Why macroeconomics is not a science

by on December 2, 2013 at 9:43 am in Current Affairs, Economics | Permalink

From the United Kingdom:

Manufacturers enjoyed a jump in demand that pushed growth to its fastest rate for more than two years and saw the sector take on thousands of new staff last month.

New orders were the strongest for almost 20 years and job creation accelerated, according to the Markit/CIPS UK Manufacturing PMI survey.

There is more here, and I will reiterate that this trend was not very well predicted by any macroeconomic school of thought, including liquidity trap theories, recent emphases on long-run secular stagnation, or for that matter the contrasting “of course there is mean reversion” approaches, which don’t tell us much about timing and which would appear to contradict the slow recoveries seen elsewhere.  Spain, by the way, does not have an equivalent degree of cheer

1 Ray Lopez December 2, 2013 at 9:55 am

Just random noise. Next month will be down. More significant was Japan’s uptick in prices perhaps.

2 Michael December 2, 2013 at 10:05 am

…one month bucks a trend and suddenly macroecon isn’t a science? Seriously?

3 prior_approval December 2, 2013 at 10:13 am

Well, some people confuse micro and macro far too easily ….

4 Alex K. December 2, 2013 at 2:05 pm

There are many arguments supporting the claim that macroeconomics isn’t a science, one of the strongest arguments being that using entirely standard assumptions, you can construct microeconomic models which can explain almost any macroeconomic fact. If this is true, then how can one explain the economy using only aggregates of dubious coherence (which is what macroeconomics is today)?

The only reasonable way to behave is to identify measures that are robust to macroeconomic model misspecification (which _will_ happen) and apply those. For instance, helping the unemployed can be called humanitarian aid for those suffering the necessary structural adjustment, or it can be called fiscal policy that increases aggregate demand — it does not really matter what you call it, so it is a measure that is robust to model specification mistakes. If someone advocates a measure that depends crucially on some specific macroeconomic model being true, then you should bring down on him the entire weight of theoretical criticism and shame him into submission.

“prior_approval” is clueless as usual, behaving as if the macroeconomy falls down from the sky, instead of being the result of many microeconomic events.

5 Careless December 3, 2013 at 2:53 pm
6 RR December 2, 2013 at 10:36 am

What a ridiculous claim.

7 Ed December 2, 2013 at 10:43 am

The Guardian article doesn’t actually make an argument on whether macroeconomics is a science or not. It quotes a few economists, without deriding any of them for not having predicted this. I don’t think that many economists spend their time on the UK manufacturing sector.

What is a “science” anyway? I’ve never really understood why mathematics counts as a science.

8 Chris S December 2, 2013 at 10:55 am

It isn’t. Its Math.

9 Careless December 3, 2013 at 2:55 pm

I don’t think that many economists spend their time on the UK manufacturing sector.

Yeah, who cares about manufacturing in the world’s sixth largest economy?

10 Chrisare December 2, 2013 at 10:45 am

Science isn’t always right; economics isn’t a science not because it gets things wrong but because it’s not conducted scientifically.

11 chuck martel December 2, 2013 at 10:53 am

Wait a minute, isn’t collecting all kinds of numbers, putting them in columns and adding them up to something, isn’t that scientific?

12 Bob December 9, 2013 at 11:34 am

It’s very hard to apply the scientific method to something when you won’t make controlled experiments.

13 Norman Maynard December 2, 2013 at 11:02 am

Granted that Newtonian and particle physics are a more reliable basis than microeconomics, why is macroeconomics different in this sense than epidemiology, climatology, or astronomy? Yet they are all generally granted the label “science.”

14 dearieme December 2, 2013 at 11:08 am

There’s plenty of people who dismiss the claim of climatology to be a science. Unless it’s come on a mile since I did some reading a few years ago, they’re right.

15 Z December 2, 2013 at 11:30 am

Falsifiability has always been the test, no pun intended. Popper’s formulation comes down to testability. If an assertion, theory or proposition can be proven false by testing or observation, then it qualifies as science. Economists, I think, would argue that their theories can be and are disproved every day. Therefore, that makes them scientific. Plam readers could make the same claim, so take it for what you will.

16 ummm December 2, 2013 at 11:57 am

correct > wrong> not even wrong

17 Nicholas Marsh December 2, 2013 at 12:26 pm

Similarly, lots of people who are universally regarded as being scientists don’t make falsiable theories. Theoretical physicists have been working for decades on string theory. Which at the moment is still not falsifiable.

18 dead serious December 2, 2013 at 1:08 pm

Neither is the existence of god/s.

Religion = science. To right-wingers.

19 Marcos December 2, 2013 at 1:14 pm

Your comment can not be more wrong.

String theorists are working for decades, trying to get falsifiable predictions out of their theory, and they do get some from time to time. Problem is, experiments keep falsifying the predictions they get, thus the theorists keep changing their theory, and trying to get falsifiable predictions out of their new theories.

That’s how a mature (read stagnated) science develops. It does not lead to the great leaps that inmature sciences pass through, but it does not make them any less honest.

20 Norman Maynard December 2, 2013 at 3:26 pm

Marcos, you said “String theorists are working for decades, trying to get falsifiable predictions out of their theory, and they do get some from time to time. Problem is, experiments keep falsifying the predictions they get, thus the theorists keep changing their theory, and trying to get falsifiable predictions out of their new theories.”

If you replaced “experiments” with “new observational data”, you could say the exact same things of 90% of macroeconomists. It’s simply untrue to claim that macroeconomists aren’t *trying* to get predictions which can be falsified with data we could potentially collect. It just doesn’t happen often enough, and perhaps the changes to the theories aren’t big enough, to produce the kind of respect string theory gets.

21 Doug M December 3, 2013 at 12:06 pm

Some varieties of string theory can be, and have been, falsified.

The problem with string theory, is that there are so many variations of string theory out there, and new variants are derived faster than the old variants can be tested.

22 Doug M December 3, 2013 at 12:07 pm

Any sufficiently analyzed magic is indistinguishable from science!

23 chrisare December 2, 2013 at 2:37 pm

Economics may produce falsifiable theories in some cases but more often theories (and the reputations that depend on them) are defended against evidence of their falseness. Because social phenomena is so messy, theories just have to be accurate on average and not in every context to have value, defenders say. In other words, just because a theory is falsified in one context doesn’t mean its falsified for all contexts.

Or, as a last resort, others say that even if economic theories may generally be wrong they help us to clarify our thinking, a sort of mental discipline that doesn’t need to be accurate in most cases to have value. As a result, Mas-Collel is still taught in most PhD programs 20 years after its publishing even though behavioral economics has found much of it to be at best not consistently right and at worst just wrong.

24 Nicholas Marsh December 2, 2013 at 12:23 pm

Yes. It’s pretty ridiculous to expect a science to have to make consistent and reliable predictions. In addition to the examples you mention people aren’t very good at predicting earthquakes, but that doesn’t stop Geology from being a science.

25 Careless December 3, 2013 at 3:03 pm

But geology isn’t about predicting earthquakes.

26 Benny Lava December 2, 2013 at 7:35 pm

You may not know this, but astronomers make predictions that are falsifiable. Astronomers are able to predict things like eclipses of the sun and moon centuries in advance. On the other hand when did economists predict economic activity?

I’m always amused when conservatives start the “what’s a science?” debate.

27 Careless December 3, 2013 at 3:05 pm

“what housing bubble? Don’t be a credit snob!” <- this blog 8 years ago

They make lots of predictions. They're just not any good at it.

28 freethinker December 2, 2013 at 11:10 am

Deirdre McCloskey says science is disciplined inquiry, so that makes even literary criticism a science

29 Bill December 2, 2013 at 11:18 am

So, whenever Tyler makes a claim for a macropolicy, we can quote this back to him that it is not a science.

A Christmas present.

30 Lee A. Arnold December 2, 2013 at 11:34 am

Plumbers in the west Los Angeles area have noticed a distinct small uptick in biz in the last 6 weeks. Of course this is a high-rent district, but I have noticed before that our signals (both up and down) predate others by only a few months. Longer-term prognostication I cannot do: I was expecting this two years ago. Why? Main reason: Because at some point, deleveraging has to subside, new consumption should ramp up. HOWEVER: all the GDP’s are setting onto a growth path at a lower level, so I’m not sure the “secular stagnation” thesis is entirely disqualified. Macroecon needs to be rewrit to reflect the fact that private-bank credit-creation, and interbank trading thereof, has become a CORE of the economy. In consequence, without the equal creation of real (non-financial) value, the global-rentier “search for yield” is leading to serial bubbles. Yet, in the intro textbooks, the financial sector is supposed to be almost entirely subservient to the production of real (nonfinancial) goods and services. Now, I have never waded into the quagmire of upper-level economics, but it sounds like the higher reaches of macro do not divest one of this quaint illusion about the innocent function of finance. Two real-world consequences are, 1. government is entirely committed to propping the whole lousy thing up, and, 2. economists have developed a valid pack of intellectual tools, but it hardly matters, they are entirely lost at sea. They have the wrong system. All in all a rather classical tale of the destruction of morals, all the way around. Well done, all!

31 ummm December 2, 2013 at 11:38 am

Newton’s laws are not necessarily correct. they are just accurate for a limited set of conditions

32 bill reeves December 2, 2013 at 1:13 pm

Macroeconomics isn’t even a pseudoscience. At best it’s a superstition made pernicious by (in it’s Keynsian form) its utility in advising politicians to do that which is already their natural inclination: spend, spend, spend.

33 Doug December 2, 2013 at 1:37 pm

I think Tyler has been leaving a lot of bait for his readers/commenters in his most recent posts…

34 Jon17 December 2, 2013 at 4:19 pm

Oh my god people. Hard science involves repeatable experiments to falsify data. Chemistry has the test tube. Physics has the collider. Biologists have the Petri dish. It’s easy to reproduce experiments and prove people wrong. It may cost a few billion for people to build another collider but it can be done.

You can make the argument that micro is falsifiable. However, Macro has so many variables its hard to know what’s significant. You also can’t put the same system in a box and run the same experiment again. Thanks to the “animal spirits” it’s easy for people to pull a Paul Krugman and slither their way out of a bad prediction. That’s hard to do in the hard sciences. It’s just too easy for others to rerun the experiment and say “I don’t think so buddy”.

That’s the diff between science and the rest.

Reproducibility, controllable variables and falsifiability.

35 Urstoff December 2, 2013 at 4:48 pm

Hand-wringing over whether something is a science or not is pointless. Philosophers of science have long put to rest (because of the underdetermination of theory, infinite revisability, etc.) the notion that there is any major distinction between “science” and “non-science”, which, of course, was just trying to demarcate along lines already set by the philosopher (Popper somehow knew that Marxist historical analysis or Freudian psychoanalysis weren’t science but physics was, so he searched for a criterion). No, the much more important question is whether the field of inquiry delivers us important truths over untutored common sense and whether the propositions of the inquiry are more likely to be true than either untutored common sense or some rival inquiry. In that regard, microeconomics, at least, is deserving of respect (and I don’t think anyone argues the contrary; when people like to argue that “economics isn’t a science” usually they mean “macroeconomics” and more usually they mean “business cycle theory”). Macroeconomics is a more difficult question. Has macroeconomics (conceived as broadly or narrowly as you like) given us any important truths? Some propositions of monetary economics come to mind as do comparative advantage and the weak form of the EMH. Hopefully for macro’s sake the list is longer than that. However, concerning the question whether macro is more likely to give us true propositions than common sense or a rival inquiry, I think the answer is yes, but it must be paired with the observation that any specific proposition in macro may not be likely to be true itself (just more likely to be true than propositions of common sense or rival inquiries). So, we’re back to where reasonable people started: macro is very hard (because of the nature of the systems being studied and limits on the ability to conduct inquiry), and one’s confidence in the propositions of macro shouldn’t be very strong. However, it’s all we’ve got. There is simply no alternative that has proven to be better than the low bar that macro has set.

36 vetr December 2, 2013 at 9:53 pm

To take TC’s bait, I am going to put it out there that there are three, and only three, aspects to science. (1) There is chess expert /bridge expert/ gambling expert science, and in that sense anything a macroeconomist says that is smarter or less foolish than what a power hungry politician says is science, because it possibly averts real and greater losses. This form of science correlates to much of what real world professionals in finance, first and second world politics, and academic economics appear to me to try to do. (2)Then there is stamp collector science, which is what is teachable, and it is generally a romanticized but detailed program consisting of persistent effort to describe or redescribe already accomplished physical or biological discoveries, or to manipulate a codified version of formerly discussed equations, or to generate gradeable algorithms seeking to, fan-fiction-like, rediscover those previously discussed equations or physical reactions. This type of science is what kids pay to go to MIT or Caltech or Harvey Mudd to learn. By the way I am very impressed by highly skilled stamp collectors. (3) Then there is real science, which people who devote a lifetime to science generally (in my semi-informed opinion, based on a half century of having known or been related to two or three “geniuses” and several dozen “near-geniuses”, i.e., “scientists” at the Professor on Gilligan’s Island level or above) do for anywhere from a few minutes to a few hours every year, and that involves discovering something in nature and creation, something that is a discovery that is not based on teacher-approved algorithms or on cataloging and approving former discoveries (stamp collector science) or on critiquing, taking advantage of, or ameliorating the lack of common sense of other people’s misguided efforts to manipulate systems with simple rules and complex effects (chess expert bridge expert gambling expert science). ….The only people who know when macroeconomists have engaged in this third type of science are other macroeconomists – but this subject, like the subject of the fabled and unrecorded conversations of the legendary conversationalists of our great grandparents’ generation, is not a subject on which accurate opinions are possible by those outside of very finite and limited (and, with the passage of time, zero set) inner circles. Back to T.C.’s comments, who knows if he is right or not?

Comments on this entry are closed.

Previous post:

Next post: