Solution to Tutorial 2

March 9, 2022

1 Question 1
1.1 (a)
First, we know from class that the optimal leisure choice is either l∗ = wT2w
+vi
or l∗ = T , depending if we
have or not a corner solution. As we have seen in class, the reservation wage is the one that makes workers
indifferent between working or not. At that wage, the indifference curve will be tangent to the budget
constraint exactly at l = T , making both solutions coincide. Thus,

wiR T + vi
=T
2wiR

so solving for wiR we obtain

vi
wiR =
T
This means that the reservation wage is increasing in the person’s non-labor income, and decreasing on the
total amount of time.

1.2 (b)
The labor force participation rate (denoted by P ) is defined as the fraction of the population that either
works or is looking for a job (i.e. the unemployed). In this economy, there is no such a thing as “looking for
a job” (i.e. unemployment is zero) so all we have to do is to compute the fraction of individuals working.
Individuals work if their reservation wage is smaller than the market wage w. That is, if

wiR ≤ w
⇐⇒
vi ≤ wT

Recall that vi (non-labor income) is distributed Uniformly on support [0, v] in the population. Therefore, a
“mass” wT of workers are working, and a mass v − wT is not working, which implies that P = wT v . You can
see this graphically:1
1 On 1
the graph f (v) = v
denotes the density of v in the population.

1
1.3 (c)
As seen in Lecture 1, total hours per person (THP) equals to P (1 − u) h, where h denotes the average hours
per worker. Since u = 0 in this economy we have that T HP = P h . Recall that a working individual with
non-labor income v works for this amount of hours:
wT + v T v
h (v) = T − l (v) = T − = −
2w 2 2w
So all is left is to compute the average hours across working individuals. The working individuals non-labor
income, vi , are distributed Uniformly on support [0, wT ]. The density function of the Uniform distribution
in support [a, b] is f (x) = b−a1
, where f denotes the density function. In our case that is f (v) = wT 1
.
Recall that the average (expected value) of a function γ (x) of random variable x (with distribution f (x)) is
computed as
Z b
E (γ (x)) = γ (x) f (x) dx
a

In our case, we are re-labeling x with v, f (v) = 1

wT , and γ (v) = T
2 − 2w .
v
Applying the formula we obtain2
ib
αxβ+1 αbβ+1 αaβ+1
Rb h
2 Here we used the fact that αxβ dx = = − .
a β+1 a β+1 β+1

2
 Z wT  
T v 1
h= − dv
0 2 2w wT
Z wT  
1 v
= − dv
0 2w 2w2 T
" 2
#wT
v
v
= − 22
2w 2w T
0
(wT )2
wT 2 T T T
= − = − =
2w 2w2 T 2 4 4
Finally, we obtain
wT T
T HP = P h =
v 4

1.4 (d)
To answer this question we compute dw , dw
dP dh
and dT HP
dw :

dP T
=
dw v
dh
=0
dw
dT HP T2
=
dw 4v
Therefore, an increase in the wages increases labor force participation, and leaves hours per worker un-
changed. The combined effect is an increase in total hours per person, all through the extensive margin.
Note that the fact that dw
dh
= 0 (average hours remain unchanged) is somewhat puzzling. For an individual
who was already working his/her hours increase because dh(v)dw = 2w2 > 0. However, the increase in wage
v

induces “high” v workers to start working, and they work fewer hours. This compensates the fact that the
existing workers are working more, causing average hours to remain unchanged.

2 Question 2
2.1 (a)
β captures the degree of impatience: how much one values tomorrow relative to today (in terms of consump-
tion and leisure). A lower value of β discount the future further, so it increases the degree of impatience.

2.2 (b)
•

maxc1 >0,c2 >0,l1 ∈[0,T ],l2 ∈[0,T ],s1 ∈R U (c1 , c2 , l1 , l2 )

s.t.
c1 + s1 ≤ w1 (T − l1 )
c2 ≤ w2 (T − l2 ) + (1 + r2 )s1 + G

To consolidate the budget constraints we first recognize that both constraints are satisfied with equality,
rearrange the second
c2 − w2 (T − l2 ) − G
s1 =
(1 + r2 )

3
 and plug into the first
c2 w2 (T − l2 ) + G
c1 + ≤ w1 (T − l1 ) +
(1 + r2 ) (1 + r2 )

2.3 (c)
We write a Lagrangean

maxc1 ,c2 ,l1 ,l2 ,λ L = log (c1 ) + log (l1 ) + β [log (c2 ) + log (l2 )]
 
w2 (T − l2 ) + G c2
+ λ w (T − l1 ) + − c1 −
(1 + r2 ) (1 + r2 )

Take FOC
• c1
1
−λ=0
c1
• c2
1 λ
β − =0
c2 (1 + r2 )

• l1
1
− λw1 = 0
l1
• l2
1 λw2
β − =0
l2 (1 + r2 )

If we combine the first two, plus β = 1

(1+r2 ) we obtain

1 1
=
c1 c2
⇐⇒
c1 = c2 = c

From the budget constraint we know that

c w2 (T − l2 ) + G
c+ = w1 (T − l1 ) +
1 + r2 (1 + r2 )
⇐⇒  
1 + r2 w2 (T − l2 ) + G
c = w1 (T − l1 ) +
2 + r2 (1 + r2 )

Now, from combining the FOC wrt c1 and l1

1 1
= w1
l1 c1
⇐⇒
c = l1 w 1

and the same for c2 and l2

c = l2 w 2

4
 Plugging back onto the BC we obtain
   
c
1 + r2 

c
 w2 T − w2 +G
c = w1 T − + 
2 + r2 w1 (1 + r2 )
   
1 + r2 w2 T + G 1 + r2 c
= w1 T + − c+
2 + r2 (1 + r2 ) 2 + r2 (1 + r2 )
  ⇐⇒  
1 + r2 c 1 + r2 w2 T + G
c+ c+ = w1 T +
2 + r2 (1 + r2 ) 2 + r2 (1 + r2 )
   
1 + r2 (2 + r2 )c 1 + r2 w2 T + G
c+ = w1 T +
2 + r2 (1 + r2 ) 2 + r2 (1 + r2 )
⇐⇒  
1 1 + r2 w2 T + G
c = w1 T +
2 2 + r2 (1 + r2 )
and finally
 
1 1 + r2 w2 T + G 1
l1 = w1 T +
2 2 + r2 (1 + r2 ) w1
 
1 1 + r2 w2 T + G 1
l2 = w1 T +
2 2 + r2 (1 + r2 ) w2

2.4 (d)
If we plug w1′ and w2′ onto the formulas
 
′ 1 1 + r2 w2 (1 + g) T + γw2 (1 + g) 1
l1 = w1 (1 + g) T + = l1
2 2 + r2 (1 + r2 ) w1 (1 + g)
 
′ 1 1 + r2 w2 (1 + g) T + γw2 (1 + g) 1
l2 = w1 (1 + g) T + = l2
2 2 + r2 (1 + r2 ) w2 (1 + g)
where the last equality is due to the fact that the terms (1 + g) get canceled out. In sum, nothing happens
to leisure (and labor supply) when the wages in both periods increase. This means that in the long run
income and substitution effects in this model exactly cancel out, leaving labor unchanged. This is peculiar
to log-utility as seen in class.
In terms of consumption. It’s clear that it is increasing in g, because income and substitution effects
w.r.t. consumption go on the same direction.
 
1 1 + r2 (1 + g) w2 T + γw2 (1 + g)
c′ = w1 (1 + g) T +
2 2 + r2 (1 + r2 )
(1 + g)c

2.5 (e)
If w1 = w2 = w the first thing to note is that l1 = l2 = l. Then, the long run labor supply curve is given by
the equation
h (w) = T −l
 
1 1 + r2 wT + γw 1
T− wT +
2 2 + r2 (1 + r2 ) w
 
1 1 + r2 T +γ
= T− T+
2 2 + r2 (1 + r2 )
i.e. it is vertical (on a graph with w on the y axis). In terms of how this relate to them empirical evidence
shown in class, perhaps this prediction is somewhat similar to labor supply in the USA over the past 50
years. However, in most developed countries h is decreasing as countries are getting richer.

5
2.6 (f )
The simplest way to answer this is to compute dl1
dw1
 
dl1 1 1 + r2 T 1 1 + r2 w2 T + G 1
= − w1 T +
dw1 2 2 + r2 w1 2 2 + r2 (1 + r2 ) [w1 ]2
" #
1 1 + r2 w2 T + G
= − <0
2 2 + r2 [w1 ]2 (1 + r2 )

i.e. leisure necessarily decreases, which implies that hours worked increase. Thus, in a boom labor goes up,
whereas in a recession scenario the response would be the opposite. This indicates that the model generates
pro-cyclical THP, which is consistent with the data.

2.7 (g)
The short run labor supply curve is given by the equation
 
1 1 + r2 wT + G 1
h (w1 ) = T − w1 T +
2 2 + r2 (1 + r2 ) w1

This yields an increasing function, just as the one in slide 22 of Lecture 2b. The reason why this model
generates flat labor supply in the long run and increasing in the short run has to do how income and
substitution effects operate. In the short run, an increase in wages only affect a small share of the worker’s
permanent income, so income effects are small. On the other hands substitution effects are at full force since
they only rely on contemporaneous wages. This is why substitution effects dominate in the short run but
not necessarily in the long-run.