Monday, April 02, 2012

Is economics BS? A response to a KNZN

I'm a Not-So-Angry Bear...

Wow, people other than the barristas at my local Starbucks know who I am! KNZN names me "The Bearish Smith". Compared to Karl that may be true, but compared to Yves I'm a ray of sunshine...

Anyway, KNZN (whose moniker, I am assuming, is an acronym for "Korea/New Zealand National", reflecting his unique heritage) goes on to say some very pessimistic things about macroeconomics:
But another post to which I was referred by the Bearish Smith’s blog leads me to think about the issue of macroeconomics as a field... 
Here’s my take: to begin with, economics is basically bulls**t... 
Economics is bulls**t because it relies on the premise that human beings behave in a systematic way, and they don’t. Once you have done enough research to convince yourself that they behave in a certain way, they will change and start behaving in another way. Particularly if they read your research and realize that you’re trying to manipulate them by expecting them to continue behaving the way they have. But even if they don’t read your research, they may change the way they behave just because the zeitgeist changes – cultural sunspots, if you will.
The last paragraph may vaguely remind you of the Lucas critique. Lucas basically said that macroeconomics (as it was being practiced at the time) was bulls**t, but he held out the hope that it could receive micro-foundations that wouldn’t be bulls**t. 
The problem with Lucas’ argument, though, is that microeconomics is also bulls**t. And Noah Smith, writing some 36 years after the Lucas critique and observing its unwholesome results, takes it one step further by saying, if I may paraphrase, “Yes, the microeconomics upon which modern macro has now been founded is indeed bulls**t, but if we do the micro right, then we can come up with non-bulls**t macro.” 
Yeah, I doubt it. Maybe we can come up with slightly better macro than what we’ve got now, but the underlying micro is never going to be right. Experimental results involving human subjects are inevitably subject to the micro version of the Lucas critique: once the results become well-known, they become part of a new environment that determines a new set of behavior. And the zeitgeist will screw with them also. And so on. And in any case, even if the results were robust, I’m skeptical that we can really build them into a macro model or that it would be worth the trouble even if we could. Economics will always be bulls**t.
Wow! After calling me bearish, KNZN out-bears me by several Standard Bearishness Units! I've written quite a lot on the less-than-ideal state of macroeconomics, but I feel a need to defend my optimism about the (potential) future of the discipline. As KNZN says, I really do think that better micro will help us do better macro.

First, let's look at this "cultural sunspots"/"zeitgeist" argument. Stuff like that may happen, but it just means we need to keep updating our data and our models as cultural conditions change. Maybe the change happens faster than science can keep track of it (in which case we definitely are in trouble) but I don't think anyone really knows if that is the case. To assume science can't keep up seems to me like unreasoning pessimism.

And maybe cultural change isn't driven by sunspots at all! Maybe it's predictable. For example, things like fertility transitions, urbanization, and the shift from manufacturing to services all seem pretty universal, and these things may have predictable effects on culture. In fact, from seeing Japan, my bet is that this is the case. And perhaps things like financial crises, globalization waves, and debt supercycles change consumption/saving/investing behavior in long-lasting and predictable ways. I would not be surprised. 

I'm just generally very leery of the idea of "culture" as an inexplicable, chaotic force that makes economics hopeless.

Next, there is the idea that microeconomics is subject to a Lucas critique. I.e., that trying to exploit models of human behavior leads to changes in that behavior. This is two uber-pessimistic claims in one: 1) the claim that psychology is also bullshit and rather than science, and 2) the claim that market structure imposes no regularity on market outcomes.

First, psychology. Maybe it's because I'm the son of a cognitive psychologist, but I am highly skeptical of this claim. People have limitations on how much information they can process; telling them that you know this is not going to suddenly turn them into supercomputers. Furthermore, people have norms and values. Telling people that you know they think something is fair will not make them suddenly think it's unfair. Yes, some behavioral regularities disappear when you try to exploit them, but not all

Case in point: the lottery. The lottery is a reliable way for a government to make lots of money. So far it has never failed. And it relies on human behavioral regularities; it's easy to imagine a species that wouldn't play the lottery. But we know that we are humans, and we know that humans like to play the lottery, so the government can run lotteries and get money again and again and again. That's microeconomic technology in action.

But microeconomists don't just study psychology, they study markets. The structure of markets makes a big difference. For example, continuous double auction markets very reliably produce the law of one price, while bilateral trade negotiations do not. That is a robust result! It doesn't go away when you tell people about it. And it doesn't seem to depend on people's preferences. Other examples come from auction theory; everyone knows that auction theory exists now, and yet that hasn't made the theory's predictions suddenly start to fail. 

These are examples of outcomes that are predicted by theory and confirmed by experiment. That's called "science"! You can say "Well, maybe these things will go away tomorrow," and yes logically that is true, but that is true for any science, including physics. There's no guarantee that what holds today will hold tomorrow. Science rests on the assumption that the mechanisms of the Universe are stable. When that assumption fails, science fails. But I see no good reason to assume that this assumption always fails for economics. If you're just a pessimist, fine. But I'll need more convincing!

And there are LOTS of areas where we have barely begun to even study micro behavior. One big example is the behavior of firms. We don't understand this very well. In most macro models that you see today, the behavior of firms is even more hopelessly stylized and unrealistic-seeming than the behavior of consumers. A second, rather obviously, is financial markets. We need a lot better understanding of both how people behave in financial markets, and how the structure of these markets might produce certain regularities (like bubbles). And a lot of people, including myself, are working on this.

Now, note that I haven't talked about macroeconomics. Can we aggregate from micro to macro? That's something KNZN doesn't discuss, and it really deserves an entire post of its own (coming soon!), but basically, there are lots of reasons to think that the answer is "yes". Even if the macroeconomy is incredibly chaotic, we can use brute-force computer power to make short-term forecasts, like weather forecasters do. If there are emergent properties, agent-based modeling will give us a clue as to how complex the economy needs to get for these to take effect. And there may be many situations in which we can do things the old-fashioned way, i.e. by math - even if representative agents don't usually work (and I think they don't usually work), we can do things like OLG models and various other heterogeneous-agent models, and see if that works!

I think that lots of what people call "economics" is, in fact, bullshit. But I think that's true to a greater or lesser extent in every scientific field. I do not believe that economics must be bullshit. I don't think a convincing case has been made for that proposition. So even if I am The Bearish Smith when it comes to the current state of macro, I am bullishly optimistic that economic science will continue to move, however sluggishly, in the right direction.

49 comments:

  1. It's bullshit.

    Simple example: those arguing "sticky wages exist" as an argument for policy X (printing money), are the same people who wail when you advocate policies to end sticky wages.

    Economists UNDERSTAND auctions work.

    But they don't advocate we auction the unemployed.

    Ultimately your world view is clouded because you don't start a company, and employ people, and worry about customers want to buy your product.

    You worry about:

    1. "externalities"
    2. people in your tribe (bureaucrats, academics, journalists, etc having less power compared to the businessmen.

    Your entire focus is about the stuff that doesn't matter to what would make MICRO a science.

    And MACRO well it's a science until there's just one currency and no one gets to print it.

    ReplyDelete
    Replies
    1. Well, I do think that economists need to get out and observe firms much more closely, to get ideas about how they really behave. The models of firm behavior currently used in macro modeling are total crapola.

      On the other hand, understanding how to turn a profit is different than understanding how to grow an economy. For example, from the perspective of an entrepreneur, getting a government subsidy for your product is awesome. But from the point of view of the whole economy it is not so awesome.

      Delete
    2. "For example, from the perspective of an entrepreneur, getting a government subsidy for your product is awesome."

      Econ lesson for Noah #1: run govt in way to minimize rent seeking.

      From that we can support two new thoughts:

      a) smaller distributed government is healthy
      b) too much capitalism does not mean too many capitalists, but too few capitalists.

      Let me put it in logic you understand:

      In a capitalist system, you cannot have big government, unless you have big business, thereby a government built to favor small business, will be tend to be small itself.

      -----

      But let's stick to basics here:

      Are you ready to demand that economists arguing "sticky wages" also favor policies to end sticky wages?

      That's a lot of DeKrugman nuts you have to kick.

      Delete
    3. OK, I don't want to be a total jerk here, but I think you seem to have some interesting ideas, and you just need to work on communicating them more effectively, because frankly I can't really tell what you're talking about most of the time...

      Delete
    4. Noah, do not be afraid of answering questions directly. I'm not being a jerk, I'm being helpful. Really.

      So again,

      Are you ready to demand that economists arguing "sticky wages" also favor policies to end sticky wages?

      I suspect you'll struggle with this answer, because you like micro policies that prop up wages AND you like macro arguments that suppose sticky wages are unavoidable.

      There is a scientific econ policy that solves for this conundrum, but before I assume you'll punt on the question, I should give you a chance to answer.

      -----

      You answer there gets us quickly through the discussion of econ as science which is why I keep mentioning it.

      Next, I will make an argument for distributism, but you are right that was a new topic.

      Delete
    5. Have you read this?

      http://noahpinionblog.blogspot.com/2012/02/lower-wages-can-be-good-thing.html

      Delete
    6. "Are you ready to demand that economists arguing "sticky wages" also favor policies to end sticky wages?"

      This does not follow unless you think this is the only market imperfection AND/OR think that the distribution that would result given present endowment is better than whatever we got.

      I think wages are sticky. I do not think wages should fall in general. See! B did not follow from A!

      Delete
    7. Hurrah on that pinion! Next up, distributism.

      nemi,

      The point is that MACRO argues from the presumption that wages are sticky, which is fine in as much as the same guys making the argument aren't the ones peeing in the pool.

      The way to solve unemployment is a Paypal style weekly deposit into a debit account ($200-$240) of all who say they want to work. A Guaranteed Income.

      And then auction their weekly labor to private bidders in a Ebay system starting at $40. Let them keep 50% of the weekly bid. and let them choose which bid they accept. don't make them travel more than a couple miles.

      But EVERYONE now has a pain in the ass boss making them dance. Winning bidders pay no taxes or outlandish worker insurance etc.

      Overnight, unemployment will evaporate. Greedy local bidders will comb through the system looking for finds, stealing good deals from one another.

      Entrepreneurs find workers to supply cut rate services in poor neighborhoods. Ghetto neighborhoods get fixed up cheap and have access to low cost services. Mommies now can afford child care to go work. And the upper and middle classes will look for personal help.

      The point nemi is that we must completely separate the commitment we make to Americans on a living wage VS. admitting what our unemployed are really worth in a global market.

      We cannot assert that our labor is special. We can assert that our needs are special.

      And it REALLY does matter what's worth $3 per hour and what's worth $5 per hour. The unemployed need that signalling. They need to have shitty bosses trying to earn a buck off them even at $1 per hour.

      The thing of it is... we shouldn't worry about deflationary forces of 30M unemployed being auctioned any more than we worry about productivity gains. If we're all in this together that's a TAX and AID issue, it is not a WAGE issue.

      So it doesn't matter if suddenly car wash attendants aren't worth $8 a hour anymore. It matters how many cars now get washed at the cheaper rate.

      Finally, this is how you increase trust and sharing in a large multi-ethnic country, you ensure that everyone knows everyone else is working their 40, and that public employee jobs have the same quit rates as the private sector.

      This improves trust in others and the institutions that represent us.

      Delete
    8. "The way to solve unemployment is a Paypal style weekly deposit into a debit account ($200-$240) of all who say they want to work. A Guaranteed Income."

      Add a low level single payer health insurance (including disability insurance) and you will get my vote.

      Delete
    9. Anonymous8:19 AM

      @Morgan Warstler

      Sticky wages do not just exist because of long-term wage contracts, but also because many workers/families have "sticky" expenses. Allowing the income of workers to fluctuate might cause many problems by itself (many personal insolvencies). In a downturn you have less unemployment, but even more people face large wage cuts that might cause millions of private insolvencies. This insolvencies can accelerate the crisis as well. It is hard to say whether that is better than sticky wages. And that is not even taking into account the insecurity many people continuously have to face about how much they are going to earn in the near future.

      As for your auction model, we basically have that already just with stickyness and without government support.

      Also in Germany we allowed firms to pay less than people need to live and subsidized them accordingly (comparable to your system, who earns more gets a higher total income).
      What happens was that wages were going down even further for low-skilled workers (which earned less than minimum standard of living). You hope that this will largely result in lower prices, but in Germany it lead to larger profits instead...

      Delete
    10. Keith Beacham8:44 AM

      Make that ($360-$400) tax free then "add a low level single payer health insurance (including disability insurance) and you will get my vote."

      Delete
    11. Anonymous10:43 AM

      @ nemi, Keith
      That is the problem with all the nice sounding basic income approaches. They look good in theory, but if you try to pay an actually amount people can live of (inc. house, insurance...) they do not work out anymore. That might change in the future, since capital income as a share of GDP goes up and up. However, as long as we are not able to access capital gains in a substantial manner in order to pay for the basic income(without having our capital flee *g*), we will never be able to implement such a system to the advantage of the majority.

      Delete
    12. Folks, one thing we cannot promise people is that they get to live single, or even perhaps with children, by themselves. Once you get over room mates, $240+ is workable.

      A tapering pay scale as bid go up is easy, so that higher bids reward the worker and decrease the gvot. subsidy.

      But everyone who wants to work will likely clear $8+ per hour.

      BUT NOTE, in poor areas where these people tend to work, all the services get cheaper - their pay goes further.

      Healthcare is EASY to accomplish, the moment you come to terms with a basic fact: like in all things the poor get less.

      The reason Obamacare is truly despised by the top half, is because it tries to put the bottom 25% into the same pool sharing the same level of service.

      This is unacceptable, and will not stand, even if SCOTUS doesn't toss it.

      Two tiers of care is unavoidable, and no one should resent it. And this offends some liberal technocrats who actually do want pick and choose who gets the MRI first.

      Won't happen, but if the left lets go of that, we can solve it.

      All pricing follows a 80 / 20 rule - you get 80% of the functionality of high end goods for 20% of the price:

      $100 feature phone / $500 iphone
      $4 lb cheese / $20 lb ccheese
      $14K KIA / $70K BMW

      For $4K per man per year we can deliver better than 80% of the high end outcomes. It means out of patent drugs, Xrays not MRIs, student doctors, big scars from surgery, etc. no choice of doctor etc.

      We'd STILL want to use a form of HSA for the low end, and let those who get a subsidy keep growing their account, so they become good shoppers.

      But we can have Soup Kitchen Care, that is basically effective and mostly single payer, and covers 100% of uninsured.

      1. no free rider problem - Obama now says, "if you can afford better, you'll want it" If you use Soup Kitchen Care system you'll be means tested and billed for work if you aren't carrying either type of insurance insurance.

      2. we keep the private sector high end that drives new tech development AND the low end clamors loosening of licensing requirements. Incentives align to serve each groups needs getting both new and faster and cheaper WITHOUT making those who can pay feel like they are getting cheated.

      Delete
    13. Anonymous4:53 PM

      @Morgan Warstler
      If you describe it like this, you probably lost your two new supporters again. ;)

      What makes you so sure that workers can claim 8$ and that prices will go down substantially? As I said we basically have your system in Germany for low income earners (except the non-stickiness). Salaries were going down and profits up.

      Your idea of a two-class society does not appeal to me. Unfortunately, many countries are moving exactly in that direction, which makes me really said. Once the society is separated, it is almost impossible to revert that. Children from your soup kitchen families will never have comparable chances to children from affluent parents. This will cement the separation over generations.

      But of course, the feelings of the people who can afford to pay are more important than actual poverty. There have been times where almost everyone in USA could live decently from his or her work. Today productivity is up, skill level is up and more people work per household. Don't you think that the people who are able to pay profited from that development? After all, wages are determined by supply and demand, not productivity, effort or fairness.

      Maybe we should tell them that, so they do not feel cheated.

      Delete
    14. As you type out feelings on their free Internet blog, note that you get to talk to all kinds of other people, and be anonymous and live or die on the value of your ideas.

      There will always be power, my system is built to keep the power held by govt. officials to down to the minimum spend and maximum outcome.

      Small aside: you suffer from blissful ignorance of what you do not have even while you lament the disparity of haves and have nots.

      You think about right now who is rich poor.

      I think about what system would have gotten poor people a computer and Internet 10 years sooner.

      As such in my medical system, I focus on doing harm to the demand side of high end treatments, we need to do to heart surgery what we have done with fake tits - better and better and always cheaper.

      Next, I want the bottom half advocating for less licensing and learning to price shop.

      Taken together this means that 10 years from now there have been more things invented and they have gotten cheaper faster.

      You do not recognize the really expensive drugs that today would cost far less if that last 10 years had been run this way already.

      You should be ANGRY that poor people don't yet have college that comes free when they buy a new TV.

      Finally: if you take 30M unemployed and auction them in $1 per hour auctions under terms I describe above - overnight you'll get $50 a week daycare for mommies who can now go to work. You'll get neighborhoods being fixed up, and housing being repaired, you'll get people saving money on haircuts and car repairs. The 30M unemployed are essentially propping up the labor wages for a bunch of folks slightly more skilled - this is not acceptable.

      The benefit is ALWAYS cheaper car washes meaning more cars being washed, more people lower down the chain enjoying something previously priced out for them.

      Delete
    15. Anonymous10:03 AM

      "There will always be power, my system is built to keep the power held by govt. officials to down to the minimum spend and maximum outcome."
      What about power in the private sector?

      "I think about what system would have gotten poor people a computer and Internet 10 years sooner."
      Even if your system were more efficient, it is well established that poverty is a relative problem. Hence, you cannot argue that poor people would have a better live, living of the crumbs of a more efficient high society. I am currently in South Africa and many white people argue that blacks had a better live under the Apartheid regime (and it is true from a purely economically perspective). Now can we conclude that Apartheid was a better system?

      Does fairness actually play a role in your system? Do you think it is fair, if I am born to be very intelligent, grew up in a very affluent family that helped me develop and could afford to send me to an elite university even without a scholarship? Now you would like to add to that, that this blessed person should only pay 1$ for car washes and other services that are subsidized by the government?

      There are a couple of countries in Europe that have relatively flat income curves, which do just fine. Of course, people would like to pay less for services, but do you think that this is decreasing technological advancements? After all, we can expect the most productivity increases, when labor is expensive. The question is just what happens to the people, which are replaced with capital. You want them to live in poverty and work for the middle class and rich for almost nothing, I want them to have a share in the productivity increase, either through higher salaries or through less working hours.

      "The 30M unemployed are essentially propping up the labor wages for a bunch of folks slightly more skilled - this is not acceptable."
      That is probably true, but how is your system different? Now the "slightly more skilled" workers earn less and have to pay much less for services. That is also a real wage increase. Alternatively, if you system eventually includes the "slightly more skilled", they also will have trouble making a living and the "even more skilled" will have huge benefits from that.

      "The benefit is ALWAYS cheaper car washes meaning more cars being washed, more people lower down the chain enjoying something previously priced out for them."
      Maybe you can give some more live like example of what people who earn only $1,000 a month would like to consume.
      In Germany, we analyzed the expenses for the bottom 20% (excluding housing) in order to calculate the minimum standard of living.
      As it turns out, most of the spending is not in actual services: 35%food/drinks,37% Energy/Cloth/household machines/health products (e.g. toothpaste),6% transport,11% free time (e.g. cinema), 7% others (includes hairdresser), 2% gastronomy
      Only the last three positions, which account for 20% are likely to have a large labor component that could get cheaper. If you include housing you probably also have lower maintenance cost. Nevertheless, you can hardly make the case that the poor are actually benefiting from getting paid less if the majority of their expenses are not getting any cheaper.

      Delete
  2. I'd say there's a continuum between pure science and pure bulls**t. Sociology is close to the pure bulls**t end. Chemistry is close to the pure science end. Economics is a step or two in from sociology. Cognitive psychology is perhaps a few steps in from economics. When you try to build something on a foundation, the building process introduces a higher bulls**t factor (kind of like multiplying two imprecise estimates), so macroeconomics built on microeconomics built on psychology ends up being pretty far toward the sociology end of the spectrum.

    ReplyDelete
    Replies
    1. I agree that increasing complexity usually makes things more difficult to understand. The less complex something is, the easier it is to study it in a lab.

      But I would only call something "bullshit" if it pretends to have predictive power that it does not have.

      Delete
    2. knzn,

      'Once you have done enough research to convince yourself that they behave in a certain way, they will change and start behaving in another way.'

      We have research that point out people don’t usually like to jump up cliffs/bridges. Once that research is published do people suddenly decide they like to jump of cliffs/bridges?

      Even animals behave in systematic ways to maximise 'utility' let alone human beings. And yes, their is research backing that up, it’s called evolution theory.

      Your argument is based on a superficial understanding of rational choice. The fact that economics cannot predict all human behaviour to it’s minutiae detail doesn’t mean some human behaviour isn't systematic and immutable.

      Delete
    3. *Correction. Jump of cliffs/bridges. Not 'jump up'. Though life would be easier if people could jump up to cliffs and bridges.

      Delete
    4. David, I don't think you can build a very useful economic model based just on the rational ways people behave that can be confirmed by experiments. You can assume full rationality, but then you get garbage-in-garbage-out, because people do actually behave irrationally. Or you can look at the irrational ways that people (according to experimental results) behave, but then if you try to exploit that behavior, people may realize that it's irrational and stop doing it.

      Delete
    5. Noah, Admittedly I haven't been very precise in my use of the word "bulls**t." But then "pretends to have predictive power that it does not have" is an ambiguous predicate as well. Surely some economists claim that their analyses have predictive power that they don't actually have. I don't claim that economics has no predictive power at all. But I don't think its predictive power will ever be robust and comprehensive enough to be comparable to fields that are generally acknowledged to be scientific. Economics is a useful and necessary field, but I know what science looks like, and I know what bulls**t looks like, and economics looks more like the latter.

      Delete
    6. KNZN:

      So I can summarize your argument as "HARUMPH GRUMP GRUMP HARUMPH"? ;)

      Just kidding. Sure, econ will never be physics. It'll probably never even be biology. But we have to still do our best! And that means giving the science primacy over the non-scientific empirics, and any empirics primacy over the assumption-making and intuition.

      Delete
    7. “When you try to build something on a foundation, the building process introduces a higher bulls**t factor (kind of like multiplying two imprecise estimates),”

      Right. And on top of that, you usually do literally multiply two (or 32) imprecise estimates (see e.g. Wietzman on climate change), have to use very farfetched proxy and instrumental variables, choose rather arbitrary functional forms, etc. etc. etc. Not to mention that the model you actually estimate/calculate/calibrate usually only has a pretty loose resemblance to the actual theory, and you will probably make assumptions about e.g. probability distributions out of convenience, and … etc. etc. etc.

      To say that the final “empirical results” actually tells us anything about anything is often quite a stretch of the imagination. Add publication bias and the resulting incentive compatible behavior of researchers and you will, if possible, think even less of the literature.

      However, I do not think it have to be like this. I think there are many ways in which the study of economics could be improved, but there is essentially no competition in the way things are done. Sorry, that is not entirely true, economic blogs is a new format for discussion – but it is not enough.

      Delete
    8. knzn,

      Without a more specific example I’m not sure if I can agree with you or not. Specifically, I’m not sure what you mean by 'very useful economic model[s]' or 'irrational'.

      However if I was to make some presumptions,

      If by 'irrational' you mean having inconsistent preferences, I’m not convinced that experimental evidence show inconsistent preferences as opposed to performance errors or after adjusting for the appropriate parameter into the preference function.

      Secondly, even if people change 'systematic' behaviour, why is that a con? (I’m presuming by ’systematic' you mean cognitive biases). If (a big if) people adjust their behaviour to reduce cognitive biases (cognitive biases btw in my mind doesn’t imply 'irrationality' as commented in above para.) then that could only make behaviour easier to predict, not less.

      Thirdly, as I’ve alluded to in my first response, people don’t just change behaviour schizophrenically. We can make plenty of 'useful' models based on their behaviour that is immutable. I agree with the notion that people can change their preferences etc. over time but that’s not an insurmountable barrier to making 'useful' models.

      I’m not fussed about your opinions of macro, but saying micro is bullshit is going to far in my opinion. Again without being sure what you mean by 'useful economic model[s]' I think on the contrary there are plenty of those in micro.

      I think what you’re not satisfied with is the ability of economics to provide a theory of everything. I’m ok with that, and I agree with Noah, economics is plentiful useful without being physics or hard science like.

      Delete
    9. But I would only call something "bullshit" if it pretends to have predictive power that it does not have.

      I recommend to your attentions the bond vigilantes, the inflationistas, and most especially the expanionist austerians.

      Cheers!
      JzB

      Delete
    10. Yes, that is bullshit.

      Delete
  3. "1) the claim that psychology is also bullshit and rather than science"

    I'm not sure if you even need to claim psychology is bullshit for microeconomics to be based on a huge amount of assumption (bullshit).

    The strongest sentence in kznz's post, in my opinion, is:

    "Economics is bulls**t because it relies on the premise that human beings behave in a systematic way, and they don’t."

    It seems to me, that you can base microeconomics on observed, independent microeconomic experiments and if you use that as a baseline. However, even if other experimenters produce lots of repetitions of it, it doesn't guarantee the human beings behave in a systematic way outside of the experiment, which is how I'm interpreting knzn's point. Hopefully he'll yell at me if I'm wrong.

    As an economics undergrad, I suppose I also have hopes for the future of the economics profession; like you, I'm pretty unsatisfied with where it's at. However, I generally agree with knzn that it's a pretty soft, deductive, intuition-based science and there's not really any amount of microeconomics that's going to change that, in my opinion.

    I entirely agree that we need better microeconomics, though, of course. But I agree with knzn that it pretty much has to be bullshit, in some sense. We can reduce the bullshit, but we're always going to have to make assumptions.

    ReplyDelete
  4. (A lot of) economics really is bullshit. I´m still a PhD student, and I have only published two studies this far, but I do believe that I really understood the methodology and assumptions I used. However, I would not bet a cent on the correctness of those results. There really is very little incentives to do things right (or maybe there is, but it is a very high risk game). As long as you simply do what everybody else is doing, you will easily get away with the stupidest things imaginable. The way of least resistance, for me at least, is to do nonsense economics for my dissertation and real economic thinking during my free time (which I have a ton of given that I no longer care how stupid my so called research is).

    One of my problems, I guess, is that I am not skilled enough to express my thoughts in the kind of rigorous consistent models that would be demanded in order to even be considered (maybe this formal requirements are stricter at relatively low ranking institutions due to low self esteem?) – and these damn economic blogs takes too much of my time. However, since I still haven’t been able to complete even one model that I am comfortable with, I am glad that I decide to do BS.

    ReplyDelete
    Replies
    1. As long as you simply do what everybody else is doing, you will easily get away with the stupidest things imaginable.

      Yes. I know this is true. But that is about the culture of the econ profession, not about the potential of economics as a science! Back in the days of alchemists - a situation I'm sure you'll agree was far worse than econ today - there was still huge potential for chemistry to be a real science and get real results. See what I'm saying? I think econ has a lot of scientific potential, but the culture of the profession means that that potential isn't being realized.

      Delete
  5. Anonymous11:30 PM

    << To assume science can't keep up seems to me like unreasoning pessimism. >>

    Yeah, maybe so, but to think that economics is science seems to me like unreasoning optimism.

    ReplyDelete
    Replies
    1. What do you think of the examples I gave in my post?

      Delete
    2. I do not know what a continuous double auction market is. I know what a double auction is, and of course LOP will trivially hold for such a game (since there only exist one price, created by a central clearing house). What do the LOP hold over in the continious version? Time, different games, cultures, countries?

      However, maybe this is the way to make economics scientific. I.e. instead of understanding the mess around us, create a system that we do know the properties of and replace the mess with this new system.

      Delete
    3. Anonymous1:31 AM

      I am not particularly impressed by your lottery example. If I believe economists and statisticians all over the Internet, it is irrational to play the lottery - see this one for example: http://www.circlemud.org/~jelson/megamillions/ - and yet, the greater the prize the more people buy tickets. Is this really an example of an outcome predicted by theory and confirmed by experiment? Isn't it the best demonstration that mainstream economics is bullshit?

      I don't know about your auction example, but let me cite another example that I take from Hal Varian's Intermediate Microeconomics Fifth Edition, p. 312: "The optimal (in italics in the original) strategy in a common auction like this is to bid less than your estimated value - and the more bidders there are, the lower you want your own bid to be." Is this prediction really a triumph of micro? Yeah, I guess, until you read the next paragraph, or you work in the real world, the one economists never step in.

      Why don't you come intern for the summer where I work? We bid on distressed bank debt in competitive auctions. I will provide room and board, you provide me with economic education. I am sure we would both benefit greatly.

      Delete
    4. PS: "The lottery is a reliable way for a government to make lots of money"

      Sure, at that level theory is rather stable.
      In my first economics class the teacher rhetorically asked, how do you think a consumer would respond to a price increase, and I already knew what the effect of a tax would be.

      When I talk about economic BS it would be about the size of the response w.r.t. to the tax increase, estimations of income and substitution effects, feed back/general equilibrium effect, dynamic effects w.r.t. e.g. technological progress, industry structure, off shoring, trade etc. – where uncertainties really get compounded.

      Delete
    5. If I believe economists and statisticians all over the Internet, it is irrational to play the lottery

      Ah, unless you are risk-loving over certain small gambles! See, if there are multiple theories, and they make contradictory predictions, and you can verify which one is right and which one is wrong by doing experiments, that is called SCIENCE!


      Why don't you come intern for the summer where I work? We bid on distressed bank debt in competitive auctions. I will provide room and board, you provide me with economic education. I am sure we would both benefit greatly.

      Sure, you're on. What's your email?

      Delete
  6. Anonymous8:11 AM

    "Yes, some behavioral regularities disappear when you try to exploit them, but not all."

    I'd have thought that was so obvious that it couldn't be missed. I'm afraid KNZN's effort to make light of economics has led him to take his own argument too lightly.

    And as to this --

    Noah, do not be afraid of answering questions directly. I'm not being a jerk, I'm being helpful. Really.

    So again,

    Are you ready to demand that economists arguing "sticky wages" also favor policies to end sticky wages?

    -- the writer seems to have mistaken a discussion about whether economics tells us anything valuable for a an opportunity to push a particular agenda. The whole bit, with its "afraid", "are you ready" and "demand" reads like patronizing claim that that unless Noah accepts Morgan's views, Noah is not "getting it". It also reads a little bit like a cult recruiting pamphlet.

    ReplyDelete
  7. Anonymous8:29 AM

    I just want to challenge the idea of microfoundations for macroeconomics. To stay with so called "real" sciences like physics, they don't try to do that either. On every level of abstraction they observe completely new things. That is probably why we put airplanes in a wind-channel instead of trying to infer information from the interaction of the air molecules.

    And I would like to add that interactions of air molecules are probably easier to understand and formalize than interactions between people.

    ReplyDelete
    Replies
    1. Anonymous10:35 AM

      Much easier, but sadly air molecules refuse to change their behavior once we tell them how they've been behaving until now. Oh, KNZN is a complete jackass.

      Delete
  8. Noah: "One big example is the behavior of firms. We don't understand this very well. "

    Data is a big issue here. Think of all the micro data that exists for consumers, and is readily available for analysis: the census, survey of consumer finances, labor force survey, PSID, National health and retirement study etc, etc.

    How many comparable data sets are there for firms? Even when the data is collected (as it is in Canada) it is barricaded and almost impossible to obtain.

    Yes, there are commercial data sets compiled from firms' publicly released financial statements available for a (usually hefty) price, but nothing comparable to surveys of people.

    I would love to see some of those advocates of corporate personhood lobbying to have a publicly released census of firms, just like there is a publicly available census of humanoids!

    ReplyDelete
  9. Anonymous10:26 AM

    You're the son of a cognitive scientist?! Then you should be ideally placed for entry into the science of the future.

    This sounds sarcastic, but isn't. Psychological regularities as elucidated recently by e.g. Thinking, Fast and Slow and Foundations of Neuroeconomic Analysis aren't going away because people are informed about them. As you well know.

    I suggest you stop wasting time on at least this particular KNZN. There is no shortage of other things to think/blog about.

    ReplyDelete
  10. sriram12:05 PM

    Late comment – but hope you read it. From following the blogs I figured out that economists fundamentally do not yet know what the economic capacity is i.e. no idea of what the output gap could be. So you can’t say if the economy is overheating or not. They don’t know if Government spending can spur private demand over and above Government purchases i.e. no idea if there is a multiplier greater than 1. They don’t know if tax cuts provide lasting benefits or just temporary. No idea if leeching the economy of debt helps or throws it into a vicious deflationary cycle. After ALL THESE YEARS, no idea even now if banks can create unlimited amounts of credit or not. I could go on. There isn’t much written about micro in the blogs I read, but you can imagine that successful micro-economics would have been first signalled by successful market strategies that exploit individual’s behaviour (rational or otherwise). There are billions of $ to be gained from understanding how individual investor behaviour can influence markets. So far, despite such fruits, we seem no-where close to understanding it. That leaves me sceptical that the economic profession has been successful in micro. And micro to macro conversion has so many caveats to be considered that again I doubt there have been too many successes there. And to top it all, you have people like Steve Williamson who says he doesn’t make predictions but only tries to understand how the world works. You can hardly blame people, especially amateur economists from feeling like the whole profession is vague and relies on foundations that shake everytime there is a new crisis.

    ReplyDelete
  11. Morgan Warsler is loco.

    an op-ed from a couple of days ago on "Physics Envy"

    http://www.nytimes.com/2012/04/01/opinion/sunday/the-social-sciences-physics-envy.html

    ReplyDelete
    Replies
    1. He just isn't used to writing for an audience.

      Thanks for the article...but I don't like it at all! Grump grump. I may have to write a post about it.

      Delete
    2. JustBob814:13 PM

      That NYT op-ed is a rather poor reflection of its authors' new book, which is relatively short and a good read if you have the time.

      http://www.amazon.com/Model-Discipline-Political-Science-Representations/dp/019538220X

      As far as KNZN and his bullshit continuum. All he's really discovered is that there is a continuum of difficulty in areas of study. If, like the drunk at his light pole, we just focus really hard on the things that are easy, it doesn't help us make any progress on the things that are hard.

      Delete
  12. Anonymous3:29 PM

    Noah,

    I'm with you. Great post as usual. I agree that writing down a model of behavior does not necessarily cause people to change their behavior and, hence, invalidate the model. However, I've always worried that this is true of finance. This causes me to question the "science" of, say, CAPM. When an economist writes down an expected utility problem, people still buy groceries the same way. However, the Black-Scholes Theorem did change the way finance is conducted. Any thoughts?

    ReplyDelete
  13. The discussion whether economics is bullsh*t has happened many times in the past, and is normally one of the ways that economics practitioners rationalize, or come to terms with, their continuing use of bullsh*t. That is, this discussion allows the implicit statement "We all know its bullsh*t, but ...". The challenge is to avoid this cop-out. How do we know this is a cop-out. Because there are some economists trying to minimize the bullsh*t but they are not well treated by the economics profession. e.g. a 1-sector, 1-good, closed IS-LM model assumes S=I, but the empirical evidence suggests that I>S, then efforts must be made to develop new models and those economists trying to do this should be accorded some respect.

    ReplyDelete
  14. Anonymous4:33 PM

    Perhaps psychology is not bulls**t but economics is.

    After all, mathematics offered a basis for physics, and physics (with its weird quantum stuff and electron orbitals) offered a basis for chemestry. Chemestry offered a similar basis for biology, and now we have biologists making inroads to psychology. So maybe someday economists too will be provided a ground, but it won't be from representative agents and Pareto optimality, and the agents won't be rational (at least in some game-theoretic sense).

    ReplyDelete