My newest NGDP paper
Yesterday the Mercatus Center published my latest paper advocating NGDP targeting. Coincidentally, on the same day David Beckworth and I presented the case for NGDP targeting on Capitol Hill. Bob McTeer moderated and the turnout was larger than usual.
Saturos pointed out (correctly) that I am inconsistent in my use of the term ‘bubble’, sometimes arguing that NGDP targeting can reduce bubbles, and sometimes arguing that they don’t exist. This has also bugged sticklers for accuracy like Bob Murphy. I plead guilty. I don’t believe in bubbles in the sense of asset price movements that obviously diverge far from a rational expectation of fundamental value. I do believe there are asset price movements that look like bubbles to the average person, and which are regarded as bubbles. Those are big increases in asset prices, followed by big decreases. Most people (wrongly) assume those big swings obviously do not represent changes in a rational expectation of fundamental value.
Generally whenever I use terms like ‘bubble’ or ‘income’ or ‘inflation’ you should think of them in terms of scare quotes; “bubbles” or “income” or “inflation,” even if I forget to use the quotation marks. I am using the terms as they are generally regarded by society. I believe in the EMH, not bubbles, I believe in consumption, not income, and I believe in NGDP growth, not inflation.
Oh and one other thing. Whenever I talk about “wages” I mean hourly wages, not wages per employee.
PS. Obviously NGDP is national “income.” I don’t believe income data for individuals is meaningful. Look at consumption.
Tags:
24. October 2012 at 05:47
I declare a caption contest!
“An unlikely revolutionary”
24. October 2012 at 06:02
Great, I wanted a bigger version of that pic.
“Bad Boys III: Targeting the Congressional Forecast”
(Sorry, others will do much better.)
24. October 2012 at 06:03
Nice paper. I enjoy the way that you engage the gold standard advocates. As I believe Tyler has noted before, the gold standard isn’t as obviously absurd as some like to pretend it is, but it is (I am convinced) inferior. It can be argued against successfully while granting its advantages, merely showing that the disadvantages are greater, as you did. I think that is a fruitful approach.
24. October 2012 at 06:11
Scott, I’m a bit confused by footnote 17 in the paper:
Did you mean to cite the same paper twice, as far as I can tell?
24. October 2012 at 06:16
Great paper. But one thing concerns me:
When I advocate to people for NGDPLT, the first question I get is ‘how’? I tell them ‘NGDP Futures Contracts’, and they ask ‘how?’
They see your mechanism for implementing NGDPLT as mere hand-waving. Its an underdeveloped part of your thesis — I recommend you develop it.
24. October 2012 at 06:30
Oh, Nick, Nick, Nick –
Scott has spent decades developing the futures targeting concept. See the papers he refers to. Or see the post “Spot the flaw in nominal index futures targeting” on this blog (check the sidebar).
But targeting futures is not essential. The Fed can simply be set a new mandate to follow a predetermined NGDP level path, say halfway back to the peak + 4%/cap growth path each year thereafter. And it can simply target its own internal forecast in order to acheive that, as Svensson has advocated.
24. October 2012 at 06:32
Scott should add this paper to the sidebar, in fact, it may even supersede ReTargeting the Fed.
24. October 2012 at 06:36
A very good paper. I particularly like the point you make about NGDP targeting involving less subjectivity and controversy about measurement than inflation targeting.
It might be useful to develop this point: inflation targeting has many of the practical problems of Old Monetarism, in that there are many definitions of inflation (CPI, PPI, RPI, core XYZ, XYZ excluding housing etc.) just as there are many measures of monetary aggregates AND to make it worse than monetary aggregates these indexes are revised over time on the basis of subjective judgements that violate what microeconomics teaches us about interpersonal utility comparisons. To put it in rather dramatic terms, you can take inflation targeting seriously or you can take the marginalist revolution seriously, but not both.
NGDP, like monetary aggregates, involves no interpersonal utility comparisons and unlike monetary aggregates does not face difficult issues of redefinition in response to financial innovations like eurodollar deposits, repos, MMFs, bond funds, equity funds and so on.
24. October 2012 at 06:40
“I believe in EMH, not bubbles”
You’ve got to be kidding me right? When you have companies at a P/E ratio of 1000 who have a negative cash flow and are doubling in price every few months, are you telling me that’s not a bubble?
You do realize that if you change one small assumption of the EMH(that the price movements are independent increments), you get very different results. The whole idea that’s based on the rational expectations theory(where the errors are falsely assumed to be white noise) is completely false.
If you make it so that the errors aren’t random and the price movements can be a little correlated, you get very different(and far more realistic behavior). Here are a couple of papers on this issue:
Hurst Exponent and Financial Market Predictability by Bo Qian and Khaled Rasheed
Nonlinear Forecasts, Rational Bubbles, and Martingales by Benoit Mandelbrot
If you calculate the Hurst exponent for something like the Dow Jones; it ends up varying over time. The Hurst exponent calculates the correlation between the price movements. If H.5, then we have a random walk; otherwise, you get very different behavior.
I suggest a much more realistic alternative, like the Fractal Market Hypothesis. Here’s a paper that discusses its use and how it’s use works well to describe the financial crisis in 2007-2008.
http://arxiv.org/pdf/1203.4979.pdf
24. October 2012 at 06:44
W.Peden, I think a GDP deflator definition of inflation is fully compatible with the marginalist revolution. Of course that’s the one included in NGDP.
Bubbles have been proven to exist in the lab. Even Scott know that Rational Expectations is false (and not in the glib sense of “all models are false”). It’s false enough that the EMH clearly has visible theoretical glitches. The question is whether it is wise (and for whom) to act (or model) as if the EMH were perfectly valid.
24. October 2012 at 06:46
The first person to really come up with this EMH type behavior was Bachelier(not Eugene Fama). When he used this theory; even he noted that it wasn’t perfect and that it would fail.
It’s funny, because I took at Graduate level Time Series class not too long ago and I remember asking my professor about how you get “fat tailed” behavior in time series. He responds, “the tails aren’t really fat, it’s just that the behavior is usually correlated”.
24. October 2012 at 06:51
I think Scott should have further emphasized that level targeting is necessary because sticky wages stay on a level path. As Nick does here: http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/08/inflation-rate-vs-price-level-in-the-srassrpc.html
24. October 2012 at 06:59
Another benefit of NGDPLT: This might be reversed. http://www.insidehighered.com/news/2012/10/24/survey-finds-professors-already-liberal-have-moved-further-left
24. October 2012 at 07:01
Saturos,
Good point. Then again, how many central banks target the GDP deflator?
24. October 2012 at 07:07
I think the main problem raised by the article, in the end, is why we should be so focussed on stabilizing NGDP instead of aiming to make wages more flexible? (The MF critique.) Is it purely because of sticky debts? Why can’t society find ways to make them more flexible? Perhaps what it comes down to is that stabilizing NGDP is (ought to be) simply much easier, if you that it needs to be done.
24. October 2012 at 07:12
W. Peden, I think what it comes down to is the best measure of the stance of policy. There is the famous Sumnerian dictum:
but for me its a lot simpler. Money is tight when its aggregate flow dries up. It is loose when the aggregate flow is heavy. And the equation of exchange tells us we can measure that flow (or the final-output part of it) by looking at NGDP. We don’t need to be so philosophical: NGDP is really really, concretely, the correct measure of the stance of monetary policy.
24. October 2012 at 07:17
Did Tyler have any role in getting this published by Mercatus, Scott?
24. October 2012 at 07:32
Saturos,
To answer your question about whether it’s wise to act as if the EMH was completely valid, I certainly don’t. The EMH completely fails under extreme events. It has no power to deal with them whatsoever. The problem is that in markets(and in most of the real world), the extreme events are what really matters. It’s the kind of randomness that Mandelbrot describes as “wild randomness” compared to the “mild randomness” that the behavior is supposed to be.
Looking back, I might have misunderstood a little bit about what Prof. Sumner said(and I apologize). However, there is a lot of psychology in bubbles. So while bubbles might represent large changes in expectations of fundamental value; it doesn’t change the fact that they’re bubbles. I think it’s critical to understand that the stock market is supposed to represent certain fundamentals and when you have large changes from those fundamentals; you’re going to have large price swings(volatility will increase). To say that bubbles don’t exist and can’t be detected is completely false.
Usually when bubbles occur, they ending being a Ponzi scheme because the only way they can be sustained is if more resources are entering than leaving. The fact that more resources are entering than leaving is what causes the asset prices to rise and what keeps them rising. In a world of finite resources and limits; more resources cannot keep entering and eventually all bubbles must collapse.
I think we have to make it very critical that the market can be wrong and all asset bubbles have to eventually collapse.
24. October 2012 at 07:37
I think we should clarify: I was talking about weak EMH; you seem to be talking about the strong version. Weak EMH is about whether the forecast is fully informed and unbiased; it doesn’t have to be right.
24. October 2012 at 07:50
Scott should have probably worn a flashier suit to Capitol Hill, like Beckworth’s: he looks too much like the college professor that he is. Since when did Congresspeople take academics seriously?
Were there any recognizable faces? Anyone from the Administration?
24. October 2012 at 07:53
‘When you have companies at a P/E ratio of 1000 who have a negative cash flow and are doubling in price every few months, are you telling me that’s not a bubble?’
Sounds like Microsoft in its early days.
24. October 2012 at 07:54
Scott has gone Humpty-Dumpty on us … “rational” and “bubble” and all the rest mean whatever he wants them to mean, and he has nothing backing it up but a dogma derived from a completely unrelated technical literature.
FAIL.
Scott writes,
“Those are big increases in asset prices, followed by big decreases. Most people (wrongly) assume those big swings obviously do not represent changes in a rational expectation of fundamental value.”
24. October 2012 at 07:59
Wow Patrick, just checked out your blog. It’s pretty good, I just subscribed.
24. October 2012 at 08:00
Greg, you should link to Taking Hayek Seriously on your username…
24. October 2012 at 08:08
In other news, The Economist endorses a Syria intervention: http://www.economist.com/news/leaders/21564840-despite-huge-risks-involved-time-has-come-west-and-arabs-intervene
24. October 2012 at 08:09
Saturos: “Scott has spent decades developing the futures targeting concept. See the papers he refers to. Or see the post “Spot the flaw in nominal index futures targeting” on this blog (check the sidebar).
But targeting futures is not essential. The Fed can simply be set a new mandate to follow a predetermined NGDP level path, say halfway back to the peak + 4%/cap growth path each year thereafter. And it can simply target its own internal forecast in order to achieve that, as Svensson has advocated.”
— How? buying risky assets? Mega-QE?
I’ll read the post, but these things aren’t clear.
24. October 2012 at 08:30
Scott, you recently said about the “money is a bubble” debate that money can’t be a bubble.
But it seems that because of differences in definitions you might agree with Karl Smith take defending that money is a bubble.
He said that what people usually call bubbles are an increase in the price of an asset driven by liquidity premium. And he said that money is only liquidity premium. It seems to me that this can be consistent with EMH.
Obviously everyone agree that money is a bubble under Karl Smith’s definition, but you think is a relevant definition for bubbles?
24. October 2012 at 08:37
Nick, you need to read the whole blog. You’re a believer in liquidity traps, as I once was. And yet no fiat money central bank has ever tried and failed to inflate.
Svensson has a foolproof mechanism involving currency depreciation on forex markets. And the Fed could buy every asset on earth if it had to (Bernanke himself once used this reductio ad absurdum). In practise it wouldn’t go that far: the Fed would simply announce clearly where it wanted to go, and the markets would do the heavy lifting: http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/10/engdp-level-path-targeting-for-the-people-of-the-concrete-steppes-.html
The main failure of the Fed has been its failure to coordinate expectations appropriately. With higher expected NGDP growth, the demand for money would be far lower, leading to more spending – a self-fulfilling prophecy. But yes the Fed could brute force it if it had to – and the markets know this. But temporary money injections never work, certainly not when you pay banks to hold on to the money. No wonder QE hasn’t boosted spending much – the Fed never intended it to.
24. October 2012 at 08:40
“Permanent money injections are always effective, even in a liquidity trap (according to well-known Keynesian Paul Krugman)”
– from the FAQs
The main difference between MMs and everyone else is that we look through the other end of the telescope: http://www.themoneyillusion.com/?p=15868
24. October 2012 at 10:35
It is great news that you were invited to give a presentation to potential policymakers – and I thank you for sticking to your guns long enough to get there. I very much appreciate everything you’ve done and the enormous effort involved.
I really like your paper too. But I have one area of curiosity when it discusses the housing “bubble” being related to above trend NGDP growth. I don’t agree with that. I put much more weight on supply-side policy issues being the primary culprit, and not so willing to point in the direction of the Fed, letting the regulatory regime and politicians off the hook, when NGDP growth had been below trend from the very late 1990’s until approximately 2005. I subscribe to a more counter intuitive theory of “bubbles” as being caused by suboptimal NGDP growth that leads to demands for interventionist policies such as those expected to stimulate investments in government-preferred sectors. It is an important distinction that separates the investment phenomena from availability of funds because it is quite possible to have what people call a “bubble” with all of the same consequences in addition to tighter than required monetary policy which is far more damaging than the assumed cause given the opportunity cost.
24. October 2012 at 10:35
ssumner:
FTA:
1. Hasn’t historical central banking been “marred” by periods of inflation and deflation as well?
2. What is so evil about peaceful, voluntary price changes that derive from peaceful, voluntary monetary orders, that justifies state violence to impose price stability? Price stability is not consistent with a private property society that respects individual preferences.
So (one of your) arguments against the gold standard is that the state interfered with it? I don’t follow. It is like arguing against freedom on the basis that freedom alone can’t guarantee it’s permanence. Gold isn’t a law that enforces itself. It can only be sought for. If it isn’t, that isn’t a flaw of gold.
The gold standard was all but abandoned by the 1930s.
You aren’t really arguing against gold, but against governments. You are saying that a gold standard cannot protect itself against meddling governments. Well of course. If every political strategist, economist, and philosopher just sat back and believed gold would take care of itself, limiting states by itself, then of course gold will collapse. We NEED intellectuals to constantly intellectually fight against government meddling, the same way most economists today, such as Keynesians and Monetarists, constantly intellectually fight for government meddling.
We are in charge of our own futures. You can’t think that history will take care of people like a warm blanket. People make their own histories. We right our own chapters. If there are no authors for limited government, gold standard, free markets, then history’s will be filled by authors who are for expanding government, fiat money, and socialism.
Unacceptable TO WHOM?!?!? The critics? Come on. This is like a gold bug criticizing fiat money on the basis that it is likely to deliver unacceptably small short term fluctuations in the price level. After all, if people really do want to spend sufficiently less money that spending falls in a gold standard, then a lack of falling demand and hence prices can be viewed as “unacceptable” by gold bugs.
So “unacceptability” by some unstated people is not a rational grounds against a theory. You can’t even say that you are assuming “the majority”, because by THAT logic, one can say that if the majority wanted to kill all red heads, then it would be “unacceptable” to the majority if some annoying individuals in the minority stopped them, and hence those annoying individuals are somehow in the wrong.
If we can count on the authorities to accept the discipline of NGDP targeting, why not instead make them obey the rules of forcing them to not interfere in a private property driven monetary system?
See what I did there?
If you believe authorities are disciplined enough to be able to engage in NGDP targeting, then why not a gold standard, or even better, a free market standard?
Or are you making an intrinsic argument that states simply cannot be trusted to enforce a gold standard? Well, that again isn’t an argument against gold, it’s an argument against states.
Nevertheless, the gold standard isn’t even intellectually justified by way of gold alone, without any human input, limiting state meddling. It is a tool to limit states. The people have to fight for gold if they want to limit the state.
It cannot be overly emphasized that while Volcker did reduce inflation in the early 1980s, almost immediately after that he engaged in significant re-inflation, which was then followed by a stock market crash in late 1987.
Even the “inflation fighting” Volcker couldn’t resist himself.
The swings of the 1980s, bear in mind, occurred under 100% fiat and no gold whatsoever. Why couldn’t Volcker be disciplined the way you seem to be saying central bankers can be disciplined? If the contrarian Volcker can’t do it, then who can? Let me guess, his name ends with –umner. I wouldn’t even trust you as an NGDP targeting theorist. If you had control of the money printing press, then you may be willing to trade off short term inflation benefits, for long term dismissal because you “broke the rules.” You may engage in 5% NGDP growth for a time, but then what if your son or daughter or someone you care about, is in danger of dying, and you have to choose between inflating and overshooting NGDP and saving their life, or remaining disciplined and watching them die?
Then, 20 years from now, neo market monetarists will come out of the woodwork and be talking about how states cannot be trusted to maintain an NGDP target, and so “pragmatically” there should be something like “NGDP can fluctuate…but only if there is a war, or famine, or “triple mandate” issues at stake…”
Then you’ll see “pragmatists” using your own irrational epistemology against you. Meanwhile, free market bugs will keep chugging along without contradicting themselves.
24. October 2012 at 10:36
Saturos,
For the record, I have no problem with the weak form of the EMH; I have a huge problem with its strong form.
Patrick,
“sounds like Microsoft in its early days.”
Here’s the price history of Microsoft. Microsoft didn’t have negative cash flow. You have to differentiate between Microsoft and pets.com. Pets.com was trading at $50 a share in the dot-com bubble. Microsoft wasn’t doubling every few months. It did increase a lot during the dot-com bubble, but it also crashed.
http://finance.yahoo.com/echarts?s=MSFT+Interactive#symbol=msft;range=my;compare=;indicator=sma(5,10,25)+volume;charttype=area;crosshair=on;ohlcvalues=0;logscale=off;source=undefined;
24. October 2012 at 11:06
Do these NOT look like two economists?
(in the hottest my complimentary way)
24. October 2012 at 11:15
John, I blame the editor. Seriously, it went through many revisions, and that second reference got added near the end of the process–I should have proofread one more time.
Saturos, You asked:
“Scott should have probably worn a flashier suit to Capitol Hill, like Beckworth’s: he looks too much like the college professor that he is. Since when did Congresspeople take academics seriously?”
I don’t have a flashier suit. That’s my first new one in decades. There were about 70 people, very few of which wore any sort of suit at all. Most were very young Congressional staffers.
I don’t know if Tyler played a role in the Mercatus gig–he’s often given me opportunities by recommending my name. Mercatus is located at GMU, so I would imagine that several faculty there might have mentioned my name.
Bonnie, I agree with you on the housing bubble, I probably worded that section poorly. I agree that supply-side problems were the biggest policy mistake. But NGDP instability can also play a role. A bubble also requires a crash. If there is no price crash (as in Australia, or NYC), then people don’t consider it a bubble. Does anyone still consider million dollar condos on the Upper East Side of Manhattan to be a bubble? No, those prices are here to stay. So the NGDP crash helped create the perception of a housing bubble, by creating a price crash.
Suvy, I am mainly talking about the weak form EMH–prices are unpredictable.
Arthur, Yes, money provides liquidity services. Those are useful. Hence money has fundamental value.
24. October 2012 at 11:33
But your belief in the strong form leads you to trust market forecasts, right?
I’ll tell you what, Scott, once you make the headlines, we at the blog will all pitch in and buy you a new suit. You deserve it.
24. October 2012 at 13:05
Small correction needed on page 10;
‘It is easy to understand the recession worries; we had been
in a recession since December 2008.’
Should be 2007.
And, it isn’t a flashier suit needed. It’s a darker suit.
24. October 2012 at 14:03
Prof. Sumner,
Okay, I misunderstood you then(my apologies). I agree that prices are unpredictable. However, asset price inflation of 10% is just not sustainable. I just disagree with the likelihood of a crash and the predictability of crashes(I think Australia is headed for a huge one by the way-the amount of leverage they have would easily amplify a small shock).
I think there’s a herd behavior effect; which causes a positive feedback loop on the way up and the way down. I think almost all crashes are predictable because they rely from extremely high unsustainable growth.
24. October 2012 at 14:03
Prof. Sumner,
Okay, I misunderstood you then(my apologies). I agree that prices are unpredictable. However, asset price inflation of 10% is just not sustainable. I just disagree with the likelihood of a crash and the predictability of crashes(I think Australia is headed for a huge one by the way-the amount of leverage they have would easily amplify a small shock).
I think there’s a herd behavior effect; which causes a positive feedback loop on the way up and the way down. I think almost all crashes are predictable because they rely from extremely high unsustainable growth.
24. October 2012 at 14:03
Prof. Sumner,
Okay, I misunderstood you then(my apologies). I agree that prices are unpredictable. However, asset price inflation of 10% is just not sustainable. I just disagree with the likelihood of a crash and the predictability of crashes(I think Australia is headed for a huge one by the way-the amount of leverage they have would easily amplify a small shock).
I think there’s a herd behavior effect; which causes a positive feedback loop on the way up and the way down. I think almost all crashes are predictable because they rely from extremely high unsustainable growth.
24. October 2012 at 14:14
I do agree that NGDP targeting is the way to go. I don’t know if you should just set it to the same value all the time, but it certainly makes far more sense than targeting inflation. With inflation, what are you targeting? Different goods respond differently to inflation in different circumstances. You could target CPI, but what about asset prices? Commodities would definitely respond differently as well.
One thing I really like about NGDP targeting is that you can set it to make sure the economy doesn’t overheat. You can simply reduce the NGDP target if you feel like an economy is overheating to make sure booms don’t get out of hand.
24. October 2012 at 15:26
Why is NYC housing not a bubble while it was in other markets? (I wouldn’t be in too much of a hurry to say Australia wasn’t a bubble)
1) Comparison with equivalent rentals. You can think of this as slow motion arbitrage.
2) This lets you divide the prices into a shelter price, an expected gain from price appreciation, and perhaps a small premium for the financial advantages of ownership. The price series tends to return to the shelter price and real housing prices have a reasonably well established long term trend.
3) When the gain term gets large and the market has a significant percentage of buyers (like flippers) who either need the price trend to continue in order to service their debt (Ponzi finance), or who have characteristics like low equity or balloon payments combined with a high loan to income ratio, you have a bubble.
4) Toward the end you see signs of drug tolerance. It takes more and more stimulation to maintain market momentum, and if maintenance of momentum is critical, the Wile E. Coyote theorem applies.
Weak efficient market theory (whether you believe in it or not) doesn’t stand in the way of bubbles. If you use formulations like:
“Excess returns (when considering transaction costs) cannot be earned in the long run by using investment strategies based on historical share prices or other historical data”
or that
“market participants not be able to systematically profit from market ‘inefficiencies'”
The phrases “in the long run” and systematically, allow for bubbles where the risk adjusted costs of a short position don’t allow a reliable profit, or where profit depends on additional information or a privileged position in the market.
By the same token there is no bar to Ponzi bubbles where there is a rational expectation of getting out ahead of the bust, and many people managed to do it. Often those most hurt were mere bystanders, who bought at reasonable prices and had problems with selling for purely cold blooded financial reasons.
There’s also the problem of marginal price inflation. If the all the housing stock is valued at the latest marginal prices, then you can get a form of inflation (which I expect has a name somewhere among economists), if it is possible to exploit this value without selling and thus depressing the marginal price.
Furthermore, this process can and does run in reverse.
24. October 2012 at 15:37
Suvy says, wrongly: I think almost all crashes are predictable because they rely from extremely high unsustainable growth.
“predictable” means within a reasonable time frame; perhaps a quarter, perhaps 6 months. Must be less than a year.
Greenspan in 1996 did NOT predict the dot.com bubble crash; the Economist in 2003 did NOT predict the housing crash. Not knowing when the crash is coming is typical of bubbles — and those who predict “the crash” are most often so premature as to be considered wrong (if more than year off).
The essence of an asset price bubble is that financial speculators/investors are buying the asset with the purpose of selling it soon, at a higher price, due to bubble increases in price. The main value of the asset is price increase value.
It’s pretty silly of you to want to claim bubbles are not much different usual investment.
24. October 2012 at 17:34
Those two handsome chaps look positively bubbly.
24. October 2012 at 18:05
Tom,
Let me state this, the harder thing about bubbles is to predict their timing. However, I think that it’s certainly possible to spot if a bubble is developing. I never said that you could predict the timing of a crash. I said that almost all crashes are predicable, I never said you could predict the timing of it. Predicting the timing of a crash is near impossible.
“The essence of an asset price bubble is that financial speculators/investors are buying the asset with the purpose of selling it soon, at a higher price, due to bubble increases in price. The main value of the asset is price increase value.”
The statement above is completely false. In every single market, most traders just buy a security to sell it later–there are also traders on the other side that sell securities short purely to make money. Most traders don’t buy or sell securities because of the assets “value”, that’s what Warren Buffet does, not most traders. The only way to make money trading is to buy low and sell high; that’s it.
However, if you have house prices increasing at 10% a year while real GDP is growing at 2-3% and price/rent ratios were constant for the previous century and then price/rent ratios double in 7-8 years; that’s the sign of a bubble.
24. October 2012 at 18:12
Tom,
I never said that you can predict the timing of a crash. I think that’s near impossible. What I did say is that it’s possible to predict a crash. Two completely different things.
“The essence of an asset price bubble is that financial speculators/investors are buying the asset with the purpose of selling it soon, at a higher price, due to bubble increases in price. The main value of the asset is price increase value.”
That’s complete nonsense. Almost all traders in the market make money by buying low and selling high(or selling high and buying low if they’re short). What you do have in a bubble is herd behavior. It’s usually also associated with leverage (not all the time, but most of the time).
24. October 2012 at 18:19
That was a cracking read Scott, one of the best things I’ve read in ages. It also does an admirable job stripping political partisanship from the economic concepts (in my view a key failure of the majority of economists); someone from any political walk of life should be able to latch onto the benefits pretty quickly.
My two big takeaways from it were this:
1. I had scarcely considered the effect of not sufficiently factoring housing price movements in the inflation calculation (P16), I understand in the U.S (much like Australia), about 2/3 of homes are owner-occupied, so it would make sense for a flow-through consideration into the inflation figures (or to simply focus on NGDP without worrying too much about inflation).
2. The power of NGDP targeting to mitigate the politically driven propensity for ‘bail-outs’ (P18/19). The more I think about this, it is self-evident & the targeting instead of NGDP would lead to the most efficient (rather than least efficient) businesses feeling the economic support, which would likely improve productivity overall.
I hope the paper is as well received as it deserves to be & the case for NGDP targeting continues to gather momentum – Tony
24. October 2012 at 18:23
As a side-note, I don’t generally agree with EMH, though I do believe markets are ‘mostly’ efficient & certainly the markets for something like ‘TIPS’ are more efficient than the markets for individual stocks – Tony
24. October 2012 at 19:45
I’d like to add, the main problem I have with the EMH is that it says that price movements fit a random walk. That’s what I have a problem with. You could still have a series where the price movements are correlated, but betting on the direction of the price movements might not be profitable because of some sort of an asymmetric payoff.
Now, I think it’s okay if you use the weak form as an approximation to reality. However, I that it fails under many circumstances (especially periods of high volatlity). A better alternative is Edward Peters Fractal Market Hypothesis which states that the market consists of different investors/traders that work on different time frames. When the you no longer have that structure(usually when longer term investors drop out of the market or behave on short term information), you get cascades and crashes.
For example, I think formulas like Black-Scholes are very useful for traders, but to set the value of an option to the current volatility could get you into a lot of trouble(if you’re the one selling it). Same thing with VaR. If you used it in 2005 and said let’s leverage up as much as we can without losing x amount, you’d have been in trouble. Using stuff like VaR and Black-Scholes can be extremely useful, but it can get you into a lot of trouble if used incorrectly.
I just think of Long-Term Capital Management. These guys thought that they weren’t taking much risk and couldn’t blow up, but they were leveraged 30:1. Those guys were bad traders that couldn’t manage their money very well. If you’re borrowing $29 for every $1 you have, you’re going to blow up because a small shock has a huge impact.
24. October 2012 at 19:53
Suvy, but prices do follow a random walk, those are the successful tests…
Bob Murphy, impossible, one of them doesn’t even believe in bubbles. (Though I guess that’s how Scott looks when he looks “bubbly” – more like mildly bemused by his surroundings.)
24. October 2012 at 19:55
Tony, of course you don’t believe in EMH, you’d be out of business if you did…
Good to see another Aussie Market Monetarist, though.
24. October 2012 at 20:20
Saturos,
Prices follow a random walk under some situations. Just calculate the Hurst exponent across an entire time series(like a 100 year history of the Dow Jones), you’ll actually see that it fluctuates across time-so in some cases, it’s valid to model a time series as a random walk. I actually have a spreadsheet where I take the historical data of the Dow Jones and I use an estimation of the Hurst exponent. Sure enough, the Hurst exponent actually fluctuates across the time series. Not only does it fluctuate across the time series, but I actually did a statistical test on the fatness of the tails and related it to the Hurst exponent. What I found was that as the tails got fatter as the Hurst exponent increased.
25. October 2012 at 02:29
Those are big increases in asset prices, followed by big decreases. Most people (wrongly) assume those big swings obviously do not represent changes in a rational expectation of fundamental value.
They represent changes in expectations of fundamental value, period. Whether you can call them rational or not, they end up being proven wrong in the end. Humans are fallible.
For some much-smarter-than-average market participant to profit from the collective miscalculation and bring the market back to its long-run fundamentals, there are practical obstacles. I mean, how do you short a million houses ?
I don’t believe income data for individuals is meaningful. Look at consumption.
Income = Consumption + Increase in net worth.
So, OK, for the second term you may quibble about nominal/real, which brings out the other I-word. Deflate by NGDP per capita if you want.
Someone with low income but high consumption is probably going into debt. That may be a problem at some point. Income is relevant.
25. October 2012 at 04:38
Here you go, Suvy – a controlled test of the EMH, supporting the semi-strong version: http://www.sciencedirect.com/science/article/pii/S0148296397000039
25. October 2012 at 04:39
Here’s the pdf, in case you can’t get it: http://ac.els-cdn.com/S0148296397000039/1-s2.0-S0148296397000039-main.pdf?_tid=861fa1fc-1ea0-11e2-90b5-00000aacb362&acdnat=1351168736_1c571d1526cb26b8a46d77f5a94de8ba
25. October 2012 at 04:56
Saturos,
Here’s a paper which calculates the Hurst exponent over different time scales. What you’ll notice is that the Hurst exponent depends on the time scales.
http://www.princeton.edu/~sircar/Public/ARTICLES/bps.pdf
http://en.wikipedia.org/wiki/Hurst_Exponent
Here’s another paper that uses the Hurst exponent and compares the EMH to the Fractal Market Hypothesis in the case of the financial crisis.
http://arxiv.org/pdf/1203.4979.pdf
The Hurst Exponent is a very robust method of measuring correlations in the data. H=.5 means a random walk, H>.5 means the data is correlated, H.5 for most. H<.5 for something like volatility because it is antipersistent(volatility can't increase forever).
25. October 2012 at 05:04
Here’s another paper on this as well which sets up a statistical test. They come up with the conclusion that H is not .5 and the time series exhibits long-range dependence.
https://editorialexpress.com/cgi-bin/conference/download.cgi?db_name=CEF2011&paper_id=128
It’s very difficult to express the extreme deviations from the data without having some sort of correlation between price movements. There are periods where H=.5, but overall; H does not have to be .5.
By the way, I can’t actually read that article. I don’t have a subscription. Let me see if I can do it from my University page.
25. October 2012 at 09:39
Suvy and Saturos,
Efficient markets doesn’t necessarily imply a random walk in prices. That is 1950s – 1970s finance. Modern finance takes into account that risk aversion changes over time.
If you were willing to buy in 2008-2009 when no one wanted to hold risk, you could expect a higher return. The price of risk itself changes over time.
That’s modern finance. It makes sense when you think about it.
Check out this thoroughly EMH consistent paper called “Predictability, The Dog that Didn’t Bark: A Defense of Return Predictability” by Cochrane
http://faculty.chicagobooth.edu/john.cochrane/research/papers/cochrane%20dog%20that%20did%20not%20bark.pdf
25. October 2012 at 14:36
Yes, happy to pitch in for a new suit.. and shirt.. and tie.. and haircut. Sorry Scott, but you have a responsibility to maintain a certain image now . See how Bernanke has lifted his game.
25. October 2012 at 14:39
Hmm, maybe try this.
25. October 2012 at 14:41
Argh! Here it is: http://www.forbes.com/sites/afontevecchia/2011/09/21/dont-expect-the-dollar-to-tank-after-bernanke-co-hit-the-dance-floor/
26. October 2012 at 04:40
Suvy, Even if Australia crashes, the bubble predicters of 2005 will have been wrong. But they’ll claim they were right. The anti-EMH side plays the following game. When there is a sharp price runup, they start calling prices overinflated. Then later they will claim that they were right even if prices never again fall below the price they claim was overinflated! The trick is to keep predicting a crash over and over again. In any volatile asset market like Australia housing there will eventually be a crash. And even if prices never again fall below 2005 levels, they’ll say “see, I was right all along.” It shows tha most EMH critics don’t understand the EMH. The EMH is perfectly consistent with asset price crashes.
26. October 2012 at 15:42
Prof. Sumner,
I’m not against the weak form of the EMH and I still think the EMH and all of the things that come with it are pretty useful(Black-Scholes, VaR, etc). You just have to be careful how it’s used (unlike the guys that ran LTCM).
The only problem that I have with it is the random walk part, but if you had modeled prices as a random walk with a stochastic volatility, that could give you correlations in the data(I’m actually working on a project where I’m trying to do this). So correlated data still might not be completely inconsistent with the EMH.
21. November 2012 at 19:47
[…] it wants more inflation, which the public perceives as bad. This is a marketing problem that was pointed out by Scott Sumner and is one more argument for “re-targeting the Fed” from inflation to nominal […]