• King-Plosser-Rebelo preferences satisfy balanced growth requirements, we have that income and substitution effects of labor cancel. Labor does not respond to a change in the wage level.
  • Greenwood-Hercowitz-Huffman preferences shut off the wealth channel: There is only a substitution effect left, hence labor co-moves with wages.
  • Jaimovich-Rebelo create the nested case with a wealth effect within the two aforementioned.

Is there any commonly used preference specification for which the wealth effect dominates the substitution effect, and hence labor supply reduces as a response to an increase in wages?

  • $\begingroup$ Just realized that I need a clarification, to be on the safe side: are you looking for a comparative static result in a given time period, i.e. something like $dL/dw <0$, ($L$ is labor) or for a result in rates of change through time, i.e. something like $\dot w > 0 \implies \dot L <0$? $\endgroup$ Commented Sep 28, 2015 at 17:52
  • $\begingroup$ @AlecosPapadopoulos I'm looking for comparative statics. $\endgroup$
    – FooBar
    Commented Sep 28, 2015 at 19:10
  • $\begingroup$ Please specify whether you are looking for a dominating wealth effect locally or for all budget constraints and what these budget constraints include (e.g. capital income). $\endgroup$
    – HRSE
    Commented Sep 30, 2015 at 1:33
  • $\begingroup$ @HREcon I'm open to local and global domination. Regarding budget constraints, I prefer the case with labor income only. $\endgroup$
    – FooBar
    Commented Sep 30, 2015 at 5:23

3 Answers 3


Note first that if an individual generally reacts to increases in the wage by a decrease in labor supply, then the individual must have maximal labor supply at a wage of 0. I will supply such an example but the example can be adjusted to a more realistic setting.

Let's start with the consumer problem. Suppose there is no capital income, such that $c=w\cdot l$ where $w$ is the real wage. If we think of a reason why labor supply is decreasing in income, then one idea is that the higher the consumption, the larger the disutility from labor. (This is actually not that far fetched, think of how lovely leisure can be with a yacht.)

So let's take a standard quasilinear utility function and adjust it to $u=\ln c+(1-l\cdot c)$. Inserting the budget constraint yields: $u=\ln (w \cdot l) + (1-w\cdot l^2)$. As the function is concave in $l$, first order conditions are necessary and sufficient for a maximum: $$1/l - 2wl = 0$$ from which follows $l=(1/(2w))^{.5}$ and labor supply is decreasing in the wage. We could generalize this function by adding an offset to $c$ and $l$ in order to create the income effect dominance only after a certain wage rate.

  • $\begingroup$ (+1). Indeed, separable utility in consumption and leisure, is totally unrealistic. The fact that "leisure" does not mean "not working and staying still" but engaging in enjoyable activities (and this always implies consumption even if only of calories and a bit of wear-off of durable goods), is a good starting point to rethink the modelling of utility from consumption and leisure. $\endgroup$ Commented Sep 29, 2015 at 17:11

In open macroeconomic models some cases of logarithmic preferences cause the wealth effect to dominate the intertemporal substitution effect.

If I recall correctly, at least Vegh's book "Open Economy Macroeconomics in Developing Countries" (2013, MIT Press) presents this case.


I think it was in one of Joan Robinson's books where a not just backward-bending but "mirrored S"-shaped labour supply was presented, referring to agricultural societies:

enter image description here

The rationale of course is that if your productivity is very low, you have to work very long and exhausting hours to even survive. As your productivity raises, what you want is not to work even more for the extra income, but thankfully to get some rest.
The empirical regularity is that most of workers in the world are away from the upper "backward" movement of labor supply - but in poor countries, they may very well be in the lower backward part. And I think this resonates with the information presented in @JohnL. answer. So chances are, you will find such preferences modeled in studies that focus on poor/developing economies.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.