I'm reading Blanchard's Macroeconomics, and on page 206 of the 5th edition, he writes: «Let $u_t$ denote the unemployment rate in year t, and $u_{t-1}$ the unemployment rate in year t-1, and $g_{yt}$ the growth rate of output from year t-1 to year t. Then under these two conditions ($Y_t=N_t$ i.e. Output=Employment, and labour force $L_t=L$ constant), the following relation would hold: $$ u_t-u_{t-1}=-g_{yt}$$

But what I get is $\frac{Y_{t-1}-Y_t}{L}=-\frac{Y_t-Y_{t-1}}{Y_{t-1}}$ which is not true... where did I go wrong, or is there a typo?

Any help would be appreciated.

  • $\begingroup$ It is hard to help because in its current form the question is not self contained, i.e. not all variables are defined nor are all equations given. $\endgroup$
    – Giskard
    Commented Aug 29, 2015 at 22:30
  • $\begingroup$ @denesp thanks for the help. Could you please tell me what's missing ? $\endgroup$ Commented Aug 29, 2015 at 22:38
  • $\begingroup$ @denesp or what I should assume to reach that conclusion? $\endgroup$ Commented Aug 29, 2015 at 22:49
  • $\begingroup$ For example, what is $N$? $\endgroup$
    – FooBar
    Commented Aug 30, 2015 at 8:27
  • $\begingroup$ @FooBar it's employment $\endgroup$ Commented Aug 30, 2015 at 9:00

2 Answers 2


I get the same equation as you do. So you are correct. But tough usually $u_t - u_{t-1} \neq -g_{y_{t}}$, for small values of $g_{y_{t}}$ $$ u_t - u_{t-1} \approx - g_{y_{t}} $$ does hold. In macroeconomics these kind of (in my opinion annoying) approximations are used every now and then. The author perhaps neglected to mention that he was using an approximation. For example the Fisher equation is sometimes written as $i = r + \pi$, even though it is actually $i \approx r + \pi$ and that only for small values of $r,\pi$.

  • $\begingroup$ Yes, I think you're right. These implicit approximations usually written as equalities bother me a lot! Thanks for the help ;) $\endgroup$ Commented Aug 30, 2015 at 12:37

Under the stated assumptions, we have that (using $N_t = Y_t$ and $L_t = L$)

$$u_t - u_{t-1} = \frac {L-Y_t}{L} - \frac {L-Y_{t-1}}{L} = \frac {Y_{t-1} - Y_t}{L}$$

The "growth rate of output" is defined as

$$g_{yt} \equiv \frac{Y_t-Y_{t-1}}{Y_{t-1}}$$

So we can manipulate the first equation as

$$u_t - u_{t-1} = -\frac {Y_t-Y_{t-1}}{L} \frac{Y_{t-1}}{Y_{t-1}} = -g_{yt}\cdot (1-u_{t-1})$$

$$\implies u_t - u_{t-1} = -g_{yt} + g_{yt}\cdot u_{t-1}$$

So we do have an approximation. Such are fine for theoretical expositions (educational or not), but whether they should be acceptable in real-world studies depends on the actual magnitudes of the two rates involved. For example with a growth rate of $3\%$ and a previous-period unemployment rate of $5\%$ we would get an approximation error of $0.0015$. The true drop in unemployment would be over-estimated ($3$ percentage points instead of $2.85$ percentage points), which appears acceptable.

  • $\begingroup$ Alecos, I'm not sure I understand you. How can the right-side be equal? $\endgroup$ Commented Aug 30, 2015 at 22:57
  • $\begingroup$ Thanks for the more elaborate exposition. p.s.: in these last few days I asked a few questions. If you find them interesting, could you help me? Either way, thanks for this answer. ;) $\endgroup$ Commented Aug 31, 2015 at 7:51

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.