The Real Business Cycle model

Spring 2013

1 Historical introduction
• Modern business cycle theory really got started with Great Depression

• Keynes: “The General Theory of Employment, Interest and Money”

• Keynesian economics and the study of aggregate demand

• You may have seen something like an Aggregate Demand / Aggregate Supply model in ECON
1B

• Dissatisfaction with the Keynesian model

– Practical: “stagflation” of the 1970s

– Theoretical: where are the “microfoundations”?
– Theoretical: are the model predictions robust to policy changes (the Lucas critique)?
– Theoretical: model not so useful for normative questions since there are no utility
functions

• “Real business cycle model” tried to remedy these deficiencies, especially in terms of “micro-
foundations”

• “New-Keynesian” models blend some of the logic of the Keynesian model with the from-first-
principles approach of the RBC

• We’ll study New Keynesian model later in the course

Key questions
• Positive: what causes economic fluctuations?

• Normative: what should we do about it?

1
ECON 52, Spring 2013 2 THE RBC MODEL

2 The RBC model

• RBC = Neoclassical model plus “shocks”

• Are market economies fundamentally unstable?

– RBC approach: no, the internal mechanisms are stable but there are exogenous distur-
bances (“shocks”) that move things around
– Maybe the existence of shocks is a fundamental feature of a market economy?
– Do we lose understanding if we model the shocks as exogenous?

• RBC is a “Real” model as opposed to “nominal”: no role for money or nominal variables

• Since RBC is a version of the Neoclassical model, First Welfare Theorem holds. What should
we do about business cycles?

A two-period RBC model

• Special case of the neoclassical “general equilibrium” model that we have studied but with
just two periods:

– t = 0 represents the present

– t = 1 represents the future

• Pros and cons of doing 2 periods instead of ∞:

– Pro: easier to solve and interpret

– Con: we need to adjust the model a little bit for the fact that the world ends after
t = 1, eg. capital fully depreciates after t = 1.
– If we want to put numbers of the model we need to be careful: 2 6= ∞!

• Households

max u (c0 ) + v (l0 ) + β [u (c1 ) + v (l1 )]

c0 ,l0 ,a1 ,c1 ,l1

s.t.
a1 = w0 (1 − l0 ) − c0 + (1 + r0 ) a0
c1 = w1 (1 − l1 ) + (1 + r1 ) a1
a0 given

2
ECON 52, Spring 2013 2 THE RBC MODEL

• FOC: Euler equation and consumption-leisure equation

u0 (c0 ) = β (1 + r1 ) u0 (c1 ) (1)

v 0 (lt ) = wt u0 (ct ) (2)

• Firms (static problem)

rtK = FK (Kt , Lt ) (3)

wt = FL (Kt , Lt ) (4)

• Arbitrage between lending and holding capital

r0 = r0K − δ
r1 = r1K − 1 (because the world ends after t = 1) (5)

• Market clearing

Lt = 1 − lt
K 1 = a1

• Equilibrium conditions:

u0 (c0 ) = βFK (K1 , L1 ) u0 (c1 )

v 0 (lt ) = FL (Kt , Lt ) u0 (ct ) (for both t = 0 and t = 1)
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y 1 (because the world ends after t = 1)
Lt = 1 − lt (for both t = 0 and t = 1)

• Now we have 7 equations in 7 unknowns: c0 , c1 , l0 , l1 , K1 , L0 and L1

• Replacing Lt = 1 − lt , we have 5 equations in 5 unknowns: c0 , c1 , l0 , l1 , K1 :

u0 (c0 ) = βFK (K1 , L1 ) u0 (c1 )

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y 1

3
ECON 52, Spring 2013 2 THE RBC MODEL

• The Williamson textbook (ch. 10) tries to do a graphical version of this system of equations,
which I find hard to follow

Using the model to talk about business cycles

• Exogenous variables in the model

1. The production function.

For instance, if
F (Kt , Lt ) = At Ktα Lt1−α

the parameter At is exogenous

Also, if the production function is

F (Kt , Lt ) = At Ktα L1−α

t + et

(i.e. it has an endowment component), then et is exogenous.

2. Intertemporal preferences, measured by β
3. Preferences for consumption/leisure. For instance, if

θ 1+
v (l) = l 
1+

the parameter θ is exogenous

4. The initial capital stock K0

• Steps in using the model to think of business cycles:

1. Define some basic values for the exogenous variables and call that “normal” (e.g.: middle
of the business cycle)
2. Compute the values for the endogenous variables: c0 , c1 , l0 , l1 , K1
3. Suppose an exogenous “shock” moves one or more of the exogenous variables
4. Compute the values for the endogenous variables: c0 , c1 , l0 , l1 , K1 in response to the
shock
5. Ask: does this look like what we see in business cycles?

• In particular, we should check whether in the model it’s true that output, employment,
consumption, investment and productivity more or less move in the same direction when
there is a shock.

4
ECON 52, Spring 2013 3 PRODUCTIVITY SHOCKS

3 Productivity shocks
• Consider a shock to A0 in the production function

F (K0 , L0 ) = A0 K0α L01−α

• An improvement in technology, that lasts only for today.

• In the equations:

u0 (c0 ) = βFK (K1 , L1 ) u0 (c1 )

v 0 (1 − L0 ) =↑ FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) u0 (c1 )
K1 =↑ Y0 − c0 + (1 − δ) K0
c1 = Y1

• Shock improves the marginal product of labour, which raise wages (other things being equal)

• Graphically, you can see this as an increase is labour demand (but remember that other
things change as well)

• Adjust by ↑ c0 , ↑ L0 , or both

• Economically, higher wages induce:

– income effect - why?

– substitution effect - why?

• Does it matter if the shock is temporary or permanent?

• Suppose:

– Shock is temporary
– Income effect exists but is small
– Substitution effect is strong

• Under these assumptions

– c0 goes up a little bit

– L0 goes up quite a lot

5
ECON 52, Spring 2013 3 PRODUCTIVITY SHOCKS

– Y1 goes up for two reasons: direct effect of shock on productivity and additional labour
input
– K1 goes up (you can see this from resource constraint) - households smooth consumption
by building capital for tomorrow

Interpretation and implications

• Success! The model gives you a simultaneous rise in

– output
– productivity
– hours worked
– consumption
– investment

• Booms are times when households choose to work more in response to temporarily high
wages

• Symmetrically, recessions are times when households choose to work less in response to
temporarily low wages

• The model can “explain” a large fraction of the fluctuations in GDP

1. Construct an infinite-horizon version of the RBC model

2. Do a growth accounting exercise with data from the US economy and find the Solow
residuals for each period (usually one quarter)

Solow Residual ≡gY − αgK + (1 − α) gL

3. Assume that the Solow residuals are an accurate measure of a series of exogenous
technological shocks
4. Simulate the response of the model when you ask it to respond to the shocks the you
estimated from the data
5. Compare GDP, etc. in the model to the same variables in the data

2
• Result: the model generates (explains?) about 3
of the fluctuations in GDP that we observe
in the data.

6
ECON 52, Spring 2013 3 PRODUCTIVITY SHOCKS

Implications for policy

• The government shouldn’t do anything about the business cycle

• By the 1st Welfare Theorem, the allocation is efficient

• If people aren’t working because there is a recession, this is an efficient response to the fact
that productivity is low today: people take time off when productivity is low and work harder
when productivity is high

• This is an assumption of the model, not a conclusion

• To be precise: competitive markets are an assumption of the model and whenever we have
competitive markets the FWT holds

• BUT the model does say that it is possible to observe fluctuations in GDP, etc. in a world
where there is nothing “wrong” (i.e. no inefficiency)

• Observing fluctuations does not prove that there is something wrong that should be fixed

• (But it’s also possible that this is the wrong model and in the right model business cycles
are indeed inefficient)

• The only thing the government should do about a recession is to try to improve productivity

– Encourage competition
– Lower transaction costs
– Promote innovation
– All of these things are desirable anyway, not just when there is a recession

Criticisms and limitations

• Where are the technology shocks?

• Moreover, where are the negative technology shock?

• Elasticity of labour supply

– Remember we talked about this being a very important number when we talked about
the effects of taxes
– A high elasticity of labour supply is an important ingredient of BOTH:
∗ the argument that taxes have large effects

7
ECON 52, Spring 2013 4 OTHER SHOCKS

∗ the RBC model

∗ (which are views often held by the same people)

• Cyclicality of wages. In general, the model does better on quantities than on prices.

• No unemployment?

• Measurement of TFP with unobserved capacity utilization

• Measured productivity has become less cyclical than it used to be

4 Other shocks
Impatience shocks: sudden decreases in β
• Has some appeal because it means households want goods now rather than later, an effect
that is important in Keynesian models, as we’ll see later

• In the equations:

u0 (c0 ) =↓ βFK (K1 , L1 ) u0 (c1 )

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y 1

• Adjust by ↑ c0 , ↓ c1 ↑ FK (which means lower investment), or all at the same time

• If ↑ c0 , this implies ↓ L0 !

• Consumption increases but people work less!

• Therefore output must fall

• If consumption increases and output falls, it means investment goes down

• This shock DOES NOT produce the “co-movement” of business cycles: consumption goes up
but output, investment and hours of work go down!

8
ECON 52, Spring 2013 4 OTHER SHOCKS

Laziness shocks
• Suppose households temporarily don’t want to work, i.e. the marginal utility of leisure goes
up

• (You could also think of the opposite: households temporarily don’t mind working)

• In the equations:

u0 (c0 ) = βFK (K1 , L1 ) u0 (c1 )

↑ v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y1

• Adjust by ↓ L0

• Y0 goes down because L0 goes down

• If r1 didn’t change, households would ↓ c0 and ↓ c1 to smooth consumption

• In order to smooth consumption, households save less so ↓ K1 goes down

• Output, consumption, investment and hours worked all go down together

• This shock DOES produce the “co-movement” of business cycles

• But do we really want a theory based on waves of laziness?

• Maybe other things that “look” like laziness: disruption in labour markets, prices not adjust-
ing...

Optimism / news about the future

• Keynes had this idea about “animal spirits”: businessmen become very confident about the
future and things start to go well

• Strict interpretation: news.

– The economy reacts to what people think about the future

– (and then what people expect actually happens, at least on average)

• Less strict interpretation: optimism

9
ECON 52, Spring 2013 4 OTHER SHOCKS

– The economy reacts to what people think about the future

– (and maybe they are completely wrong, but we still care about how their expectations
affect things)

• Good news about future wealth:

Y1 = F (K1 , L1 ) + e1

and we ask what happens if e1 changes

• In the equations:

u0 (c0 ) = βFK (K1 , L1 ) ↓ u0 (c1 )

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) ↓ u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y 1

• Adjust by ↑ c0

• Therefore ↓ L0 and ↓ K1

• Future wealth makes households consume more, but also work less and invest less, so it
DOES NOT produce the “co-movement” we observe in business cycles.

• Good news about future productivity:

Y1 = A1 K1α L11−α

and we ask what happens if A1 changes

u0 (c0 ) = β ↑ FK (K1 , L1 ) u0 (c1 )

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) =↑ FL (K1 , L1 ) u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y1

• Effect on current consumption depends on income vs. substitution effects

• If ↑ c0 , then ↓ L0 (and vice-versa)

10
ECON 52, Spring 2013 5 CONCLUSION

• Either way, the shock DOES NOT produce the “co-movement” we see in business cycles.

Uncertainty
• Uncertainty is also about how people think about the future

• Instead of pessimism (“the future will be bad”), uncertainty is about (“I don’t know what the
future will look like”)

• A simple way to model uncertainty is to say

Y1 = F (K1 , L1 ) + e1

and e1 , instead of a known number, is a random variable

• Under standard preferences, uncertainty will lead to higher expected marginal utility of
consumption, i.e.: E [u0 (c1 )] rises.1

u0 (c0 ) = βFK (K1 , L1 ) ↑ E [u0 (c1 )]

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )
v 0 (1 − L1 ) = FL (K1 , L1 ) u0 (c1 )
K1 = Y0 − c0 + (1 − δ) K0
c1 = Y 1

• Adjust by ↓ c0 (precautionary savings)

• But then ↑ L0 and ↑ Y0 and ↑ K1

• Again, no “co-movement”

5 Conclusion
• Productivity shocks and laziness shocks could, in principle, generate “business cycle” co-
movement

• Disagreement about how reasonable they are as theories

• Evidence is somewhat mixed

• Other types of shocks tend to generate the wrong co-movement

Mathematically, this depends on whether u000 (c) > 0, i.e. on the third derivative of the utility function. Most
1

evidence suggests that the reasonable case is that it is positive.

11
ECON 52, Spring 2013 5 CONCLUSION

• The key is in the equation

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )

Unless something happens to v or FL , consumption and hours of work will move in opposite
directions! Consumption and leisure are normal goods.

• The alternative explanation is that there are some market imperfections missing from our
model

• Other shocks combined with market imperfections could be part of the story.

• In particular, we have to look for market imperfections such that

v 0 (1 − L0 ) = FL (K0 , L0 ) u0 (c0 )

does NOT hold.

12