Unemployment and Efficiency Wages Prof. Eric Sims University of Notre Dame
Fall 2010
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
1 / 15
Unemployment
In models up until now, there have been two ways for households to use their time: work and leisure Choice to “consume” leisure is completely voluntary In reality there is a third state, called unemployment A person is unemployed if they would like to work but can’t find work. In this sense it is involuntary To get unemployment in a model you need some kind of a friction that prevents labor market from clearing
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
2 / 15
Terminology and Definitions
n: number of people employed and working UN: number of people looking for work but not working L = n + UN: labor force (all people either working or looking for work) NL: people not in labor force (not working or looking for work) POP = n + UN + NL: total population u=
UN n+UN :
unemployment rate
participation =
Sims (ND)
L POP
Unemployment and Efficiency Wages
Fall 2010
3 / 15
Unemployment Rate in the US Unemployment Rate 11 10 9 8 7 6 5 4 3 2 50
Sims (ND)
55
60
65
70
75
80
85
Unemployment and Efficiency Wages
90
95
00
05
Fall 2010
4 / 15
Basic Facts
Mean unemployment rate: 5.6 percent Standard deviation: 1.5 percent Countercyclical: correlation with HP filtered output is -0.6
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
5 / 15
Equilibrium and Unemployment
There is no unemployment in our equilibrium neoclassical model There are fluctuations in employment, but there is no involuntary non-employment. People choose to not work at a given real wage To get unemployment as it is defined here we have to think seriously about frictions in the labor market We’re going to study efficiency wages. Essentially will fix the real wage at a given level and allow for disequilibrium in the labor market.
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
6 / 15
Efficiency Wages
Assume now that firm has some market power. Instead of being a pure price-taker, it can choose the real wage it pays workers Assume further that there are two components to the firm’s labor input: time spent working (n) and “effort”, e. Total labor input = en Assume that effort is some increasing function of the real wage: e = e (w ), e 0 (w ) > 0 Why? Moral hazard: firm can’t perfectly monitor what workers do during the day. The higher is the wage, the more costly it is to “get caught” slacking off at work (i.e. you get fired). Hence effort increasing the real wage
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
7 / 15
Firm optimization
Firm production function is now: y = zf (k, e (w )n ). Objective is to maximize PDV of profits subject to capital accumulation. Now can choose w : maxn,n0 ,I ,k 0 ,w zf (k, e (w )n ) − wn − I 1 + z 0 f (k 0 , e (w 0 )n 0 ) − w 0 n 0 + (1 − d )k 0 1+r s.t. k 0 = I + (1 − d )k
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
8 / 15
Characterizing Solution Just focus on FOCs with respect to n and w : n : e (w )zfn (k, e (w )n ) = w w : e 0 (w )zfn (k, e (w )n ) = 1 Both just MB = MC conditions Combine the two to get: e 0 (w )w =1 e (w ) This implicitly defines an optimal wage rate, w ∗ , that is independent of everything else in the model except the functional form of e (w ) Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
9 / 15
The Labor Market
Call the optimal real wage from the above condition w ∗ . This is now an exogenous variable Assume that this is greater than the real wage that would clear the labor market With the real wage “too high”, quantity of labor demanded will be less than the supply People who want to work at that real wage but don’t get hired (firm won’t go off its labor demand curve) ⇒ unemployment
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
10 / 15
The Labor Market Graphically ns
w u w*
nd
n0
Sims (ND)
n0s
Unemployment and Efficiency Wages
n
Fall 2010
11 / 15
The y s curve
Since we’re now long on the ns curve, the y s curve will be different Modified definition: y s curve is the set of (r , y ) pairs where the firm is on its labor demand curve and the production function holds (i.e. y = zf (k, e (w )n ) Labor demand unaffected by r ⇒ y s is vertical Mathematically, y = zf (k, e (w ∗ )n (w ∗ ) What will shift it? Only z, k, or w ∗ y d and money market same as before
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
12 / 15
The Efficiency Wage Economy ys ns
w
r
u w* r0
nd
n0
Sims (ND)
n0s
n
Unemployment and Efficiency Wages
yd
y0
y
Fall 2010
13 / 15
Comparative Static Exercises
Changes in z 0 , g , and g 0 will have no effect on output since supply is vertical. Will only affect real interest rate, composition of output, and unemployment What about change in z?
↑ z ⇒ increases in y , c, and I . r down. n up, u down. p down Very similar to features of the data, except that real wage is not procyclical
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
14 / 15
Implications for Policy
In RBC theory, there is no welfare justification for active use of economic policy Here there might be In general, policymakers would want to achieve the level of output that obtains in the base neoclassical model. Why? Because this is the “efficient” level of output The only way to do that is to try to lower w ∗ . w ∗ too high is the source of the inefficiency. How to do this? Does not make sense here to use “countercyclical” fiscal policy
Sims (ND)
Unemployment and Efficiency Wages
Fall 2010
15 / 15