s work is done here
-> tee hee, some joke about doing God’s work
-> (optional) God opposes what Scott Sumner is doing
-> continue until no longer funny
Come on! Delegation was Bob’s duty!
I am constantly amazed that so many highly intelligent economists and finance-types seem incapable of understanding something as simple as an asset price bubble…
I think [A] makes sense if you read it as saying that a believer in EMH would not be skeptical of bubbles ever existing, but would be skeptical that anyone at the time could reliably detect a bubble. The only way to detect a bubble is post-hoc by which time the knowledge is useless.
Suppose for example I said, “I’m a believer in randomness hence skeptical of the idea of winning the lottery.” and then when asked to explain what winning the lottery is about I would say, “The definition of winning the lottery is putting down some numbers, and then those numbers come up.”
Later on I might say, “Why is it so difficult to understand that if winning the lottery requires you to select the numbers that come up then a world without lottery winners would violate probability? Is that inconsistent with my earlier statements? I don’t think so.
Seems a bit nit-picky. I’m sure that you could find this kind of inconsistency in any blog that has the volume of posts that Sumner’s has. Does this have any bearing on the correctness (or otherwise) of his NGDP targeting proposals ?
There is also a Door #3: Sumner could say that when he said he was skeptical of bubbles “as the term is usually interpreted” he meant “as the term is usually interpreted by economists.” After all, it’s not like the average man on the street spends a lot of time thinking about whether the Swiss Franc is over valued.
Implicit term restriction is a standard part of English. If you’re looking in my fridge for something to drink and I say “I drank all the beer last night,” you’re not going to prove me wrong by taking me down to Wal-Mart and showing me the shelves stocked with Bud Light.
Exactly. THis is playing,’Ignore the context’ gotcha games.
Weren’t we told by Sumner in (B) how bubbles are usually interpreted? Namely, a sharp rise in prices followed by big declines?
If in (A) he defined bubbles with the important caveat that they have to be correctly predicted, then (A) conflicts with (B) and (C), since in (B) and (C) bubbles are defined as large price movements only, with no predictions included.
If in (A) he defined bubbles without that caveat, that bubbles are just a particular pattern of price movement, then (A) also conflicts with (B) and (C), since in (A) he was skeptical they even exist, but in (B) and (C) he is talking as if they do exist.
I know I don’t see how these three statements can be reconciled in any way other than Sumner’s definition of a bubble changing from 2011 to 2012.
BTW, it’s not a big deal for someone to change their definitions over time. After all, definitions are stipulative, not objective. I used to define inflation as rising prices until I started using the original definition of rising money supply beyond the rate of precious metals (which is what I assume will be the money of choice in a free market).
I don’t see what the big deal is about Sumner’s changing definition of bubbles nor why you and BA feel compelled to question Murphy on this point. It’s really rather obvious.
Two comments:
1. People use the term bubble in two distinct ways. In statement B I meant that if you define bubbles as a big jump in prices followed by a decline, then bubbles exist. I agree it was poorly worded.
In statement A I meant that the term ‘bubble’ is used by most economists to imply a violation of the EMH. I don’t agree that that sort of bubble exists.
It’s always been true that most economists see bubbles as violation of the EMH, and most people see bubbles as big prices rises followed by sharp falls. That was true 11 months ago, and it’s still true.
2. You have way too much free time
I totally agree. He should use it more wisely on something else rather than spend time critiquing your nonsense.
IT would be petty to say ‘I toldya so’. So I ll just say Blackadder told ya so.
2. You have way too much free time
About as much free time as some people have in digging up old DeLong quotes and pointing out inconsistencies, am I right?
OK, I have two comments as well:
(A) Scott Sumner 11 months ago says, “I don’t believe in bubbles, the way most people mean the term, because I believe in the EMH,” and now says, “I can’t understand why people think bubbles–the way most people use the term–are somehow a violation of the EMH, because clearly they aren’t”–and somehow I’m the one at fault here? I’m just nitpicking?
(B) Sumner, Blackadder, Ken B., et al.: So let me get this straight. It took Scott Sumner, Chicago School PhD and brilliant blogger, to explain to us that the EMH is consistent with asset prices rising and falling (a lot)? This is what the argument among certain people was supposed to be, that Scott was swooping in to solve? There were actually a bunch of people–whether economists or man-on-the-street, you can pick–who were saying, “I think markets are rational, and therefore if you tell me the stock market sometimes goes way up and then way down, I think you are lying to me” ?
No fellas, if you are siding with Sumner on this one, something is wrong with you. Just say, “OK he screwed up.” I’m not blowing up the EMH here, I’m blowing up Sumner’s blogging style and the people who think he’s a genius.
(BTW I love Scott as a person. He’s really cool, and that’s why I am so harsh to him in print, because over email I’ve made sure he “gets” my jokes.)
It took Scott Sumner, Chicago School PhD and brilliant blogger, to explain to us that the EMH is consistent with asset prices rising and falling (a lot)?
To quote George Orwell: “We have now sunk to the point where restatement of the obvious is the first duty of intelligent men.”
Blackadder give me a break. You’re saying that there are people out there who, if you said to them, “Hey, did you realize house prices went way up and then came way down a few years ago?” will say back to you, “Impossible, I don’t believe it! I believe markets are efficient.” ?
I guess these are the same people who don’t know commercial banks lend out 90% of their deposits?
Blackadder give me a break. You’re saying that there are people out there who, if you said to them, “Hey, did you realize house prices went way up and then came way down a few years ago?” will say back to you, “Impossible, I don’t believe it! I believe markets are efficient.” ?
No, but there are people who if you tell them about the EMH will say “that’s absurd! Just look at the big drop in housing prices.”
The EMH is absurd, but not because of price changes, no matter how rapid…
Bob,
I won’t answer for SS, who seems quite capable of doing that himself, as your bruises attest. I am pointing out a simple and frankly rather common error on this blog (more often committed by commenters than you). It is a failure to understand conjectural argument, anything that in French might be in the subjunctive. As Blackadder and I noted there is an easy, direct way to understand SS’s comments that makes them consistent. If it makes them right I have no idea. I think just to be fair and honest one has to read them that way therefore. I gave examples on the other thread with religious talk.
The problem as I say seems to come from not understanding when Sumner was implicitly granting something ad arguendo if I grant X ad arguendo I am not granting X.
I repeat I do not know if and am not arguing that SS is right. I am arguing your posts, and the comments of your minions, misinterpret him, and rather obviously do so. Blackadder’s beer comment is a funnier way of making the point I made on the other thread, and is completely correct as we now know from Sumner directly. And you did it again with ‘people’ in your point A. Different people.
Ken wins for most obnoxious comment I’ve read on this site in a while.
Dr. Murphy do you remember the post where Ken said you and him had discussed a certain aspect of WWII but you actually never did? He linked to a previous post that he thought proved it but when you actually go to the post he linked to you find he was the only one who brought up the point and there was no further discussion of it. He chastised you about a conversation that never took place and has the balls to call out anyone on this site for misinterpreting another person.
Quite distorted. The point at issue was Murphy falsely suggested I was blithely unaware lend-Lease could be controversial. RPM admitted he knew I had long debates on the issue on his blog. What that says about his intial charge I leave to others to decide. Whether i linked to the right comment in your opinion is beside the point.
And yes Dan, RPM did equivocate on ‘people’.
Actually that isn’t a “distortion”.
At the time, I too went back to see your accusation that Murphy was engaged in a discussion with you, and I didn’t see it either. You were caught red handed misrepresenting Murphy, and now you’re claiming the moral high ground? You’re obtuse.
Oh and Dan, I’m not ‘calling out’ anyone. I am pointing out — and not just me — where RPM really DID misinterpret Sumner. As we know from Sumner himself. Strangely I feel entitled to point out errors evn if you think I once made one myself. I’m just uppity I guess.
Ken, you like to think you are just pointing out errors but you do it in this sanctimonious way when you have no business acting like that. If you had simply said you disagree with Dr. Murphy on this and stated your reasons why then I wouldn’t have even responded. I wouldn’t have agreed with you but I wouldn’t feel compelled to respond.
It would be like Paul Krugman pointing out that another economist was a political hack. Even if he was right I would still call him a hypocrite.
Except there is no ad arguendo, not when we’re told first that bubbles don’t exist the way most people understand it because that would violate EMH, and then we’re told that they do exist the way most people understand it but don’t worry it doesn’t violate EMH.
Conceding that you do not recognize it is an odd way to argue you aren’t missing it …
All of Sumner’s statements are consistent with the view that EMH can accommodate large rises followed by large falls in asset prices (respective of whether you call such large rises and falls “bubbles” or not.)
It would be a way more interesting discussion in my opinion to debate if this is a tenable view rather than nit-picking on the somewhat inconsistent wording he used at various times in the past year to express it.
It wouldn’t have been a huge problem with the inconsistency in the way he used the term bubbles, if Sumner himself didn’t nitpick others in the way they used the term bubbles.
Murphy is nitpicking a nitpicker. Not all that out of line.
You know guys, the more I think about Scott’s statements quoted in this post, and then his response to it, the more astounded I become. Here is what Scott is saying:
When he wrote this:
“I am constantly amazed that so many highly intelligent economists and finance-types seem incapable of understanding something as simple as an asset price bubble…[My argument] means that if bubbles are big price advances followed by substantial declines, then a world without bubbles would violate the EMH. Which means asset price bubbles do not violate the EMH.”
…Scott was using the term “bubble” in the way that the man-on-the-street uses the term. Because, by Scott’s post 11 months ago and his clarification today, if he used the word “bubble” the way economists and finance types use the term, then clearly bubbles don’t exist (if we believe in the EMH).
So to paraphrase the italicized words above, here is what Scott’s brilliant point was a few days ago:
BOB’S UPDATING OF SCOTT’S INSIGHT, IN LIGHT OF HIS CLARIFICATION TODAY: “I am constantly amazed that so many highly intelligent economists and finance-types seem incapable of understanding something as simple as an asset price bubble…If I assume they are using the term “bubble” to mean what economists are always assumed to mean by the term, then they are perfectly correct. But, if I assume that economists use the term “bubble” in the way that non-economists use the term, then the economists are wrong. I am so amazed by their error and confusion. Plus, I get annoyed at people who nitpick me over definitions.”
OK come at me, guys. Explain how I’m out of line here.
Last point: I doubt you will EVER find me saying on one day, “The regression theorem proves Bitcoin can’t exist,” and then 11 months later saying, “The regression theorem proves Bitcoin must be possible.” And if you somehow did catch me redhanded with such contradictory posts on the implications of my economic theories, I wouldn’t say, “Wow you’ve got a lot of time on your hands.” No, I’d say, “Holy cow, that’s embarrassing. Let me think about what happened there, because I was really confident when I wrote both of those posts.”
Bob,
1) Lots of economists think the existence of bubbles refutes the EMH.
2) Many of the same economists think that you can identify bubbles by looking at large price swings after the fact.
Scott’s point is that large price swings are perfectly consistent with the EMH, so if that’s all you mean by bubbles then of course bubbles are consistent with the EMH.
You can say this is an obvious mistake, but it’s a mistake people make all the time.
Blackadder, I really can’t keep doing this. You are telling me that if you pinned down a PhD economist who is skeptical of the EMH, he would say this:
(A) “Yes, I think that the EMH, if true, would mean that when I buy an asset, I am sure it will keep going up. If I thought the asset might go down in price, then that would violate the EMH.”
I submit that no one actually believes statement A, who has any familiarity with the EMH and is a trained economist. Now, maybe some people wrote things in the heat of the moment on their blogs or in an op ed, that actually would back them into a corner of saying (A). In that case, they wouldn’t double down on (A). They would say:
(B) “Oops, yeah, I was a little unclear there. Please be charitable in how you read my statement. Did you really think I’m that stupid? Boy, you’ve got a lot of free time.”
Now are you guys starting to get it? You somehow think I’m attacking the EMH or something, when no, my modest point is that Sumner is being really slippery and confusing in his discussions of the EMH and bubbles, and then has the audacity to accuse others of being slippery and confused in their treatment of the EMH and bubbles.
Scott’s point is that large price swings are perfectly consistent with the EMH, so if that’s all you mean by bubbles then of course bubbles are consistent with the EMH.
So how could Sumner have been skeptical of the idea of bubbles as in (A)?
“OK come at me, guys. Explain how I’m out of line here.”
I think you are ignoring Sumner”s main point which is to explain how asset price falls must sometimes follow asset price increases if the EMH is to be true and instead highlighting a fairly minor inconsistency in the use of the term “bubble” that Sumner has already acknowledged.
To me it just seems out of line with the level of argument normally associated with your posts, that’s all.
OK Blackadder, i have let this sit for a few hours. I am calmer now. Some thoughts:
(A) I think things would have gone much easier, if you would have just admitted upfront, “Wow, yeah Sumner totally contradicted himself there. That’s confusing as hell. But, if we look at the full context of his posts, we can see a common element there. His underlying position hasn’t actually changed, and he’s not just stating tautologies. A lot of economists out there really are making a basic mistake when they think about asset bubbles and the EMH. So you’re right that Sumner executed it poorly, but let’s not throw the baby out with the bath water.”
(B) I am actually one of the economists Sumner has in mind. When I look at the housing bubble (as the layman uses the term, so it obviously existed), I also see a housing bubble (as an economist uses the term, meaning it violates the EMH). But when Sumner looks at the housing bubble (as the layman uses the term), he doesn’t see a housing bubble (as an economist uses the term). By the way, do you see how awful this nomenclature is? And I’m the jerk here?
(C) There is a flaw in Sumner’s reasoning in his latest post. To paraphrase, he is saying something like, “The existence of the housing bubble in the layman’s sense doesn’t prove a housing bubble in the economist’s sense. Suppose you thought otherwise. Then, that means if housing prices happened to be up 100% in 3 years, that you would be sure they couldn’t go down, say, 80% over the following 3 years. But then that means everybody would pile into housing, because it would be a sure bet. There’s no such thing as a sure bet, because of EMH. Therefore, EMH must be consistent with house prices falling 80% over the next 3 years.”
==> I submit that there are some screwy things with this argument. First, the people who think bubbles in the economist’s sense can exist, would already be saying “this is a bubble” when house prices went up by 100% in the first 3 years. Yet Sumner is taking that as the starting point of his argument. This is really odd. The critics of EMH (like me) are saying that if it were true in the way that many of its proponents think it’s true, then house prices should never have been bid up that much “for no good reason” in the first 3 years. So it’s a really screwy argument.
Second, even if we are OK with the premise of the argument, the conclusion itself relies on an application of the EMH. Sumner is invoking the EMH when house prices are at their peak, to say that we can’t know at that point whether they’ll keep going up or might possibly crash. But the whole purpose of this argument, is to determine whether Sumner is correctly using the EMH when he talks about empirical asset prices. So it’s weird to invoke it in the middle of the argument.
Of the two points above, my first is the stronger. To boil it down: A guy like me will say, “If you want to use the EMH in any but a tautologous way–if you want it to actually influence your views about how actual asset prices behave–then the massive boom and crash in house prices clearly violates it.”
Sumner then comes along and says, “No, because if the EMH allows for a massive boom in house prices, then the EMH must also allow for a massive crash.”
Right, but that doesn’t mean the EMH (if it is to have any empirical content, and not just be a neat way to look at the world) is consistent with a boom-and-crash. That was the original, skeptic claim.
Bob,
On the substantive point I think you are right. I don’t think the argument Sumner made in that post was very compelling.
I would also agree that Sumner could have phrased his point differently so as to remove the possible ambiguity (Scott himself said as much).
However, I don’t think the two quotes are contradictory, and I think it’s pretty clear they are not contradictory if you look at the context.
Blackadder, yeah, I’ve sent the recall codes to my long-range bombers. If you care, here’s what threw me: I thought in the second post that Sumner was saying, “I can’t understand why so many economists who believe in the EMH are convinced that it renders bubbles impossible. In fact, the EMH means bubbles *have* to occur once in a while.”
Since Sumner himself argued 11 months earlier that the EMH means bubbles are impossible, I fell out of my chair when reading this.
But, since you clarified that he had in mind economists who did NOT believe in the EMH–because they thought it was incompatible with bubbles, and because bubbles existed in their view–now it all falls into place.
Last thing: I still maintain there is a formal contradiction among all the posts, taken at face value. All you’ve done is got me to realize he wasn’t being slippery, just confusing; i.e. he wasn’t objecting to people who made the same claim he himself had made 11 months earlier (as I first thought). But look at what was going on here: Sumner was objecting to *economists* talking about the EMH and bubbles in a certain way, when his argument only makes sense if those economists are using “bubble” in the way that Sumner reserves for the layman (even though we’re talking about economists!!). At the same time, Sumner himself–when saying bubbles don’t exist because of EMH–means the ‘economist’ notion of bubble, even though he himself is on record saying “I use bubble to mean a rise in fall in asset prices.” So this is supremely, magnificently, confusing as heck.
Ahem.
Oh, one last thing Blackadder: Yes, if you ran 200 million iterations of a computer simulation of a decade of housing prices, and plugged in some empirical values for mean and variance in changes of house prices historically, then I’m sure you would find the exact pattern we just observed…maybe in 0.03% of the cases. But you would also find much smaller housing “bubbles.”
I haven’t actually done the work, but I am pretty sure that if we look at the actual record of historical home prices, I don’t think you will see a normal distribution of such bubbles. I think you will see general trend growth, with occasional falls maybe during severe recessions. And then, from the late 1990s through 2008, you will see…
…A GIANT FREAKING BUBBLE THAT POPPED.
I love writing computer simulations. Sketch it out and I’ll put it to code.
How would you suggest modelling the positive feedback where a person looking to invest will tend to buy into a neighbourhood where house prices are rising (thus causing house prices to continue rising) vs the negative feedback of people wanting to move into an area looking for a job and being discouraged from doing so by the high prices?
Some sort of statistical ensemble of owner occupiers, owner investors, and a rental market?
A collaborative effort with my name alongside Bob Murphy would be a massive boost to my ego, and probably my future career prospects as well.
Sorry Tel I think you are reading too much into what I was saying. I just meant, if you were going to model house prices as a random walk. So you’d look at the history of house prices from (let’s say) 1950 – 1995, and take their mean and variance or whatever measure of dispersion the analyst wants. Then you’d say, “OK, if house prices are a random walk with drift, calibrated to the 1950-95 experience, then what is the ex ante probability we’d see something like the 1996 – 2006 experience?” And I think it would be pretty low. So Sumner has pointed out that it’s not impossible, but again, I am saying I think it is a pretty damning refutation of the EMH, if we think it is even possible for data to ever refute the EMH.
Last thing for anybody who cares: I have no problem if someone uses the EMH as a neat way to view the world, just like “supply and demand” is a neat way to view the world. But if you take that approach, don’t kid yourself that you are using an empirically-based theory. In other words, if you believe in the EMH and you don’t think your belief is a matter of faith, but rather a matter of the evidence, then I think you are fooling yourself. If the housing bubble didn’t make you drop the EMH, then what would?
But you have changed ground now, from SS iscontradicting himself to SS is wrong.
You’ve got a lot of free time Ken B.
I waste little on church.
Completely off topic, but this quote was so awesome I had to reproduce it.
”I can be friends with non-anarchists, but only if they keep their mouths shut.” Stephan Kinsella
Should you be talking about ‘bursting’ and ‘climaxes’ at the same time on an economics blog? I suspect many many several spectacle may have fogged up as they read the title.