Let $L=E+U$ where $L$ is labour force, $E$ is number of employed, and $U$ is unemployed people.

Let $u = \frac{U}{L}$ and $v = \frac{V}{L}$.

Given $m(u,v)$ as a matching function that determines the flow of matches between workers and firms, and assume $m$ has constant return to scale (CRS) in the two arguments $u$ and $v$.

The elasticity of the arrival rate for workers is given by $$\theta q(\theta),$$ where $\theta = \frac{v}{u}$ represents "tightness" in the labour market.

Elasticity ($\epsilon$) = $$\frac{\mathrm{d} f(x)}{\mathrm{d} x} \frac{x}{f(x)} = \frac{\partial m}{\partial \theta} \frac{\theta}{m(1,\theta)},$$

where $\theta q(\theta) = m(\theta^{-1},1) = m(1,\theta)$ (by using CRS property of $m$ in the two arguments $u$,$v$).

The rate at which unemployed workers find job: $$\theta q(\theta) = m(\theta^{-1},1)=m(1,\theta),$$ since the matching function, $m$, is constant returns to scale in the each argument.

Since $\theta q(\theta) = m(1,\theta)$, $$\frac{\partial m(1,\theta)}{\partial \theta} = q(\theta) + \theta q'(\theta).$$

So the elasticity of the arrival rate for workers, $$\epsilon = \frac{\partial m}{\partial \theta} \frac{\theta}{m} = \left [q(\theta) + \theta q'(\theta) \right ] \frac{\theta}{\theta q(\theta)} = 1 + \frac{\theta q'(\theta)}{q(\theta)}$$.

Since $q'(\theta)<0$, $\epsilon<1$, but how can I formally show that it is between $0$ and $1$?

  • $\begingroup$ What does the function $q(\cdot)$ represent? $\endgroup$
    – Giskard
    Feb 28, 2016 at 18:48
  • $\begingroup$ $q(\theta)$ is the rate at which jobs get filled. It is positive and decreasing in $\theta$. $\endgroup$
    – OGC
    Feb 28, 2016 at 18:59
  • $\begingroup$ @denesp So I'm not sure if $\theta>0$ I thought it can be $0$, but that's the extreme case. $\endgroup$
    – OGC
    Feb 28, 2016 at 19:02
  • 1
    $\begingroup$ That is not a problem. Given $\theta \geq 0$ and $q'(\theta) < 0$ even if $\theta = 0$ you have $$\epsilon = 1 + \frac{\theta q'(\theta)}{q(\theta)} \leq 1.$$ Basically you want to show $$\epsilon = 1 + \frac{\theta q'(\theta)}{q(\theta)} \geq 0.$$ For this I think you will need more information about $q$ as it does not hold for all functions. $\endgroup$
    – Giskard
    Feb 28, 2016 at 19:05
  • 1
    $\begingroup$ Afraid not. You need the specific form of either $q$ or $m$ or some additional properties. $\endgroup$
    – Giskard
    Feb 28, 2016 at 21:36

1 Answer 1


Let $m$ cobb-douglas. This means $m(u,v) = A u^\alpha v^{1-\alpha}$ for $\alpha < 1$ and $q(\theta) = A \theta^{-\alpha}$.


$$\frac{\theta q'(\theta)}{q(\theta)} = -\alpha > -1$$

To what extent this holds for other functional forms I'm not sure, but I haven't seen anything but Cobb-Douglas being used as the matching function.

  • 1
    $\begingroup$ Got it! Works perfectly for this special case, i.e, $\epsilon > 0$. However, not so clear if $q$ is general. $\endgroup$
    – OGC
    Feb 28, 2016 at 20:00

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.