# Cyclicality of Real Wages? Counter, pro or acyclical?

Is the real wage counter or pro- cyclical? I've read in some textbooks that the real wage is mildly pro-cyclical, but I've also found some papers stating that it's acyclical.

So, which is it? Also, if the real wage is pro-cyclical, does that mean that there's some sort of price rigidity, i.e., the economy is not always on the Price-Setting curve?

• Also, if the real wage is pro-cyclical, it means that there's some sort of price rigidity, i.e., the economy is not always on the Price-Setting curve. Does it? The standard RBC model generates strongly pro-cyclical wages without any price rigidities. – FooBar Jan 1 '16 at 21:55
• @FooBar I thought so, since, if the real wage is pro-cyclical, could mean wage setters have used their increased power (due to a increase in output), to increase their wages, but price setters haven't still caught up with the change in costs (wage), so, there would be some price rigidity. If I'm wrong, could you explain me why? thanks ;) – An old man in the sea. Jan 2 '16 at 10:05
• Sure. I've cleaned up the question a bit. You should post the removed part as a separate question if you're still interested, right now it was a bit broad and required too long answers. – FooBar Jan 2 '16 at 11:45

### Are Wages Procyclical?

Here's the entry from Table 1 in King and Rebelo (2000; Resuscitating Real Business Cycles). The variable of analysis is log-wages, detrended with HP filter. It has a std, relative std, first-order autocorrelation and correlation with output of 0.68, 0.38, 0.66, 0.12.

This leads the authors to conclude that

The real wage is much less volatile than output.

You can see that for their long US period, 0.12 is a positive number, but pretty close to zero. This is why different authors take this number as either acyclical or weakly procyclical.

### Price rigidity required for procyclical wages?

Assume that some sort of fluctuation in TFP "drives" the business cycle. Furthermore, if markets are competitive, the wage rate should be set by the marginal product:

$$w = AF_L(K, L)$$

If $L$ stays constant over the business cycle (e.g. with inelastic labor supply), this setup will generate wages $w$ that vary with $A$, and hence any change in $A$ will result in a one-for-one change in $w$ and in $Y$: $Y$ and $w$ are perfectly correlated, while $L$ is fixed.

This is pretty much the prediction of the RBC model, and it is the opposite of what we observe in the data: $L$ is volatile, while $w$ is fixed. If you'd assume that $L$ responds strongly to changes in $A$, $w$ might be pretty constant over the business cycle. This is what we observe in reality.

And most of the labor-side of RBC research is to come up with ways that force $L$ to respond so much.

King and Rebelo (2000) should be a key paper of any introductory reading list for macroeconomics and I strongly recommend reading it.

### The case of less competition

If there is literally zero competition, then firms will at all times only pay the household his outside-option value, making him indifferent between working and not working. If that outside option is constant over the business cycle (for which we have to some extent theoretical and empirical arguments), wages will be acyclical.

However, to the extend that there is no perfect monopsony, firms have to compete for workers and offer higher wages whenever workers are worth more. Hence, wages will be procyclical even if, to some extent, firms are not perfectly competitive on the labor side. The more competitive they are, the more pro-cyclical wages should be.

### Alternatives and extensions to the neoclassical model

The above model was just a simple one to generate a case in which pro-cyclicality of wages can be generated without price stickiness. There is much more to the subject than an answer here can provide. Here's some more references:

• Barro (1977): Long-term contracting, sticky prices, and monetary policy. *Wages are long-term contracts, hence if firms are less risk-averse than workers, firms will insure workers against wage changes over the business cycle and we can observe (optimally) constant wages
• Beadry and DiNardo (1991) is a follow-up on that discussion
• Haefke et al (2013, JME) is also a very recent follow-up
• Bewley (1999): Why wages don't fall in a recession
• Shimer and Hall (two separate papers, both in the same 2005, AER) are the standard references on whether the more modern and reasonable frictional DMP model can match labor (and wage) volatility over the business cycle.
• Thanks for the answer. I've already asked the other part of the question. Check it out if you're interested. ;) – An old man in the sea. Jan 2 '16 at 18:21
• Also thanks for the paper. I'll definitely read it as soon as possible. ;) – An old man in the sea. Jan 2 '16 at 18:24
• FooBar, would your conclusion, or a similar one, still be valid if we were no longer in a perfect competition setting? – An old man in the sea. Jan 4 '16 at 11:54
• @Anoldmaninthesea. I've added some more text and references – FooBar Jan 4 '16 at 12:13
• Foo, Many thanks for your very complete answer. ;) – An old man in the sea. Jan 4 '16 at 12:19

It really depends on what model you use.

Classical and Keynesian economists believe that real wage is counter-cyclical. In Lucas' New Classical model, he believes real wage is acylical. In the Real Business Cycle model, economists say that real wage is procyclical after WW2, in America.

The general consensus now is that

The most recent evidence shows that the cyclical behaviour of real wages has changed, from being countercyclical during the interwar period, to procyclical in the postwar era.

This article gives a good rundown of a bunch of the literature.

Intuitively, what do you think? If the economy is doing well, and companies see positive real demand increases, the demand for labor is higher, and you'd expect real wage to increase. Then again, during recessions, labor hoarding may prevent an immediate fall in wages or layoffs (companies do not like to give wage cuts, because of morale, and they do not like to layoff right away in case the recession ends quickly).

• An empirical phenomenon should be independent of the model that you believe to be true. – FooBar Jan 1 '16 at 20:32
• Of course. The model should fit reality, not the other way around. My impression was empirically that wages were procyclical, so that's what I argued for at the end. – Kitsune Cavalry Jan 1 '16 at 21:04
• As curiosity, what's the rationale for the Keynesians to say that the wage is counter-cyclical? thanks for your answer ;) – An old man in the sea. Jan 4 '16 at 12:21
• I'm currently uncertain. – Kitsune Cavalry Jan 4 '16 at 19:04