Is economics an empirical science ?

by John Quiggin on July 14, 2004

Tyler Cowen[1] lists a number of economic propositions which he formerly believed, but has abandoned in the light of contrary evidence. Most of these propositions were elements of the economic orthodoxy of the 1980s and 1990s, variously referred to as Thatcherism, neoliberalism, the Washington consensus and, in Australia, economic rationalism. They include the efficacy of monetary targeting, the beneficence of free capital movements and the desirability of rapid privatisation in transition economies.

Following in the same spirit, I thought I’d list a couple of propositions on which I’ve changed my mind in the face of empirical evidence. These are elements of the Keynesian orthodoxy of the 1950s and 1960s, on which I was trained. Following Cowen, I’ll list them as false claims I used to believe

* There is a long-run trade-off between unemployment and inflation

* Keynesian fiscal policy is a powerful and reliable instrument for stabilising aggregate demand

On both these issues, I’ve come to accept that Milton Friedman was largely right, and his Keynesian opponents largely wrong.

On the first, I think Friedman’s victory was total, although the supposed implication that there exists a “natural rate” of unemployment at which inflation remains stable has proved equally unreliable.

On the second, Friedman was absolutely right in talking about “long and variable lags” and rejecting the idea that it is possible to “fine-tune” the economy. This doesn’t mean that fiscal policy is of no value. In particular, the fact that budgets naturally go into deficit when the economy turns down provides a measure of automatic stabilisation. And when a deep recession lasts for more than, say, a year, there’s time to bring discretionary fiscal policy into play. The suggestion, by Nick Gruen and others, of some form of independent body to manage discretionary fiscal policy, analogous to that of the Reserve Bank in monetary policy, has a lot of appeal for me. Still, this is a long way from the kind of mechanical Keynesianism I was taught a few decades ago.

Combining my concessions and Tyler Cowen’s, it’s evident that there is some process of convergence in the beliefs of economists, though with a lot of oscillation. The breakdown of Keynesian economic management during the crisis of the 1970s, and the general ‘fiscal crisis of the state’ that occurred at the same time, validated many of the criticisms made by Friedman and others of the Keynesian and social-democratic orthodoxy of the postwar period. But it produced an overreaction, most notably in the extreme claims made by advocates of rational expectations macroeconomics, but also in overblown claims about the merits of privatisation and deregulation. These have gradually lost favour as their empirical weakness has been exposed by events.

fn1. Jason Soon has also noted this piece



Keith M Ellis 07.14.04 at 10:07 am

Yep. I think it’s a sign that economics is becoming a science, where empiricism is slowly replacing ideology.

What I think is interesting about this is how closely economics is associated with politics—which is all about ideology. This implies a burdgeoning rift between the politicians (and their ilk) and the economists. And, indeed, we see such a thing happening. A worrying possibility is that the politicos will throw the economists overboard. Economists may find themselves as irrelevant to public policy as political scientists.


Giles 07.14.04 at 11:10 am

Folk beleived that there was a long-run trade-off between unemployment and inflation
and that Keynesian fiscal policy was s a powerful and reliable instrument because , empirically, that was what they observed in the middle of the 20th century.

So its not clear that changing these beleifs refelcts a tilt twoards economics becoming an empirical science.


Bob 07.14.04 at 12:41 pm

I’ve long regarded questions about whether economics is a “science” as a subsidiary issue since it all depends on what we mean by “science” and the term is used in different ways in different cultures. We tend to regard geology as a “science” yet much of its focus is on understanding the natural causes for rock formations and the like, which we can observe. Is the intent there distinctively different from that of economists and economic historians in attempting to understand how economies function?

Hammurabi’s law code from Babylon in 18th century BC contains elements of a statutory prices and incomes policy. Implicit in that is some notion of how markets function as well as a prescriptive intent to modify expected natural or unregulated outcomes to better achieve (undisclosed) “efficiency” and equity objectives. Much the same can be said of the economic policies of all governments since so it makes sense to unravel the implicit modelling and prescriptive analysis for closer scrutiny and (preferably, open) debate.

As for Keynesian fiscal policy, the central focus of his seminal book, The General Theory (1936) is:

“it is an outstanding characteristic of the economic system in which we live that, whilst it is subject to severe fluctuations in respect of output and employment, it is not violently unstable. Indeed it seems capable of remaining in a chronic condition of sub-normal activity for a considerable period without any marked tendency either towards recovery or towards complete collapse.” – p.249 at:

The implications for activist fiscal policy is a byproduct of Keynes’s endeavour to understand how markets economies could slip into and then sustain an apparently stable state with persistently high levels of unemployment with all the social misery that can inflict. The effectiveness of fine-tuning fiscal policy to modify expected aggregate demand to achieve higher employment – or dampen the otherwise expected change in price levels – depends on the dependability of forecasting models. In turn, that depends on accurately modelling the downstream consequences of policy interventions, including trade-offs between inflation and unemployment in the short-run and beyond. The extent of such trade-offs, if any, are empirical factors in the modelling which will likely change over time and differ as between national or even regional economies.

Famously, Keynes, in his time, was sceptical about the dependability of econometric modelling but its methods have improved since. Even so, by definition shocks cannot be predicted and a perennial problem in modelling is the extent to which policy interventions affect the expectations of households and business, thereby prompting these agents to modify their spending behaviour.

However, with all the concerns over the dependability of forecasting models, Keynes wished reject a conventional wisdom which had concluded that nothing could or should be done through fiscal policy to address situations where an economy had stagnated, resulting in a persistently high level of unemployment.


Cranky Observer 07.14.04 at 1:14 pm

> Famously, Keynes, in his time, was
> sceptical about the dependability
> of econometric modelling but its
> methods have improved since.

Well, I guess that depends on what you mean by “improved” and what you consider the absolute level of usefulness to be. If you can find a practicing engineer who thinks that econometric modeling works I would be interested to meet him/her. Econometrics compares unfavorably with global warming climate models for scientific accuracy as far as I can tell.



q 07.14.04 at 1:18 pm

RE: fine-tuning – A clear distinction between “Keynesian fiscal policy” and Keynes’s fiscal analysis is always wise.


Rob 07.14.04 at 2:12 pm

Actually econometric models do quite well across a variety of variables since there tends to be a great degree of persistence that can be used. That is completely different from whether estimated theoretical models produce good results. Because in general they don’t.


Barry 07.14.04 at 3:16 pm

The idea that eonometrics should be judged by the opinions of practicing engineers is pretty funny, when one thinks about it. I guess that it’s part of the legacy of the internet, when the majority of people using it were ‘hard science’/engineering/CS studs.


PEmberton 07.14.04 at 3:41 pm

Giles post is inadvertently a great rebuttal of Quiggin’s anti-rational expectations claim. There was a long-run Phillips curve in the data because government did not try to exploit it. Once they did try to play the curve, and As the private sector caught on, the trade-off disappeared.


Alex R 07.14.04 at 3:51 pm

This is a bit off topic, but you make a side comment here: The suggestion … of some form of independent body to manage discretionary fiscal policy, analogous to that of the Reserve Bank in monetary policy, has a lot of appeal for me.

I’m very curious about the details of this idea, because as stated, it sounds profoundly undemocratic.

“Discretionary fiscal policy” is another name for policy to decide how much government can tax and spend, on those matters where it has a choice. But nearly everything that government *does* is determined by what it *spends*. In what sense can decisions about government spending be made by an “independent” body, if decisions about the purpose, function, and actions of government are to be made democratically, or by the democratically elected representatives of the people?

One could imagine a structure in which some constraints on spending are set by an independent fiscal policy body — say, “Spending shall exceed revenue by an amount in the range X to Y” — but even this seems to take too much away from the powers of the people’s elected representatives. Independent bodies may quite reasonably analyze date, make forecasts, and even make policy recommendations, but deciding fiscal policy seems inappropriate.


Brian Wilder 07.14.04 at 4:02 pm

Keynesians, at least the American variety that came to power on the Kennedy Council of Economic Advisers (a stellar group, if there ever was one) never believed that there was a “long-run tradeoff between unemployment and inflation” Orthodox theory said that there was a short-run “L” — if you pushed fiscal policy beyond full employment, you got inflation for as long as it took to drain away the excess.

In the U.S., Republicans always favored higher unemployment rates, for their redistributive effects — a moderately high unemployment rate (6% instead of 4%, on the U.S. scale) holds down wages relative to income to capital. In the 1980’s the bastards actually had economists discussing why 4% was simply not possible. Linking unemployment to inflation is just political hypocrisy — you can’t say outloud that policy is to hold down the wages of the working class in the long-run.


Bob 07.14.04 at 4:23 pm

To inject a UK perspective on econometric forecasting, there are periodic assessments by an independent party of leading forecasting models of the economy to compare predictions against outturns, out of which the Treasury’s own model usually comes somewhere among the best few out of many, most coming from financial and other private sector institutions. The Treasury publishes regular surveys of the best-known two dozen or so published forecasts.

Whatever the problems with forecasting, it is challenging to argue that it is an entirely worthless endeavour. The main orientation of debate is whether a dedicated policy commitment to regular fine tuning of aggregate demand by fiscal policy – as during the 1950s and 1960s in Britain – does more harm than good. The prevailing consensus, derived from experience, is that fine tuning by fiscal policy does harm, partly through generating destabilizing expectations – the trade-off between unemployment and inflation evaporates. Since 1992, the onus for short-term stabilization policy has shifted to the Bank of England applying monetary policy to target the inflation rate. But this has not removed the need for forecasting since the Bank needs to forecast where the inflation rate is heading and assess the likely downstream impact on inflation of changes in its monetary policy, which can take two years or so to work through.

Of course, such a policy framework is feasible only so long as Britain remains outside the Eurozone. Once the Euro is joined, so to say, interest rates are set by the European Central Bank to target the average inflation rate across the Eurozone. However, that leaves open the possibility that interest rates set centrally could lead to unacceptable inflation or recession in national economies participating in the Euro because Eurozone economies have not sufficiently converged and differences in languages and social security systems continue to inhibit cross-border consumer purchasing and job searches.

In those circumstances, the case for an activist fiscal policy by national governments comes back into its own. Political pressures upon governments force public debate if unemployment rates (or inflation) in national economies reach levels widely deemed unacceptable. The policy debate needs to refocus then on the extent to which national unemployment is the result of “deficient” aggregate demand or due to structural factors and malfunctioning (or “sclerotic”) markets in the national economy. If Eurozone experience is anything to go by, such debates tend to be politically painful for governments when policy analysis concludes structural factors or sclerotic markets are mainly to blame since neither are usually remediable by quick fixes, fiscal or otherwise.


Ian Montgomerie 07.14.04 at 4:36 pm

This question is not useful. You can’t deny that economics is an empirical science, because economists do at least sometimes look at the actual economy and adjust their results to it.

But there are often big differences in just how “empirical” various branches of the sciences are. It seems to me that in the social sciences, it is very common to have “foundation theories” that affect how research is conducted, what questions are asked, how those questions are tested, and so forth, that may be used for decades but be completely wrong. This is because while researchers look at evidence to test individual propositions and predictions, they are loathe to challenge the foundation theories because without them they’re not quite sure how they would conduct their research.

In modern economics the foundation theories are that the economy can be studied using a static equilibrium analysis assuming rational actors. It was obvious pretty much from the beginning that the real world is not static, very seldom in equilibrium, and does not contain any rational actors. So one might expect that a theory based on such things would be seen as a temporary simplifying model, and that results that it couldn’t explain well would cast doubt upon the model. But the problem was that a generation of economists was trained on how to use general equilibrium analysis and so forth, and essentially if they rejected it they’d lose all their theoretical tools and the ability to create results in a framework that would be acceptable to other economists. So these foundation theories are still what one overwhelmingly finds in economics research, driving _how_ the empirical data is interpreted, what kinds of theories are created, and so on, even though we know that empirically speaking the foundation theories are simply wrong and there is a slow, painful effort in progress to create alternatives.

There has been a similar situation in experimental psychology, which from a naive first glance would seem a highly empirical enterprise. After all, modern psychologists often don’t have grand theories about how the mind works, just specific ones to explain their neat results. But when you look at the field you find foundation theories that are generally unstated unless someone disagrees with them, and which dramatically affect how research is conducted. One of them was that genetics and heredity, “human nature”, plays no significant role in psychology – if you find a correlation between an individuals behavior and part of their environment not obviously caused by their behavior, obviously the environmental factors caused the behavior. Another was that the basics of psychology are so universal that you can base entire research programs on studies of American university students, it being nice but not necessary to check against the general population or other countries. And a third assumption, common in cognitive psychology, is that the human brain is such a general-purpose reasoning machine that you can study reasoning and problem-solving independently of context. If a person reasons a certain way with toy problems, they’ll reason the same way with logically equivalent problems in the real world.

Despite psychology being _highly_ empirical and experimental, entire research programs spent decades coming up with results that were systematically wrong because all tests actually done were guided by the foundation theories, and the foundation theories weren’t themselves rigorously tested. When people did first mount serious challenges, it resulted in some of the ugliest political controversies in the history of science. Any step away from environment-is-everything universalism was a step toward fascism and racism. After a couple of decades of challenge, though, the challengers have clearly won the day and there is nothing left but a quixotic rear guard action by the old die-hards. Evolutionary psychology has shown that an analysis based on human nature can come up with strong results in areas traditional psychology never really even thought to look at. (Incredibly, when evolutionary psychologists started studying love, mating, and sex differences in them in detail, they weren’t so much overturning earlier results as studying an area which had been consistently ignored). Genetics-and-development analyses have shown that virtually all studies showing parental behavior influencing children were wrong – genes account for most parent/child associations, and peer environment has much more influence on children than expected. International studies have shown that many “fundamental” results of social psychology actually vary considerably between eastern and western cultures. Evolutionary psychology has shown that context matters a great deal – the human mind has distinct reasoning systems that can be applied to problems, and different contexts can trigger different systems to be used on problems with the same logical structure. (A result of interest to economists is the stuff about moral vs. cost/benefit reasoning, which use different parts of the brain. People can act like economic cost/benefit maximizers much of the time in their personal lives, but when they come to a moral problem they start to use different parts of the brain optimized for rules-based thinking. And the cost-benefit part of the brain isn’t optimized for morals, so in situations where people are encouraged to “think economically”, they think more selfishly and amorally).

Some of the resistance in psychology, again, comes not from politics but from practical “how do I conduct research under the new scheme” questions. Evolutionary psychology requires a decent exposure to evolutionary theory, something most psychologists are clueless about. Doing developmental studies in an era where genes matter requires either using behavioral genetic techniques such as twin studies, or using a “population sample” with various levels of genetic relationships and applying complex statistical techniques to it. Studying reasoning using real-world types of problems requires a lot more attentiveness than just coming up with any old test that is logically equivalent to what you want to look at. And international comparisons require developing new professional relationships with far-off countries where psychology is less well funded. Comparisons to non-industrial societies even require inter-field cooperation with anthropology.


Jack 07.14.04 at 6:30 pm

I always thought science was an approach not a result. Economists have to make do with short data sets in changing and uncontrollable environments. If they waited for the kind of datasets other sciences hope for we would be here forever.

I imagine progress will come from finding rules that are performative (Black Scholes formulae for example) rather than correlations that obey Goodhart’s law (Phillips curves apparently).

Sorting the wheat from the chaff can’t possibly happen very quickly. Would the most scientific economist armed with the biggest computer and the largest army of researchers make much progress before seeing some of the theory put into practice?


Luis 07.14.04 at 8:06 pm

Barry, ‘what do the engineers think’ is the legacy of engineers having to actually prove things or risk having people die when things collapse, whereas the soft sciences tend not to have to prove anything in any kind of rigorous way (at least by comparison.) Trying to do any kind of engineering calculations with the large tolerance for error and ludicrous base assumptions that are very common in the social sciences would, in most applied engineering disciplines, at best get you laughed out of the field and at worst leave you in danger of being put in jail. Economics is better than, say, political psychology in this regard, but not so much so that most engineers wouldn’t laugh at the so-called ‘standards of proof’ employed.


CalDem 07.14.04 at 8:15 pm

What’s really interesting is when you apply engineering type analysis to answer economist’s questions-like the various estimates of the costs of pollution control under marketable permits. The engineering numbers are almost ludicrously wrong in these cases-because they don’t have the theoretical or empirical tools to consider the questions.


Zizka 07.14.04 at 8:15 pm

Not completely on topic, but I’m just reading (as much as I can) Sen’s “Rationality and Freedom”.

What he consistently seems to do is take rigorous, but not useful or empirical, definitions of rationality, social choice, and “welfare”, and replace them with more commonsensical, but less rigorous definitions. The effect is to embed the “scientific” part of economics in a larger, ecclectic, open discourse which is more philosophical and normative.

The effect is to give up the claims of economics to absolute scientific truth, while also disavowing some of the more objectionable conclusions of social choice theory, welfare economics, and rational choice theory.


Jack 07.14.04 at 8:52 pm

Ah, but CalDem, how do you know they are wrong?


PanJack 07.14.04 at 9:33 pm

Economist have always been state-of-the-art in their beliefs about how to do “science.”

For instance, by using utility theory economists are start-of-the-art 1830s philosophers.

By their belief in the proper link between theory and empirical work economists are start-of-the-art 1930s Popperians.

I’m glad of this because nothing of interest has been done in philosophy since 1830 and nothing of interest has been done in the theory of theory-evidence link since Popper.

You go Economists!


Cranky Observer 07.14.04 at 9:34 pm

> ike the various estimates of the
> costs of pollution control under
> marketable permits. The engineering
> numbers are almost ludicrously
> wrong in these cases-because they
> don’t have the theoretical or
> empirical tools to consider the
> questions.

That’s interesting. I have done a lot of engineering economic analysis over the years and have never seen a result that is “ludcrously wrong”. Could you provide links to some case studies? I would like to learn more.

However, I suspect that the prices you are talking about (wholesale electricity, pollution credits) were “set” in markets where Enron and Dynergy were major participants from 1995 – 2002. If so, based on what we now know about those players I don’t think you take the price set by those “markets” as “correct”.



CalDem 07.14.04 at 10:15 pm

Check out the ex-ante, ex-post SO2 marketable permit studies.


agm 07.15.04 at 1:00 am

Whether economics is a science hinges on a key property: falsifiability. Falsifiability is necessary for something to be a science. Consider, as bob did, geology. Geology is alot more than observation, you can also test your explanations experimentally, you can apply chemistry (to get mineral X, you must have temperatures so high, with these substances around, under pressures so much, etc) and you can apply physics (density differences causing differences in behavior for land versus seafloor), you can create new tools and theories and test them. Not only that, but the tests are very common and very definite in determining the validity of your ideas — based on our understanding, we predicted an oilfield at a certain place at a certain depth; is it there? Thus geology has the element of falsifiability.

The same concern exists in the hardest science. Most branches of physics can build an experiment to test a model or theory, the exceptions seem to be astro-, space, and atmospheric physics. Falsifiability is much more elusive since you can’t go out and build a universe or solar system to test your theories, the best you can do is create numerical and theoretical models and test them against observations and throw satellites up to gather data. Thus the old guard doesn’t really regard them as branches of physics, and these endeavors are often run in separate departments.

Thus the issue is whether economics is falsifiable. If not, if you can’t say, “Under these circumstances my model says X will happen”, and then go and check whether that happens in those circumstances, then it cannot be science. Conversely, if you can tease out pre- or postdictions from your model and compare against what actually happened, even if your model did a horrible job it’s still science. That’s how, in the span of hundred years, cosmology progressed from “turtles all the way down” to become a very hard science indeed.


Zizka 07.15.04 at 2:49 am

Not to be too annoying, but to me the question of whether economics is a science or not is pretty empty. It’s like some kind of USDA seal of approval, giving economists license to strut around a little more arrogantly than before.

To me the questions are, “How good is economics at what it does?” and “If you don’t like the way economics does things, what improvements do you suggest?”

There seem to be apples-and-oranges comparisons being made between the results of two different kinds of study. Thus, maybe the reason historians and economists get “worse” results from a scientific point of view, is that the actualities studied by these disciplines are, for whatever reason, ontologically not susceptible of deterministic explanation (or whatever the criterion is).

My guess is that economics is a lot more like history than physics, FWIW.


Nicholas Gruen 07.15.04 at 2:51 am

I posted this on John’s blog, and then realised he’d posted on CT – where there is more action. So here’s my post. I’ll also respond to Alex R in the next comment.

It is nice to see the proposal to try to develop independent fiscal policy institutions, along the lines of monetary policy institutions staying on the radar screen. Alan Blinder mentioned it a few weeks ago at a conference of the Federal Reserve in Boston (including mention of its Australian incarnation via the BCA Paper). It got a similar mention in a major think piece by CESifo commissioned by the OECD recently. This was reported by Alan Kay in the Financial Times at (email me for either or both papers). The OECD also showed some interest in the BCA paper when it was first published and written up in The Economist. And the book ‘Imagining Australia’ which was only launched on the weekend mentions it as one of four proposals to “strengthen our institutions and the macro economy”. (I think an extract was in this Weekend’s Australian, but I don’t have the link).

Pleased to see you like it John, I didn’t know if you would. Those who instinctively favour active fiscal policy often see constraints on fiscal policy (in the name of fiscal responsibility) as constraints on flexibility. The debate on fiscal flexibility itself is often seen as one between those arguing for fiscal responsibility and those arguing for fiscal flexibility – which is usually code for fiscal expansion. The dichotomy is silly of course, and the idea of developing our fiscal institutions along the lines of monetary policy is based on the idea that fiscal responsibility is a precondition for fiscal flexibility. Clearly party politics has difficulty delivering this, and so the idea is to build out institutions as we have with monetary policy – where we have improved our management of a very similar tradeoff. Ideally more independence in monetary policy improves monetary responsibility, and in purchasing credibility enhances flexibility.


Ian Montgomerie 07.15.04 at 2:56 am

“Whether economics is a science hinges on a key property: falsifiability.”

I think as I was mentioning, the problem here is specific propositions vs. the broad underlying theories used to produce them. Whenever economics actually predicts something specific about the economy, it can be falsified. However, the predictions are usually generated using theoretical approaches that are effectively not falsifiable.

If you ask the question “is general equilibrium theory falsifiable?”, the answer is clearly “in practice, no”. The thing is, general equilibrium theory has been known to be wrong for some time – the actual economy is simply not in equilibrium. So very quickly, economists using equilibrium theory (which is still the great majority) took the approach that it should be treated sort of like a heuristic or approximation. So long as general equilibrium theory produces useful results in general, it’s a good tool despite being, strictly speaking, wrong.

Once economists took this approach, they had the ultimate defense of equilibrium analysis – there can be no such thing as a clear falsification of it. Any one problem with it may simply be one of a minority of circumstances in which it isn’t a good approximation. So economists are free to continue assuming that equilibrium analysis is accurate for any specific case unless proven otherwise – exactly as one would do with a falsifiable but as-yet-unsalsified theory – even if many cases have been found where it doesn’t apply.

The practical reality that this creates is that no matter how many errors equilibrium analysis has, it can never be falsified, it could only be replaced by some theory that was able to so thoroughly beat it at its own game that it could be shown a priori to likely make a better prediction in any given situation. Disproving any specific prediction of equilibrium analysis isn’t treated as disproving the technique.

And of course, it’s even worse because of a massive sampling bias.

There are two ways to properly analyze how useful a heuristic is. One way is with a proof, if you know enough about the real system being modeled to prove that the heuristic is accurate to within certain limits. This doesn’t work in economics because we don’t know enough about the underlying real systems. The other way is to get a decent, representative sample of problems, and test the heuristic’s predictions on them. If the sample is decent, you get a decent test of the heuristic’s usefulness. This also doesn’t apply to economics – economists don’t work by taking a representative sample of questions one might want to ask about economic issues, applying equilibrium analysis to them, and then reporting whether it succeeded or failed. Instead they naturally start with the theoretical tools they have, apply them preferentially to those problems where they seem likely to yield decent results, and preferentially report positive results that can be interpreted as supporting an equilibrium analysis.

But then, perhaps some economist disagrees with me. If so, perhaps they can explain how one would in practice falsify general equilibrium theory, the “homo economicus” rational agent model, etc?


Nicholas Gruen 07.15.04 at 2:58 am

Here is an extract from a paper of mine which addresses the question Alex R raised. It and my original paper on the issue are posted at

Democracy and an independent fiscal stance

It might be argued that the arrangements for more independent fiscal policy would be less democratic than existing arrangements. However, as we have seen there is little in the ideas being explored in this paper for which there are not precedents elsewhere within the OECD. A wide range of government bodies and authorities in our society are distanced from – but still ultimately accountable to – representative democratic institutions.

The ideas explored here would be incapable of implementation – and nor should they be capable of implementation – without the democratic sanction of legislation. The degree of independence given to any government policy-making body is itself ultimately legislated by Parliament and so accountable to the people. In this sense, Governor Ian Macfarlane of Australia’s Central Bank, has argued, the issue of ‘independence’ is best seen as ‘a discussion about the optimal degree of delegation, including the circumstances in which the delegation could be withdrawn’ (1996).

In this context it must be remembered that the kind of power to be delegated is of a precise and closely circumscribed kind. An independent fiscal authority would have none of the kinds of power of the legislature to do any favours or impose specific costs on any section of the community. It could not introduce any new taxes or tax concessions. Politicians would continue to perform their democratic duty in deciding who pays what rate of tax, what concessions there are, how much money is spent and on what.

But an independent body would have an important role in influencing the management of the total tax take. It would take the entire tax system from existing legislation and, by calibrating that system across the board, have influence over one economic variable which is critical to both long and short run economic performance – the relationship between outlays and revenue.
Beyond the fact that that it could deliver better performance, fewer and milder recessions, and lower unemployment, the philosophical justification for fiscal policy independence is analogous to the philosophical justification for monetary policy independence.

The fact that we have an intermediary in the area of monetary policy and do not have one in the area of fiscal policy is a product of the different history of the relevant institutions rather than differing principles. But its not hard to appreciate a very sound political philosophical justification for some independence in monetary policy management. One political generation’s undermining of the value of the currency to underwrite an unsustainably high rate of economic growth for itself is not only unfair. It is undemocratic. It represents oppression of the minority.

And protecting the minority from oppression does on occasion require departures from strict representative democracy – most obviously concerning the independence of the judiciary for instance. But such protections from oppression of the minority are, in many senses, things which are necessary to the checks and balances which constitute and protect democracy. They are building blocks of democracy rather than derogations from it.

Clearly if the government of the day set short-term interest rates it would be more democratic in the simple sense but so too would electing the Auditor General or the judges of superior appellate courts. In that light, it does not seem entirely foolish to say that current protections against debasing the currency in democratic countries are, like property rights, part of the apparatus of checks and balances that underpins democracy and makes it more rather than less secure. Precisely the same could be said of checks and balances against one generation running up government liabilities for the next generation to pay in higher taxes.


q 07.15.04 at 4:17 am

_Whether economics is a science hinges on a key property: falsifiability. Falsifiability is necessary for something to be a science._

Yes this is one interpretation. I recommend that you read Gellner, Kuhn and Feyerabend on scientific processes to expand the framework of your debate.


Bob 07.15.04 at 1:47 pm

The extent to which parts of economics are “scientific” in a Popperian sense is doubtless a fascinating topic but governments are continually pressured into making a huge variety of decisions on macro stabilization, monetary systems, competition policy, regulation and trade restrictions.

Regiments of lobbyists are out there championing a wide variety of causes on behalf of contending vested interests. Does anyone seriously suppose lobbyists or their clients worry about the niceties of whether the beguiling rationales for the causes they hawk are “scientific”? What matters to them is whether the rationales are sufficiently plausible to start a bandwagon rolling. What unites them is discrediting nit-picking economists who find fault with their touted nostrums.

As Adam Smith put it: “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices. It is impossible indeed to prevent such meetings, by any law which either could be executed, or would be consistent with liberty and justice. But though the law cannot hinder people of the same trade from sometimes assembling together, it ought to do nothing to facilitate such assemblies, much less to render them necessary.” [Wealth of Nations (1776), Bk.1 chp.10]

Markets and governments existed long before economics emerged as a separate discipline. Much, if not all, early writing about markets and associated policy for their control was avowedly prescriptive – such as the “just” price, prohibition of usury, mercantilism and the rest of dirigisme. Indeed, the early title of Political Economy aptly described the battlefield. The parading of policy issues in government is relentless and inescapable. The challenge for economists is how to sharpen up their discipline.

Alan Blinder has proposed the illuminating Murphy’s Law:

“Economists have the least influence on policy where they know the most and are most agreed; they have the most influence where they know the least and disagree most vehemently.” [Hard Heads, Soft Hearts (1987)]

Keynes, again: ” . . the ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed the world is ruled by little else. Practical men, who believe themselves to quite exempt from any intellectual influences, 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 ill.” – from:


Barry 07.15.04 at 4:01 pm


“Barry, ‘what do the engineers think’ is the legacy of engineers having to actually prove things or risk having people die when things collapse, whereas the soft sciences tend not to have to prove anything in any kind of rigorous way (at least by comparison.)”

Well, when surgeons screw up, people die or are injured. However, the idea that surgeons should be the judges of econometrics is ludicrous.

And the (auto) engineers I worked with generally couldn’t prove much with ‘rigor’, because there wasn’t that much really good data which was available. They tended to make incremental modifications to what had worked, or to run things through test rigs. They could run surprisingly complex calculations, but the models all depended on poor environmental/manufacturing data.

“Trying to do any kind of engineering calculations with the large tolerance for error and ludicrous base assumptions that are very common in the social sciences would, in most applied engineering disciplines, at best get you laughed out of the field and at worst leave you in danger of being put in jail.”

I’d put it that the engineers that I worked with did have to work with large tolerance for error and ludicrous base assumptions.

Another way to say what you said above is that economics (and the rest of the social sciences, and business) have to work with bad data on a regular basis. It’s not surprising that things which are more lab-simulatable can be engineered better.

“Economics is better than, say, political psychology in this regard, but not so much so that most engineers wouldn’t laugh at the so-called ‘standards of proof’ employed.”

Actually, most of the *experienced* engineers I’ve worked with wouldn’t have laughed. They’d think that the data s*cked, and might be glad that they didn’t have to deal with that stuff.


Jonathan Goldberg 07.15.04 at 5:25 pm

I suggest that this discussion stop turning on Popper and start using a better philosopher of science, namely Quine. “The Web of Belief,” Quine and Ullian, is a good place to start.

As a member of the Quine or Kitchen Chemistry school, I attempt to clarify some points.

One: falseificationism is at best an interesting partial view. The reason for this is that is simply transfers the question to when something is falsified. With enough special purpose assumptions and escape clauses any single belief can be preserved. The question is when do you quit. (Lakatose tried to capture this with his noting of healthy and declining research programs).

It’s easy to come up with historical examples of things that were thought to be falsified but turned out true; for instance, the Gordon-Kline equation. On the other hand, in a famous paper on the weak nuclear force Gell-Mann and Feymann cited about a half dozen experimental results which disagreed with their theory and must therefore be wrong. And indeed they were.

Francis Crick has gone so far as to elevate this into a principle: he holds that no theory that explains all the facts can be correct, because at any time some of (what are thought to be the) facts are wrong.

Two: Kitchen Chemistry. Quine’s view (and mine, FWIW) is that there is no essential difference between the thought processes or logical principles used in the kitchen and in the science of chemistry. The latter is simply more careful and more successful.

Three: what is really being argued over is not whether economics is a science (as Richard Rudner, my revered philosophy teacher, once remarked, if snience works and science dosen’t I’ll do snience), but whether it can produce strong results. This is not dependent on a procedure such as “make only falsifiabile statements.”

Ask yourself whether anyone, however scientific and intelligent, could have produced quantum theory in Aristotle’s time. Sometimes the background is simply not there. Then, we grope around trying to produce it, or abandon that field of study altogether. If you don’t want to do that, you work with what you have, as well as you can. And that is science.

It’s true that this leaves one without a clear criterion by which to seperate science from non-science. That is because no such criterion exists. Sorry about that.


HP 07.15.04 at 10:18 pm

As a disinterested layperson, I’ve always gotten the impression that economics is a branch of some larger, unnnamed and undescribed science. Just as mechanics is a branch of physics, it seems that economics is a branch of [something], with that [something] dealing with large-scale theories of production and distribution across not only market economies but feudal serfdoms and moieties and termite mounds and the builders of Stonehenge and prides of lions and household chores when there are teenagers in the house and ….

Is there a such a [something] that economics builds on? If not, why not?


dan 07.15.04 at 11:47 pm

re q: I dunno about Feyerabend, but I’d agree that Kuhn was a big improvement over Popper or Lakatos. I work in math psych, and I find myself doing more-or-less exactly what he describes in Scientific Revolutions. You start with a basic framework of assumptions which you use to design experiments and build models. After a few successes (this is psych, not physics, so we have frequent revolutions) the framework starts to collapse under its own weight, and you build up a new one that can handle the new data. Is economics much different to that?


q 07.16.04 at 1:07 pm

Association for Information Systems- Discussion of Design Research in IS including March and Smith (1995)… “constructs, models, methods, and instantiations” would have an applicability across many disciplines, including economics, and I suspect psychology.

March and Smith “Design and Natural Science Research on Information Technology.” Decision Support Systems 15 (1995)


Bob 07.16.04 at 2:19 pm

“Is there a such a [something] that economics builds on? If not, why not?”

Historically, economics evolved out of philosophy, especially political philosophy, and from the exigencies of government, conspicuously in making laws and regulations. As mentioned before, most, if not all, early manifestations reflect prescriptive – or proscriptive – intent from which the implict analytical or modelling content has to be inferred [1].

This becomes evident from notions of “the just price” in medieval Christian theology, prohibitions on usury in several leading religions, and from ascendant mercantilist doctrines, in the 16th through 18th centuries, as recipes for enhancing national affluence. Much of Adam Smith’s Wealth of Nations (1776), regarded by many as the starting point of political economy as a distinctive discipline, is focused on unpicking the justifications for mercantilism. It also staked out the explicit claim that private pursuit of private interests, within a framework of law and the (minimalist) functions of government, conduces to maximising social welfare through “the invisible hand” of the market. However, the central notions of mercantilism still have resonance – hence recurring advocacy of protestionism – and the claim made on behalf of the optimal outcome from the private pursuit of private interests remains an area of active controversy – hence the continuing arguments about the superior social benefits ensuing from state ownership of business, planning, intervention and regulation.

Someone relatively famous once suggested to me the best prior disciplines from which to embark on studying economics are history and mathematics: history because to understand where we are we need to know how we got here; and maths as a tool for modelling, testing hypotheses and for decision-making, especially in situations of risk and uncertainty.

[1] Illuminating and retrospectively amusing insights can be gleaned from early economic legislation in England. The ravages of the Black Death (plague) 1348-51 created a shortage of labour relative to demand at going wage rates and thereby changed the bargaining power in markets of peasants and craftsmen. Real wages increased in consequence so the state intervened in an attempt to restore wage rates to pre-plague levels:

The dissolution of monasteries by Henry VIII, in pursuit of conflicts with the Catholic church about his divorce arrangements, removed an institution which had charitably provided for poverty abatement. As a result, his daughter, Elizabeth I, found it expedient to introduce legislation empowering local justices and then overseers to raise local taxes to provide care for the poor and needy, thereby establishing the foundations of what eventually became the welfare state:

In the official categorisation of the poor in 1563 legislation between the deserving poor, the infirm and the idle, the Elizabethans appear to have grasped at least a basic understanding of moral hazard.


Bob 07.17.04 at 10:41 am

Another thought.

It is often proclaimed by critics of economics that the subject isn’t and will never be an exact science like physics. If so, how come this?

“After nearly 30 years of arguing that a black hole destroys everything that falls into it, Stephen Hawking is saying he was wrong. It seems that black holes may after all allow information within them to escape. Hawking will present his latest finding at a conference in Ireland next week.

“The about-turn might cost Hawking, a physicist at the University of Cambridge, an encyclopaedia because of a bet he made in 1997. More importantly, it might solve one of the long-standing puzzles in modern physics, known as the black hole information paradox.

“It was Hawking’s own work that created the paradox. In 1976, he calculated that once a black hole forms, it starts losing mass by radiating energy. This ‘Hawking radiation’ contains no information about the matter inside the black hole and once the black hole evaporates, all information is lost.

“But this conflicts with the laws of quantum physics, which say that such information can never be completely wiped out. Hawking’s argument was that the intense gravitational fields of black holes somehow unravel the laws of quantum physics.” – from:


Frans Groenendijk 07.18.04 at 2:04 am

Since trackbacking does not seem to work I humbly invite you manually to read my followup on my own site. ( )

Comments on this entry are closed.