This must be the period of soul searching, with the Economist engaging upon multi-article exegeses on where mainstream macro went wrong [1], [2], [3]. Alternatively, I think this is a happy time for some economists outside the (perceived) mainstream, who can now chortle “I told you so”. One recent example is by Mario Rizzo.
The objective facts are far easier to handle in the models than the shifting, subjective expectations of people trying to deal with radically uncertain futures. This is what may get reflected in financial markets. Attempting to understand all of this requires conceding that some knowledge will be imprecise and will lie outside of the box (model). The model is simply a toy that can be thrown out when it no longer suits. This means that it is indeed possible to have valuable knowledge outside of hyper-models (although, of course, all thinking proceeds in terms of assumptions and simplifications).
But this will give the “scientists” among us headaches. As John Maynard Keynes famously said about the econometrician Jan Tinbergen, “[H]e is much more interested in getting on with the job than in spending time in deciding whether the job is worth getting on with.”
As long as this is the dominant attitude, macroeconomics will remain “other-wordly.” Instead, the way to greater realism is through more attention to the methodology of science and to whether “the job is worth getting on with.” Paradoxically, greater philosophical sophistication would put economists is closer touch with the real world. (Or so I hope.)
Lean, mean DSGE machines?
Reading the recent characterizations of Ph.D. education in our top departments, one would conclude that all one ever learned in a program is how to write out and calibrate dynamic stochastic general equilibrium (DSGE) models, or for the older among us, calibrate a real business cycle model. I have to say that this all seems a little like an all too convenient caricature (and, as I have said repeatedly in the past, these types of models have led to important insights for issues besides crises [4]).
I won’t deny that in the past 20 years, I haven’t seen more than a few models that struck me as pretty irrelevant for analysis of real world issues. But I think that some mathematical training, and the use of models, is essential to economic analysis. After all, one can think of completely irrelevant frameworks for looking at the world even without a model, just as one can with a model.
Furthermore, perhaps my experience in a Ph.D. program is atypical but I don’t remember being forced into a particular mode of analysis in writing my dissertation (University of California, Berkeley, 1985-1991). In macro/international/econometrics, my teachers included Roger Craine, George Akerlof, Jeffrey Frankel, Andy Rose, and Richard Meese. We studied Euler equations as well as the market for lemons. We knew what Arrow-Debreu markets were, but we also learned about the Great Depression (from Bernanke’s paper as well as Friedman and Schwartz). The time series econometrics taught did not presuppose optimizing behavior. We even studied models with sticky prices (gasp!). Doesn’t sound too doctrinaire to me.
So what was a common theme in the curriculum? For me, the defining feature in thinking about what model to use was whether the analysis answered the question posed, and whether the question posed was of interest. Now, whenever I read a dissertation prospectus, the key question I ask the student is: “What is the question being asked?”, not “What is the methodology?” (Admittedly, the subdisciplines have different “characters”, as alluded to by Paul Krugman; my focus was open economy macroeconomics, rather than macroeconomics/monetary economics.).
How monolithic?
I wonder if indeed the macroeconomic mainstream is as monolithic as conveyed by various observers. For instance, one certainly perceives a certain homogeneity amongst Ph.D.’s trained at certain universities. And there’s a certain similarity in the mode of analysis preferred by economists in financial firms. Since the financial press tends to focus on Wall Street economists, one gets a misleading impression regarding the degree of uniformity of views.
To make this more concrete, let’s consider whether academic economists differ in their views regarding the economy, as compared to those in the financial sector. I have some indirect evidence, pulled from Dilbert’s survey of economists in the American Economics Association (see also [5]). Scott Adams, with the assistance of Joshua Libresco of the OSR Group, was kind enough to have the stats pulled. Last summer, academic economists believed that a President Obama administration would promise more progress on the economy than a McCain administration, by a 2 to 1 margin (n=314); in contrast financial sector economists were equally split. The sample in the latter case is quite small (n=29) (I dropped the undecided/no difference responses). Nonetheless, a difference in means test (recalling the variance of a binomial is (1/4)/n) rejects the null hypothesis of equality at 6%, using a two tailed test.
(As an aside, this finding further suggests that when the WSJ says most economist oppose a second stimulus, that probably characterizes Wall Street economists better than all AEA economists. Even for Wall Street economists, it’s interesting to note that a majority of economists feel the stimulus package has already improved economics prospects, and will have a bigger impact in subsequent months. Hence, opposition to a second stimulus among WSJ-surveyed economists is not necessarily rooted in skepticism about the aggregate demand enhancing effect of the ARRA. (One would need a cross-tabulation of responses to verify that assertion.))
Concluding thoughts
If my conjecture is correct, then the supposed failure of macroeconomics is more the failure of macroeconomics as described in the popular press, rather than of the discipline itself (after all, Joseph Stiglitz is as much of the economics discipline, if not more, than Eugene Fama.) My conclusion: Not quite time to jettison the apparatus of modern macroeconomics.
For a less personal perspective, see Brad Delong and Paul Krugman, as well as my April post on “macroeconomic schisms”.
Update 9am 7/21: See also Mark Thoma’s observations.
Technorati Tags: macroeconomics, DSGE,
RBC, rational expectations.
Why is no one interested in how India had a conservative Macro and Bank policy and avoided all of the problems. Perhaps Mr. Reddy is the macro economist with the results to show for his good ideas…
I had a good laugh when I heard that AER has offered authors with papers scheduled for publication the chance to withdraw them if they predict a strong economy — and no one accepted. Getting published is more important than getting it right.
The macro profession failed to see the crisis coming, and I think it’s because modern macro is indeed an apparatus, a body of techniques that’s first of all for tackling academic puzzles and publishing, rather than a body of theory that deals with the behavior of real humans in any important way. Summers’ “The Scientific Illusion in Modern Macro” deserves a re-reading.
Menzie,
I trust that you don’t have the time one your hands to do this, but a truly rigorous analysis of the effectiveness of your peers would be as follows:
-Identify the number of papers that would have been useful in predicting the bubble bursting.
-Divide those by the total number of papers produced.
-Compare that fraction to the number of papers that would have MISLED someone to not predict the bubble bursting, divided by the number of total papers.
That comparison will give you a more accurate picture of how the economics profession handles itself.
That’s a thoughtful piece.
I think macroeconomics is fine. The problem, I think, is more with macroeconomists. These concerns run along two lines.
The first is the ‘peacetime military’ syndrome. We hadn’t had a really bad recession in such a long time, the skills went rusty, both on an individual and institutional level; indeed, I think we really didn’t know how to think about the issue. I, for example, was running strategy for a small investment bank at the time, and I didn’t think to sit down and calculate the oil price which would dump the economy into recession. It’s not really a big deal, and having lived through both oil crises, I was certainly aware that oil could be a problem. But I didn’t think to do the actual calculation. (I have since.) The failure was not analytical, it was one of mindset. And I don’t think I was alone in this.
The second issue is one of character, and in particular, risk assumption. For example, the Econobrower recession index is at 99.5% when an economist on Bloomberg today called the trough for Q2. I also think the trough was Q2. Jim has defended the index as based on historical data. Fair enough. The index, on that basis, is accurate. But is it right, and it is useful? Probably not, if your orientation is forward-looking. But it comes down to an appetite for risk. Do you want to be defensible or do you want to be actionable? Is an economist a historian or a sentinel?
If there was a failure of the profession, it was that too few economists thought of themselves as sentinels at a time of impending crisis.
So, to my way of thinking, the issue is not one of analytics, it is one of mission. The objective function itself, and not just its optimization, is worthy of consideration.
Relax…economists are professional scapegoats.
The profession exists to provide behavioral excuses on the way up, and cover on the way down. If any one of you could predict economic movement EVEN ONE DAY ahead, you could place bets, make a fortune,
and retire. So relax…it’s time to earn your pay…that’s all. Baaah.
Sorry to be less than polite: I think most economists are epistemologically / methodologically very naive and unprepared. How many PhD courses have Kuhn as required reading?
My take on things is that what you are looking for is ***explanations***. Models are a tool for checking that the mechanics that underpin the explanation function properly. But all the incentives in the profession are skewed towards modelling. (When was the last time you saw a paper with ideas and no equations? the last one I recall is “Exogeneity” by Hendry, circa 1982…)
David Pearson: I think your exercise is ill conceived; shouldn’t I be interested in growth, or productivity, or trade balances, even when crises occur occasionally? An analogy would be — should I divide the number of papers that deal with diabetes by all the papers in the medical literature to assess whether doctors are studying the right things?
Steve Kopits: I think economics, as a social science, has multiple missions, only one of which is to be a “sentinel”, as you put it. A related point is that most of economics is not focused on prediction/forecasting, even though a reader of the financial press might think that is the case.
KnotRP: You are mis-apprehending the goal of economics; it’s not to make money (we leave that to business schools). The main objective of economics to understand how scarce resources are allocated.
JJ Saenz: You set up a false dichotomy. To me, if you have a framework that’s internally consistent, even without equations, that’s modeling. Equating models with equations is, dare I say it, naive. By the way, one doesn’t have to read the entire Kuhn book to understand the author’s thesis (unless one is dense).
Your analogy is telling: diabetes is a major killer, and the bursting bubble has resulted in the biggest recession since the Great Depression. Surely it is relevant to find out whether doctors/economists focused on predicting or preventing both conditions.
Also, major economists, including those resident at the Fed, analyzed the appropriateness of housing prices based on elevated terminal growth rates in an NPV calculation. This was an elementary error — assume house prices rise forever more, and high house price valuations are justified. Then there were spurious analyzes showing that rising subprime delinquencies were not a risk because regressions showed that they were driven by unemployment, which of course was low. Then there were the number of economists that insisted that the banking system was better able to withstand credit cycles as they were so well-capitalized and geographically diversified.
You see, I could go and on. I know because I was actively betting on a housing collapse and wider, global, credit contagion from the summer of 2005 on. In my research I explored what economists — academic, think tank, wall street — had to say, to help me reach conclusions on the situation. I have to say that almost every piece of research I came across made me want to abandon my investment posture. In other words, I succeeded by ignoring economists. That, at least, I know to be a fact.
I recommend this article.
It is titled “No-one saw this coming?” Balderdash!
http://www.debtdeflation.com/blogs/2009/07/15/no-one-saw-this-coming-balderdash/
My problem with modern macroeconomics as I understand it is that there is no involuntary unemployment in the models. What can these models say to explain 4.4 million individuals unemployed for 27+ weeks.
The fact many economic time series display turbulent characteristics gives a good answer to the problems of scientists trying to predict the economic future.
Turbulence is characterized by fat tail (non-gausian, broader than gausian) distributions (of e.g log returns on stock prices) in small time scales (relative to some driving information/energy input) process time scale) that include large deviations from average as a MUST.
To me the problem of predictive force of economic science seems to be that turbulence is by definition interaction of rational and irrational (observational, fad like) driving forces , and none of these is possible to eliminate without eliminating the markets themselves.
The current fast upswing in stock prices will most likely illustrate the force of observation over the force of econometrics- as it already happened after Lehmans collapse, just in opposite direction.
What economists are afraid of, is to recognize clearly that their short term predictive abilities are as bad as those of physicists trying to predict exact values of turbulent motion in small time scales because of the nature of the subject. Its impossible to separate it from the irrational subjects, Lagrangian particles having their own destiny but observing others and reacting.
But they should not be afraid since it is exactly such processes that will always require soothsayers.
KnotRP: You are mis-apprehending the goal of economics; it’s not to make money (we leave that to business schools). The main objective of economics to understand how scarce resources are allocated.
To the first response — yes, you leave the money making to business schools…and the business execs use economists in the way I described.
To the second response….misallocation cannot be identified before it’s happened and misallocation is not typically reversible, so what’s left to understand, and why is it useful when there is no means to apply the understanding?
This sounds more like self-entertainment than a profession, and I expect it’d pay at that rate if
business execs didn’t need a scapegoat handy.
Accept that any form of economics, with or without numbers, is a junior branch of history and this little difficulty disappears. As historians, the fact that economists en masse have near zero forecasting ability doesn’t matter.
No one expects historians to predict what will happen next year. But historical insights are of great value in understanding future possibilities. And some individual historian may hit the jackpot with a specific forecast. Picking which historian will be right ex ante is a little tricky, of course.
Why are Economists beating themslves up over a banking crisis? After all, the bankers fired those economists that pointed out the risks, and promoted geeks with no common sense to tell them how clever they were, or better still, geeks whose maths they didnt understand but which made their strategies sound very clever indeed.
Only for the geeks to find out that when one billion rush for the exit door it isnt big enough.
KnotRP – I wish it were true; in my experience a winning run merely scares ones bosses and leads to professional isolation.
Real sciences are about humility and probity one may fake his results but his assumptions.
As regards economics,what best representation of an organised social amnesia (unable or unwilling to draw lessons from the past)than this document?
http://www.nber.org/chapters/c2313.pdf
The economists temptations to bypass reality is reflected in the affirmed social definition of the NBER a non profit membership corporation for impartial studies…..
National Bureau
of Economic Research
BULLETIN 50
APRIL 18, 1934
A NON.PROFIT MEMBERSHiP CORPORATION FOR IMPARTIAL STUDiES IN ECONOMIC AND SOCIAL sciences
Do not despair it will happen again and again!
David Pearson, I like the medical analogy… People die every day from diabetes, but we do not call it failure of modern medicine.
Menzie, in answering your question, you focus too much on your peers… PhD economists. While it is people with no more than an undergraduate degree that make most of the decisions… to buy or sell some AAA CDO for their firm’s pension fund… Perhaps it is a failure of economics and financial education… I think Shiller feels this way.
David Pearson,
To take your “rigorous analysis” a little further it would be instructive to take papers submitted per discipline compared to those submitted by discipline that actually predicted the crash. Then discard those who have a low score and study those who have a high score.
How did Austrian economists rate against Keynesian economists, or how did monetarist economists rate against communist economists?
DickF: In conducting your proposed experiment, one would want to also tabulate “false positives”, as well as true positives. Put in plain English (since I know you dislike the language of statistics), “even a broken clock is right twice a day…”
malcolm: I suspect your reading of modern macroeconomics is a little too circumscribed if you think all models assume zero involuntary unemployment.
Ivars: Most macroeconomists I know are aware of the limited ability to forecast. So I think we acknowledge the fact that you assert we’re afraid of. Those who work on asset prices are typically even more humble about the ability to predict than others. On the other hand, there is a group of economists that are paid to forecast; they of course do not have an incentive to publicize the fact that some variables are hard to predict, but I suspect they are cognizant of the difficulties.
MikeR: I accept that your point may explain in part why the system behaved the way it did. But I wanted to focus on the argument that Ph.D. economists were so narrow in their training that they all failed to predict a crisis.
As I understand it, the goal of (many) modern macroeconomic studies is to write down an internally consistent framework that is simultaneously compatible with balanced growth facts over the long run (stable ratios of C/Y and K/Y and trendless hours worked per-capita) and yet also compatible with the business cycle facts (C, I, hours pro-cyclical, with C less volatile than Y, I more volatile than Y, and hours as volatile as Y).
It took a while for the profession to figure this out, but now most macroeconomics has this as a goal. Including macroeconomics at Princeton (Pat Kehoe)! Read the debate — can’t remember where published, but around 1985 — between Ed Prescott and Larry Summers. It’s very enlightening.
What makes this study difficult is that whatever mechanisms cause occasionally big fluctuations to macro variables in the short run have to basically wash out over the long run. This puts many restrictions on preferences, i.e. how people trade off utility from consumption or leisure today with utility from consumption and leisure in the future.
I’m sympathetic to the idea that our models are missing some key ingredient that would have been helpful in thinking through this current episode. It could be that all major recessions etc start out with a banking panic (Finn Kydland’s Nobel speech I believe made mention of the fact that he thought some “technology shocks” were reduced form for banking panics). Fine. Then let’s figure out an internally consistent way of including banking panics in our growth models. But then we need to test these models: They must deliver balanced growth in the long run, cycles in the short run.
I gets calls from journalists all the time that have no formal training in economics or finance. I have no sympathy, then, for journalists telling me that what I call macroeconomics is a waste of time. I suspect this sort of thing is why most newspapers are going broke: The market is speaking loudly.
Morris A. Davis
Wisconsin School of Business
Department of Real Estate and Urban Land Economics
Humility is always a good lesson to learn, in or out of academia. At least some of those who needed to learn it, did.
But the Economist’s summary of two main debates, on efficient markets vs animal spirits, and Keynesians vs “purists”, left me disappointed. On the former, is it so hard to see how bubbles are made by rationally selfish behavior? On the latter, isn’t it clear that now, as in the late 70s, the US is faced with the ugly necessity of needing to adjust real wages downward, and is going through a phase in which it tries to politically engineer some escape from that necessity, which only drags out the process and delays recovery?
Professor, this train wreck was very easy to foresee, and required nothing more than simple tablulation of household debt to GDP over ’22-’29-’33-’39 and comparing it to ’81-today.
Some economists did that simple tabulation, but not many.
As an outsider looking in and a reader of a range of economists’ views daily, my sense is that economists have little appreciation for pre WWII history and statistics and have little appreciation for anything but Keynesian and Monetarist-Keynesian mechanics and models.
I came to appreciate the utility of complex mathematical tools in my U. of C. coursework, but have come to the conclusion that big simple concepts — e.g., household debt to GDP — appear to be the real driver of things.
So, like David P., I have an all-in bet that this mess lasts years, until household debt to GDP comes back in line with historical norms.
So far, I’ve been proven right, and relish that I have bested Ben ‘This is contained to subprime’ Bernanke, posterchild for the Monetarist-Keynesians.
Professor, I’ll be forward, and recommend that you expand your horizons; I recommend Rothbard’s ‘America’s Great Depression.’ Fascinating, and frighteningly prescient. It will be a quick, easy read for you.
Who cares what journalists think. Consider a majority community of investors, business owners, consumers, and financial professionals wondering what kind of bizarre metaphysical language macroeconomists are using to describe phenomena that has been predefined by convention and free revelation for decades, if not centuries. Nobody is against measuring things. The trouble is with the macroeconomists’ voluntary selection of variables and constructs that are, as it turns out, wildly deconfigured from the conventions (“facts”) they propose to measure. The notion that an entire generation of firy-nostriled pit traders and clueless MBAs have spent the last couple decades urinating in the profession’s general direction tends to say more about macro than it does the subject they claim to study. The market is speaking loudly indeed. And macroeconomists continue to ignore it, ruminating on “internally consistent frameworks” as though anyone in the business community hasn’t already provided the framework thousands of thousands of times over, publicly and privately.
Macroeconomics is a sunk cost dilemma of wondrous proportions. The investment we cannot bear to cut loose is a grotesque cesspool of taxonomy that insures against its own improvement by innovation. Seasoned businessmen are forced to keep their economists locked out of any conspicuous decisionmaking for sake of shame, while the nation’s most promising young talent has its finances held hostage and its noses pressed incessantly into canonized literature, for years upon years of esoteric indoctrination. At some point the only thing macroeconomics will have to do with business and human behavior is its effective identity with the infested upper ranks of too-big-too-fail corporations. Constantly identifying problems and answering them by restating the questions. This farce of a self-critique has been going on for decades. When will it stop. Too many brilliant minds have been lured into this ridiculous sham of a “science,” with its analogies to medicine and technology. The pride is appalling. The defiance is incredible. It is like watching heroine addicts rationalize their behavior. WOW!
Menzie,
So let me make sure I understand you plain language. You are saying that those who have a bad track record concerning the current crisis are worse than stopped clocks? Doesn’t that mean that their analysis is actually destructive, I mean a stopped clock is right twice a day yet they are not even right once?
Enough of the cuteness. If you understand classical economics you know exactly why we had a crisis. Following Paul Krugman’s advice too much money was pumped into the system. Following Barnie Frank’s advice this excess liquidity was diverted to create the real estate and credit crisis.
Classical economists saw it coming. Demand siders are still stealing savings so the government can substitute for “consummers.”
“I won’t deny that in the past 20 years, I haven’t seen more than a few models that struck me as pretty irrelevant for analysis of real world issues. But I think that some mathematical training, and the use of models, is essential to economic analysis. After all, one can think of completely irrelevant frameworks for looking at the world even without a model, just as one can with a model.”
Can you be more specific about this? By that I mean just what insight about the real world can mathematical modeling really create . Saying that nonsense unrelated to that can be found is nowhere near being good enough. Just because interpretations of literary works have been misleading without mathematical modeling of literary works, sure doesn’t prove that mathematical modeling really would be helpful in analyzing literary works. You have to provide positive examples.
The Chairman of the Federal Reserve Board is much more influential in directing and creating conventional wisdom about the trajectory of the U.S. economy than all the other micoreconomist combined. The question is not why did microeconomists not see this coming; the question is why Alan Greenspan and Ben Bernanke did not see this coming? They are the people supposed to guide the economy – or at least, explain to us what is happening and why they are not intervening.
Ben Bernanke’s major contribution to conventional wisdom which misread the warning signs was his 2005 speech on “Global Savings Glut”. In it he cleverly deflected attention from the growth in global liquidity by labelling it something else and he transferred responsibility for the U.S. trade deficit to our trading partners. Very clever. Just what Bush wanted to believe. I think that speech helped elevate him to head of the FRB.
Misleading suggestions from Bernanke was only a part of the problem. Long before either Barnanke or Bush came on the scene, Bill Clinton signed into law one of the worse bills ever passed by the Congress, the Commodities Futures Modernization Act of 2000. It assured participants in the shadow banking system that the U.S. courts would enforce the contracts they signed such as credit default swaps, combined with a prohibition of the Federal Government being able to require participants in the shadow banking system to provide information about what they were doing.
We did not know the extent and complexity of the contracts in the shadow banking system, therefore, economists and the rest of us were preventing, by lack of information, from connecting the dots that led to the collapse of the U.S. private financial system.
Don’t blame economists for not knowing what they were prevented from knowing. Blame the people who pushed the 2000 Act through the Congress. And blame Alan Greenspan and Ben Bernanke for providing a false picture of what is going on in the U.S. economy.
Keynesians are classical economists where it counts. Amazing how obsessed with monetary politics the anti-monetarists are. Stare too long into the eyes of a monster…
If it is not clear by now that money was and is endogenous to the system, then economics as we once knew and loathed it is truly finished. There are no prizes for second place rallies by the heterodox schools whose claims to safety outside the herd is belied by their history. It is like watching a dysfunctional family of heroin addicts deny relation to one another, in some kind of bizarre reputational prisoner’s dilemma. Austrian notions of some deistic interest rate providing the essential morality tale for mortal men of enterprise make about as much sense as anything the Keynesians have put forth to accommodate perceived crises. Pitting supply against demand in some kind of battle for moral priority makes about as much sense as politics and little more.
Keep obsessing about bank debt and the Fed as though there were some kind of prestige in it, as though it is not clear by now that the free market doesn’t have hundreds if not thousands of other types of money and credit and leverage at its disposal to extend, exchange, and eventually withhold, to and from humanity. Each has its own limit in terms of enforceability, fraud, and title. What further ancient, banal concepts can we avoid admitting are still the foundation of our discourse by constructing even more byzantine networks of mathematical scaffolding and pop-logical festoonery? What obsolescence for the “Economists” will be assured as soundly as the Ephors or the Scholastics now find for themselves, as Paul Krugman, Ben Bernanke, and a hundred other angels of metaphysics dance on the head of a pin.
This is nothing more than pride and laziness masqueraded as mystery, to continue these attitudes of detached superiority, a la “we don’t care about profit, we seek to explain allocation of resources” and such, in a system that is so obviously designed to carried out the same — every day right in front of everyone’s face — via the profit motive, income, and marginal returns on credit. We cannot quit because we have to satisfy our advisors and pay our loans…Unbelieveable crypto-nihilist soap opera! lol
Menzie: with regard to “methodology”, I still believe the profession is underdeveloped. When I was starting my postgraduate studies in Economics I was research assistant in a neurobiology lab and the level of “methodological” reflection present there made me feel a bit embarrased by the innocent use of the word “rational” and all the other points where Economics seems to have a blind spot the size of an elephant. Anyway, things have been improving a lot lately on that front so maybe I shouldn’t complain too much…
MikeR: good point… possibly a lot of problems arise from some people making decisions based on an “Introductory Microeconomics” worldview…
Menzie,
Your message to DickF above is disingenuous. Very, very few economists even fit the “stopped clock” criteria. That is, not many were right even “twice a day”. Presumably you are thinking of the Austrians, or just some Austrians, but again, they are a tiny proportion of economists.
The reality is that doctors shouldn’t be judged on whether they all spend all their time reseraching diabetes (MikeR), but whether they, collectively, spent SOME time researching it, and whether in that time they make sloppy conclusions from inadequate models with too-small data sets.
My direct criticism of economists is that they did not catch the simplest errors in their analysis.
Too-small data sets (delinquencies since 1994)? Everywhere you look.
Making huge going-in assumptions about key variables that define the model output (such as “house prices never fall”)? Again, guilty.
Ignoring feedback loops (“subprime is only $150b — a drop in the bucket!”)? You know it.
There are many more examples. I think the problem was over-reliance on econometric models, and blinders for everything that didn’t fit in those models.
In Wonderland the less connection between marvelous theories and drab reality the better.
I think the whole Keynes-ian macro ‘invisible hand’ ideas bespeak a great deal more about the difficulty of trying to pidgeon-hole anything ‘macro’ distilled down to ‘macro-variables’.
Maybe if the models went deeper, even macro models, we might see more semblance of a match to reality.
Adele
The Economists magazine has no interest in the role that excess liquidity or cheap money played in encouraging risk taking. Could that be because the excess of imports to exports in the U.S. created an overly large supply of dollars owned by U.S. trading partners?
Of course macroeconomics has failed. Any form of economics that attempts to largely ignore the real world while claiming to be dealing with it will fail. I read the entire three part Economist article. Look at it and ask how often it mentioned the need of people to make a living. The economists parallel Wall Street in how very little effort they put forth in understanding Main Street. How much attention is really paid to the destruction wreaked on real human beings by this disaster? Not much. Consider this possibility. GDP should no longer be considered the measure that determines where in a recession, depression or recovery we are. The purpose of an economy is to provide the ability to make a living to the people of a society/nation. This is in reality its only purpose. Everything else is gravy. Do any of the economic theorists who have screwed up so badly for so long really keep their eye on this ball? No, they do not. This is why they deserve to be beaten about the head for their failure.
Macroeconomics is not on trial so much in the Economist blog as is its prevailing paradigm, the rational expectations hypothesis (REH), which has been enthusiastically embraced by salt water as well as fresh water economists, not just by RBC theorists. The REH may or may not be per se inconsistent with the existence of a bubble. But, it is completely inconsistent with the notion that the bubble’s bursting was unexpected–the widespread idea that “no one saw it coming”.
North American macroeconomics more generally, not just the REH paradigm, is now on trial for having swept under the rug with the rubric of “The Great Moderation” some very inconvenient, presaging events, such as the panic and near financial collapse from the demise of the single hedge fund LTCM and the Lost Decade in Japan after its famously exhuberant real estate market fell flat. These were studied but their broader implications seem not to been incorporated into the conventional wisdom. No doubt the same will prove true of the 2008-2009 financial crisis and its economic aftermath. It too will become, like the Great Depression, a single, isolated exceptional macroeconomic event with no implications for the present.
I vote for a combination of the viewpoints of The Economist (there needs to be a fusion between financial and macro economics), and ReformerRay (eliminate secret proprietary transactions from financial markets).
From my perspective, most Americans are scandelously ignorant of how finance, contracts, and banking works. This makes them easy prey for the loan-sharks of the mortgage, credit-card, and installment-purchase industries. It starts by not requiring basic finance & business eduction in the core curriculum of high schools in America; and ends in doctorate degrees conferred on researchers with little practical knowledge of how large-scale capital markets and finance work.
We’re all marks for the banksters if we don’t wise-up, get educated, and take control of our lives.
Count me among those who side with the Economist, rather than the economists.
Macroeconomics journals are full of very elaborate and sophisticated mathematical models that simply are not justified by the quality of their underlying data. They are like physicists trying to do precision quantum physics with 19th century lab equipment. There’s nothing wrong with the math, but its irrelevant. More generally, macroeconomists are fundamentally operating like physicists or physical chemists, trying to develop mathematical models of the economy, when they should be acting more like biologists or geologists or palentologists, developing and expanding descriptive data sets that help show what is actually going on.
When economists aren’t making models, as often as not, it seems, they are doing sterile game theory type experiments with college students involving poker night class stakes, when they should be doing ethnographies, to see what the players at the fulcrum of the economy are really doing.
There are exceptions, but the basic research plan of the discipline is wrong. I personally considered pursuing an economics PhD at three different times over the past decade and a half, once going so far as to send GREs that would have been good enough to get in to multiple programs. But, ultimately, I decided that the core curriculum that I would have to endure before doing my own research in the discipline was too far out of touch with the real world to be worth considering.
The microeconomic program of the discipline has been very successful in a wide variety of areas and has also had great impact on my profession, business and property law. But, the staff journalists at publications like the Wall Street Journal and the Economist, who don’t have PhDs (and often don’t even have graduate degrees), are publishing much better work on macroeconomic questions than the PhDs; not just more accessible work, but work that is on the merits more accurate, more insightful and more evidence based.
Widespread acceptance of mainstream, widely accepted academic macroeconomic theories in areas like the efficient market hypothesis and option pricing theory, by regulators, outsourced consultants for credit reporting agencies (and the credit reporting agency managers who accepted their reports) and Wall Street derivatives professionals (the nerds of New York) is indeed a big part of what led the markets to get things so horribly wrong on such a widespread basis prior to the financial crisis.
Yes, academic economists look at the Great Depression. This is an 80 year time horizon. But, how many seriously examine the panics that took place around the globe in the century or two before then. Everybody knows about the tulip bubble, but academics often don’t know much more. The people who have a more balanced and proportionate view on the macroeconomy’s potential fluxuations and characteristic failings of the market are not the macroeconomists, but the economic historians who a reading dusty newspaper articles, memoirs, household ledgers and archived business records.
Notably, both Adam Smith and Karl Marx built their own powerfully influential theories not on mathematical detail, but on the synthesis of vast amounts of raw descriptive data and historical examples of the phenomena they sought to describe.
I am in general sympathy with much of what appeared in the articles in The Economist, although not everything. However, the points that current DSGE models do not adequately account for behavioral effects, heterogeneity of agents, along with an ability to model endogenous bubbles of the Minsky type, strike me as very pertinent (and, Menzie, you were rather lucky to have that early behavioral macroeconomist, George Akerlof, among your profs at Berkeley). I think the assumptions of rational expectations and homogeneous agents are going to be very hard to fit with these issues, whereas the supposedly non-orthodox sticky wages and prices can be easily put into the DSGE context by allowing for information costs. Will assuming these, along with learning, save the day for the supposedly new DSGE models?
I saw this argument fought out last week at the 15th SCE conference on Computing in Economics and Finance in Sydney, Australia. There were representatives of some of the heterodox views, wtih Australia’s “Dr. Doom,” Steve Keen, pushing Minskyian positions (which he has been providing mathematical models of for some time), and plenary speaker, Cars Hommes of the University of Amsterdam, advocating a combination of behavioral and agent-based modeling. America’s “Dr. Bounceback,” Jim Morley was there also, talking a lot about Jim Hamilton’s work, but being a bit more cautious as he averaged in an AR(2) model in wiht his “bouncier” model.
However, what struck me was the large number of people at this conference who came from central banks all over the world. To a person, they were fans of DSGE models. I had it reconfirmed that this is the only approach being used in the basement of the Board of Governors of the Fed. Albert Marcet was a plenary speaker also, and these folks were eating his stuff up. So, I got told that all the DSGE models need is to “include some learning” and all will be well. I beg to differ.
Oh, let me add regarding Mario Rizzo, that he is indeed a leading Austrian economist. He makes some good points in the quoted piece, but he also ultimately rejects the use of mathematical modeling. I think he is barking up the wrong tree on that one, although a greater degree of “philosophical sophistication” would not hurt one bit right now.
“academic economists believed that a President Obama administration would promise more progress on the economy than a McCain administration”
‘promise more progress’? You can promise the moon on a stick but it’s meaningless. And what is “progress”? Higher growth? A more uniform income distribution? Etc.
According to Blanchard, even the New Keynesian Calvo pricing based models have no involuntary unemployment. They may have employment fluctuations but they are not involuntary.
j: If you’d bother’d following the links provided, you’d find the question here.
malcolm: Yes, Blanchard is (of course) absolutely right — in all the DSGE’s I’m aware of, there’s no involuntary unemployment ex ante. But it doesn’t mean that all models (of which DSGEs are a subset) have price-clearing. Shapiro-Stiglitz’s shirking model is one with unemployment that can’t be cleared by price adjustment.
Barkley Rosser: Yes, I count myself lucky to have Akerlof as a teacher. But other Berkeley faculty did not adhere slavishly to rational expectations. Jeff Frankel had an early paper of chartists and fundamentalists interacting to explain long swings in exchange rates.
On a separate note, you must be talking to a different set of people than I am. I don’t believe the only model they’re running in the “basement of the Fed” is a DSGE (and there are many different variants of DSGEs). For better or for worse, they’re still running the older style macroeconometric models. The same is true, to my knowledge, at the other international policy institutions and the central banks; that is they appeal to a multiplicity of models. If you want to say that most of the resources devoted to model development are focused on DSGEs, there I’d agree.
ohwilleke: Sure, I’ll grant you Marx had some important insights. But do you really believe in the labor theory of value (which does involve some math)? Oh, and better tell Peter Garber he doesn’t know anything else besides tulips…
Menzie,
I am sure you are right that some of the older style macroeconometric models are still being run, but they are certainly not the focus of the research efforts, as I guess you basically agree. Furthermore, I am not aware of most of those models dealing particularly well with the issues raised by the articles in The Economist, although perhaps some of them can be interpreted as doing so to some extent, at least implicitly.
Agreeing with many points above, a personal summary:
I LOVE maths and modelling, but the next true breakthrough will not come from improved DSGEs, from using cellular automata or any other modelling efforts. It will come from understanding the world better.
(Namely: i. devising a new set of distinctions ii. using them to create a new narrative iii. THEN making good models and all the rest)
That heterogeneity and homogeneity of agents are effectively identical is not just a DSGE problem. The inability to model endogenous bubbles is not just a DSGE problem. Rational expectations was not invented by modern macroeconomists, nor does its persistence as a market phenomena start or stop with DSGE. Charles Schwab is not going to be sued in 2009 because economists forgot to account for sticky prices. Home prices and balance sheets didnt plummet overnight because of someones miscalculation of information costs. Unemployment shocks dont happen because someone forgot to distinguish between voluntary and involuntary. Corporate bond prices havent been deflated for nearly two years because of reserve balances kept behind the public money window.
It is quite likely that inputs and nodes of transformation through which modern business cycles flow have nothing at all to do with DSGE or anything even remotely commensurate in modern orthodoxy. The news delivered from the SCE conference bodes ill for the American academy. It is clear that the models are so rife with universal qualifying assumptions, economists cant even trace which ones have been falsified. It is simply too painful to consider that the entire enterprise is a flexible Texas Sharpshooter Fallacy from the ground up. It will be simply bad manners to point out that the vast majority of agents in the real world may be utterly unaware of these strange assumptions and relationships assigned to them by these alien beings, economists. It can be nothing but a dirty and filthy act of ill will to point out that, in fact, when agents occasionally become aware of these models, their sole utility is as levering instruments for highly gainful prevarication. Incredible.
Disclosure: I once had a summer job shuffling deck chairs on the Titanic. I also kept Samuel Untermeyers water glass half full during the Pujo Committee investigations. I am a 136-year-old investment professional with $23 Trillion in puts on the modern macro industry.
Menzie,
I posted on a previous thread about potential output gap measuring methodology about a week ago.
Lo, and behold, the latest St. Louis Fed Review arrived at my mailbox the next day with a whole issue dedicated to potential output gap measurements.
I thought you might be interested (It’s available free from the STL Fed’s website): http://research.stlouisfed.org/publications/review/current/
Menzie,
Of course you are right that Frankel was ahead of the curve on advocating use of chartist-fundamentalist type heterogeneous agent modeling for forex markets. This is now very much the workhorse for many of the agent-based modeling of financial markets, such as by Cars Hommes more generally and by Frank Westerhoff of forex markets more particularly, along with a bunch of other folks, some of whom were in Sydney.
I suspect the main recent use of those older macro models was to get dragged out to estimate impacts with multipliers of recent fiscal policies, given that most of the DSGE models (maybe there are some exceptions among the wide variety, which you would know better) simply assume away such multiplier effects, or have them very low.
And I would agree with Hephaestus that Calvo-style price-wage rigidity is not the source of what has been going on lately. Indeed, older Post Keynesian critics have argued that this aspect of the DSGE models is a joke, and only marginally Keynesian. Keynes did allow that such effects might occur and might cause problems, but emphasized that his more serious arguments held in a flexible price world, and quite a few PK economists have shown models backed up by empirical evidence that flexprice systems may actually be more macroeconomically unstable than fixprice ones, against the textbook and DSGE stories.
lance: Yes, thanks, writing up a post on that right now… (actually, and on more).
Menzie, it’s a good thing that you limit your application of your statistics tools to the relatively inoccous areas of economics and politics, rather than say, medicine!
Your attempt to use the Dilbert survey to consider “whether academic economists differ in their views regarding the economy, as compared to those in the financial sector” fails miserably on the basis of design and interpretation: you are attributing a difference in choice of the two candidates solely to the economists’ views regarding the economy, while there are many other factors it could be attributable to (e.g.: views regarding the candidates, or their policies, or even views on the candidates parties, etc.). And indeed, these alternate explanations are strongly suggested by the study report.
I have to comment that I also adore the subtle shift in connotation as well, from the Scott Adam’s initial concept (“which candidate has the best plans for the economy”), to the survey’s objectives (“understand which Presidential candidate…is believed…to be best for the economy overall, over the long term” and “which candidate is expected to make the most progress dealing with the issues that economists see as most important”), to your post’s summary (“believed that a President Obama administration would promise more progress on the economy than a McCain administration”): a subtle but significant slide towards generalization which is unsupported by the study.
Of course, this to me illustrates a key problem germane to this exact discussion, the failure of macroeconomics (or is that macropolitics?)….
-f
Dr. D: OK, I’ll keep working on inoccous areas.
True. This is a question of semantics. The RBC/
New Keynesian models are the dominant paradigms today and I find that frustrating. Are the efficiency wage (Akerloff/Stiglitz) insights built into a macro model? Akerloff and Shiller wave there hands alot in Animal Spirits but it is just that kind of a book. Shapiro and Stiglitz carry out there analysis at the firm level as did Solow if I remember correctly.
My point remains. It’s sad that involuntary unemployment has no place in the dominant paradigms taught to graduate students.
malcolm: Efficiency wage models are not typically built into DSGE’s but I’ve seen at least one circulating. If you want to see technical expositions by Akerlof, see here. Regarding Shapiro-Stiglitz, there are representative agents and firms, and the model has economy-wide implications for unemployment.
BTW, one way that flexprice models can be more unstable than fixprice ones is if one allows bubbles. Of course, this is one of the major problems with pretty much all of the current pack of DSGE models, unless I am mistaken. They do not model bubbles, and, gosh, it turns out that bubbles are important. Their flexible prices are always zooming pretty quickly to that good old optimal GE. Introducing learning may modify this or slow it down, but does not obviously to me really bring in the bubbles that can have flexprice systems more unstable than fixprice ones.