« Hair of the Dog | Main | Job creation, job destruction, and the Detroit Three bailouts »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I like that theory (and not just because it makes economists sound a bit less stupid than some of the other theories :-) ). Another way to put it is that economists are insufficiently normative. We rarely say "the price is wrong". Or rather, we only say "the price is wrong" when we have a good theory to explain why the price is wrong, like externality, or monopoly power, or something.

I think the best alternative theory to yours is the "economists are too compartmentalised" theory of Martin Wolf.

I knew that house prices in England (and the level of mortgage debt) seemed too high, but only because I visit England yearly, and have a personal interest in property prices there. I had vaguely heard that US house prices seemed to high, and some Canadian prices also. But other than warning my Mother, a few Brits down the pub, and a couple of people moving to Calgary, I said nothing. I wasn't 100% convinced anyway. Maybe with all the Poles moving to England to get jobs houses really were worth 5X income. And it's almost impossible to build more there anyway.

But I never knew that house prices were so high in so many countries, and I never realised how big an impact it would have on financial markets. That would have required a lot more knowledge of banks' balance sheets than I had. To take one small example: I didn't know that US mortgages were nearly all "non-recourse". I didn't even know what those words meant. It's not so much the depth, but the breadth of knowledge required. We didn't have it.

For example, I've not been able to make head nor tail of stock market data since 1995.

It might make a lot more sense if you look at it a bit differently from what you've learned. Even the current market turmoil is pretty easily explainable when you look through the right lens.

Nick, I think the compartmentalism hypothesis amounts to the same thing: no-one can specialise in everything, so if a model's prediction goes awry, our reaction is to assume we missed something, and that it would all make sense if only we knew more about other fields.

The fatal flaw in this post is that you think that economists are not arrogant. I didn't see a lack of opinion on the housing bubble or any kind of diffidence---I saw a cacophany of opinion, with very little to back up any one of them.

What I didn't see was any real attempt to address the underlying problem of growing inequality. I might respect economists more if I saw a real effort on their part to address growing inequality as a central problem, and addressing the problem of the housing bubble with respect to that. That would have been a science that might have been relevant to ordinary people, but otherwise...it was an arrogant cacophany and remains so.

The world would be a utopia. You guys fucked up bad.
At the time (I started studying stocks right around 1995) it was easy enough to see that flat USA taxes meant a growing amount of capital that was once invested at a safer (pension) risk profile was being controlled by those who could afford to take risks. In retrospect, since 2000 a third element has been the presence of derivative-type traders who have borrowed lots and used technical derivative strategies to....kill time.

Next time someone questions your predictions mail them a mousetrap.

“A sound banker, alas, is not one who foresees danger and avoids it, but one who, when he is ruined, is ruined in a conventional way along with his fellows, so that no one can really blame him.”

I'm sure you know where this comes from.

Application to economists?

I'm inclined to partly agree with Mandos on the "cacaphony of opinion" on house price bubbles. In doing my own post about a month ago on whether Canadian house prices were too high, it rally wasn't clear what the answer was. My gut tells me they will come down, and they are high (in some places) compared to average history, but crunching the numbers doesn't give you a totally clear answer. The rates of return on housing were lower than mortgage interest rates, but higher than bond interest rates, so whether houses are a good or bad deal depends on how you are financing the purchase. And even if I were 100% certain about the theory applied there, and that nothing was excluded (and I'm not), the answers were very sensitive to small changes in the data and the assumptions about how those date would look going forward (e.g. future rents and interest rates).

I disagree on inequality though. Here, economists seem much like the rest of the population. Some people think inequality is the most important public policy problem, but most people don't. Economists are the same. So only a minority of economists spend all or most of their time working on that question.

Phillip: could you explain why you believe this: "....flat USA taxes meant a growing amount of capital that was once invested at a safer (pension) risk profile was being controlled by those who could afford to take risks."?

anon: Keynes of course! I think economists (at least, academic economists) have different incentives to bankers. Unless you are one of the very small number of economists who already has a big reputation, your best strategy is to make a very wild prediction, totally different from the crowd. If you are lucky, and get it right, you win fame and fortune (and you can win some fame and fortune just by making the prediction, while everyone's waiting to see the answer). And if you lose, no big harm done, because you still keep your academic job and drift back into public obscurity. I'm surprised we don't see more books published with titles like "The 2015 crash/boom in snake oil and how to profit from it".

Krugman, Gordon and others wouldn't be asking this question if the bubble in consumer durables (not just houses) had morphed into the worst financial and economic crisis since the Great Depression.

Plenty of economists correctly identified the housing bubble but few if any identified the collapse of the financial sector. It is not clear to me that economists should be anymore gifted at identifying major structural breaks or 'Black Swan' events than any other professional group. As Nick Rowe suggests, specialization and compartmentalization is a factor but there is more.

Inadequate theory? Economists are most comfortable with markets that are out-of-equilibrium or in disequilibrium for long periods of time. Exchange rates are a classic example. Only the most naive or mule-headed New Classical macroeconomic ideologues believe that markets are always in equilibrium or market-clearing behaviour is always instantaneous.

Furthermore, behavioural economics and finance have furnished us with plenty of alternative stories in recent years.

Then there are the incentives or rather the lack of incentives of professional groups of economists to stand on roof tops and warn people that they need to protect themselves from themselves.

Public sector economists? For the most part independent thought and action have been beaten out of them a long time ago. Either that or they quit and moved on. Many voters augment their private wealth during bubbles. Why should elected political leaders risk interrupting the party?

Business sector economists? They are busy making money during booms and bubbles.

Academic economists? The system rewards original thought and research. Peer review and criticism is typically reserved for in-camera situations away from the eyes of the public.

Then there are the underlying ideologies and world views that economists like Paul Krugman share with many others. It is generally materialistic, pro-growth, still favourable to open-access regimes and still very favourable to aggressive colonialism.

I would argue that you can almost draw a straight line from the 1967 war of conquest and occupation to Intifada I to the miserable, doomed-to-fail Oslo Peace Accords to Intifada II to the September 11 low-budget box cutter cuts on the World Trade Centre to the invasion and occupation of both Iraq and Afghanistan to negative real federal reserve interest rates to record-high oil prices and now the current collapse.

Krugman has gone on at length about the folly of invading and occupying Iraq but has never breathed a word about the territorial expansion of Israel. Does that mean that Krugman exhibits a high willingness to pay in foregone wealth and blood for the Occupied Territories? Or does it suggest that he is being 'careful' to avoid a 'delicate subject'? (I suspect that it is the latter.)

"The fatal flaw in this post is that you think that economists are not arrogant. I didn't see a lack of opinion on the housing bubble or any kind of diffidence---I saw a cacophany of opinion, with very little to back up any one of them."

Most people confuse the economic counsel of stock traders, investment advisors and "financial gurus" with that of actual economists.

20+ years of falling inflation lowers interests rates, increasing the value of homes
The mortgage market changes from loans banks keep to loans banks sell for fees
Running out of good buyers, banks lower standards to keep the fees coming in
Eventually gimmicky loans allow buyers to bid up home value past income ratios
Market falls as marginal buyers try to sell into rising interest rates and recession
Personally, I think it’s mostly a self inflicted wound of the banking industry

I'm indifferent to the fatal flaws of economists. The substantive flaws are manifested by people who take their advice without understanding a word they've said. Decision-makers can either insist on greater clarity or economists can work on popularizing their largely mystifying system of thought.

Are you saying that you were wrong to conclude that Allais' Paradox invalidate the expected utility (EU) model of preferences? Are you saying that the EU model is right and that people are making mistakes or are irrational when they don't follow it? Models are not there to show what people should do, they are there to try to describe/predict what they do. Maybe you shouldn't be reconsidering only models, maybe you should be reconsidering key assumptions. For example, people may care about more than just financial wealth (e.g. status, competition, relative performance)... or what proxies they use for financial wealth when making decisions.

Well, that's the thing, isn't it? When we looked at the US housing market, our instinct was to reconsider the assumptions of a model that couldn't explain what we saw. We were more likely to say "Maybe I'm missing something" than "Millions of people are crazy-stupid and it'll all blow up horribly."

How about the fatal flaw of believing that the measure of the thing is the thing itself? As a former BLS price index economist, I've been arguing for most of a decade (and especially after 3 years of under 1.5% of Fed Funds Rates) that neither CPI nor PCE are accurate measures of inflation, and that many monetary measures themsleves are flawed. Look at indexes of home sale prices, equity prices, debt-asset prices and healthcare costs for the past 10 years and try to reconcile with what CPI has been telling us. They cannot both accurately descsribing the same thing. As Mr Greenspan said in an (often misunderstood) speech:

"At different times in our history a varying set of simple indicators seemed successfully to summarize the state of monetary policy and its relationship to the economy. Thus, during the decades of the 1970s and 1980s, trends in money supply, first M1, then M2, were useful guides... Unfortunately, money supply trends veered off path several years ago as a useful summary of the overall economy... One factor that will continue to complicate that task is the increasing difficulty of pinning down the notion of what constitutes a stable general price level... As the century draws to a close, the simple notion of price has turned decidedly ambiguous. What is the price of a unit of software or a legal opinion? How does one evaluate the price change of a cataract operation over a ten-year period when the nature of the procedure and its impact on the patient changes so radically. Indeed, how will we measure inflation, and the associated financial and real implications, in the twenty-first century when our data--using current techniques--could become increasingly less adequate to trace price trends over time?..."

"How do we know when irrational exuberance has unduly escalated asset values, which then become subject to unexpected and prolonged contractions as they have in Japan over the past decade? And how do we factor that assessment into monetary policy?... We should not underestimate or become complacent about the complexity of the interactions of asset markets and the economy. Thus, evaluating shifts in balance sheets generally, and in asset prices particularly, must be an integral part of the development of monetary policy... Doubtless, we will develop new techniques of price measurement to unearth them as the years go on. It is crucial that we do, for inflation can destabilize an economy even if faulty price indexes fail to reveal it..."

In the past 15 years, we have not lived up to the challenge proposed by Mr Greenspan's most famous speech.

We need an index that accurately is designed from the top down to be used for monetary inflation measurement, whether that inflation occurs in consumer prices, asset markets, debt, or eslewhere.

We need measures of money and money substitutes that reflect the reality of the globalized financial world since the mid 80's.

We need a predictive mesaure of money demand for the monetary authorities.

You may be right that the unwillingness to acknowledge that people don't act 'rationally' doesn't help economists make good predictions, but I think you need to look deeper to find the fatal flaw in economics.

As for the housing bubble, I thought it was pretty obvious to all, no? I mean, the Economist did their cover story on the Global Housing bubble in 2005 and it seemed like a pretty open and shut case.

The fact (in Vancouver anyway) all media housing price forecasts still come from those same people who've been wrong all along suggests another way in which the incentives don't seem to be very well aligned for identification of bubbles.

One additional problem for economists who want to identify a bubble is that it is so hard to get the timing right which lowers the incentives to point out the obvious, knowing that it might take a while for what has to happen to happen.

Finally, to address your post more directly, there is a certain arrogance in believing that only mathematical models can explain or predict economic events, is there not?

Mario Sanchez: I would much prefer to argue about whether CPI (say) is a *useful* measure of inflation than argue over whether it is an *accurate* measure of inflation. The first debate leads us into good directions, where we recognise that an index that is useful for one purpose may be less useful for a different purpose, and it leads us to recognise that we first need a theory of how (say) asset prices affect the economy before we decide what index number would best serve as an empiricical proxy for the concepts in that theory. The second debate leads us into a semantic/definitional quagmire.

So, should central banks target assets prices, consumer prices, or a mix? If we can answer that question, then we can set about constructing a useful index number for central banks to target (and whether or not we call it *inflation* doesn't really matter).

Declan: this is what always puzzles me about identifying bubbles (in real time): if we could identify them, we could presumably justify our reasoning, and demonstrate to anyone that house prices are in a bubble, then why would the bubble continue? I know there are some theories out there of "rational" bubbles, where everybody knows it's a bubble, but I don't (from my limited knowledge) find them very convincing.

On your point:"....there is a certain arrogance in believing that only mathematical models can explain or predict economic events, is there not?" I hate maths as much as the next maths-challenged economist, but I don't see how you can talk about house prices without doing at least some maths. Prices, rent/price ratios, price/income ratios, rates of return, etc. all seem to matter. There is a certain arrogance in believing we can hope to explain or predict economic events, admittedly, but that arrogance seems to be the price of admission to any social science. If we didn't believe that, why would we bother?

From what I can tell, this supposed excess of diffidence is a complete fiction.

example: the great oil bubble of 2008. When oil was at an absurd price ($145) Krugman just completely refused to believe that the price could be out of line with supply and demand, because according to the theories that he accepts as bedrock-truth, there would have to be massive hoards of oil for the price to be driven up beyond what supply and demand justified. It was an obviously a bubble, a mispricing that requires no hoarding, but Krugman's belief in economic models completely blinded him to this. He has still yet to retrospectively discuss what actually happened during the oil bubble, because the explanation does not fit with the models he was sketching up in his blog. It seems that even the best economists in the world are blinded by their models, and still have no way to differentiate a bubble inflated price from one that is justified by the balance of supply and demand. IMO between the reality before their eyes ( a feverish peak-oil hysteria/bubble) and a simplistic model that offers some form of certainty, even one of the best economists in the world will choose the hopelessly wrong but 'certain' model every time, rather than attempting to understand things like psychology, culture or politics that are complex and uncertain.

Nick, yes certainly we need to use math to help analyze if there is a bubble or not, all the ratios you suggest and more, but what I was getting at was that these mathematical tools are only useful in conjunction with history, psychology, keen observation and other tools can't really (as far as I've ever seen) be factored into mathematical models. The arrogance I see is the belief that a (mathematical) model, built using our current analytical toolkit and not supplemented with the other areas of knowledge I mention above, can capture enough of the situation to make any sort of useful predictions.

It was basically, the same point that 'rrr' makes in the comment above mine.

Speaking of semantics.... prices can overshoot long-run equilibrium prices without being in a so-called 'bubble'.

Oil prices were not in bubble territory this past summer despite having overshot well beyond what long-run fundamentals would support. The first piece of evidence was the reluctance of equity markets to bid up the market values of exploration and production companies to reflect oil prices greater than US$120/barrel. The second piece of evidence was the persistently modest if not tight forward capacity in higher grade oil markets.

As Rowe asks, should central banks take into account asset values into the indicator mix? Perhaps.

Should central banks target a stable low inflation rate or average price level (e.g., 0 to 1% headline inflation)? Most definitely. But such a policy is only workable if broader society agrees that a policy of growth at any cost is fraught with peril. We are not there yet.

Declan, I don't see how your point contradicts the post. Models oblige us to make explicit the assumptions that are required to reach our conclusions. Wouldn't it be even more arrogant to refuse to make those assumptions clear, and to claim that one's (unarticulated) understanding of these other things is somehow reason enough to accept the conclusion?

If you're saying that the models aren't perfect and need to be revised, well, that's the instinctive reaction of a typical economist. But that still requires making explicit the assumptions that you're willing to make.

What I'm saying is that economics *can't be modelled* (in the sense of 'modelled' that economists seem to generally use) and the arrogance of economists is that they think it can.

That was based on me interpreting your post to mean that economists were too diffident to believe that their models were right even though people don't behave like their models say they should, and that if only they had had more faith in their models, they could have made better predictions (e.g. seen the bubble coming).

Re-reading your post, I think I may have misinterpreted and what you really meant was that if people's behaviour doesn't match what the model predicts you simply throw up your hands and don't make any comment at all.

That certainly doesn't square with my experience of the economists who work for the IMF and for central banks and as government advisors and who fill the pages of the newspapers, but even if I accept that that is the majority behaviour patttern amongst economists, it still seems to return to a misguided belief that economics = models.

Except now we have defeatism (diffidence if you prefer :) replacing arrogance as the reaction to the clear reality that the models don't work very well on their own - and the unwillingness to combine mathematical metrics like the ones Nick listed with knowledge and approaches from other disciplines to make a complete assessment of the economic situation remains as the primary problem.

Have you considered that the reason you can't make heads or tails of the market is because you attempt to connect economic indicators to market movements, when they have only an abstract relationship?

Dr. Gordon-

I believe that diffidence somewhat captures what might be the phsychological state of some economists: Their lack of self-confidence coloured by their distrust of market mechanics. However, I strongly assert that all the information necessary to predict the current world-wide banking crisis has been readily available for years. Any competent economist should have been able to connect the dots. Failure to publicly communicate the dangers and impending destruction reflects either gross incompetence or ethical depravity.

In my opinion, most professional economists feel that they are rewarded by supporting the Status Quo. The big money is made in securitization, asset leveraging, and asset inflation. Far be it that professional economists should suggest that the table be cleared, just as the orgy is getting into high gear.

What I'm saying is that economics *can't be modelled* (in the sense of 'modelled' that economists seem to generally use) and the arrogance of economists is that they think it can.

I don't know how to respond to this. To my mind, all alternatives to models are variations on the theme of "Because I said so, dammit." Moving in that direction cannot be considered becoming less arrogant.

Incorrect use of models is worse than no models at all. It is "because I said so, dammit" dressed up with a false appeal to science. "Because I said so, dammit" at least allows us to examine the political and moral motivations of a proposal. A model in this context just becomes a smokescreen, or something like a squid squirting ink.

I suppose. So the correct response is to come up with a better model. Show us how it's done, sport.

No, because it is unmodelable under present understandings. We don't have the mathematics to model complex systems like these. If we did, we'd also have made enormously more progress in developmental biology, etc, etc than we have.

Economics is just an extension of cellular automata, as is the study of any complex system of very many autonomous but interdependent agents. We can't actually deduce from starting position how the ending position will line up. The model has to *already have been run* before you can say anything about what happened. In some cases, you can develop statistical models by running the experiment millions of times. Then, you can make *very* simple predictions.

We don't have the luxury of running a real-world economy many millions of times. That is the economist's folly.

Consequently, when I read the writings of an economist, I ask first, "Who does this serve? What does this mask?"

In other words, it is not a science, it is a moral argument about winners and losers that can only be approached via very simple "common sense" understandings. For instance, wherever possible, left unchecked, the owners of capital will seek to pay lower wages.

At some point, an economist will tell us about Arrow's Theorem. Yawn.

Quadruple post (one was redundant, of course)!

By "unmodelable", I mean that there is no relevant means to evaluate the merits of one of these sophisticated models over another, not that you can't build a model as such. You can build a model of anything if you want to. Goodness knows there have been attempts to model things like, eg, poetry. It doesn't mean that it's useful.

there is no relevant means to evaluate the merits of one of these sophisticated models over another

A very odd claim to make on a blog that has an entire category set aside for econometrics. The ability to explain/predict observable data is my criterion for model evaluation.

Not merely the ability to explain and predict observable data but:

1. To do so reliably.

2. To be able to predict what changes in the system will lead to what later outcomes.

Very few sciences dare to claim that their models respect 1 and 2 in any but the simplest cases, especially not the "cellular automata" type sciences. No respectable evolutionary biologist, for example, would be willing to talk about what would have happened if the asteroid had hit a million years later, or not at all. Even physics makes robust claims for 1 and 2 only under very simple idealized scenarios involving very small numbers of objects. Certainly, give any nontrivial system and most physicists throw up their hands.

Economists, however, not only seem to believe that they have models that respect 1 and 2, but also believe that because of this, they have a greater authority to prescribe policy alternatives. That's the height of arrogance. Perhaps economists can make statements about very simple ideal systems in the way that a physicist can, but that's about it. Any claim beyond that is reliably a form of obfuscation of a political and moral claim, and treads very close indeed to malpractice if taken as a form of policy engineering.

"Phillip: could you explain why you believe this: "....flat USA taxes meant a growing amount of capital that was once invested at a safer (pension) risk profile was being controlled by those who could afford to take risks."?"

Nothing Earth shattering. Flatter taxes in the USA since Reagan have meant the rich get richer. When the poor have investments they generally have it in a risk-averse pension; usually heavy on bonds, bills, and dividends and almost no growth stock exposure. Rich people can afford to diversify their own trading accounts. You need $60000 to play options game and even more to play hedge game. Saying poorer people have more pension exposure and rich people are more likely to use their networths to trade. This increases voliatility in markets. Maybe effect is minor compared to globalization and decreased derivatives oversight. I googled around a bit to find a paper and couldn't use the right keywords. Maybe a thesis topic for a University of Ottawa student: do flatter taxes decrease fundamental trading strategies and increase volatility?

I guess some professional economists are called on to make forecasts for budgeting purposes and this explains, at least in part, all these accusations of "arrogance" and similar.

Others test in-sample predictions with a view to verifying hypotheses and some lay people apparently fail to grasp the difference.

The only macroeconomic fluctuations model with any out-of-sample predictive power is the yield curve forecasting technology, and even then that seems to only work for horizons of 9 to 12 months, and not always. To make matters worse, yield curves furnish ordinal predictions.

Me? I like to toss caribou shoulder blades and see how they land in the sand. I simply make up the necessary assumptions as gravity pulls the bones to earth. Nobody is the wiser, above all this gaucho. (Needless to say, no disrespect is intended for the Montagnais and Innu brothers and sisters.)

Others test in-sample predictions with a view to verifying hypotheses and some lay people apparently fail to grasp the difference.

Alright, this sounds like a fancy way of saying "economists explain history". This post, I believe, is about the relevance of economics to policy discussions---a prognostic role. I'm OK with a discipline that claims to try to explain history. Biologists often claim to do the same thing. Historical linguists, just so.

Mandos. Let's put the ball in your camp.

Does 'risk management' have a role in policy? And if so, how does forecasting, the relative accuracy of forecasting, and similar considerations play into 'risk management'?

In very limited ways. For instance, we know historically that in some places, there is a risk of earthquakes. The cost of earthquakes in some places can be high (particularly in lives). Consequently, we should have policies that ensure earthquake-aware building techniques in those places where the risk of earthquakes is high.

Considering that some very respectable people online (Roubini, DeLong, Krugman---who I often like, Mankiw, and so on and so forth and, heck, Bernanke) seem to have very divergent views on what to do about economic crises, well, it's not clear to me that economics as a discipline is really analogous to the limited forms of prognostic risk management I mention above. I really do doubt that, eg, seismologists are going to argue that parts of California are at risk for serious earthquakes.

Is there an analogy to earthquake risks that I'm missing in economics? It's entirely possible.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad