Every other macro blogger seems to be taking a crack at this question. I like what they have to say. But I have a much simpler theory.
Let's suppose you wanted to design an experiment to test the effects of monetary and fiscal policy. And suppose you had the power to do whatever you wanted, and couldn't care less about getting clearance from the Research Ethics Board. What experiment would you design?
Probably something like: get 100 countries, then for each country toss two dice, one for monetary policy, the other for fiscal policy, then watch what happens for a couple of decades. That should settle the question.
Now, just for a laugh, imagine you wanted the worst possible experiment design. An experiment design so bad that no researcher would ever be able to figure out the effects of monetary and fiscal policy by looking at the data. What would you do?
Probably something like: make monetary policy negatively correlated with fiscal policy, and negatively correlated with any other shocks you observed hitting the economy. So it would be impossible for the researcher to disentangle the effects of monetary policy, fiscal policy, and any other observed shocks. If you were feeling really mean, you would do your best to try to set monetary policy and/or fiscal policy so that nothing ever did happen to the economy. So all the researcher would ever observe would be a few small fluctuations in the economy due to shocks you didn't observe (and he probably can't observe either), plus a few small fluctuations because you over-compensated or under-compensated for the shocks you did observe, and even here the researcher won't be able to figure out even the sign of the effect of monetary policy because he won't know whether tight money caused inflation or whether you just didn't tighten monetary policy enough in response to an inflationary shock.
That's the world we live in. We live in something very close to the worst possible experiment design for testing macroeconomics. We live in a house controlled by Milton Friedman's Thermostat. It's sometimes a wonky thermostat, that under- or over-compensates for the shocks it can see, but it's still a thermostat. For all its faults, it does not play dice with the amount of oil going into the furnace. Even if it did play dice, you could never be sure that it was playing dice, or just responding as best it could to some shock it sees that you don't see.
You want us macroeconomists to figure stuff out better? Sure. No problem. Just lend us 100 countries for a couple of decades, and let us play dice with monetary and fiscal policy.
Worthwhile Canadian Initiative 
One thing that I find odd is that you usually don’t get bitter philosophical disputes between theoretical physicists. If the data just isn’t there, then people will say “I don’t know” or “I’m just guessing.” The other thing is that physicists seem to get less offended at professional disagreements.
One other cultural difference that I’ve found is that most theoretical physicists are really excited when they are proven wrong. You’ll find that most particle theorists are a little depressed because all of the data coming out of LHC is consistent with our current models. If 99% of the things fit your model, you look for the 1% that doesn’t, because that 1% might be telling you something important.
The way that things would work in economics if people did things the “physics way” is that if you find a general rule (say market economies work better than centrally planned ones), then you’d go out of your way to examine the situations where the general rules don’t work. If the general rule works everywhere, then you start getting suspicious that you missed something important.
The other thing that I’ve found is that economists overuse math and formal models, but they have definitions that are extremely vague (i.e. what is a socialist economy anyway?) Also economists make a claim of being objective and rational when objectivity and rationality aren’t that particularly valued in physics. Theoretical physics values emotion and subjectivity.
Twofish,
I think that the problem is that the word ‘socialist’ is ambigious, which is slightly different from its being vague. ‘Bald’ is a vague word, but not an ambigious word; ‘tartarean’ is ambigious, but none of its meanings are notably vague.
Some of the definitions of ‘socialist’ are vague, but others e.g. “state ownership of the means of production” are not. When ‘socialist’ is required to do major theoretical work, as in the socialist economic calculation debate, it’s the ambiguity that often leads to problems e.g. in Mises’s sense, he says that there CANNOT be a socialist economy, but the Eastern Bloc seem like obvious counterexamples if one uses a different definition from him and thus commits the fallacy of equivocation.
(I’m an analytic philosophy student, not an economist, and our occupational sin is pedantry!)
Interesting exchange. Physics was my best subject at school, but I realised my math wasn’t good enough, and it didn’t excite me enough. I did philosophy (mostly analytic) BA, with a sort of unofficial minor in economics. Then switched to economics for MA and PhD, because even though philosophy was great, I wanted something a bit more solid I could get my teeth into.
Twofish,
I’m not so sure the actual practice of economics is so different from the “physics way” you describe. After all, there is a broad consensus among economists about general principles (the 99%, or 90% or whatever, of things that fit their model) – poll after poll of academic economists, for example, show an overwelming acceptance of the proposition that free trade is potentially welfare improving (this is the economic equivalents of “apples fall down”). The debates, or at least the interesting debates, revolve around the exceptions to the general principals (which is why, for example, much of the interesting work in international trade is on instances where free trade may not be welfare improving, e.g., infant industries, externalities, distributional considerations, etc.).
Where there is an important distinction between economics, as a social science, and the “hard” sciences, is that it is also subject to normative debates. A physcist may describe how the big bang happened or how protons decay, but he or she is never likely to have a discussion of whether or not that’s good or bad – it just is. For the same reason, you’ll never read an op-ed column (except, perhaps, in The Onion) on why the government should/shouldn’t be trying to stop proton decay.
On the other hand, 99% of economists may agree that free trade is potentially welfare improving (and in many/most instances, actually welfare improving – doubt you’d find any economist who thinks that imposing inter-provincial/state/EU barriers welfare improving), but if “welfare” has an inherently normative element to it (as it surely does,for example, “conservative” and “socialist” economists – and I use those labels only for convenience – likely disagree as to the impact, if any, of inequality on welfare), economists might disagree as to whether it is welfare improving in a particular instance or what additional policies need to be adopted to ensure it is welfare improving (say, a green tariff, or redistribution from the “winners” to the “losers”). Since economic policies, like free trade (as the last century demonstrated), are generally alterable by human agency (unlike the big bang or proton decay) these normative debates are inherent, and relevant, in any (interesting) economic discussion, in a way that they aren’t for physicists.
Every physicist would (presumably) agree that Newtonian mechanics is a good enough theory for figuring out how to drive a car.
Not all economists agree on basic stuff like whether monetary and/or fiscal policy can and should be used to steer an economy.
Nick,
Fair enough, but then again Newtonian physics has been around for 300+ years (and it built on the earlier work of people like Kepler a century earlier – Newton’s line about having stood on “ye shoulders of giants” is telling). The same can’t be said of macroeconomics. It’s easy to forget that macroeconomics is a relatively young social science. It’s a conceit to think that, because we’re modern, macroeconomists should be able to figure out the building blocks of their science faster than physicists did.
This also ties back to Mike’s point about the relative paucity of data points for macroeconomists. By the time Newton came on the scene, people had been accumulating reliable data on the motion of the planets for centuries, and if so inclined, they could collect 365 observations a year. Newton’s predictions could readily be falsified if they were not “good enough”. Macroeconomists have, what, a century’s worth of, maybe, monthly or quarterly observations, and then only from a handful of developed countries (many of whose economies are closely integrated throughout the relevant period)? How long would it have taken to confirm Newton’s predictions of planetary motion if we could only observe the planets once a quarter?
Macroeconomists haven’t found their Newton yet, or maybe they just need more giants.
Bob: and I’m remembering that Asimov(?) story “Nightfall”(?). where a planet had multiple suns, so they couldn’t figure out the basic laws of motion. In economics we have the n-body problem. Plus our “bodies” are a lot more complicated, so we need something a bit more complicated than F=ma and the gravitional law to describe them.
Agreed, physicists had it easy!
Nick Rowe,
Didn’t you study at Stirling? I’m working as a teaching assistant there right now in my gap year.
W Peden: Yep. Ask Sandra Marshall if she remembers an arrogant little first year student who thought that act-utilitariansism said everything that needed to be said about moral philosophy! (She convinced me otherwise, and I’ve been trying to figure it all out ever since.)
Nick Rowe,
I haven’t met her yet, but I’m sure she’ll be very happy to hear how well you are doing.
The fact that economists try to make normative value judgments is part of the problem, but I really don’t see why it’s so essential that economists do that. It would be seem to be more useful to just argue that if you have policy X you end up with result Y. One problem in trying to make value judgments, I’ve noticed that economists define values in ways that make the math easier. For example, you have the classical economic arguments that free trade maximize welfare, but “free trade” and “welfare” are defined in ways that make this tauntologically true. The fact that economists will often define terms to make them tauntologically true in a toy world that is completely unrelated to our own in fact makes these debates uninteresting.
The other thing is that when you are talking about value judgments and everyone agrees, then I think you have a problem. If 99% of the economists believe that “free trade” improves “welfare” then is that more relevant than the fact that 99% of the Catholic priests in the world believe in the Holy Trinity? It’s not obvious to a person that is about to lose their job that their “welfare” is being improved.
Also economic policies are often not alterable by human agency, and even they are there are timescale issues. You might be able to get the EU to have a coordinated fiscal policy, but not in five minutes. The other thing is that economic policies have feedback issues. If you implement policy X, you will get political result Y, which influences what you can do next. One major criticism of macroeconomists responding to the financial crisis was that their policy solutions often assume a “blank slate” in which you have all powerful politicians, when in the “real world’ policy is subject to actual political and legal constraints.
Nick Rowe: Every physicist would (presumably) agree that Newtonian mechanics is a good enough theory for figuring out how to drive a car.
It’s not. Even trivial things like wind resistance are things that you cannot calculate within a purely Newtonian framework. Most cars have electronic parts and anything with an electronic part is something that you just can’t figure out using Newtonian mechanics.
Bob Smith: Macroeconomists haven’t found their Newton yet, or maybe they just need more giants.
This is exactly the sort of thing that I was complaining about. Any working physicist knows that for all but the most trivial problems, Newtonian mechanics is either irrelevant, unusable, or more often than not, just plain wrong. You just can’t model even easy things like wind resistance or burning gasoline in a Newtonian model, and rather than trying to do that, real physicists burn gasoline and put things in wind tunnels to discover how they work. Sometimes you just give up and use real world data.
What I was complaining about was that economists often try to copy a cartoon version of what physicists do and how physicists think that has no actual relationship to how real physicists do and think. This is a perfect example of this.
Twofish: “The fact that economists try to make normative value judgments is part of the problem, but I really don’t see why it’s so essential that economists do that. It would be seem to be more useful to just argue that if you have policy X you end up with result Y.”
I disagree. Suppose for example we are considering a policy that gives people more choice. I may have no idea how people will choose if they are allowed to do so. So I have no idea what the results of that policy would be. But under certain conditions (people making rational well-informed choices, no externalities, etc.) I would say that policy is a good one. A lot of economics is like that.
W Peden: Stirling’s philosophy department was a very good place for me to be as an undergrad from 73-77. I hope it’s still as good. Those were very different times though!
Twofish: a lot of what we do involve choices affecting people’s lives. After graduation, my first job was to work on a model to help our provincial health ministry to decide on “objective grounds” how to spend the department budget in the “most effective” way. Even if we had been succcesful, it would have meant that some children would have been tested for some genetic disease while some elderly people would not have received supplementary meals. Or the reverse. You don’t get more subjective than that.
Also economic policies are often not alterable by human agency, and even they are there are timescale issues. You might be able to get the EU to have a coordinated fiscal policy, but not in five minutes. The other thing is that economic policies have feedback issues. If you implement policy X, you will get political result Y, which influences what you can do next. One major criticism of macroeconomists responding to the financial crisis was that their policy solutions often assume a “blank slate” in which you have all powerful politicians, when in the “real world’ policy is subject to actual political and legal constraints.
Which plays right into another of my economics peeves: modelling. In any economy with savings and investment (time-forward and time-reverse consumption) the minimum equation that can describe that completely is at least second-order differential. A second-order differential cannot change instantaneously and still be consistent; in an economics context if you try to change too quickly (uncertainty) you get default. Which is to say that an economy can never be completely consistent with itself and still adapt to uncertainty, there is always a cost of default. Any sort of information or planning is costly because of uncertainty.
Twofish, read my comment again, carefully. I didn’t say that 99% of economist believe that free trade is welfare improving for everyone (they don’t, everyone knows there are winners an losers), I said that it is POTENTIALLY welfare improving, and I made that distinction consciously. The potential welfare improvement doesn’t depend on value judgements, but how you weight the winners and losers and how you (re)distribute the gains does.
In any event, normative discussions are inherent the social sciences because human activity, unlike the activities of protons or black holes, has an (important) normative aspect. To say policy x gets you result y isn’t very helpful if you don’t know if result y is good or not, or how to weight that against alternative results.
Also the statement that economic policies are not alterable by human agency is fascinating – policies, by definition, are the product of human agency. If you don’t believe that, pray tell, who do you think sets them? The policies themselves might or might not achieve the desired result or achieve them at the desired times, but they are inherently the product of human agency. The fact that you go on to discuss political or legal constraints on policies only reinforces the point that policies, like politics and law, are the products of human agency.
Nick Rowe: Plus our “bodies” are a lot more complicated, so we need something a bit more complicated than F=ma and the gravitional law to describe them.
As of point of matter F=ma is false for any sort of non-trivial system. Also people sometimes have bizarre ideas about what physicists actually do. Most physicists work on quite complex systems, and the goal is to create a model that captures the “essence” of a system. It’s like a painter or a poet uses pictures or words to describe an emotional truth. Physicists look at systems and then try to figure out how to describe the “essence” of the system. Much of that involves trying to figure out what is important and what is irrelevant.
And sometimes you can describe something complex with something very simple. A good poet or novelist can take incredibly complicated feelings and write three lines that conveys the “basic truth” of a situation. Physicists try and often succeed at doing the same thing. The other thing is that sometimes a lot of complexity makes things simpler. If you look at three atoms, you end up with a mess of equations. If you look at a trillion atoms, the complexity disappears and you end up with a simple gas.
But a lot of this involves observing things. You look at a star or piece of metal or economy the same way a painter would look at a face and try to figure out what is important and what isn’t. And you might have to try a few times before you get it right. You paint a picture or write a poem, and it just looks and sounds wrong, and so you go back and figure out what you need to do to get it right. If you are lucky, you will stumble on some “general truth” but most of the time, you are lucky if you just describe the system. This fetish that macroeconomists have toward abstraction and universality is something that I’ve never really understood.
These turn out to be very useful skills for economic analysis and this is the reason that for some types of work, banks hire physicists. For example, after the 2007 crisis, the equations that described interest rates changed, and having equations that just change on you is something that physicists are used to handling. You go back and look at the data and think deeply about what exactly changed and why, and that involves going into the very nitty-gritty. One big thing is that before 2007, people assumed that a default by a major bank was impossible, and after the impossible happens, people start putting it into their thinking. Trying to figure out how to model that involves looking into the details of contract law, the institutional dynamics of the Federal Reserve, currency flow rules, and a lot of other factors, and it turns out that you can put that into the model.
Personally, I don’t think that even if you can perform random experiments on countries that you will get useful data. The thing is that countries aren’t black boxes and the fiscal and monetary dynamics of the United States are totally different from those of Germany or that of Peru. The fact that the dollar is the world’s reserve currency made the US behave very differently in 2007 than Argentina in the early-1990’s and anything that treats countries as country A and country B just won’t work. Also, technology means that economies behave differently in 2010 than they did in 2000 than 1990. One reason there is a lot of work for physicists in banks is that the fundamental economic rules change every three years or so on a global scale and they also change constantly at a local scale. Central Bank of Country A changes currency policy, and all the equations suddenly change.
There seems to be a “pseudo-physics fetish” in which economists try to invent abstract and universal rules, whereas physicists when given the same situation will just not even try to be abstract and universal, and just try to understand the dynamics of a particular situation knowing that it probably will change suddenly.
Nick Rowe,
Being a student in the mid-1970s must have been hard generally, with a fixed income and 20% plus inflation. I don’t know how my parents managed it.
Do you remember Alan Millar? He is still very active within the department, now as a professor. The Stirling department has gone from strength-to-strength over the last 30 years, like the rest of the university. I think it actually does better than Edinburgh now on student satisfaction.
Twofish,
I also get the sense from reading some physics and philosophy of physics that physicists are a bit better than economists at being explicit about their ceteris paribus assumptions, but that could just be that it’s easier for me to spot an implicit ceteris paribus assumption in economics than physics.
Rowe: But under certain conditions (people making rational well-informed choices, no externalities, etc.) I would say that policy is a good one.
The trouble is that that assumes a definition of “good” which is a problem if people have different definitions of “good” than you do.
I don’t have a problem with economists making statements like “policy X will improve personal incomes”. I do have a problem when it’s assumed that “improving income” is automatically good. A lot of economics involves trying to characterize a number as “good” and maximizing that number, and that involves implicit moral decisions which I think should be explicit.
What worries me is when people define something as “good” because it happens to be easy to calculate and quantify. If someone wants a moral principle that something unquantifiable should not be used to make moral choices, that an interesting idea, but it’s something to be discussed rather than assumed.
Bob Smith: I said that it is POTENTIALLY welfare improving, and I made that distinction consciously.
The problem with saying things like POTENTIALLY is that it makes falsifiable difficulty to impossible. I can just as well say that import substitution and Marxist central planning is potentially welfare improving, and at various points in the past, those philosophies had as much empirical justification and mathematical credibility as the neoliberal model did before the 2007 crisis.
If I say that policy X will improve “welfare” and we go down six months and it doesn’t, that’s good. I look at how my model mismatches reality, and we can make policy corrections. If you say that the policy potentially improves welfare without any qualifiers, then we do it for six months, and it looks bad. When do you figure out that you just got it wrong and that for whatever reason, comparative advantage just doesn’t apply in your particular situation.
Now if you say policy X will improve “welfare” if you include policies Y and Z subject to condition A, they we are getting somewhere.
Bob Smith: To say policy x gets you result y isn’t very helpful if you don’t know if result y is good or not, or how to weight that against alternative results.
But I don’t think that it’s the place of “experts” to make those decisions. I tell you that policy X will lead to result Y, you tell me if you think that’s good or not. If you tell me that you want a policy that results in Y, then I’ll come up with solution X. The thing is that I don’t see any reason to think that physicists or economists are any more qualified to decide what “good” is than random people with no technical skills.
Bob Smith: Also the statement that economic policies are not alterable by human agency is fascinating – policies, by definition, are the product of human agency.
Some are. Some aren’t. Most basic economic policies are specifically designed so that they cannot be altered easily. You can put in booby traps so that a policy can’t be altered at all (ask the people that designed the Euro.)
There are timescale issues. For example, it’s possible for Europe to have a unified fiscal policy. It’s not possible to have that done in one year or probably even five years. So if you have a banking crisis in which the time scale is hours, you can’t change the policy fast enough. It takes a few weeks to get an Act of Congress, and if you have hours, forget it. For some basic legal things, you need decades, maybe even centuries to change something.
Also a lot of institutional infrastructure ends up being either by accident or design to be difficult to impossible to change. If you want to turn the US into a Marxist society, that’s not going to happen without lots of difficulty.
These things matter particularly in a crisis, when time matters. I read a paper by a Treasury official that said that he was seriously annoyed at some economists for suggesting policies that just could not legally be implemented.
Bob Smith: If you don’t believe that, pray tell, who do you think sets them?
It depends…..
A surprisingly large amount of stuff “just happens.” The thing about highly complicated non-linear systems is that you end up with a “ghost in the machine” and things happen in which no human actor can be said to be the cause. Markets often work well because they can process information faster than any individual human being can, and they have dynamics that are independent of any particular component part.
If you believe in chaos theory and quantum mechanics then a lot of stuff is just random. Some electron jumps, a butterfly flaps its wings, and then empires rise and fall.
It’s complicated, it’s situation dependent, but its the type of thing that physicists just love to figure out.
Let me just describe how I see the situation.
1) the standard model of comparative advantage says that free trade (as defined in the model) improves welfare (as defined in the model). This model has mathematical definitions of all of these terms and just states a mathematical fact
2) therefore in situations in which the model is an accurate description of the reality and in which the terms mean what we want them to mean, social welfare will be improved by free trade
The tricky part is to figure out whether the model is in fact a useful and accurate description of a particular situation. That requires pretty deep knowledge of the situation you are trying to model, and you can just try it and see if it works.
Categorical statements like “free trade increases social welfare” is something that I want to avoid. Making statements that sound like “universal laws” is a bad idea, IMHO.
What I can say is that “in this particular situation, the dynamics of the system is accurately modelled by comparative advantage, and therefore we expect if we reduce tariffs than personal incomes will increase.” If X doesn’t happen then you go back and figure out what that means. If you reduced tariffs and personal incomes didn’t increase, then you made a mistake somewhere. This might be a good thing. A model that produces an obvious mistake is better than one that can explain everything, since something that explains everything, explains nothing.
Also if you reduced tariffs, and personal incomes did increase, then you still need to dig some more.
Maybe you got the right answer for the wrong reasons. For example, maybe the reason things got better wasn’t that people are doing comparative advantage, but rather because you can create larger markets with larger economies of scale. Maybe in reducing tariffs, you decreased rent seeking which increased internal economic efficiency. Getting something right for the wrong reasons is dangerous because you might end up pushing things in the wrong direction (i.e. all of developed nations that went for Soviet-style central planning in the 1960’s).
All models are false, some models are useful. Newtonian dynamics happens to be a useful model for some very, very limited situations, but in most interesting situations it’s either incomplete, or useless, and you have to figure out what are the relevant dynamics of the situation, which more often than not have only limited connection to Newton’s laws. (Also there is a tauntology here, if you have a situation in which Newtonian mechanics is a good model, physicists will be uninterested and give it to engineers to handle.)
It’s no different than the efficient market model or the comparative advantage model, and I find a lot of this “Newton worship” to be bizarre and a little dangerous. One of the most dangerous things that you can do is to take a model and push it past its limits.
This is basically a philosophy point. Different physicists have different philosophies and philosophies of science can change over time. The idea of a “clockwork universe” or science as a quest for “universal laws of nature” was popular in the 19th century and early 20th century, but I find it bizarre, and I find it even more bizarre that I see traces of it in a lot of social sciences.
What I’m seeing is that social scientists seem wedded to a philosophical view of “science” that was popular in the 1950’s, but which most physicists have abandoned because it just doesn’t work at understanding complex systems. Something that I find weird is that there seems to be an agreement that “pseudo-Newtonian” models don’t work very well in economics. Okay fine. But maybe the solution is to just abandon the Newtonian view of the world and find something else, just like most physicists have done.