« Is "Y=C+I" a PPF? | Main | Health Spending Update: Into the 21st Century »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I take it that Ruth is a Senator's fan?

I thought the same thing as Nick - although at first I was wondering why an Ottawaian would be giving to the Sens, until I realized it is the anticharity option. It is brilliant because it is not only the cost of the cash, but the mental anguish from having the money support something that you don't like that enforces your behaviour.

I think that the micro is an incredibly valuable tool to not only understand why people make certain decisions, but also to understand what kind of decisions are rational or efficient (leaving aside the argument that nobody can know that because it is measured in everyone's individual utils and is not objective). As an example, behavioural economics can explain why people tend to overinsure consumer goods and small items (loss aversion), but micro can teach you when it makes financial sense to insure something - and also about asymmetric information, signalling and lemon problems that might also be important to understand in the insurance market. Complements, not substitutes.

Nick, yes, though she experiences some conflict when the Canucks come to town.

Whitfit, I like that insurance example, it's a good example of the complementarity of the two approaches.

I'm currently taking intermediate micro (technically "Managerial Economics," but it boils down to intermediate micro), and have been thinking more about the relationship with behavioural economics.

It seems to me that indifference curves, at least in my class, are taught as being consistently shaped, and therefore easy to work with. Research on actual behaviour seems to indicate that our preferences change depending on how the question is framed, but perhaps a better way to put this is that our preference - and therefore our individual indifference curves - are constantly changing depending on what other things are near the top of our mind when a decision needs to be made.

Our decisions still make sense in traditional economics, but that doesn't mean that a person will make the same decision repeatedly, even if a researcher were to control everything else. Brains are too complicated and dynamic to allow us to hold preferences as a given.

Neil - you make a good point about the way that indifference curves are always shown with that nice shape, without people really interogating where indifference curves come from. Try creating an excel spreadsheet, coming up with a bunch of bundles of say milk and cookies to which you're indifferent, and getting these to form a nicely shaped indifference curve - it's really remarkably difficult!

Your reaction is very typical of the reaction of my students when exposed to behavioural econ, which I'd take to be "behavioural econ makes sense, rat choice, doesn't, let's just dump all of this indifference curve b***s***."

But at the same time, there are limits to which people will respond to framing. When the price of gas goes up, people work out ways to consume less gas, no matter how hard the marketing department works to reframe gas purchasing decisions. Econ is about the kind of hard-edged instrumental rationality one needs to survive in a harsh world. Framing only matters so much - people do what they have to do to survive, and create new frames when necessary.

Ooh, if you like StickK I predict you will love Beeminder, which is like StickK for data nerds. (Commit to keeping all your datapoints on a "yellow brick road" to your goal and if you ever go off track, we take your money.)

Love your take on the economics of it.

Danny of Beeminder

Dreeves - interesting. Wondering why one might work better/worse than the other, and found this in the testimonials:

"This system has really helped me to stick to my goals. I tried StickK but it didn't work for me. It was too rigid and too long term. This is day to day, but I can't change what I do today. I love that. It really makes me stick to my goals."

Since these systems are really designed to help people solve problems of procrastination, prevent people from sacrificing long-term benefits for short-term pleasures, etc., I can see that a system that punishes deviations immediately would work better than a system with distant, far-away punishments. After all, if someone engages in hyperbolic discounting (basically ignores any cost that's in the future) they'll ignore the far-off penalty under the StickK plan.

I can see it now: Apple iProd! Forget NFC, the killer peripheral is taser probes! Camera, GPS, trasmit to monitoring center in China. Non-compliance is punished instantly.

Thanks Frances! There's another way we think Beeminder improves on StickK, besides bringing long-term consequences a bit nearer. Namely, by making the commitment contract revolve around a visualization of your data we can give more flexibility in changing the commitment. The behavioral economics behind the idea is that hyperbolic discounting means that your decision-making is distorted by immediate consequences. We take "immediate" to mean "within a week" and let you make any changes you like with a one-week delay. We call it the Akrasia Horizon. More about that at http://blog.beeminder.com/flexbind or check out the discussion on LessWrong: http://lesswrong.com/lw/7z1/antiakrasia_tool_like_stickkcom_for_data_nerds/

Danny - Another thing I liked about Beeminder is that it's free, so it's something I could assign my students to participate in as part of a class project - to do a behavioural economics experiment on themselves.

Well, free if you stay on your yellow brick road! :) As we put it on http://beeminder.com/money

If you're disciplined and self-motivated to reach your goal then Beeminder will be free for you forever. (Beeminder is obviously not targeted at people like you!) Stay on your yellow brick road and all is roses and ponies. If you veer off the road, you can walk away and abandon that goal — no further obligation. Maybe you decided that that goal sucked and you don't need Beeminder because you don't want to reach that stupid goal anyway. No problem! Beeminder was free for you too.

However, if you really meant to stay on the road, and you want to try again, you do so by committing to pay $5 if you go off the road a second time.

A woman wants to decide whether to have a child. She just adds the NPV of all the associated costs approxixately estimating future rates, including loss of income, child raising expenses, wear and tear on the body (based on differential cost and return on same), value of changes in marital utility (like trade offs between increased bonding and economic stress on probability weighted divorce costs with allowance for possible remarriage and the effect children would have on it) and the probability weighted costs of having a sub-optimal child. She compares this against the increased emotional utility (using some utility to money conversion that allows for changes in conversion with age and status), the NPV of the child at maturity and other misc. gains. All this has to include tax calculations based on demographic projections of future rates.

Women as economic agents do this every day, right?

Any program that tries to base a rational macroeconomics on aggregating rational agents faces some problems:

There is no reason to believe that evolution would select for economic rationality or that this rationality would involve simple linear combination of all sensed utility to some single basis.

Experiments have clearly shown that homo sapiens is rather different from homo economicus

The assumption that aggregating the results of individual actors working on individual parameters can be simply mapped into representative agents operating on aggregates of these parameters is demonstrably false, and has been known to be so for some time. The response has been to use the Friedman defense that all that matters is that it seems to work, but how you can combine this with a belief in rigorous micro-foundations is beyond me, since the two are antithetical.

It might be better to divorce micro from the need to be a role model for macro and gain some freedom of action. Better to die free than live a slave.

Peter N - child bearing is a strange example for someone critiquing rational choice theory to pick, as there is overwhelming evidence that decisions about whether and when to have children respond strongly to economic incentives.

Economy booms, birth rates go up; economy tanks, birth rates go down. Governments create policies that make it possible for women to combine work and family (Scandinavian model), birth rates go up. No child care, no parental leave, difficult to achieve work family balance - birth rates go down.

Economic forces certainly don't explain all of the variation in when/whether people choose to have children. But anyone ignoring economic incentives does so at their peril.

On micro/macro - the typical micro economist, in my experience, devotes very little mental energy to macro. Needing to be a role model for macro has never, ever, once crossed my mind!

"There is no reason to believe that evolution would select for economic rationality or that this rationality would involve simple linear combination of all sensed utility to some single basis." See http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/07/is-your-dog-rational.html

Evolution very much favours the kind of instrumental rationality modelled by mainstream micro.

Rational expectations holds that "predictions of the future value of economically relevant variables are not systematically wrong in that all errors are random".

This doesn't itself imply anything about rational behavior. People could decide everything by coin flips and errors would be random.

However economists tie this in with a number of assumptions.

1) That there are simple markets that clear and concerning which "individuals take all available information into account in forming expectations".

This isn't right. It isn't even wrong, because it isn't well defined. In reality a market is a lossy network with delays. It can't clear, because it's time to clear is much longer than the time between transactions, and the transaction costs (including information costs, risk premiums) differ between parties. Also the pairwise costs and delays don't always have locality. A has high costs to do business with B and low costs with C. This tells you very little about whether B's cost of doing business with C.

The price you get depends on where you are located in the network and what you perceive the information cost of finding the best price to be.

2) That people perform logical analysis of all available information.

People's economic behavior makes evolutionary sense. In the wild there aren't usually very many plausible choices at one time, and the information cost of obtaining more choices is large. On the other hand the ability to make fast decisions through habit and heuristics is life saving. People find logical evaluation of too many choices stressful, and faced with stress, they fall back on the fast mechanism.

3) That because, in theory, all preferences can be expressed in money through transitive indifference curves, economic behavior can be explained by financial incentives.

Unfortunately, to get any good answers by comparing money value requires evaluating a large set of personal and organizational preferences and beliefs. Armchair deduction from supposed first principals won't work. For instance, sticky wages aren't a mystery people who run businesses.

As for economic incentives working, I never said they wouldn't in some cases. But that's not rational micro, that's statistical inference. And the design of the incentives is at least as important as their money value).

"Evolution very much favours the kind of instrumental rationality modelled by mainstream micro." where it has proved successful in the past, not everywhere all the time considering all available choices. Try giving your animals more choices. Eight should be more than enough, and make the decision be important to the animal.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad