March 2, 2008

Labour Supply Curves

Filed under: Economics, Vitiated by Ignorance — duncan @ 5:56 pm

As I try to educate myself in the basics of economics, I’m uncertain how much I should post here. If I put up every thought that crosses my mind, the blog would be unutterably tedious, and I’d never get any reading done. But I often want to express bafflement or disagreement – and that’s what the blog’s for.

Very quickly, here’s a passage from my textbook (Krugman, Wells and Graddy, ‘Economics: European Edition’, 2007):

“Economists believe that the substitution effect usually dominates the income effect in the labour supply decision when an individual’s wage rate is low. An individual labour supply curve is typically upward-sloping for lower wage rates, as people work more in response to rising wage-rates.” (p. 319).

I just don’t see how this can possibly be right. Putting aside the more general problems with labour supply curves (e.g.: we mostly don’t choose how much of our labour to supply – the amount of labour we end up selling is massively determined by social structures we have no control over: the five-day week; the ‘nine to five’; ‘voluntary’ overtime; and that’s just in economies characterised by a reasonable degree of personal wealth and leisure time)… putting aside such problems, and taking the labour supply curve as given… I just don’t see how it could be accurate to say that “[a]n individual labour supply curve is typically upward-sloping for lower wage rates”. Surely at the lowest wage rates people work most. Those who survive on subsistence wages spend their whole lives working for those wages. That’s the problem. Krugman et al. can’t seriously be suggesting – can they? – that in the lowest income brackets, the lower the wages, the longer the ‘leisure hours’.

Is the backward-bending individual labour supply curve really an economic orthodoxy? If so, why?


  1. there’s always going to be shit jobs (unskilled, waiting, boring, tedious, difficult hours). and people are always going to be lazy in school, the dropouts. unskilled workers —> unskilled jobs.

    unfortunately, that’s just the way things are. human nature. no amount of social engineering can change that.

    Comment by wanderer7 — March 2, 2008 @ 6:07 pm

  2. Okay – but my problem here isn’t with social engineering, it’s with economic analysis. Why does the backward-bending individual labour supply curve – which, if my text-book is to be believed (?) most economists agree accurately represents the behaviour of the labour market – why does this supply curve suggest that at the lowest wage rates people work the shortest hours (at least until you reach the very high income levels at which people can live their whole lives as one big party)? Surely common sense suggests that at these very low wage-rates people have to work incredibly long hours, just to earn a subsistence wage. I mean – I don’t seriously think that economists haven’t thought of this issue. 🙂 It’s just that my textbook hasn’t addressed it (yet).

    Comment by praxisblog — March 2, 2008 @ 6:36 pm

  3. Economic “orthodoxy” is not actually meant to represent reality – not even to economists. Instead, it teaches the abstract principle that all transations can be analysed on the margin. Marginalist analysis is an incredibly powerful tool, but anyone who’s taken more than a few undergrad econ courses knows that these fundamental assumptions are very quickly mitigated by a slew of important factors.

    The main problem I have with these methods of imposing marginalist orthodoxy is that they don’t address actual situations, but theoretical ones. And of course, here the theory is oversimplified by huge leaps and bounds. Why aren’t there qualifiers about poverty in economics textbooks? That seems a question of ideology, doesn’t it?

    Comment by Dave — March 4, 2008 @ 7:07 pm

  4. Dave – thanks. Very quickly, ’cause I’m typing on the fly.

    1) Of course I totally agree with you about ideology and economics’ abstractions. I keep running across remarks in economics blogs favourable to the orthodoxy, to the effect that of course the stuff in elementary textbooks is nonsense; once you get onto the advanced stuff, you get taught all the qualifications that makes the stuff make sense. I’m both mortified and enraged by these comments. My feeling is a) they’re probably right – when I criticise economics textbooks, I’m criticising stuff that no economist actually believes; but also b) they’re probably wrong, because if economists didn’t at some level believe this stuff, it wouldn’t be in the textbooks. My feeling is that the simplistic, wrong stuff in the textbooks serves as a useful ideological base to which economists can retreat when threatened by reality; but that when the simplistic stuff is criticised, economists can return to their subtleties and qualifications, and wash their hands of the elementary textbooks’ falsehoods. But I just don’t know enough to be sure what’s going on.

    2) But w/r/t the individual labour supply curve – I just don’t see how this can be explained away by saying that it’s simply a tool for teaching marginal analysis. (Not that I’m suggesting that you’re trying to explain it away like this. 🙂 ) Surely the fact that the individual labour supply curve in my textbooks slopes upwards at low income levels is simple an empirical claim – and a false one (?). I don’t see how this issue comes down to degrees of abstraction.

    Could be wrong; typing very fast; may not even make sense. Thanx for your comment!

    [Also, in belated response to wanderer7 above, surely the lowest wages are earned not by school drop-outs, but by people who never went to school in the first place – principally, people in the third world…]

    Comment by praxisblog — March 4, 2008 @ 8:25 pm

  5. well, economists believe those things to some degree – I recently read some rather successful economists agree with a statement to the effect of “Most of the things we actually need to use we learned in Mankiw” (as in N. Gregory Mankiw, who wrote the dominant intro econ textbook used in the U.S.). But what they believe in is the principle of the way these work on the margin, and that the model works close to perfectly under certain definite but rare real-world circumstances.

    The labor supply curve you’re talking about works perfectly well under certain conditions where the security of the wage-earner is assured. BUT of course in the vast majority of conditions in the world as it is such security is not assured. Yet, welfare or unemployment actually makes the curve conform more effectively to the theoretical model. There’s an interesting discussion of S-shaped labor supply curves at this blog post that might be a good addendum to you (curiously unfootnoted) orthodox economic text. My guess is that the qualifiers – e.g., the income and substitution effects – will be brought up in the next chapter of your textbook, when the authors discuss the labor-leisure tradeoff and say something like “we were kind of lying to you just then. sorry.”

    Comment by Dave — March 6, 2008 @ 12:29 am

  6. Dave – sorry it’s taken me an age to reply. Thank you for the link. The [inverted]S-shaped labour supply curve in that post does indeed make a whole lot more sense. I’m still a little baffled as to why my textbook wouldn’t just start from this curve, but at least that post has put my mind at rest to some extent. [BTW – my textbook is the European edition (rewritten by Kathryn Graddy) of Paul Krugman and Robin Wells’ ‘Economics’. I didn’t mean to be obscure – sorry. I’ve added the proper reference to the post.] (Krugman et al don’t own up to the lie, by the way – at least not in the next two hundred pages or so, which is how far I’ve gotten.)

    Re: Mankiw – I haven’t read his textbook but I sometimes read his blog. I discuss one of his posts on intergenerational social mobility here.

    Thanks again for your comments. It’s incredibly helpful for me to be corrected and educated by people who actually know what they’re talking about. 🙂 Take care.

    Comment by praxisblog — March 15, 2008 @ 7:08 pm

  7. When Krugman says that at lower wages the substitution effect is greater than the income effect, he means low wages, not super low wages.

    You are right that if someone was making 20 cents an hour and you cut their wage to 10 cents, they would probably work more hours (to the extent that they could choose their hours), to eat better.

    But by low wages, Krugman was more talking about say a U.S. Taco Bell worker making $8.00/hour. If you raised his wage to $12/hour he would likely want to be scheduled for more hours, say 45/week instead of 40. But if you raised his wage to $1 million/hour, then you can really see the income effect kick in and get big. He would want to go from 40 hours/week to something like 40 hours/year!

    Comment by Richard H. Serlin — May 15, 2008 @ 2:52 am

  8. Thanks Richard. Dave, above, drew my attention to the ‘S-shaped’ labour supply curve, which ended my bafflement. So I now see where Krugman’s coming from (essentially – he’s ignoring poverty…) I ought to do some more reading about this – I’d be interested to see some actual data on the relation between wages and working hours. But at least the immediate mystery’s solved…

    Comment by praxisblog — May 15, 2008 @ 6:53 pm

  9. You are a college student so this explanation will probably work. While you are in school you can choose to pick up a part time job. If the best paying job you can find pays $5.00 an hour you might say the heck with it and solely focus on your studies. If the best paying job pays $40 an hour you might quit school and work full time. After working full time at the $40 per hour job they decide to give you a raise so that you make $60 an hour but they want you to work 50 hours a week. You are like ok that is a lot of money I’ll take it. Now they decide to give you another raise and pay you $200 an hour but you half to work 60 hours a week. You are like wow that is a lot of money I will take it. Then they decide to give you a raise to $400 dollars an hour but want you to work 80 hours a week. You think that is a lot of money but I will be at work the whole time and cannot spend any of it. You turn down their offer and find another job. That is the backward sloping labor supply curve at work. P.S. I agree with Dave about the poverty thing. Remember though in the U.S. we have Welfare and minimum wage so the poverty idea isn’t so relevent.

    Comment by jeffrey — October 16, 2010 @ 11:48 am

  10. ‘praxisblog’ here (the post author). Jeffrey – I in fact understand the logic behind the backward sloping labour supply curve. My objection to the Krugman et al. passage was that it is empirically wrong for lots and lots and lots of low wage earners. That’s correct – Dave, above, says so too, when he writes that Krugman et al’s statement does not apply “in the vast majority of conditions in the world as it is”. When you say “I agree with Dave about the poverty thing”, you’re saying that you agree with what I’m saying in the post.

    As a separate point, I’m really profoundly sceptical that the U.S. welfare state and minimum wage provide anything like enough of a income safety net to invalidate my objection to the broad validity of Krugman et al’s claim, even for the U.S. I don’t have time to follow this up now, though.

    Comment by duncan — October 17, 2010 @ 2:34 am

  11. While you are in school you can choose to pick up a part time job. If the best paying job you can find pays $5.00 an hour you might say the heck with it and solely focus on your studies

    And while I am in school how am I paying for my schooling, my rent, my food, my transport costs? Has the money just fallen from heaven? In fact some people who have part-time (or full-time) jobs while they study do so in order to afford schooling – and in order to pay rent, in order to buy food. This is the point. A large percentage of the population has to work to afford very basic things. If this is your situation, as it is for many many people, the idea that ‘substituion effects’ dominate seems prima facie extremely misguided: one needs work; one gets the best work one can; if the only work one can get is low-paid, one needs to work a lot in order to get the money one needs to pay for the goods one needs. The example you use to bolster Krugman et al’s description of labour-supply curves is inadequate because it assumes a position of privilege. (It assumes I can comfortably devote myself full time to my studies without having to earn any income). If you have some actual data about labour supply curves I’d be interested to look at it – but seriously, your example doesn’t speak to my objection in the slightest.

    And I’m not a college student. So that’s one thing you’re definitely just wrong about, empirically.

    Comment by duncan — October 17, 2010 @ 2:47 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Create a free website or blog at

%d bloggers like this: