« Evidence for peer effects | Main | What use are nation states? »

June 25, 2008



Science, properly so called, has a Royal Road to truth - the controlled experiment. It's crucial not only for the familiar reasons, but (I have decided) because it acts to keep scientists honest. If you fake it and someone else repeats your claimed contolled experiment, you're nicked. Observe that "cosmology" and "climate science", without that constraint, seem to be generous with the Tall Tale and the Lie Direct, respectively.


Dearieme is right. Anyone can have a bright idea, but its the expermental evidence that counts.

Stephen Gordon

Experimental data would be nice, but it's not necessary. Astronomy, climatology, evolutionary genetics and economics (among others) have to make do with non-experimental data. What really matters is predictive power.

Bob B

From long experience, debating whether economics is "a science" is totally pointless. Those out to discredit the generality of economic analysis, come what may, are never, ever convinced.

Economics is what economists do - as Kenneth Boulding once put it long ago. There's a regular market for economic forecasts in business and government for glaringly obvious reasons.

Both need to know what the likely downstream consequences are if tax rates are increased in terms of the likely revenue impact and what will happen to business investment decisions. Both seek advice on the potenial consequences of policy changes.

Years ago in an online debate, someone made the shrewd observation that most populist statements about economics tend to be prescriptive and tend to have little or no regard for any previous history of economic analysis and little regard for the strictures economists place on the validity of claims that a particular policy change will enhance economic welfare. Little wonder. Such prescriptions are often motivated by self-interest and the proponents resent informed critical comment from economists. How surprising!

The relevant issue is not about whether economics is a science but whether any piece of economic analysis is valid or not. Economists debate a lot about such issues - but then, as S&M points out, there are an awful lot of unresolved and contentious issues in "pure" sciences like astronomy, physics and medicine. It is difficult to impossible to run laboratory experiments in geology, which is why exploring for oil and gas reserves is often so costly. I can't understand why so much is made of economists debating issues when politicians, lawyers and journalists make professions out of public disagreements.


"And no-one gives economists $10bn to study our ideas."

Markets (banks, futures, stocks, mortgages, plain-f***ing-gambling) turn over trillions of dollars. And you, Chris, are suggesting that nobody in that environment performs economics research? Not even 0.1% of turn over?

Possibly in support of Chris and his argument about science: I work in IT services at a University. Increasingly, there is an overlap in the software tools used by economists ("not really" a science), psychologists (sociology mixed in with empirical science) and epidemiologists (empirical and theoretical science). The stuff in brackets is not intended to be a serious classification, so don't shout at me; epidemiologists, for example, study "alternative therapies" which are very far from science IMO.

But the fact that three different disciplines choose to use the same tools (implying same methodology and standards) suggests that those divisions of economics should be given the same weight as epidemiology. The converse is also true, alas; perhaps when the doctor recommends a drug next week, you should ask what a non-stats economist would advise.


"What really matters is predictive power." But if you have no nice, clean, controlled experiments to test your predictive power, what are your predictions worth? Only if your system is really simple (e.g. historically, astronomy) is mere observation much good. The moment you go from effectively a two-body problem to a three-body problem, your predictive ability degenerates dramatically, as Poincare showed a century ago. Only when you reach the other extreme of huge numbers of essentially identical objects - e.g. gas molecules - do you get predictability again. Meantime, economist only a few months ago were vapouring about the world's financial problems being "contained": yeah, yeah.


Failure to predict next year GDP goes to show that something about the current economic model regarding GDP is wrong. And if this is the case, then the way to go is not to say "hi, it's really complicated stuff - what did you expect?". Rather, you need to question the validity of the model, of its assumption or of your ability to provide a meaningful prediction.

"Economics" purports to give quantitative predictions (about measurements that it deems as meaningful) while the underlying system is not understood well enough (and sometimes the measurements don't say that much, either). This is the main problem.

There are fields in science which are poorly understood. In these cases, however, the researchers do not usually go out and supply an "engineering-grade" model. Economy does not have this luxury - because it faces constant demand for such models and predictions.

Making a hard-science out of economy might be impossible task. The fact that statistical and quantitative research principles were adopted is a good step forward. However, this just the first step.

Bob B

Total and accurate predictability of economic events is highly unlikely because of chaos theory effects:

In the 19th century and before, folks understood that unforeseen inclement weather - or flooding! - could lead to bad harvests and initiate unanticipated and unwanted movements in the trade cycle.

They knew about the corn-hog cycle - a good harvet leading to low priced corn which, in turn, would lead to a rapid expansion in the farmed pig population inducing an increase in the price of corn and a glut of pork on the market. One sage even attributed the trade cycle to the cycle of sunspot activity.

A more interesting debate is about whether business cycles are exclusively due to events in the real economy or whether (perhaps unintended) changes in monetary conditions can have (unwanted) consequences in the real economy.

These are challenging issues and will remain so regardless of whether some journalist in search of a topic to write about believes economics to be a science or not.

Try EH Carr: What is history?

Some academic establishments in mainland Europe regard the study of history as a scientific discipline.


Again I'm not sure this is correct. You see my father was a scientist and I started work as an economist (now I work in IT). I couldn't hack the manipulation of results to suit preconceived ideological positions.

Now I know this happens in science too, but the culture is very much against this (theoretical cosmology excepted). Data drives the theory not the other way around. The problem in economics begins with the questionable nature of the data to start with. What we really should be interested in - social welfare - is not measurable (almost by assumption). And so idealogical stories are made up to show that proxies are appropriate.

And I think you know this!


Bob B.
I thought the hog cycle had an easy solution - futures markets (assuming that leveraged speculators don't wreck them).

And as for whether monetary policy affects the business cycle, the results are in and it does. Whether there is a "real" business cycle independent of monetary effects is still an open question.


Bob B.
Further, if you think economic systems are chaotic (and I don't necessarily disagree with you - although the data doesn't look that way for long periods of time - but then Florida's weather is nice - mostly) then that is a radical department from a methodoly built on partial equilibrium analysis. So essentially you are saying economists have it wrong with their basically linear methodoly.


From your post, it sounds (seems) like physics is a one-to-one equivalent of economics. I would not be so sure that they are similar at this stage. But I am sure that, despite human aspects, which you stress so much, economy is a physical system obeying some "laws" - empirical long-term equilibrium relations between measured variables.

Hilary Wade

Dearieme - I think you can now do controlled economic experiments in World of Warcraft and similar games, in fact I think I've read about some man who is doing exactly that.

Bob B

Reason: "I couldn't hack the manipulation of results to suit preconceived ideological positions"

That does go on but so long as there is open debate debate, does it really matter? Economists, in my experience, are not natural authoritarians - the individualistic ethic of microeconomics is anti-authoritarian. Problems for economists arise when some lay critics disagree with their analysis but have difficulty in rebutting it by rational argument. Some then evade the issue and, instead, claim that economics isn't a science to divert attention.

Also, ideological positions aren't easily predictable. In recent years, some Conservative politicians and sympathisers have taken to opposing proposals for road pricing to alleviate congestion on some such grounds that this interferes with free markets or will increase the burden of taxation falling on motorists.

Ho hum, I says, but Alan Walters, Mrs Thatcher's personal economic adviser, built his academic reputation as an early pioneer of the theory underpinning road pricing and here is an impressive list of citations to his academic papers and publications to support that claim. Sadly, those who don't like his analysis aren't usually up to criticising it.

"I thought the hog cycle had an easy solution - futures markets (assuming that leveraged speculators don't wreck them)"

Yes - but I was talking about the 19th century. Agricultural cycles aren't likely to have significant economic impact on most major national economies when agriculture is now such a small part of total economic activity and local shortages or gluts of produce can be abated through trade - which is one reason for governments in developing economies to foster mobile phone technology and access to IT.

My point was that in earlier centuries, when agriculture was such a large part of GDP, folks regarded variations in harvests and ensuing trade cycles as due to acts of God or sun spots. Keynesian economics fundamentally changed the way we look at business cycles. Previously, self-flagellation to propitiate the deities was often regarded as the best recourse.


For some reason in one of my posts 9:02 my spelling went haywire
department = departure
methodoly = methodology

General Specific

I see several errors or misconceptions when skimming the commentary above (not the original post, just the comments).

There are several aspects to science: (1) methodology; (2) explanatory power; (3) predictive power.

People are too obsessed by predictive power because of science such as particle physics(I'm a physicist). Predictive power is not the essence of science. Some science tells us we can't predict. Some would call that a prediction. Fine--but it's lack of predictive power all the same.

By most measures, I'd call economics a science.

My only complaint about economics: I think the political in political economy is often de-emphasized by those who are possibly ideologically motivated, hence they propose "the path" for society based upon the "the one true way" while ignoring the uncertainties.

Particle physics is not often used as a bludgeon in public policy discussions. Economics is--and often incorrectly--or misleadingly. I cringe when I hear an economist or pseudo-economist start a speech with the phrase "as an economist I" because I know what often follows is not always science but opinion clothed in economic language.

That is what give me heartburn about economics. Those who malign it as a religion or a non-science are not using scientic methods themselves in making these claims. They're angry--perhaps about misuse, as I am sometimes.

I do wonder whether economists work hard enough to address the misuse.


General Specific,
I think in general you are correct. But of course many economists are also angry about this. Part of the problem is as I said, what is chosen to be measured, and misinterpretation of what can be inferred from these measurements.

Another is that "theoretical" results are often presented as fact when in fact the theory is partially interpreted (qualifications are ignored) and contrary empirical evidence is also ignored (the classic here is the claim that progressive taxation always comes with a cost to growth in spite of the fact that taxation is known to have an income effect and a substitution that work in opposite directions and the fact that empirical evidence is weak at best).

Bob B

"I think the political in political economy is often de-emphasized by those who are possibly ideologically motivated"

How much economists can say about whether policy A would lead to an improvement in welfare as compared with policy B was explored at length and sceptically in IMD Little's seminal book: A Critique of Welfare Economics (OUP, 2nd ed. 1957). But I agree, we do need to winkle out the assumptions made in any analysis of policy change and - in particular - whether it is proposed to actually compensate those who loose out.

I'm bugged by a rather different sleight of hand - namely the implicit bland assumption often made that a hike in income tax rates will have negligible effects on incentives and risk bearing.

Of course, huge parts of economics are not concerned with testable predictions but with predicting the likely range of outcomes from "what if" scenarios. A classic case was the Treasury's assessment, published in June 2003, of the likely (adverse) consequences that would follow if Britain joined the Eurozone. I must say that seems to have withstood the test of time so far rather well:

The Treasury invested much effort in assessing the potential consequences of joining the Euro, much more effort than was, by accounts, ever invested in assessing the upside and downside risks of Britain joining the ERM in November 1990.


"And scientists’ discussion of climate change seems - to this uniformed observer at least - to be as shrill as the Keynesian-monetary disputes of the 1970s."

The Keynsian-monetary disputes of the 70's would have been accurately described as all but a literal handful of people in the field being in general agreement and discussing various details, while non-economist industry shills and politicians argued the opposite to the public slowly retreating toward the economists consensus over decades?

The problem with macro economics as I see it is that large parts of the field fail to focus on improving assumptions. Instead they just make assumptions which are theoretically interesting and proceed, for decades, building castles in the sky. This is very analogous to psychology up till the past few decades: it was mostly a philosophical arena where theories of the mind competed on the basis of internal consistency, elegance, politics, etc more than consistency with reality.

The outer edges of physics seem to be having the same problem to a lesser degree, but at least they have the excuse that the problem they have is that they are starting to have trouble designing experiments which are possible to perform.

Really every field that existed before the scientific era started out in that way and except for mathmatics, for which that pattern actually works (because of the incredible degree of internal consistency it can achieve), transitions over some time period into a real modern science. Physics around 1000 years ago, chemistry around 500 years ago, biology bit by bit 500-200 years ago, etc. Before the line one sees all sorts of fanciful theories rising and falling, most of which are amusingly ignorant in retrospect, after the line the core of the science is rooted in reality and the core theory is tweaked and expanded upon rather than being replaced (eg Newtonian physics is found to be an approximation not applicable in all circumstances, but it isn't "wrong").

Economics does continue to improve, however. There is experimental work going on that seems to me to be rather analogous to the behavioral experimentation which turned psychology into a far more powerful science in the last century.

It may be that this process is taking longer with economics in part because of the extreme political pressure it is under compared with other sciences.


The author: "And scientists’ discussion of climate change seems - to this uniformed observer at least - to be as shrill as the Keynesian-monetary disputes of the 1970s."

Jeff: "The Keynsian-monetary disputes of the 70's would have been accurately described as all but a literal handful of people in the field being in general agreement and discussing various details, while non-economist industry shills and politicians argued the opposite to the public slowly retreating toward the economists consensus over decades?"

Jeff, you gotta unnerstan - political activists=scientists=industry shills.

To the observer who is both uninformed, and incapable of figuring out that there is an overwhelming consensus in a scientific field. Which sounds like a piss-poor observer to me, but who am I to judge?

Bob B

The fact is that it took until 1996 for that august institution, the IMF, to write the obituary on "monetarism":

" ...instability of monetary demand, especially in the context of supply shocks and declines in potential output growth, complicated the task of monetary authorities. As a result, during the 1980s most central banks – with some notable exceptions – either abandoned or downplayed the role of monetary targets."
IMF World Economic Outlook, October 1996, p.106.

The UK government formally abandoned monetarism in the autumn of 1985. Among the good political reasons for doing so was that the government had never managed to keep the growth of the officially adopted measure of the money supply within the targets the government had set for it.

Several of the malign consequences ascribed to monetarism in popular mythology were attributable to other factors - such as the rapid appreciation of the Pound in the foreign exchange markets when Britain became a net oil exporter after the oil price hike of 1979 - the uncompetitive cost structures of much of Britain's manufacturing industry after the 1970s when productivity had actually declined in several leading sectors - very stringent budgets in 1980, and expecially in 1981, reflecting the government's determination to squeeze inflation out of the economy.

However, it is worth recalling that the Bundesbank in West Germany had been much more successful since WW2 than Britain, France or Italy in curbing in inflation and the Bundesbank achieved those results mostly by operating to control the money supply. But there are legitimate arguments about how much of that success was due to the Bundesbank's greater independence from political interference and how much to the Bundesbank's success in generating a climate of expectations that inflation would be kept low.

Nowadays, the Bank of England and the European Central Bank are committed to policies of setting interest rates to keep predicted inflation rates on or within a target range. Money supply figures are relevant in so far as these influence the predictions of inflation.

Btw on checking, Britain entered the ERM in October 1990, not November as stated above.

Dan | thesamovar

For me, I'm less interested in whether you want to classify something as 'science' as to whether the knowledge produced by it is in some sense useful. Now for knowledge to be useful, it seems to me, it has to make some sort of prediction about what will happen in certain circumstances, or at the very least elucidate another theory that does so.

Take heat for example. Now it's impossible to predict the exact movements of every particle in a gas say, so it looks as though physics is useless. But, you can be extremely confident that for that gas PV=nRT.

Physics has managed to predict something. And that's important. If all it ever managed to do was say "oh, we can't predict that" people wouldn't be interested. So the question for economics is, does it manage to predict anything of use? It doesn't have to necessarily predict what people want it to predict, Chris is right that that's an unfair thing to ask of it. But, it does have to predict something pretty darn accurately, otherwise it's at best a prescientific field. FWIW, I think my field (theoretical neuroscience) is prescientific. That may have something to do with the fact that it's full of ex-physicists who try to apply the tools that worked for physics in a domain which is much less clean.

Chris said: "on the rare occasions I’ve read papers in the BMJ, their statistical evidence has struck me as light reading compared to any economics journal"

Pretty irrelevant. We shouldn't be judging scientific work on the basis of how complicated their statistical methods are. If we did, we'd laugh our heads off at how bad particle physics is, and deeply revere people doing research on IQ.

"But this is because physicists have it easy, not having to study the crooked timber of humanity."

Sure, economics is difficult, but just because the problem is hard doesn't mean we should give any greater respect for failed attempts at solutions.

Incidentally, it seems amazing that nobody has mentioned Popper's criterion of what constitutes science: falsifiability. How does economics do on this? (A genuine question btw, I don't know.)

Bob B

"Incidentally, it seems amazing that nobody has mentioned Popper's criterion of what constitutes science: falsifiability. How does economics do on this?"

Popper's criterion certainly informs discussion among economists about methodology and is much referenced in the text literature. But is Popper's criterion itself "scientific" in this sense?

As said above, much policy advice from economists relates to "what if" scenarios, such as whether Britain should join the Eurozone or invest in the rail network to improve average line speeds.

Issues like those are unavoidable in government and have to be addressed regardless of whether Bryan Appleyard regards economics as a science or not. Much theoretical debate in economics has focused on building analytical frameworks for tackling such issues and S&M here helpfully documents and illuminates such debates.


It is very indicative that this discussion is always initiated by economics side. There is no discussion among physicists is economics a science or not.
By definition, it is NOT a science.

Bob B

"It is very indicative that this discussion is always initiated by economics side."

Untrue. The discussion was initiated by Bryan Appleyard, a journalist who is not an economist. I'm an economist (ex-GES etc) and I've said debates on whether economics is a science or not are mostly totally pointless.

Economic analysis has to stand on its own merits in open debate in terms of the rigours of the reasoning, the statistical measures employed and the evidence invoked. The individualistic ethic of mainstream microeconomics is anti-authoritarian.

In my experience, arguments about whether economics in general is a science or not tend to crop up whenever someone wants to divert attention from the topic under discussion.


"... whether economics is a science or not are mostly totally pointless."

Absolutely agree. There is no ultimate argument, which can prove economics to be a science or not. It is profession attitude only. that's why physicists do not discuss (current) economics inside own profession. The answer is obvious.

"Economic analysis has to stand on its own merits in open debate in terms of the rigours of the reasoning, the statistical measures employed and the evidence invoked"

I would be very cautious to adapt this statement as a true one. Do you really propose to introduce some new paradigm in scientific knowledge, which is orthogonal to that in other hard sciences? How do you think it will then interact with empirical knowledge, where we obtain our entire understanding of reality and also methodology of analysis? I guess that there are only two possible outcomes of this approach - failure or full disqualification from the family of hard sciences


All data and evidence presented in science, economics is an estimate. All disciplines are striving to derive more accurate approximations of reality and causality. But they will never achieve 100% accuracy, and will often be superceded or revised, improved or otherwise.

For me - scientists think they know what it is before they look at it. Economists don't know what it is before they look at it


I'd just forget about the initiator of this (entertaining and enlightening) discussion: Appleyard wrote arguably the worst book about science in living memory, so I don't see much value in his bletherings about its comparability with other projects.


Funny. Problem is not Appleyard's insult. Problem is that economists feel real injury. They are too open and sensitive to external criticism. The original post contains just a flavor of abuse, but the reaction is too far over.

Bob B

For comparisons, do read this on: The rise of the Health Care professional: don't say I didn't warn you:

As John Maynard Keynes once put it in the different context of economic policy:

"Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist. Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back. I am sure that the power of vested interests is vastly exaggerated compared with the gradual encroachment of ideas. Not, indeed, immediately, but after a certain interval; for in the field of economic and political philosophy there are not many who are influenced by new theories after they are twenty-five or thirty years of age, so that the ideas which civil servants and politicians and even agitators apply to current events are not likely to be the newest. But, soon or late, it is ideas, not vested interests, which are dangerous for good or evil."

The comments to this entry are closed.

blogs I like

Blog powered by Typepad