Hacker News new | past | comments | ask | show | jobs | submit login
Economics Has a Math Problem (bloombergview.com)
169 points by gedrap on Sept 2, 2015 | hide | past | web | favorite | 195 comments

I was always interested in both Physics and Economics and did studies in both disciplines.

As a physicist, the models that I encountered in economic theory always left me quite disappointed. The main reason for this was that -like the article says- most of the models used in Economics are based on abstract reasoning about "ideal" markets and actors (or more recently on game-theoretic ideas) instead of experimental data. One reason for this is of course that when many of those models were developed (in the 50s or earlier) there simply was no reliable (micro-)data available that one could develop a theory against. Another big problem that kept Economics from taking a more experimental stance towards model generation is of course that until recently it was very hard or outright impossible to conduct large-scale experiments, which are the main instrument to validate (or better, not falsify) a given theory in other disciplines such as Physics or Biology.

That said, the recent computerization of all aspects of business and the creation of virtual economies -like Eve Online, World of Warcraft- and "transparent" markets -like Bitcoin- should provide ample data to develop "real" models of economic behavior against, and I think that many researchers actually already make use of this data.

The theories that will result from this will probably be more like those developed in statistical mechanics though -i.e. making statements about the aggregate behavior of the system- rather than those developed e.g. in electrodynamics, where we usually can predict the behavior of even a single particle. Would love -and be at bit scared- to be proved wrong about this of course :)

> most of the models used in Economics are based on abstract reasoning about "ideal" markets and actors (or more recently on game-theoretic ideas) instead of experimental data.

Essentially it's like pre-Enlightenment Natural Philiosophy, in that it's a discipline that claims to explain the way the world works, but actually explains who has the most sway in getting their ideas accepted.

> ...should provide ample data to develop "real" models of economic behavior against, and I think that many researchers actually already make use of this data.

I suspect reality will get about as much traction against idealogues of e.g. the Chicago School as actual climate science does agains useful idiots like Bjørn Lomborg. It's about providing views that happen to be useful to moneyed interests, not reality.

I suspect reality will get about as much traction against idealogues of e.g. the Chicago School

This is an odd criticism. I could see it applied to Chicago's cousin, the Austrian school, which clearly relies more on ideas (viz. the work of Hayek and Mises, both of whom considered themselves political philosophers as much as economists).

But Chicago is just the opposite. It's been criticized for being too tied up in the mathematics of it all. Chicagoans love numbers.

It's my understanding that your right that Chicagoan's are known for their love of math but not necessarily their love of empiricism. However they almost exclusively use RBC(real business cycle) models which are created based on stylized assumptions about human behavior(always rational, lives forever, always able to borrow) and then the parameters are played with until it fits some data in the real world(usually the parameters are completely unrealistic). And they make pretty terrible predictions. To my knowledge they aren't ever used to actually make any predictions about the economy. People who do predictions like the(CBO, FED, IMF etc..) use old school Keynesian-style aggregate models of the economy.

Larry Summers said of RBC models

1. RBC models use parameter values that are almost certainly wrong,

2. RBC models make predictions about prices that are completely, utterly wrong, and

3. The "technology shocks" that RBC models assume drive the business cycle have never been found. (They assume that during recessions we just forget how to be efficient.)

And they make pretty terrible predictions.

That's a fair criticism, as far as it goes: it's been said (of Austrians, anyway) that they've predicted 8 of the last 3 recessions.

The thing is, qualitatively, we can say that "old school Keynesian-style aggregate models" are just as bad. The difference is that, rather than modeling based on coldly-rational actors, they're modeling based on (and I quote) "animal spirits". And those models told us that the end of WWII and curtailing war spending was going to usher in a new depression worse than the Great Depression; and that the Sequester and "fiscal cliff" were going to plunge us back into the depths of 2008.

The fact is that both approaches like to pretend that they understand what's happening, but neither really do. Anybody claiming they have an accurate model of macroeconomics is trying to sell you something.

HOWEVER, this is all talk about Macro.

If you take a look at the Micro world, you'll see that Chicago really are the winners. Microeconomics is a solved problem, at least to a first approximation. And the progenitor of that was pretty much Gary Becker - from University of Chicago.

So it's not fair to criticize Chicago when (a) in the micro world, they really have been triumphant; and (b) in the macro world, they're just a different kind of wrong than Keynesianism.

The fact is that both approaches like to pretend that they understand what's happening, but neither really do. Anybody claiming they have an accurate model of macroeconomics is trying to sell you something.

Thank you. Simple and true.

The claim that the U.S. would have grown at the same speed if the sequestration had not gone into effect is a minority view not held by the FOMC, the IMF, or the CBO.

Have there been experimental or empirical studies to support that micro is a solved problem?

I'm not sure how one would do a study to test a statement so sweeping. But I was probably too glib in any case.

It would be more accurate to say that the framework within which microeconomic phenomena is pretty well understood. While there's still plenty of room for research to fill out all the details, there's very broad consensus about how the answers tend to look. Concepts like price elasticity are well understood and non-controversial.

That stands in sharp contrast to macroeconomics, where even when talking about the same stuff, say, ZIRP, there's fundamental disagreement about what factors we should be researching. (and when I say "we", I mean people that aren't me, 'cause I'm not an economist)

> But Chicago is just the opposite. It's been criticized for being too tied up in the mathematics of it all. Chicagoans love numbers.

Empiricism often uses numbers and math, to be sure, but love of numbers and math is not love of empiricism: you can use math to tease out implications of your abstract, first-principles, non-empirical model as well as you can use apply it to real world observation in an attempt to confirm or refute an empirical model.

A criticism I've seen of the Chicago school has certainly been that it is more concerned with the mathematical implications of idealized assumptions and less concerned with how well those implications reflect real-world results.

"New York University economist Paul Romer recently complained about how economists use math as a tool of rhetoric instead of a tool to understand the world."

It's possible to love numbers and still not be interested in empirical data. One thing you may notice about much raw economic data is that it was selected because (a) it was easy to collect, or (b) it gives "reasonable results" according to the accepted theory.

See also How To Lie With Statistics.

I see an awful lot of from-a-state-of-nature fairy tales in economics, which is usually employed to justify a bunch of garbage moralizing about how markets are supposed to be. It's almost as bad about it as political philosophy—and oh, god, when political philosophers write about economics...

Great comment. Incidentally, have you come across Deirdre Mccloskey, I've been reading her book 'Bourgeois Dignity: Why Economics Can't Explain the Modern World' and finding it really interesting. Economics is a social science, taking measurements / accumulating data and theorising on that data causes changes in behaviour. Just sitting around talking about economics, or your weekly budget, drives changes in behaviour.

As a layperson with a bit of an interest in economics I don't really see how, as the article says "econ is now a rogue branch of applied math". More like economics uses applied maths, or something like that. It's a 'social science', we can't reveal the laws (let's call them 'habits') of economics like we do The Laws of Nature, because circumstances change -both globally and locally (macro / micro)- and new economic phenomenon emerge.

It'd be interesting to see what future historians have to say about present day economics.

Much denser, but if you enjoyed mccloskey, I recommend The Romantic Economist by Richard Bronk. It also repudiates the rigid but unrepresentative models of the average economics lesson, and explores the role of imagination and creativity (novelty) in the dynamic nature of economics that you've described.

> Economics is a social science, taking measurements / accumulating data and theorising on that data causes changes in behaviour. Just sitting around talking about economics, or your weekly budget, drives changes in behaviour.

Funny - you could say the same about physics. For example, knowing that the coming car may kill a person causes the human behaviour to change, namely the driver will stop the car. Does that invalidate Newton's laws?

Yes we probably cannot predict universe totally. But we can do scenarios (of complex emergent systems of interacting agents with bounded rationality), just like in physics. In fact the whole point of doing that analysis is to change behaviour of people, one way or another.

> Funny - you could say the same about physics. For example, knowing that the coming car may kill a person causes the human behaviour to change, namely the driver will stop the car. Does that invalidate Newton's laws?

The difference is that whether a person will try to move out of the way of an oncoming car is not considered within the domain of physics (so them using their knowledge of physics to predict that a car is a danger may change their behavior, but not the accuracy of what physics predicts within its domain), but whether or not a person will try to move their money out of the path of a predicted stock market collapse is within the domain of economics.

This doesn't actually make empiricism any less applicable to economics, but it does compound problems in the empirical study of economics.

No, it's an artificial distinction (or perhaps better would be to say, there is an artificial distinction where you decided that physics is not a social science). In both cases, you can analyze two scenarios:

1. Car continues and hits the person / investor won't move money out before stock market crashes

2. Car is stopped / investor will move the money out

In both cases, the physics or economics is no less applicable whatever the circumstances of the actual decision. And just like physics cannot tell you whether the car stops, economics cannot tell you whether investors will actually move the money out. But what economics can tell you, under certain assumptions, say, investors want to make money with this and this horizon, whether or not are they likely to move out.

Let me give another example. Water flows downhill due to gravity. Yet, people can decide to build water pump and aqueduct to get water uphill. In doing so, they didn't break any physical laws. But if you ignore the pump, it may seem that the water flow downhill anymore and so the laws are incorrect. What happened is that now our model of the situation should include the pump, and within the expanded model, physical laws are again preserved.

Similarly for instance, if the society decides to regulate markets, the behaviour of the people can be changed, but they don't necessarily have to break some universal economic law. The new behaviour can have a perfect economic explanation, it's just that the frame of the model changed.

It seems that this "infinite regress" confusion (also called Lucas critique, if I am not mistaken) is caused by ignoring the fact that every model of reality only includes a portion of it and is never a perfect description (for example we ignore the physics of human brain when we model the car hitting the pedestrian).

Also, you should note there are situations where knowing more doesn't actually change the behaviour (under assumption of certain rationality) - for example, getting to know the exact odds of casino wins won't entice me to play.

I was not aware of this work actually, thanks so much for sharing this!

All theories are abstract, that's the whole point of a theory. Without abstraction there's only observation and no reasoning. That's why I'm highly sceptical of the claim that it is possible to understand something simply by "letting the data speak" without "committing to a theory".

I don't say that having abstract theories is bad, after all one of the most successful theories in Physics, general relativity, was formulated without much data or even experimental hints to work with. The difference is that this theory was later put to scrutiny by performing experiments and by observing natural phenomena (e.g. light bending by gravity), which have failed to falsify the theory so far.

So what I'm saying is that a theory that cannot be falsified experimentally is not very useful.

The real danger here is that these theories -most of which are supported by no or very slim experimental evidence- are used by governments to formulate economic policies, which in turn affect the life of millions or billions of people. For me, this is a bit like developing an "idealized" theory of nuclear physics and then trying to build a nuclear power-plant based on it. I, personally, would not want to live near one of these ;)

In your nuclear power example, the counterfactual is don't build the power plant.

Suppose governments did not formulate economic policies with these imperfect models. What would they do instead?

I wonder to what extent politicians formulate policies based on models or make high level policies based on ideology or perceived voter preferences and then select the economic models to justify what they have already decided.

Edit: this isn't a comment about economics but how economics is used in a political context.

Of course they will have to take some decision even if they don't have a solid theory to back it up, but the problem today is that many governments try to make reality fit their idealized theories while it should be the other way around. I understand your point though, the lack of quantitative understanding of societal phenomena -on a national as well as international level- is a huge problem.

I don't disagree, I was criticising the idea stated in the article that one can understand the economy without using theoretical models.

> Without abstraction there's only observation and no reasoning

This is what the real science is. Observation and only observation. Reasoning is secondary and unimportant.

> This is what the real science is. Observation and only observation. Reasoning is secondary and unimportant.

"Observation and only observation" is how you build an inchoate collection of useless factoids, it's not how you do science.

Observations are a very important part of the scientific method and you have to be either rooted or anchored in it, but it's only one part of science, and it's by no means the most important product of it. Observations are seeds (from which patterns are recognised and hypotheses formed) and filters (data gathered to test the predictive power of hypotheses).

> collection of useless factoids

We call it phenomenology. The very base of any proper science. The rest is just a mechanical work of eliminating the noise from this data.

Science should never produce any knowledge which is not found in the original data (either passive observations or controlled experiments - does not really matter). The scientific method is all about techniques of extracting the most compact representation (i.e., the information) hidden inside this massive data sets. It turns out that the axiomatic theories are very common and compact representations for many areas of science, but not necessarily the universal representation.

Call it what you want.

What you disregard as 'just a mechanical work' is nothing of the sort. It has required all the ingenuity of the greatest genius in history to get modern science out of the data, and any future advances will require even greater insights.

If anything, the data gathering is the 'mechanical work' part of science.

Being "mechanical" does not mean anything "easy". It's hard, nearly impossible, simply because the most formal definition of this process boils down to an exhaustive search in a huge (but finite) morphological box. That's why we need the greatest minds that are able to find suitable search heuristics.

But yet, this process does not produce any new knowledge on its own. All the knowledge is there, in the data. Waiting to be extracted.

Except when the new knowledge is in the form of pure mathematical research.

Then, some decades later someone finds the practical application, and the data that matches the model.

No, this is not a new knowledge, and not a new information (as defined by the algorithmic information theory), it's a non-falsifiable piece of, well, probably art. For this very reason mathematics is not considered a science.

Not true. With reasoning, one can develop a model, and this model can be tested against reality by observations, thus closing the circle.

Mere observations give no interesting and no useful results. Basically, you'll end up with an ever-growing list of facts.

But all the reasoning should always be a direct consequence of the data. Not the other way around.

Science, essentially, is nothing but a process of extracting information (as in the algorithmic information theory, not in Shannon sense) from the data. Reasoning is just a tool, one of the ways of representing this information.

The OP makes a fundamental error:

Traditionally, economists have put the facts in a subordinate role and theory in the driver’s seat. Plausible-sounding theories are believed to be true unless proven false, while empirical facts are often dismissed

The error is the assumption that facts are, well, facts. In real life, it's not nearly so black-and-white. Remember back in high school physics, when we did experiments with Newtonian physics, but our answers never quite worked with the theory - not just because of Einstein, but mostly because of (a) measurement error and (b) we were (knowingly or not) assuming point-masses on frictionless planes, etc.

As long as we were willing to write off our errors against those two sources, we could never discover that Newton's theories were incomplete, and we needed relativity to explain differences. We only got to that conclusion via models.

The way that you choose to measure fundamentally drives what your measurements are. And you can only decide what and how to measure if you start off with some model of the world.

Of course, this process of extraction of information from a data is lossy, almost never lossless (in some purely classification sciences it may be discrete and lossless, but it's a totally different story). Data is not perfect, and the required precision is always finite. Once you take it into account, you can think of the entire process of scientific discovery as a mechanical, dumb data compression.

So, the corrected wording for that paragraph should be the following:

Plausible-sounding theories are believed to be true as long as their numeric predictions conform to all of the experimental facts within this theory supposed range, up to this theory supposed precision.

And a theory falsified by a new set of facts does not become immediately "false". It just get its range cut down to that range of facts where it is still giving an acceptable prediction precision.

Back to physics analogy, the phlogiston theory is still correct and true, as long as it is applied within the range of the phenomenological thermodynamics. But this does not happen in the classic economics - all those crazy theories are coming from the depths of an insane mind, not from any particular set of facts. They never had any specific range where they were known to be applicable. Therefore, phlogiston is science, and economic theory is not.

But if we were working entirely from observed measurements, we'd never have discovered relativity. It wasn't evident from the measurements you'd get from normal experimentation.

I mean, it is possible to measure the effect. But you need to first design the experiment in order to make the measurement. It was necessary for Einstein to sit and think for a long time, to come up with the idea, which others were able to design experiments around.

You can't just say that somebody would have noticed the discrepancy sooner or later. Really, doing experimentation that way is a statistical error, of the sort that's really plaguing the sciences these days.

As I explained elsewhere in this thread, the special relativity theory is a direct consequence of all the experimental evidence that existed on the electromagnetic phenomena. Formally merging Maxwell theory with the classical mechanics yields special relativity.

General relativity is an application of a well known extrapolation heuristics - the simplest possible generalisation of a model. The other possible generalisations were much more complex, and the nature for some reason tend to favour simplicity, so it was reasonable to go that way (as well as it is very reasonable now to go out and test all the possible extensions of the Standard Model, starting with the simplest ones).

I think it works both ways. Theories help us decide what to observe, or where to look and what we're looking at.

Observation is useful to develop a theory (induction, or "educated guessing"). However, having a theory guides us towards newer, often subtler observations as we attempt to falsify it. Karl Popper called this the theory-ladeness of observation [1]

[1] http://plato.stanford.edu/entries/popper/

> Theories help us decide what to observe, or where to look and what we're looking at.

Of course. But it's merely some heuristics helping to cut down the search space, not an inherent, fundamental property of the universe.

> But all the reasoning should always be a direct consequence of the data. Not the other way around.

In order to test a model, it is useful to collect new data.

No. Science requires induction or deduction (or both) which are 2 forms of reasoning.

The best theories on economics date back much further than the twentieth century. The history of economics is far more interesting than trying to run experiments. A talented economist has to be able to come up with explanations for things they observe, which can then be proven out with further data.

The mistake is in thinking it should or could be a hard science like physics.

Exactly. I'm also doubtful that taking a more real science approach to economics (hypothesis -> experiment -> result -> rule) will work. It's a science about societies and societies change.

For example, the financial markets are the most measured markets out there, with the most microscopic details recorded. One interesting thing that happens after financial crashes are crises of theory. For example prior to 1987 American options didn't show [volatility smiles](https://en.wikipedia.org/wiki/Volatility_smile). After the crash the smiles appeared, making the markets a very different place. The same happened after Subprime and the LIBOR scandal, when the notion of a "risk-free rate" suddenly disappeared: governments now had to compensate for their default risk and market quotes were unreliable. Academics still struggle with the whole notion, although it seems that the current practice is settling around OIS discounting and CVA/DVA.

The point is, rules change. Hence in economics you can't just empirically deduce a rule and let it be, you constantly have to recalibrate it.

Astrology has also developed over a couple of millennia.

The fact that some people use astrology to play the markets, and seem to get results that are as good as those of economic theory, says a lot about both.

Economics is entirely a faith-based pseudo-science, which plays the same role in modern culture as the nonsensical theistic "explanations" of the world put out by the Church played in medieval times.

It's a form of political/rhetorical persuasion and social control, with an implied morality that's never seriously questioned. Even if you accept that it's politically descriptive and not politically predictive, it still fails.

It's not a science in any useful way - it simply pretends to be one, because that makes it appear more authoritative.

The real scandal for me is the fact that skeptics and debunkers like the amazing Randi have spent so much time and energy on trivial topics like the paranormal, while leaving the epic intellectual fraud and imposture behind mainstream economic theory unchallenged.

This is just a train wreck of misinformation.

First of all, not a lot of people actually use academic economic theories to "play the markets". Ask a trader how often does (s)he use the General Theory or monetarism in their daily decision making. There are macro-strategies, used mostly by fund managers, but it's not such a big portion of the market.

Secondly, anyone found to use astrology in their market activities would be fired in nanoseconds. Whichever institution. You can do it privately though.

The fact that one can get good results speculating while using astrology as a decision generator (or in fact any other random mechanism, like a monkey throwing darts) is actually known and used in the financial theory. It is called "the random walk hypothesis" and is the foundation of a lot of assumptions. A good introduction to it would be "A Random Walk Down Wall Street" by Burton Malkiel.

So I urge you to actually build an informed opinion before succumbing to an Alex Jones brand of reasoning.

>This is just a train wreck of misinformation.

So are the markets.

>First of all, not a lot of people actually use academic economic theories

Maybe not, it you don't consider Black-Scholes and it's endless brood of offspring to be serious academic theories.

Not many people use nation-state Keynesian or other big macro policy theories to play the markets - which is a different point.

But it's a relevant different point, because it highlights the political reality of using macro as a hand-wavey excuse for policy which benefits market operators and the owners of the money behind them at the expense of the rest of the population.

>It is called "the random walk hypothesis" and is the foundation of a lot of assumptions.

Quite. Does it not worry you that this is true, or that even with this insight manic-depressive boom-bust behaviour is apparently still considered a feature, not a bug?

>So I urge you to actually build an informed opinion

I suspect my opinion is at least adequately informed. Perhaps you're mistaking being heterodox for being wrong?

The fact that some people use astrology to play the markets, and seem to get results that are as good as those of economic theory, says a lot about both.

So meteorology isn't a real science either, I assume? Having a perfect model doesn't mean you can predict the behaviour of a system; conversely, not being able to predict the behaviour doesn't mean the model is flawed, or non-scientific.

That said, I do agree that economics is taken way too seriously for its incipient state.

> That said, I do agree that economics is taken way too seriously for its incipient state.

I think this is the most spot on statement on economics I've read.

Have you seen the documentary on the amazing Randi? It's called an honest liar. It shows that he actively deceived people in order to push his own agenda of what 'truth' is. So I didn't think he was a debunker as much as a persuader.

FYI - I met Randi personally about eleven years ago. He was a wildly charismatic person to talk to. It doesn't show as much when he's on stage or on an interview speaking to an audience.

Doesn't economic theory say that you can't play the market? That's confusing.

Nobody believes that markets are strong-form efficient. There is and always will be differences in information between players in the market. It's just a useful way to talk and think about how a market with perfect information would behave. That's why semi-strong and weak form efficiency are usually taught in the same lecture :)

> Doesn't economic theory say that you can't play the market?

That depends which theory; under most that can be taken as even semi-plausibly as models of the real world, no, unless you assert some simplifying assumptions (which then make them no longer semi-plausible models of the real world).

The simplifying assumptions are common in intro-level Econ classes, for the same reason that, e.g., assuming the absence of friction is common in many parts of intro-level Physics classes.

The thing is that we don't have as many people in the public media trying to sell policy using explanations of physics that leverage the fact that most people that know anything about the subject do so through hazily-remembered intro-level classes as we do for economics...

That is not economics in the macro/micro sense, that is more financial/ financial markets theory. It really is a different branch of thinking than what economists deal with.

Finance != Economics

Shhhhh, that would require to actually be familiar with it.

>The fact that some people use astrology to play the markets, and seem to get results that are as good as those of economic theory, says a lot about both.

Economists hardly ever try to use the results of economics to predict the stock market...in fact economics is the field that has theories that no strategy exists that can beat the market reliably.

Broadly agree, but I don't entirely blame the economists, the politicians need justifications for what they do, and appealing to economics is a great one. Most political movements that have been successful have had some economic backup (eg Friedman and the libertarian right in the 1980s). In practise much of policy is just trying to change stuff using financial incentives, which in many cases has unintended consequences or is too complex.

Only the real science approach ever works. If your "science" is using any other approach, it is not a science, period.

And there is a proper, scientific approach in the economics: behavioural economics. Instead of hand waving about some "ideal", non-existant agents acting fully rationally upon a 100% correct and impossible information available, as the classic bullshit economics does, the proper, behavioural economics starts with studying the real individuals behaviour (which is very much doable, no matter what the others are saying), and only then drawing conclusions about the herd behaviour in a statistical scale.

I'm sorry, but I don't want to debate your personal definition of the word "science". Per convention economics is called a "social science" and that's a convention I will adhere to.

In addition to what JupiterMoon says, I'd say that individual behaviour was never a real problem in economics. A lot of the theories are built on stupid assumptions about rational behaviour, but that's just it: you always can change the assumptions and see what happens. Behavioural economics is just that.

The real problem of economics is scale and inability to cleanly experiment on that scale. To imply any kind of large scale "herd" behaviour from individual behaviour would still require a tonne of a priori, non-falsifiable assumptions. Even if the assumptions can be shown to be valid now, it would be impossible to prove that they are stable and valid for a long period of time.

Yet, the "spherical cow in a vacuum" expression came from a hard science. Einstein didn't exactly develop his relativity theories by making measurements and drawing statistical conclusions either.

Pure theory has its place in science, as long as one is honest about its limitations.

Relativity theory fits well into this description. It (the special relativity) was derived from a formal attempt to unify classical mechanics and Maxwell electrodynamics, which was only possible if you ditch the Galilean relativity. Both classical mechanics and electrodynamics, in turn, came from a huge body of experimental data.

General relativity is a different thing, it was a pure hypothesis, an extrapolation of the simplest possible generalisation of the special relativity [1]. And it held this hypothesis status up until it was sufficiently confirmed by the incoming experimental data. Before that, a multitude of alternative hypotheses existed which had the same numerical predictions covering the available experimental data (e.g., https://en.wikipedia.org/wiki/Anatoly_Logunov#Relativistic_t...).

So, yes, it's important to distinguish theory from a hypothesis.

[1] this "simplest possible generalisation" method works most of the time, but not always, so it's mostly valid to do so in the hard sciences, and should never be used as a justification for hand waving in the social sciences.

Isn't his point that there is a sort of uncertainty principle in economics? For example when you study people they behave differently and furthermore as soon as they think that you think that they will behave in a certain way they may well behave differently -- especially if there is financial gain at stake for changing their behaviour. As you say this all derives from the biggest problem with all the social sciences people are not rational actors and any mathematical model has to encapsulate real human behaviour.

> For example when you study people they behave differently and furthermore as soon as they think that you think that they will behave in a certain way they may well behave differently

That's not an inherent limitation for applying science. Signal theory has models that represent circuits with feedback loops, where the output ; similar models could be developed to study people who know are being studied.

Also, if you use data collections that were compiled for other purposes, you can study the behavior of people who don't know are participating un such study. Big data is essentially that.

> As you say this all derives from the biggest problem with all the social sciences people are not rational actors and any mathematical model has to encapsulate real human behaviour.

There are theories that don't assume perfect rational agents, and can still predict behavior by studying the common human bias. See the work of behavioral economist Dan Ariely for an academic approach.

The mistake is in thinking it should or could be a hard science like physics.

That's fair, so far as academic research goes, as there are plenty of academics doing respectable work in in non-scientific fields such as music and poetry.

But it's another matter to decide if and how to ethically apply the results of soft-science research (likewise music and poetry) in ways that affect society.

Those theoretical models have held up to quite a bit of empirical scrutiny, and almost perfectly so in highly competitive markets like crops, energy (production, not distribution), housing, etc. I would say that the models have remained simplistic because economists have been trained to recognize the conditions when the theoretical models break down.

Competition (or lack of it) is absolutely the number one area where the simplistic models break down, with psychological phenomena being a close second. Even where there is demonstrated and measured irrationality in individual decisions though, in aggregate the distortion effects tend to be minimized.

Outside of these known areas of completixity and modeling deficiency, there have been several major breakthroughs in modeling and understanding other seemingly puzzling factors. Matching problems, search frictions, moral repugnance, game theory, asymmetric information, and discrete choices. There have been huge advances since the 1950's, it just takes some time to catch up with it all.

Experiments are pretty damn crucial to the entire scientific process, right? What if you couldn't do controlled experiments? That's where economics is right now. Until we either get enough data or economists can run actual experiments (ie those that don't just rely on fabricated mathematical models) I don't think we'll have real rigor in economics.

In the meantime, we still have to do our best. People still want ways to predict the future which suffer less from human interference (which is ideally what a mathematical model will let you do). That's still better than just asking people what they think (in some ways).

A lot of simplifying assumptions have to be made to say anything useful about a system which includes billions of intelligent actors trading trillions of dollars. There's obviously room for improvement, and I think that'll happen once people are given the proper tools to make it happen.

We are almost to the point where an entity like Facebook or Amazon could do controlled economic experiments. Or they could gather so much data, that when filtering them based on the experimental criteria, whatever remains could show a significant result.

Aside from the technical possibility of doing economic experiments, there is also the ethics of doing them. You are not only involving humans, but doing so on a grand scale. To prove a point, you would not only need to possibly send an economy into a recession, but you would have to do so dozens of times to get significant results. The people you are experimenting upon might get a little murdery.

Greatcomment, very insightful.

I think another reason is that in business you need models you can reason about.

One example is from a bank that advises customers on investment. They have to algorithms doing machine learning on the market, a decision tree and a deep neural net. Over time they can see that the neural net outperforms the decision tree, yet they cannot use the results to advise customers because it is a black box.

Why should you buy this investment? "Because the computer said so" isn`t an answer that makes it sound like you know what you are talking about.

Assume a cow is a uniform sphere of milk...

And sometimes that's how economics feels. Absolutely agree with your comment. Coincidentally, I was thinking about this yesterday as I watched one of many economists explain the market on TV. At some level I feel they grab onto one of N plausible ideas and go with it. They have to say something, right? After all, they are economists.

I haven't failed to notice that there are very few (not one?) massively wealthy economists. That has to say something. If you look at the people who have made fortunes they are all practitioners, not theorists. In other words, they "touch and feel" the economy every second of the day for years and understand how it moves, at least at the relevant micro level, to gain advantages.

This one thing has always bothered me about economics. Lots of economists publishing books and claiming all sorts of events and effects and none of them have significant financial success to show for their understanding of the economy. Fake Jedi's who talk about the force but can't use it.

I'd like to see an economist on TV who starts their opinion of the economy in a Trump-esque way with: "I am very rich. I understand the economy and because of that I make $400 million per year".

In terms of using models and simulations, well, we do a lot of FEA, mostly for thermal and fluid dynamics. I had a poster printed and hung in the lab. It reads: "The only person who believes the results of these simulations is the one who wrote the code". I stole that from my friends in aero who always talk about how careful you have to be with aero "codes", particularly at the extremes (low and high RN).

Of course, FEA has gotten better and better over the years, yet it is true that you have to be very careful with how you interpret results. That's why we verify expected results experimentally by building a prototype every N variants of a design based on various criteria. I remember thermal management design to cool over 1,000 Watts of high power LED's concentrated within a small surface area. It took eight months of FEA and dozens of prototypes to get it right. Some of the proto's didn't behave anything like the simulation. Some behaved better. It's an art.

And so my point is that we are likely to have a new crop of economists who will base their opinions on simulations without, perhaps, understanding just how imperfect they can be.

Plenty of economists work as traders, in banks, and have money and success to show for it. But they're not making economic policy, nor creating theories in academic settings. They're just using complex models to predict things, they have no interest in creating policies. In general, the economists who are successful are mostly ignored.

An example of an economist who is successful would be Jan Hatzius, Goldman Sach's head economist.

I think the economists you are talking about become experts at some manageable and very small sub-segment of economic activity. I am speaking more about those who are quick to analyze such things as the global stock markets and, years later, they are still "CNBC contributors" and not independently wealthy.

I always go back to when I read "The Big Short" and realized that ALL of these economists missed a herd of elephants in the room. If I remember correctly, the two or three people who called it (and were yelling and screaming about the impending doom) and ultimately shorted the market and made truck-loads of money were not economists but rather intense observers and practitioners. I think one of them had Aspergers (It's been a few years since I read the book).

That's some unrealistic criteria you have there. No one is an expert in every facet of physics or maths. Yes economists specialize. As do engineers. Or doctors.

As for pundits, there's a reason they're pundits... Is Sergey Brin or John Carmack going to host a tech segment on TV? That's silly, as silly as thinking TV economists would be as successful as people like Jan Hatzius (BTW, he did correctly anticipate the crisis of 2008, and GS bet against the market and made a killing during that time).

I am only using TV economists as an example. The estimate is that there are well over 15,000 economists in the US. I think, in general terms, their performance is poor. This isn't about being experts as you suggest.

> As a physicist, the models that I encountered in economic theory always left me quite disappointed. The main reason for this was that -like the article says- most of the models used in Economics are based on abstract reasoning about "ideal" markets and actors

I recall my family's PhD. economist relating an anecdote about an colleague explaining that he's making the usual set of assumptions like, say, "people live forever", before doing some math for students. :)

Assume a spherical cow: https://en.wikipedia.org/wiki/Spherical_cow

Samuelson attempted to develop a sort of thermodynamics of economics. As far as I can tell it was a miserable failure.


I'm a fan of using physics models for predicting behavior of large groups of people, but economics .... they do occasionally come up with nice statistical tools. Sort of like psychological researchers; those fellows whose papers can only be reproduced around 1/3 of the time.

I would be very surprised if banks and big trading companies aren't doing this already, in order to predict markets. Sharing this information, however, would eliminate their competitive advantage, so I assume they just do it secretly.

There are several problems with social sciences not just economics. Recently, I've read an interview by Yanis Varoufakis[1]. Here's an interesting excerpt:


Mariana Mazzucato, University of Sussex – How has the crisis in Greece (its cause and its effects) revealed failings of neoclassical economic theory at both the micro and the macro level?

Varoufakis: The uninitiated may be startled to hear that the macroeconomic models taught at the best universities feature no accumulated debt, no involuntary unemployment and, indeed, no money (with relative prices reflecting a form of barter). Save perhaps for a few random shocks that demand and supply are assumed to quickly iron out, the snazziest models taught to the brightest of students assume that savings automatically turn into productive investment, leaving no room for crises.

It makes it hard when these graduates come face-to-face with reality. They are at a loss, for example, when they see German savings that permanently outweigh German investment while Greek investment outweighs savings during the “good times” (before 2008) but collapses to zero during the crisis.

Moving to the micro level, the observation that, in the case of Greece, real wages fell by 40% but employment dropped precipitously, while exports remained flat, illustrates in Technicolor how useless a microeconomics approach bereft of macro foundations truly is.


I really don't know at what level of complexity someone must settle, in order to have a viable mathematical model to predict financial crisis... How many variables do you need? Then again a lot in economics depends on 'perception'. A FinMin will never discuss devaluation of his currency, the moment he does... The currency will drop. How can a mathematical model 'predict' such behaviours?

[1] https://theconversation.com/varoufakis-in-conversation-with-...

In my macro economics studies we absolutely included involuntary unemployment in models, this in itself broken into different forms like frictional vs structural unemployment. Similar inclusions for debt and pricing. As for crisis, then what are macro economic stress tests? Economists can be a overly theoretical in places, they have to be in a world of unlimited and often irrational variables, but this is a bit simplistic.

I have quite admired the way Yanis has handled his brief position in Greece. He's clearly an intelligent big picture thinker that was out to serve his peoples interests. This quote seems out of context to my expectations of his usually we'll considered statements. I wonder if there is some loss or context or other soundbite missing here.

I think, if you want to truly understand where Yanis Varoufakis is coming from in that comment, read the book Debunking Economics by Steve Keen. Although Varoufakis is probably even more critical to economic theories than Keen, the two guys agree a lot and admire each other.

Besides, Keen (and other postkeynesians) truly shows how the mathematical economics should be done, that is, by having a fully dynamic macro models with time and debt/money. I am afraid machine learning the article talks about is a similar distraction as econometrics. Unless mainstream economists will be willing to rework the theoretical foundations (give up on equilibrium, at least), and base it more on empiricism, mathematical approach, no matter how fancy, will always be a quandary.

But as other have said, the problem is that many economists are pretty much the prophets of the free market (or status quo in general). If you want to somehow claim a moral basis of society, then you can't have a model which allows multiple different solutions, and so you need to stick to equilibrium instead of allowing a dynamical solutions.

you know not all theoretical models are equilibrium based? guess not if keen is your primary source

i can't find any coherent argument in your final paragraph. can you rephrase it in a way that has some meaning?

> you know not all theoretical models are equilibrium based

Depends what you mean. I would love to see a macroeconomic model, which is not just an assumption of equilibrium subjected to external shocks. I have only seen such models from postkeynesians. (Austrians do not count - they pretty much throw hands to the air and say "it does whatever it does, we cannot know anything, therefore it's OK".)

> can you rephrase it in a way that has some meaning

Well, assuming you want to show that some self-organizing system should be a moral guide, for instance, claiming that free markets lead to "efficient" (whatever that means) outcomes, then necessarily this self-organizing system has to have a stable unique outcome (which pretty much rules out chaotic behaviour that dynamical systems love to exhibit). If it had two different outcomes, then it couldn't be a unique source of morality (someone would have to come and decide which outcome is better).

For example, that's why economists, who believe in free market, cling so much to idea that there cannot be endogenous unemployment (and instead explain it as some sort of reaction to external "shock" which will eventually pass). If existing unemployment was endogenous, it would mean that at some point in the past there was a choice (bifurcation), which would set us to path of either unemployment or full employment, and the self-organizing system of free market "made" the morally wrong choice, or at least it was indifferent (meaning it is not a good moral guide).

The Danish ministry for Finance has an official macroeconomic model that is used to evaluate policies.

To the best of my knowledge, it includes a small (I think fixed) amount of structural unemployment but anything else is assumed to be transient. We might have some unemployment now, but it's always assumed to right itself after a few years.

This interview is almost incoherent.

Note the contradiction: "no involuntary unemployment and, indeed, no money" followed by "real wages fell by 40% but employment dropped".

The latter statement only contradicts Keynesian economics, which don't have relative prices (Keynesian economics has only real and nominal dollars of GDP) and uses money combined with sticky prices to explain involuntary unemployment.

But although in one sentence he seems to think the Greek experience contradicts Keynesian economics, he turns around and appeals to Keynesian economics a few minutes later: "the 3.5% primary target for 2018 would depress growth today".

Actually, Keynes called his book General Theory precisely because it included unemployment, in the context of Keynesian economics usually called the output gap. So "no involuntary unemployment and, indeed, no money" is a insult from a Keynesian against neo-classic models.

Second, "real wages fell by 40% but employment dropped," this is precisely what a Keynesian expects. The core insight of Keynesian economics is, that the employees are also the consumers, so that falling prices lead to falling demand.

No, Keynesian economics proposes that high real wages are what cause unemployment - if an employee's real output drops but his real wage stays flat, he might become unprofitable to employ. Since nominal wages are the real sticky quantity, Keynesians propose inflation as a sneaky way to reduce the employee's real wages.

Usually Keynsian economics is understood as economic thought that is based on aggregate demand, this is for example the view Wikipedia takes. [1] From there it is straightforward to argue that higher wages lead to higher employment, a view that makes Keynsian economics popular on the left.

The view, that too high wages cause unemployment, is a staple of supply side economics, the view that the economic process is driven by the supply side.

Krugman argues a lot about sticky wages recently, but not because this is a central tenet of Keneysian economics. Krugmann argues about sticky wages because it is a good argument against the troika's handling of the Greek crisis, since the internal adjustment in the Euro area would either mean inflation in Germany or deflation in Greece.

[1] https://en.wikipedia.org/wiki/Keynesian_economics

Did you read the article on wikipedia? Start from the heading "Concept".


Supply side economics is not based on wage levels at all, but rather based on tax-based disincentives to employment and investment.


> Note the contradiction: "no involuntary unemployment and, indeed, no money" followed by "real wages fell by 40% but employment dropped".

That contradiction quickly dissolves once you see that he first he talks about macroeconomics as taught in the university ("no involuntary unemployment and, indeed, no money") and then about microeconomics ("real wages fell by 40% but employment dropped").

It's still nonsense. Macro does discuss involuntary unemployment and money. Also, the specific microeconomic phenomenon he discusses is exactly the microfoundation for a specific macro theory (Keynesian economics).

>The latter statement only contradicts Keynesian economics

Uh, no, it contradicts neoclassical economics. Keynesian economics views unemployment as a deficit of aggregate demand and most CERTAINLY does not pretend that it doesn't exist.

>But although in one sentence he seems to think the Greek experience contradicts Keynesian economics

A quick skim suggests he doesn't contradict Keynesian economists.

Probably he contradicts new keynesian economics somewhere, but then again, New Keynesian is an abortion that would have Keynes turning in his grave.

According to Keynesians, if real wages fall, employment should rise (relative to a counterfactual of fixed real wages). According to Varoufakis, "real wages fell by 40% but employment dropped" which is the exact opposite of Keynesian economics. Another phrase for "Keynesian economics" is "macroeconomic models taught at the best universities", which is what Varoufakis is criticizing.

>According to Keynesians, if real wages fall, employment should rise

AGAIN, that's what neoclassical economists believe, NOT Keynesians.

>According to Varoufakis, "real wages fell by 40% but employment dropped" which is the exact opposite of Keynesian economics.

No, that's what Keynesian economics predicts which is the exact opposite of what Neoclassical economics predicts.

>Another phrase for "Keynesian economics" is "macroeconomic models taught at the best universities", which is what Varoufakis is criticizing.

Neoclassical seems more favored these days despite its atrocious predictive capabilities. This is what Varoufakis was criticizing.

You don't seem to know what Keynesian economics is. Please google it. Other key phrases include "sticky wages".

If you think I'm wrong, please show how to use a Keynesian theory to predict "real wages fell by 40% but employment dropped". Good luck.

The problem is that a huge swathe of mainstream and formerly mainstream macroeconomic viewpoints are classed as "Keynesian", and the mainstream "New Keynesian" macroeconomic model you speak of which focuses on manipulating inflation to overcome resistance to real wage cuts and assumes a minimal role for fiscal policy is rather different from the traditional "Keynesian" (and "Post Keynesian") emphasis on increasing net government spend to stimulate consumption in a downturn.

Pretty much any undergrad who's read their introduction to national accounts and Keynesian concept of aggregate demand and multipliers can devise a set of simultaneous equations which will yield both a real wage fall and an unemployment rise in a basic stylised macro model if you also slash G and raise T (especially if the assumption is - as with Greece - that the expected debt servicing will largely be repatriated to overseas bondholders)

>You don't seem to know what Keynesian economics is.

I'm really not the one who is confused here :)

>If you think I'm wrong, please show how to use a Keynesian theory to predict "real wages fell by 40% but employment dropped". Good luck.

Simple. Reduce aggregate demand enough (e.g. by slashing government spending like Greece has done) and employment and real wages will both drop.

It's neoclassical economics with its idiotic presumption of 'full employment by default' that can't account for real wages and employment dropping at the same time. According to that theory, wages will just keep dropping until everybody is employed.

Step 1: Reduce aggregate demand enough (e.g. by slashing government spending like Greece has done) and

Step 2: ???

Step 3: employment and real wages will both drop.

How do you fill in step 2? Remember, I asked if you could show how Keynesian theory predicts this, not assert that it does.

Aggregate demand is the quantity of all goods demanded in the economy.

1. Aggregate demand drops. This shifts the curve to the left. These leads to a drop in the price and quantity of goods exchanged in the economy.

2. If firms are producing less goods they need less workers to produce these goods.

3. If they need less workers, the demand for workers shifts to the left.

4. This means the price and quantity of workers drops(wage cuts and increased unemployment)

Price and wage stickiness is one tool that New Keynesian economists use to try to explain business cycles using microfoundations.

Keynesians don't need it they use other routes to explain why aggregate demand exists(Pigou's wealth effect, Keynes interest rate effect, and the Mundell-Fleming exchange-rate effect). The downside to the Keynesian approach is the Lucas critique[0]. That relationships based on historical data are vulnerable to change when monetary policy changes. This started the New Keynesian/NeoClassical/Keynesian dvision.

[0] - https://en.wikipedia.org/wiki/Lucas_critique

Aggregate demand is not a quantity, it's a relationship between output and price level. But anyway, following your story, you get reduced employment.

But you don't get involuntary unemployment - people employed at a price $P as well as people unemployed but willing to accept work at the same price. All your story gives us is that there is a new market clearing price for labor at a lower price level, e.g. $P2 < $P. Fewer people are willing to accept labor at $P2, so they voluntarily go without work.

This is pretty much just basic supply&demand. You don't even need to go as far as Keynesian theory to get this out. If this is Varoufakis' story, it's hardly clear why he is criticizing classical econ 101 at all.

You're right that if we shift the aggregate demand curve its all econ 101 from there to wage and employment drop.

The difference between schools is that even though Keynesians believe in an aggregate demand curve( a relationship between price level and output) the neoclassicals(i.e. Chicago school) do not.

Varoufakis is criticizing the neoclassicals because they don't believe in an aggregate demand curve. Instead all of their modelling is done on the supply side. And its very hard to tell a supply side story where both employment and wages fall.

Involuntary unemployment is an easy jump from falling wages and employment.

It's not very hard to tell a "supply side" story where employment and wages fall at all. Just repeat your supply & demand example but substitute a specific good for aggregate demand. Or assume a decrease in the demand for labor due to automation, a reduction in labor hoarding, better ability to sort employees by productivity, etc.

Also, Varoufakis is criticizing "macroeconomic models taught at the best universities". Keynesian economics has been taught at universities for nearly 100 years.

I'm not quite sure how any of those supply side effects could cause unemployment or wage changes on the scale of what is happening in Greece. But if you could explain how those effects could have led to the unemployment and wage drops we've seen in Greece I would love to learn.

It seems to me there are two possibilities.

1. When Varoufakis talks about macroeconomic models taught at the best universities feature no accumulated debt, no involuntary unemployment and, indeed, no money (with relative prices reflecting a form of barter) he is talking about Keynesian.

Keynesian theory includes involuntary unemployment, accumulated debt, and money. So Varoufakis would have to have been mistaken. And he is shown to be especially confused when he later remarks

Moving to the micro level, the observation that, in the case of Greece, real wages fell by 40% but employment dropped precipitously, while exports remained flat, illustrates in Technicolor how useless a microeconomics approach bereft of macro foundations truly is.

This means he believes Keynesians are especially focused on microfoundations and ignore macro foundations. When this is actually true of the Neoclassical school. Then there is the issue you brought up. Varoufakis critiques Keynesians at the beginning and then appeals to it at the end. This implies he does not understand the basics of Keynesian theory. This is very surprising because he thinks of Keynes as one of the economists that personally influenced him the most.

2. Varoufakis is actually talking about Neoclassical economics.

This makes a lot of sense. Many Neoclassical economic models lack money, don't have involuntary unemployment, and lack accumulated debt. In addition it is very hard to tell a story about large wage drops, and growth in unemployment from microfoundations(basically supply-side). This means he is being consistent when he later appeals to Keynesian economics.

Varoufakis is obviously a very intelligent man who has spent 28 years studying economics. I would be incredibly surprised if he was confused about the basics of the two most influential schools of thought in macroeconomics.

>Also, Varoufakis is criticizing "macroeconomic models taught at the best universities". Keynesian economics has been taught at universities for nearly 100 years.

That still doesn't make the models he was actually describing any more Keynesian.

It it taught, although it has fallen out of favor somewhat. The preferred school for teaching these days is neoclassical, which is the school he was criticizing.

Keynesian != Neoclassical

Just in case you didn't get it the other 11 times, he wasn't talking about Keynesian economics.

Step 2 is:

* Austerity slashes public sector worker pay and pushes public sector workers into unemployment.

* Since public sector and private sector workers compete in the same job market, reducing the wages of one reduces the wages of the other through increased competition.

* Since the public sector is the largest single customer of the private sector, reducing its spending reduces the income of the private sector. This leads to lay offs and retrenchments in pay as well.

* Increasing unemployment also puts downward pressure on wages. The more desperate people are to have a job, the lower the wages they will accept.

* More downward pressure is put on wages by slashing social security, too (usually austerian target #1). The more unpleasant you make unemployment, the more desperate people will be to take a job. The more desperate people are to take a job, the lower the wages they accept will be.

This is a purely classical (pre-keynesian) story - demand goes down in many individual markets, and the market adjusts by reducing prices and quantity supplied. It has no output gap, no involuntary unemployment (read: wages at a price $P, but people willing to work at $P and unable to find a job), etc.

You've just shown that classical economics explains the Greek situation perfectly, contradicting Varoufakis.

No, neoclassical economics would have us believe that wages would go down until a clearing price was hit and everybody was employed.

This is, indeed, exactly the type of mind numbing idiocy that Varoufakis was talking about.

That's exactly what happens in your story. Wages go down from $P to $P2. Fewer people are willing to work at $P2 than at $P, so employment goes down. That's a standard market clearing story.

The classical claim is that don't have involuntary unemployment - prevailing wage levels equal to $P (in some segment) and simultaneously people unable to find work at $P.

Your story agrees with this.

Its not Keynesian theory but basic economics. Imagine demand for labor dropped, and supply stayed the same.(demand shifted to the left) Assuming that labor demand is downward sloping this would mean price and quantity demanded would both fall.

I have the same suggestion for you. Google it.

Why would a self-described 'erratic marxist' even be concerned about contradicting Keynesian economics?

Economics has an honesty problem. Financially motivated obfuscation is a problem in other sciences: global warming denialism, sponsored studies claiming that cigarettes aren't carcinogenic, all manner of food health and safety claims, and so on. Generally the evidence is overwhelmingly against them and the battle is mostly over PR and communicating to the public.

Economics is different; arguments that are very convenient to people who are wealthy under the status quo are very difficult to challenge, resulting in failed austerity programmes, "trickle down", and so on.

I partially agree with this view; however, to be fair, this is not a problem with academic economics, but rather with "pop economics".

There is no lack of academic economists who can (and do) prove how austerity programs and trickle-down are wrong; in fact, they're the only ones really leading the charge on this matter in the political world at the moment.

The problem is that policy leaders not trained in advanced economic studies (most of them lawyers by trade, as you would expect in the business of writing laws) are awash in "pop economics", cheap literature peddled by suspect gurus sponsored by interested parties. This creates a hegemonic but distorted view of the field that is then very hard to shake.

It's not just lawyer-politicians, but economic organisations like the IMF and the ECB. Stiglitz' book on the subject is excellent, including the anecdote about how we know that their advice is the same whatever the country and situation because someone once forgot to search-and-replace the country name before sending it out.

Then there's the Rheinhart-Rogoff excel error: http://www.nytimes.com/2013/04/19/opinion/krugman-the-excel-...

Those organisations are sclerotic because they ultimately answer to politicians, not to academics; they are political constructs first and foremost, serving specific national interests, with heads nominated by political leaders and (surprise) often lawyers themselves (e.g. Lagarde). In fact, during recent events in Greece it became clear how their own researchers disagreed with the policies being pushed from the top. It's like developers were forced to write websites in PASCAL because their employers told them to.

The Excel error is also exemplary of what I referred to as "pop economics": a single argument was "divulged to death" and transformed into mainstream orthodoxy even though it was still being confuted in academic circles.

> There is no lack of academic economists who can (and do) prove how austerity programs and trickle-down are wrong; in fact, they're the only ones really leading the charge on this matter in the political world at the moment.

The thing is you can pretty much find an academic economist to back up whatever argument you want to make.

The membrane between academia and pop economics (i.e. the suspect gurus) is exceedingly porous.

Most austerity program fail because there is no true austerity. Outside of Greece which has had to make real cuts simply because it hasn't the money the vast majority of countries which claimed some form of austerity never actually cut spending, some kept increasing it.

In the US it is all a shell game where the deceit of baseline budgeting allows Congress and the President to claim cuts when spending increases. The US could balance its budget in a manner of years if budgeting were honest and they actually reduce spending increases below inflation.

Plenty of economists argue against things like austerity, trickle-down, etc...

The problem is that politicians aren't interested in implementing the most ideal economic policies, they're interested in policies that give the effect they want - which is often at odds with the common good.

To summarize, "I represent truth in the service of mankind. Those with whom I disagree if not stupid, must be acting out of the most base self-interest."

> trickle down

I was just reading the convention speech of the Democratic candidate for president in 1896, William Jennings Bryan. In his speech ( http://historymatters.gmu.edu/d/5354/ ) he says "There are two ideas of government. There are those who believe that if you just legislate to make the well-to-do prosperous, that their prosperity will leak through on those below. The Democratic idea has been that if you legislate to make the masses prosperous their prosperity will find its way up and through every class that rests upon it." So some Democrats had to counter this trickle down nonsense even back then.

As an aside, in the speech he says "we shall fight them to the uttermost, having behind us the producing masses of the nation and the world". Can you imagine a Democratic presidental nominee saying that today, even if by some miracle it was Bernie Sanders? It's impossible - the heirs, and rentiers, and LPs and VCs and "job creators" and other parasites on those of us who work and create wealth have gained too much control of the system. And for the past two weeks signs are around that it could be entering one of its periodic crackups.

The Democratic idea has been that if you legislate to make the masses prosperous

Assuming that were even possible. Whether it is or not is a pretty big open question. More to the point, can you do it without needing the hammer of government sanctioned violence (or threat of violence).

Thats a much better argument, trickle up, than people make today.

It is safe to say, that the economic theory a person subscribes to is highly related to their income.

Unless, of course, you're a politician, in which case it is related to vote count.

>It is safe to say, that the economic theory a person subscribes to is highly related to their income.

This is an overly pessimistic view that ignores the philosophical aspects of economics. E.g. I lived on under $15k per year for a couple of years before I moved into software development, but during that time I still supported a very liberal approach to economics. This is because I hate violence (I can't even bring myself to crush spiders that wander into the house), and view redistributive economics as necessarily violent, as in a roundabout way it's forcing people to give goods/services/tokens of exchange to me that they may not want to give (since they're forced to give it via taxation lest they face imprisonment, I cannot assume it's voluntary). Other people may not care so much about this and take a more utilitarian approach, focusing on maximising some particular metrics (such as GDP or income equality), so they'll reach different conclusions.

Assuming people subscribe solely to the theories (economic or otherwise) that maximise the financial benefit to them is ironically a rather economics-style view of the world, and I'm fortunate enough to say that most of the people I'm close to don't fit this characterisation.

In terms of actual data, there's a correlation, but it's far from being the strongest factor: http://www.theatlantic.com/business/archive/2012/11/does-you...

It's interesting that you view taxation as a violent fundamental wealth redistribution. Taxes do suck and I'm not going to argue that point, but it's interesting because without taxes we would be living in a fundamentally violent society (since there is no government support a monopoly on violence - e.g. Somalia - everyone then may be violent to achieve personal goals).

How about we all try getting rich (or just keep our wealth out of violent neighboring hands) without violence or a government. Try to even significantly increase any meager income in that situation.

My point is that without taxes and the government they support, there is nothing but violence to support any type of venture. This has historical bearing with places that were lawless yet aplenty with resource for ventures (e.g. North American colonies, Wild West, etc...).

If in particular you meant a progressive tax policy that increases the percentage of taxation with your income level is what you're describing as violent then I can see your point a bit more clearly. Why punish the people who are being more productive, right?

Well the thing is that the owners of capital get the most extreme benefit from government. Never the other way around. It makes sense to me that the biggest recipients become the biggest supporters.

For a brief anecdote, poor people who earn wages get social programs to help sometimes, which are then largely obfuscated from public comprehension so you'll need a social worker just to get started. Rich people who own capital get business subsidies in the form of cash on the books. They are both paid from taxes and supported by the government. Who do you think benefits most from the government here? The business owner or the wage earner? Who is being stopped more readily from using violence with this support from government?

So taxes support all of this precarious balance in a capitalist society because the government is the only thing we have found to solve this problem of violence.

How about we all try getting rich (or just keep our wealth out of violent neighboring hands) without violence or a government.

There's a pretty broad range of what you could call "government" in that discussion. The question of how to reign in would be aggressors isn't fully settled, but it certainly hasn't been proven that it takes the kind of government we have today. Especially given that the system we have today doesn't stop aggression, it just shuffles it around and renames it. See, for example, "civil forfeiture".

On violent taxation: when you don't pay your taxes, eventually someone with a gun comes to take your freedom (prison for tax evasion). So the threat of violence is always held over the head of the taxpayer. Taken a step further, the primary relationship of a government to its people is the threat of violence for not following the laws.

Please don't let your observation of present society confuse your ideas about a free society: The poor use the violence of the government to take money from others and give it to them via social programs. The business subsidies are a form of government privilege that is a corruption of free markets in which the rich use the violence of government against competitors.

Free market thinkers challenge the assumption that government should be granted a monopoly on force but rather there should be competition in all areas of society. As it stands, we see abuse of power in the courts, police, legislatures, and bureaucracy, all for the enrichment of those who have a legal monopoly on force.

> Taken a step further, the primary relationship of a government to its people is the threat of violence for not following the laws.

This is true in pretty much the same sense that it is true that the primary relationship of a people to its government is the threat of violence by the former against the latter for failure to deliver on demands. (The extent to which either is actually truly the primary relationship in practice is pretty much precisely the extent to which the government has lost legitimacy.)

> Free market thinkers challenge the assumption that government should be granted a monopoly on force but rather there should be competition in all areas of society.

The government is simply whatever entity (including a collection of entities) in practice holds a monopoly on the legitimate use of force. You can question what the nature, scope, composition, and role of government should be, but questioning whether it should have a monopoly on the legitimate use of force is exactly like questioning the assumption that a triangle should be a three-sided polygon.

Well, I studied economics and this is where my economic-style view of the world comes from. :)

Yes, it's quite a pessimistic view of things, I agree. I speak mostly from personal, anecdotal experience. That may of course vary.

Careful, I think your politics is showing.

I think there's a fundamental misunderstanding of what it is theoretical economists are doing.

A medical scientist runs experiments on rats, to gain insight into what the effects of a certain phenomenon might be on humans. Rats aren't humans, of course, but we think they share important characteristics with humans. Of course, human trials would be better---butbecause there are ethical and practical difficulties in performing human experiments in controlled environments, we are happy to accept the insights that come from studying rats.

Theoretical economics studies the interactions of idealized 'rational' actors. These aren't humans, of course, but like rats, we think they might share important characteristics with humans. Like rational actors, humans have things they want, and do respond to incentives. Economics is largely the study of systems of incentives.

You can gain insight into the systems of rational actors---and insofar as you believe the analogy between rational actors and humans, into human economic systems---by writing equations and proving theorems.

I'm sick of this theme that theoretical economists are these dogmatic lunatics who write equations on the board that have nothing to do with reality. They are in the business of crafting powerful arguments about human nature by analogy, which yes, aren't always accessible to those of us who haven't steeped ourselves in the math, but I think that is no reason to dismiss their insights as nonsense and saying they have a 'math problem'.

As a former cognitive neuroscientist with a master's in psychology, I will just point out that psychology departments look at the massive amount of data demonstrating human irrationality, and conclude that economists are mistaken or wildly naive to extrapolate from "rational actors" human beings.

You don't have a "math problem", you have a "data/science problem". And saly, at least at the micro-level, economists could have disabused themselves of the accuracy of rationality decades ago (e.g., Kahneman and Tversky's work in the 70's).

Much modern economics is like the elaborate math of epicycles, close enough, but still missing key insights. And this won't change until data is valued more.

The issue is that assumptions such as perfectly rational actors can diverge so far from reality that it renders the resulting idealized conclusions meaningless from the perspective of actual human economies. The abstract study of incentives is of course interesting in its own right, but may have little to do with economics in practice -- i.e. the important practice of how actually should we run an economy.

What theoretical economists do is not 'nonsense' but can become nonsense when it naively forms the rationale for real-world policy.

Rats are, in some ways, very different from humans too. But if it's impossible to run human trials, or there are severe deficiencies in the way we are able to execute them, studying rats is still the best argument.

If there is a convincing empirical economic study that widely contradicts a theoretical consensus---yes, that evidence should be preferred. But that is not often the case, and otherwise the theoretical argument is simply the best argument there is.

What drives me bonkers, though, is the people who dismiss the math and its conclusions as being 'meaningless', but then what they substitute is not superior empirical arguments of their own, but their own emotional sentiments and undisciplined intuitions (typically decorated by cherry-picked historical observations).

The "rat is to human" (from a pharmaceutical perspective) as "rational actor is to human" (from an economic perspective) I think is a very fragile analogy. Rats and humans are very similar in their response to pharmaceuticals (and so serve as reasonable first model), but the difference between rational actor and human is so vast (at least in terms of how the real world works) that it calls into question the value of rational actor models for real-world policy.

In that case, the "theoretical argument is simply the best argument" is wrong; we should study economic history and human behavior rigorously and let that inform our policy. Theory can serve as a useful model when it is validated by some sort of observation and data -- but even when we cannot perform controlled studies there is not a knock-down argument that "theory" must be our best guess.

I agree that a person's interpretation of economics is very likely to be influenced by their idea of how the world should work; but this is more a problem of our characteristic lack of critical self-reflection. And of course, we should have cogent arguments for our interpretation of economics, ideally rooted in objective evidence.

mathematical formalism was introduced precisely to allow the formulation of "cogent arguments". you can't evaluate the evidence without economic theory

One real world example is whether there are educational benefits to deworming a child population in developing world countries.

If you use economist math you see a benefit. If you use epidemeologist math you don't.





Meanwhile machine learning is the high interest credit card of technical debt, what could go wrong? http://research.google.com/pubs/pub43146.html

I entered stats through econometrics. I love econometrics. Econometrics is seriously hampered relative to other applied math fields by the inability to apply large scale casual experiments. For this reason parsimony is soooo important in modelling. Machine learning is the opposite of parsimony. Machine learning is wonderful for many applications, it lets us shrug off a first error of model selection, but it's not a panacea to the math problem and I think it actually might make it much worse.

While pursuing PhD in Economics, I've supported myself working as a data scientist and, like ThePhysicist says, it's fairly disappointing that many are oblivious of datasets, APIs, libraries,..

A good example of what can be done (http://econprediction.eecs.umich.edu/)

On the other side, the claim "literary types who lack the talent or training to hack their way through systems of equations" is groundless and mostly fallacious. The "literary types" have grasped many politicized inconsistencies and dedicate their research by deductive reasoning. Most of them are not 9/11 truthers.

My insight in college taking some economics courses side by side higher level physics and engineering courses was that economics has mostly gotten stuck in algebra and has thus far failed to develop its calculus.

The easiest analogies for me to make are to electronics. Talking about cash/debt is like talking about instantaneous voltage. It's sort of interesting, but it isn't doing any actual work. In electronics it is the flow of that voltage, the current, that is more often more important and more interesting and more directly applicable to the "work" of a circuit...

I think you get a very different view of economics if you model it more like electronics circuits than just about any of the other economics models I see discussed. I've often wondered what would happen if someone actually bothered to attempt such a model and apply some deep scientific rigor to it.

Economists were modeling the economy using circuits in 1949 (see MONIAC), before the Integrated Circuit was even invented.

Thanks for the interesting reference to look up. I had not heard of the MONIAC before. Looks like this is also a sideways referent in Terry Pratchett's Making Money which draws in some interesting dots in that book.

comments like this are amazing. you are completely ignorant of the current state of economics yet wonder if economists have bothered to construct dynamic models of the economy. i guess your intermediate macro course didn't get that far...

Do you have any links for what you consider to be the current state of economics? I will the first to admit that I am not an economist, but I do try to at least keep an ear out for what may be current and certainly in terms of what I've been hearing from a layman's perspective, for the most part my comments remain valid against current economics work. That said, it's also why I prefaced my comments that they were primarily centered from several years ago back when I was in college and that the analogy was mostly personal opinion. If it is personal opinion founded in ignorance, I am happy to be proven wrong or shown cool things, if you are interested in a deeper discussion beyond insults.

Certainly, the models that we see in day to day political discussions and popular/lay culture have not seemingly progressed much beyond the macroeconomics courses I took in college, but again I'm willing to admit I certainly have my biases/blinders. Also, I'm sure that dynamic models exist, I was more curious if there are economists that have played with models more closely aligned with my physics analogies, specifically.

Thus far the one of most interesting things in economics right now, to myself at least, is I have done some reading on some of the work of theorists in Modern Monetary Theory (MMT) and I liked where it seemed to be heading (not to mention its ties to turn of the last century Social Credit theories/works). That certainly seems close to what I'd expect/like to see more of in economics, but correct me if I'm wrong, it still seems "fringe" in the current state of economics.

>That would make the techniques less interesting to many economists, who are usually more concerned about giving policy recommendations than in making forecasts.

Therein lies the real reason for the panoply broken economics models. They are more often used for the purpose of lobbying rather than for impartial prediction.

Much of the profession (particularly the elites) only exists as an intellectual pretext for maintaining and perpetuating existing power structures. It's about as scientific as the departments of Marxist economics were in the former Soviet Union or the Vatican in 15th century Italy.

If economists could manage to get themselves thought of as humble, competent people on a level with dentists, that would be splendid.

My beef is when I read economists claims about perfect markets I think, 'Wow! With markets I should be able to solve the traveling salesman problem in polynomial time'

Well, I am not a big fan of today's "more debt, more gov't spending = more prosperity" but markets only parallelize a process and the best ones at this process survive while the worse ones drop out of the market.

So, in a sense, if you assume ideal markets to have an "infinite number of participants", then yes, you can solve TSP in polynomial time but otherwise they will just evolve like bitcoin miners which as of yet still didn't solve the "find a decent hash" problem an order of magnitude faster even though there are many of them.

My beef with economists is that they themselves are not subject to markets so the bad ones and the good ones survive and you end up with lots of noise.

>Well, I am not a big fan of today's "more debt, more gov't spending = more prosperity"

Because austerity is working out just brilliantly everywhere it's been tried?

What people tend to forget about government deficits and debt is that it is just one side of the economic coin. It's exactly the same thing as saying "the private sector saves too much".

There are problems that can only be avoided before the fact but not solved once they occur.

Austerity in the last few decades would have avoided the crisis but once you have more debt than you can ever afford to repay neither spending nor saving will do the trick and the only possible way forward is default.

Since you refer to Greece in your other comment - I am more concerned about the politics than about the economics, given they don't default even though they have a clear democratic mandate.

>Austerity in the last few decades would have avoided the crisis

No it wouldn't. The 2008 crisis and the European Crisis were both caused by an unsustainable build up of private debt. The levels of public debt had nothing to do with it

Even in Europe where there was multiple sovereign debt crises, the debts all started out in the private sector and were later shifted on to the public balance sheet. Same story in Spain, Ireland, Slovakia, etc. (even Greece to a large extent).

>once you have more debt than you can ever afford to repay

"Afford to pay" means something different to governments than it does to the rest of us. Governments that print their own currency cannot run out of it. The only decision they have to make regarding the deficit is "is the current level of inflation too much?"

Austerity in the UK and the US is thus kind of like saying "well, we must stop spending money because what if we run out of little green bits of paper". It would be funny how stupid it is if it weren't so devastating.

Similarly in the EU inflation is rock bottom and the ECB could kickstart every ailing economy easily. They just don't want to.

>Since you refer to Greece in your other comment - I am more concerned about the politics than about the economics, given they don't default even though they have a clear democratic mandate.

Greece has unfortunately handed over control of their payments system to the ECB and wresting back control is not something that can be done in under 9 months.

Thus it doesn't really matter what the Greek government says or what their electorate wants - if they don't follow Troika orders they flip a switch and the economy seizes up completely. Instant 3rd world status. No jobs, no money, no food, no life saving drugs.

This is why successive Greek governments have done the opposite of what they were elected to do. They have a gun against their heads and Schauble is just itching to pull the trigger if they don't fall in line.

'Because austerity is working out just brilliantly everywhere it's been tried?'

Measured on what timescale?

Well, the IMF and ECB's time scale for predicted recovery in Greece has been 18 months ever since austerity was implemented in 2009.

So that's roughly four complete and total fuck ups in a row.

How much more time do you need?

There's always a chance, of course, that austerity is utterly destructive towards the Greek economy but very good for German financiers (with the ear of the ECB) who would like to pick up a port or state electric company on the cheap and who like the flood of cheap labor the austerity crisis has instigated.

The one in which we're not waiting until we're all dead?

The private sector saving too much or too little is explicitly NOT the issue.

The question is whether that savings goes into G or I - i.e., whether the money is spent on prisons or a new Tesla factory (to borrow two examples that contradict the standard mood affiliation).

>The private sector saving too much or too little is explicitly NOT the issue.

Quite. This is what I was saying.

Public sector deficit 'too high' = Private sector saving 'too much' = Not a particularly concerning issue.

Deficits matter only insofar as they cause inflation.

>The question is whether that savings goes into G or I - i.e., whether the money is spent on prisons or a new Tesla factory (to borrow two examples that contradict the standard mood affiliation).

As far as pushing aggregate employment higher and escaping from the liquidity trap is concerned, you could do either. Just so long as it employs people.

The highest multiplier for any government spending was probably FDR's new deal public works program.

ROI is a different question and not one that can really be looked at in purely monetary terms. If the government paid people to write open source software they could end up creating staggering amounts of value but it wouldn't show up in any economics statistics.

Actually, calculating equilibrium prices of markets is PPAD-hard (similar to NP-hard, but not quite) in general. See for example http://arxiv.org/pdf/0904.0644.pdf

When you ignore externalities....

Really, because I always think of dentists as inflating their own professional stature in order to promote their supposedly indispensable remedies, much like economists.

Pure economics should be about applying abstractions to help us understand the world. The world is pretty sophisticated, so I'm fine with mathematically sophisticated abstractions. Even when those abstractions fail intuitively, they can be useful and extended for other subareas of economics or are simply philosophically fascinating. More often than not though, a lot of prominent theory seems impracticable or intractably limited. This is a valid criticism.

Noah Smith is slightly misdirecting his complaints though. If we're looking at applied economics or policy, the amount of math is not intimidating to anyone who has taken a couple of years of non-introductory statistics. The most important econometrics papers of the past few years apply basic regressions with only a few additional valid statistical techniques.

I'd think economics should be about describing systems in terms of cost, and finding how much a certain action will cost or profit the system. There is tons of economy to be found in the natural world, and it is mostly shunned by economists, and left to the naturalists. I don't know of any other science that ignores the natural world like that. And yes I see a problem with that. The probability of being wrong becomes huge. What good is an abstraction if it's wrong?

reminder: noah has never published an academic paper in economics and it looks unlikely that he ever will. he is simply not a competent guide to the field. this entire thread is a perfect example of the blind leading the blind: a journalist (noah) presents an entirely one sided view of the field, and intellectually lazy posters take it as a cue to dump on an entire academic field while freely admitting their ignorance of any of the details.

newsflash: most economists are acutely aware of the imperfections in their models. sure, you can find the blinkered and dogmatic, but that is unsurprising in such a large and varied field. the subject encompasses a serious variety of subjects and methods you (the hn poster) simply know nothing about. for example, machine learning techniques are really nothing new. theorists have been aware of the kahneman/tversky result for decades. please bear this in mind before you lazily declare the intellectual bankruptcy of the entire field.

p.s so-called econophysics was a direct attempt to apply models from hard science to economics and it has been a complete failure, since it lacks an underlying model of human behaviour. turns out this is quite important...

I find it interesting that many economists tend to ignore the limits of their own theories.

For example, the dogma that markets tend toward efficiency.

This is true -- as long as certain axioms are not violated. But they seem to forget about the axioms on which their conclusions are built and march forward as if their conclusions are absolutely true and apply that assumption to problems where they simply do not hold because the axioms on which they are built are violated.

To continue with the 'markets tend toward efficiency' conclusion: it does not hold when there are negative expectations for a party for not making a transaction. And they invented a fudge factor to get around this a bit (that still doesn't work for all cases) which is to make sure the equations when taken in aggregate do not generally lead to negative expectations for not making a transaction. This holds in many more cases but still not all.

Where does it fail and why does it matter? Well, the axioms underlying the 'markets tend toward efficiency' work great in capital markets. They tend to completely fail when pain or death is potentially involved in a transaction (severe negative expectations) but they may look like they should work and you can find limited cases that do work. So, basically war or crime or relationships or corruption or or poverty or healthcare or any other of those human interactions pain or death tend to come into play. A huge swath of sectors where politicians and special interests and the economists that inform them try to craft policy to fit a misapplied theory. Yet, economists march on pretending the theory holds from the axioms to the conclusions and on to the further conclusions built on those hold because 'mathematics' -- who can argue with mathematics?

I don't think a model based purely on empiricism will work. Because people will read about the model and try to correct for their biases.

On the other hand a purely game theoretic model is not satisfactory either because a lot of people wont read about the model, and wont act optimally.

Maybe a two-tiered model, with "fish" and "sharks" could be a solution.

Your comment demonstrates beautifully how economics is not an empirical science. Much of what is taught in economics is founded on a flawed premise.

google the "lucas critique"

This is really the important part of the article:

> Their overview stated that machine learning techniques emphasized causality less than traditional economic statistical techniques, or what's usually known as econometrics. In other words, machine learning is more about forecasting than about understanding the effects of policy.

And it's true. Economists care more about forecasting than 'understanding the effects of policy' because the money is to be made forecasting while working for a bank or a large corporation.

Economics can't be codified the way physics can because economics models human behaviour (which is constantly changing), not physical laws than can be tested and retested.

I shouldn't even comment on this BS article but here we go, duty calls ( https://xkcd.com/386/ )…

1) First of all, let's make clear that there are no "the economists" and there is no "economics" that could be identified with whatever peeve du jour of "the journalists". Economics as a field, understood as the study of exchange between decision-making agents in more or less large systems[1], is marked by an extreme diversity of approaches and opinions on even basic questions. From my superficial understanding of other fields, this seems to be a big difference between economics and the sciences, and on the other hand rather similar to the humanities. If you're going to criticize something, better make it specific and name names.

2) It is fashionable for "technical" people to scoff at the mathematics used in economic theory (basically, analysis, optimization, linear alg, and measure-theoretic prob), pretending that the problems arise because the tools are too primitive. First, perhaps they should remember that a lot of the foundations of mathematical economics were laid by people way above their paygrade. I'm talking about von Neumann, for example, or Fischer Black. Usually, when smart people do something that does not make sense to you, you take a step back and ask what you might be missing, and usually there is a nontrivial answer to that question. Paul Romer, sorry, is a good economist but he has a BS in math. I would say that's my definition of a "lightweight when it comes to equations". That's not to say he shouldn't criticize, he should! But let's not pretend that there was no reason to "go formal" in economic theorizing or that there were/are easy modeling choices.

3) In my experience, most of the criticism of economic models comes from ignorance of the goal and context of a model and a too literal reading of the math. For example, if you take Markowitz's portfolio optimization (mean vs variance of a linear combination of a multivariate normal), it is correct that individual returns have non-exponentially decaying "heavy" tails, the copula is not (unconditionally) Gaussian, and risk is therefore not captured in variance and correlations (not least because they may in fact be undefined). But that is completely beside the point of the model, which is simply to express the idea that the risk of a portfolio is not necessarily additive and that there is a tradeoff between risk and return. The point being, although it is expressed in mathematical language, it is actually a qualitative model. Once you realize this difference to more descriptive models in the sciences, economic theory starts to make a lot more sense.

4) The real problem of economic theory may be the disconnect between how many people have a stake in it and how many people have the time and leisure and inclination and background to understand what the theorists are actually saying.

5) Side comment and half-reply to ThePhysicist's complaint: The reason to go with the "perfect gas"-type models in economic theory, I think, is that you are dealing with self-interested, utility-maximizing particles or "agents" -> acting particles. The thinking being that if you put in constraints of some form, say a short-sales constraint in a financial market model, you make the model very special. But in the messy real world agents would find a way around this particular constraint eventually. So the unconstrained general equilibrium models are trying to give you a big picture, "this is where the market tends to" type of result. There is so much more to be said on this point but I'm already way above my allocated time for this "duty call"...

6) Last but not least, let me state my opinion that there is nothing inherently noble about science. It is a method to gain knowledge, and it is contingent on the affordances of the field to which it is applied. In economic theory, your basic problem is lack of data. Now that may change in some subfields, and that's great. So the scientific method can be applied in those subfields eventually. But we need answers or opinions today for practical problems. I see economics as akin to philosophy how Russell [2] understood it. Let me quote him:

“Philosophy, as I shall understand the word, is something intermediate between theology and science. Like theology, it consists of speculations on matters as to which definite knowledge has, so far, been unascertainable; but like science, it appeals to human reason rather than to authority, whether that of tradition or that of revelation. All definite knowledge—so I should contend—belongs to science; all dogma as to what surpasses definite knowledge belongs to theology. But between theology and science there is a No Man’s Land, exposed to attack from both sides; this No Man’s Land is philosophy. Almost all the questions of most interest to speculative minds are such as science cannot answer, and the confident answers of theologians no longer seem so convincing as they did in former centuries.”

One should be aware of the limits of this approach but one should also not assume that there was a choice between science and this. If there was, it wouldn't even be a contest! The realm of science may get bigger over time, as more and more data is gathered, but again, we need answers today and in real time, not in 100 years, to decide on regulations, interest rates, portfolios or budgeting decisions. In almost all the big economic questions, your sample size is one and the possibility to experiment is nil. The world is arguably nonstationary on time-scales that matter, economic mechanisms change, and every statistical test and computational experiment is always a joint hypothesis test of your assumptions, about which I said in (1) above that there is and perhaps can be no consensus.

In short, it's not that people don't know there is a problem. It's that there has so far been no better solution.


[1] I'm just making this definition up on the fly but if someone has a better one, I'm all ears.

[2] B. Russell. A History of Western Philosophy. George Allen & Unwin, 1945.

Economics is like fantasy football league. I predicts the future based on isolated personal preferences and ideologies of the "scientist", optionally sprinkled with small scale observations of data with dubious accuracy.

Considering financial policies are the de-facto social discussion around the Globe, and progress of the human species is being discussed and perceived in financial terms, it is no wonder the role of philosophy has fallen to economists.

One can hope that data-driven economics could improve predictions, but keeping the theory discussion alive is vital, if we are to ever understand our goals as something more than increasing production and consumption (quantitative).

Using the correct set of data, has long been the main issue with economic theory conflicts, and data-driven economics can only make a difference (and a huge one) if that changes.

The maths models used in economics are rarely checked against real-world data, but when they are they often turn out to be wrong.

A major instance of this problem comes from models used for trading, which often assume Gaussian distributions and Brownian-motion behaviors. According to those, extreme moves such as the ones occurring during economic crises such as the one of 2007 are several sigmas away from the mean -- making them supposed to happen perhaps once every 10^10 years. Yet they happen every 20 years. Taleb's book "Black Swan" is an interesting resource on this topic, for those interested.

gosh, i wonder why those traders use those simple, outdated models to make extraordinarily expensive decisions every singly day when they could just read the "black swan"

I like Mark Blyth's brief summary[1] of this problem. Not only is it important to base your theories on actual data, it is also vitally important to remember that the map is not the territory. All models are an imperfect representation of reality, so they should always be viewed with appropriate scepticism, even when the math seems to work out elegantly.

[1] https://www.youtube.com/watch?v=hmWbkPezgtU

There is a main difference between Physics and Economics. Physics tries to explain the laws of nature which do not change over time. But Economics tries to explain the laws of a system based on human knowledge and believes which is anything but robust. Using modeling techniques borrowed from Physics/Mathematics in the field of Economics can work very well but only if the laws you are investigating are robust over a given period of time.

Edit: typo

If anyone is interested in something other than shitting on economics...

The other half of the article seems interesting. Does anyone have handy references (besides the Athey and Varian papers in the article) to how machine learning is being used to establish causation rather than focusing on prediction? I've heard several places that there have been big strides lately in this area, but I'm not sure where to look.

Here's a link to the slides that she used for the NBER talks. I've flipped through them and there seems to be some references to working papers (that Googling should be able to surface):


Doesn't (over-applied) machine learning just lead towards a variation of exactly the same problem?

Instead of "this model fits the curve and sounds attractive and sciencey" it becomes "this model fits the curve and was done by computer!"

Not really the "empiricism coming to economics" that the author claims, but I guess it could support better outcomes. For a while.

Economics is fascinating because it is where ideology meets math.

Economics has a math problem only if you're not honest about your mathematical models reflecting an ideological position rather than "reality" (whatever that's meant to mean in the context of economics).

I love watching smart people argue, so I love economics. Economics is a great source for jokes. "Economics is astrology for people who know calculus", "How many economists does it take to reach a conclusion?", and so on.

As much as I like picking on economists and watching them bicker, there's actually a method to the madness. (Except for the shills, I'm looking at you, Krugman[1]). When done correctly, economics becomes kind of a cross between a courtroom and a bullshitting session. People present various theories and sets of measurements. Lots of discussion around instrumentation and epistemology takes place. At the end of the day, the discussion wraps up to be something like "Given that most of us agree that these kinds of initial conditions mostly lead to these kinds of results, do most of us think that these kinds of initial conditions currently exist?" (Note that the "most of us" can very well be different sets of folks)

Oddly enough, it turns out that this kind of discussion has lots of real-world value.

Machine learning and deep learning especially has a pretty cool new role here, but it will never take of this dynamic ongoing discssion.

So no, economics doesn't have a math problem. It doesn't work like that. Perhaps a better way of putting is this: people who expect the world to be deterministic have a problem with economics. That situation is unlikely to change.

[1] I pick on Krugman because his shtick is already knowing the answer, no matter what the question is. He also likes showing us his Nobel. Somebody said once that every Krugman column should begin with "Damn fools!" In this way, this isn't economics; as far as I can tell it's personal aggrandizement using economics as a prop. A huge ego is a terrible thing to waste. But damn, it sure is entertaining!

[Several edits for clarity]

Good comment - I remember while studying economics (a subject I have always enjoyed) I got into an argument with some fellow students who felt that it was too 'fluffy' and 'not enough meat'. They wanted to do equations and get hard answers.

The truth is that Economics is the study of choices in a world of limited resources, and that the process is exactly as you have said : observing initial conditions, material decisions and subsequent outcomes. It is, as laid out by Adam Smith, also dependent on certain initial inputs (theory of moral sentiments) which produce rational actors.

The amusing thing for me is that, despite repeated tests and outcomes, people continue to think that they can wander off the reservation and get better outcomes with the same humans by trying different things to try and control the flow of limited resources. The results are rarely pretty, but the siren song of being able to improve things and be hailed a hero draws people into thinking the rules do not apply.

Krugman wandered off the reservation a while back and hasn't been worth reading since at least 2008.

The point is, you don't have to agree on economic policies to agree that it's not a hard science and shouldn't be treated as such.

2008? That's a coincidental year. Around about the time of the last crisis and the beginning of austerity politics. The worst economic period since the Great Depression arguably. The rich getting ever richer. Real wages for most people stagnating or falling. Massive unemployment and under-employment. Quantitative Easing not creating real jobs and fixing infrastructure. Low interest rates (still!) Multinational corporations with so much money through tax evasion schemes that they literally don't know what to spend it on. The ongoing privatisation of public services which rarely if ever benefits the public. Essentially, ongoing class warfare and levels of inequality not seen in practically a century. Regulatory and political capture and political collusion. Trade agreements on the horizon being negotiated in secret that would exacerbate every issue already mentioned.

And you say Krugman has wandered off reservation? Rubbish. He's got to keep banging the same drum because nothing's getting fixed.

We recently had a parliamentary election here in Denmark, and economic models featured rather prominently as arguments on both sides. One particular party would consistently refer to the economic model they used as a "calculator", which obviously in most people's mind is something that cannot be wrong. This went unquestioned and even seems to have caught on.

Indeed, what is funny that while resources are limited (and declining), economics are still considering as resources were unlimited. This video has a great re-interpretation of what economics really are, just managing energy in a productive system: https://www.youtube.com/watch?v=3w6ruZ_5nPE

The best part about Krugman is reading real economists try to figure out how to deal with him. Mostly they try not to engage, but every now and then you'll see several try to band together to "patch up" some statement he's made that's out of line (in their minds). Always a mistake.

It is not a hard science, but out of all of the soft sciences, I think economists understand just how tenuous their situation is. This is probably because they are using very hard tools to get a handle on a very fuzzy situation.

ADD: One of the bad things over the past 10-15 years has been the use of economics as a political weapon. It was bad enough when professors got stuck in one system or another and you had to wait an entire generation for new ideas to take hold. Now you're seeing a lot of laymen using economics as a way of arguing politics. Doesn't matter to me one way or another what kind of politics people have, but at least if you're an academic you have some notion of the realities of this field of study. In political discourse it's all just throwing dog poop around and supporting your team. That's going to make the serious study of economics even more difficult in the years ahead.

What your advocating is precisely what the article argues is wrong with economics as it's praticed today. Debating mathematical models is one thing, and debating the application of mathematical models while ignoring empirical facts is another.

> Machine learning and deep learning especially has a pretty cool new role here, but it will never take of this dynamic ongoing disucssion.

Your pagerank is strong.

There are worse economists out there than Krugman. Larry Summers, Mankiw. Anybody from the Chicago School.

The use of machine learning seems promising. One of the troubles with using regular mathematical equations is that at the end of the day economics is about human behaviour and that doesn't fit. Take a recent phenomenon like the 2008 crash. Banks increased lending, bankers made big bonuses while passing the risk on to various other parties and used some of the money to lobby politicians and keep the whole thing going until it suddenly collapsed leaving the other parties with the bill and the bankers with their bonuses. You're not going to capture that in a simple equation but running machine learning on say 50 previous crashes might get wise to what's happening.

I'm guessing you could run something like face recognition software but instead of faces it could recognise effects like the above and flash some politicians being paid off to turn a blind eye to high risk - give 'em a kick, warning or similar.

I feel like Nassim Taleb (Fooled By Randomness, Black Swan, Antifragile) has been making a similar point for years now.

Unfortunately I don't know enough about economics to comment any more than that.

Any recommendations on economics books with just the right amount of math, for non-economists with a college degree?

I'm looking for the economics equivalent of Feynman's lectures on physics.

You might be interested in "Christian Economics in One Lesson", a modernization of Henry Hazlitt's "Economics in One Lesson" by (in my opinion) the greatest living economist of our age, Dr. Gary North.


I'm working my way through "Thinking Fast and Slow" by Kahnemann. It doesn't have much (or any) in the way of heavy math, but it references many experiments, and the tone in general is about how easy it is to "disprove" standard economic theory.

I like it.

The book I got recommended from Steve Keen, when I asked him a similar question, was J.M.Blatt: Dynamic Economic Systems. Although he has beautifully simple approach, it is really outside the standard economic theory and apart from Keen and perhaps few other individuals, sadly no one really follows in Blatt's footsteps.

keen is a complete charlatan and the laughing stock of the economic world. he doesn't even understand the basic economic theory he attempts to criticize

Are the believers in so-called "economic theory" so desperate that they have to resort to name calling?

I sincerely hope someday you would read Blatt's book with an open mind. It has nothing to do with Keen, and on few pages with a bit of matrix math he obtains far better model than economists with their supply and demand curves can ever hope. (In fact, John von Neumann apparently had fingers in it too - shame he didn't live longer, we could have a proper economic theory today.)

If you want readable economics, there is Keynes. No maths at all though.

If you want to learn why math does not belong in economics at all you should read Human Action by Mises.

Capitalism is a greedy and cruel system where 2 + 2 = 4. Socialism is a a blessed system claiming that in a name of social responsibility and justice 2 + 2 should equal 7 (at least)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact