The Basic Real Business Cycle model

1 The model
Current macroeconomic analysis is carried out mainly through the construc-
tion, calibration, estimation and simulation of dynamic general equilibrium
models, both stochastic and deterministic. These models represent an uni…ed
theoretical framework that can be used for the study of the economy both in
the short term (business cycles) and in the long term (economic growth). The
key characteristic of this typology of models is that they are micro-founded.
For the construction of this micro-founded macroeconomic model, we set up
a macroeconomic environment where both households and …rms takes eco-
nomic decisions. This basic framework can be extended by considering other
economic agents, such as the government, the central bank, the …nancial sec-
tor, etc. Here, we will solve a basic dynamic general equilibrium model. This
is a deterministic version of the Real Business Cycle (RBC) model that has
been extensively used in macroeconomic analysis. The model represents an
economy composed of two types of economic agents: households and …rms.
The model includes as an additional variable, the total factor productivity
(TFP), which represents aggregate productivity of the economy or neutral
technological progress in the sense of Hicks.

Households The household’s maximization problem is given by:

X
T
t
max Et ln Ct (1)
fCt gT
t=0 t=0

where Ct is consumption, 0 < < 1 is the intertemporal discount factor, and

Et ( ) is the mathematical expectation on future variables. Since we consider a
context without uncertainty, that is, with perfect foresight, we can eliminate

1
the expectation operator from the maximization problem, given that the
value of all the variables in the future is known at the present time.
Households maximize the weighted sum of their pro…ts subject to the
budget constraint
Ct + It = Wt Lt + Rt Kt (2)
where It is investment, Wt is wage, Lt is labor, Rt is returns to capital, and
Kt is the capital stock. Investment is accumulated in the form of physical
capital from the following process:

Kt+1 = (1 ) Kt + It (3)

where > 0 is the rate of physical depreciation of capital.

Finally, we need to determine the stock of initial capital, K0 , as well
as the stock of …nal capital, KT +1 , in the case in which the life cycle of the
household is …nite. We consider either a …nite life cycle (for the case in which
we use the Solver tool in Excel), or an in…nite life cycle, for the case in which
we compute numerically a linear approximation to the model.
The household maximization problem can by de…ned by the following
Lagrange auxiliary function,

X
T
t
L= ln Ct t (Ct + Kt+1 W t Lt (Rt + 1 )Kt ) (4)
t=0

where households take as given the relative prices of the productive factors.
First-order conditions, for t = 0; 1; 2; : : : :; T , are given by:

@L 1
: t t = 0 (5)
@Ct Ct
@L
: t+1 [Rt+1 + 1 ] t =0 (6)
@Kt+1
@L
: Ct + Kt+1 (Rt + 1 )Kt W t Lt = 0 (7)
@ t
Solving for the Lagrange’s multiplier (the shadow price of consumption),
the following optimal consumption path is obtained,

Ct+1 = [Rt+1 + 1 ] Ct (8)

2
The …rms The other economic agent that we consider are …rms, which
represents the productive sector of the economy. We assume that …rms max-
imize pro…ts subject to the technological restriction represented by a constant
return to scale technology. We assume a competitive environment. These as-
sumptions results in zero pro…ts, and hence, the factors will be remunerated
based on their contribution to the production process.
We assume that the function of aggregate production (technology) has a
Cobb-Douglas form:
Yt = At Kt L1t (9)
where Yt is the level of aggregate production of the economy, At is the total
factor productivity (TFP) and is the elasticity of output with respect to
capital. We assume the TFP is determined exogenously from the following
process:
At = A t 1 " t (10)
where < 1 is an autoregressive parameter that measures the persistence
of shocks that a¤ect the TFP, and "t is a disturbance term, which we can
considered either stochastic and deterministic. In our case, we will consider
it as an exogenous deterministic variable, whose value is one, except at the
moment in which a technological shock occurs, taking a value di¤erent from
one (higher than one for a shock that increases aggregate productivity and
lower than one for a shock decreasing aggregate productivity).
The problem solved by the …rm consists in maximizing pro…ts, such that:

max t = At Kt L1t W t Lt R t Kt (11)

The …rst-order conditions of the previous problem are:

@ t
: At Kt 1 L1t Rt = 0 (12)
@Kt
@ t
: (1 )At Kt Lt Wt = 0 (13)
@Lt
resulting:

At Kt L1t Yt
Rt = = (14)
Kt Kt
1
(1 )At Kt Lt Yt
Wt = = (1 ) (15)
Lt Lt

3
Equilibrium of the model Competitive equilibrium consists of …nding
sequences of the fCt ; It ; Kt ; Rt ; Wt ; Yt ; At gTt=0 variables such that the condi-
tions that de…ne equilibrium (households maximize utility, …rms maximize
pro…ts, and the feasibility condition of the economy holds) are satis…ed. In
summary, the model of our economy would be composed of the following
seven equations:
Ct+1 = [Rt+1 + 1 ] Ct (16)
Yt A t Kt 1
Rt = = = A t Kt (17)
Kt Kt
Wt = (1 )Yt = (1 )At Kt (18)
Y t = At K t (19)
Kt+1 = (1 )Kt + It (20)
Ct + It = Yt (21)
At = A t 1 " t (22)

Solution: The dynamic system The previous system can be reduced to

a system of two dynamic equations, one for consumption and another for the
stock of capital, plus the equation that determines the behavior of the TFP,

Ct+1 = At+1 Kt+11 + 1 Ct (23)

Kt+1 = (1 )Kt + At Kt Ct (24)

Steady state Steady state of the economy is de…ned by:

1 +
R= (25)

1
(1 + ) 1
K= (26)
A
1 1
Y = A1 (27)
(1 + )
1
(1 + ) 1
I= (28)
A

4
 1

1
1 1 (1 + ) 1
C=A (29)
(1 + ) A

A=1 (30)

Log-linearized model The log-linear approximation to the model is given

by,
ybt = b kt (31)

1 + (1 + )b
ct + b
b kt+1 = (1 )b
kt + kt (32)

b
ct+1 b
ct = (1 + )( 1)b
kt+1 (33)

bit = 1 +
ybt
1 +
b
ct (34)

b
at = b
at 1 (35)
bt = ln Xt
where x ln X.

Solution: Log-linear dynamic system Operating, the log-linearized

model can be represented by the following system of linear di¤erence equa-
tions: " ( 1) #
( 1)
b
ct (1+n) (1+n) b
ct
b = 1 (1+n) b (36)
kt (1+n) (1+n)
kt
where

= 1 + (37)
= 1 + (1 ) (38)

5
Eigenvalues Once we have the model in log-linear terms, we can proceed
to analyze its stability and obtain the eigenvalues associated with it. As we
can see, we have transformed a system of non-linear dynamic equations into
a linear dynamic system, in terms of the deviations (in logarithmic terms,
that is, in percentage) of each variable with respect to the steady state. For
this, we calculate:
" #
( 1) ( 1)
Det 1 =0 (39)

From the previous system, we obtain the following second order equation:

2 ( 1) 1 1 ( 1) ( 1)
+ + =0
(40)
or equivalently:

2 ( 1) (1 ) ( 1)
+ + =0 (41)

The solution of the system is a saddle point. Eigenvalues are,

r
2
( 1) (1 ) ( 1) (1 )
4( 1)

1; 2 = (42)
2

Jump forward-looking variable To compute the model, we need to cal-

culate the short-term e¤ect, that is, the variation in consumption (which is
the "jumping" variable) just at the moment when a disturbance occurs. As
we have seen in theoretical terms, when a disturbance occurs, the consump-
tion is adjusted immediately until reaching the stable saddle path.
The dynamic equation obtained previously for consumption is:

( 1) ( 1) b
b
ct = b
ct + kt (43)

On the other hand, the stable path is de…ned by the trajectory:

b
ct = 1b
ct (44)

6
 Matching both expressions results in:

( 1) ( 1) b
b
ct + kt = 1b
ct (45)

The value of the jumping variable (the forward-looking variable, i.e., con-
sumption) in the instant when a shock hits the economy is given by,

( 1) b
b
ct = kt (46)
( 1) + 1

where 1 is the stable eigenvalue, in order the economy be at the stable saddle
path to the new steady state.

2 Taking the model to Excel

The model can be solved using two alternative methods. Once the model
has been solved analytically and the corresponding dynamic system has been
obtained, then we will solve it computationally in Excel using the "Solver"
tool. As can be observed in the Excel …le "RBC-1.xlsx", we need to de…ne
…rst the value of the parameters of the model, which appear in cells "B4"
to "B7". From these parameters and the steady-state expressions calculated
above, we can obtain the steady-state values for the model variables, which
appear in cells "B10" to "B15". If we place the cursor in cell "B10", the
expression that appears is:

=PTF*((1-Beta+Delta*Beta)/(Alpha*PTF*Beta))^(Alpha/(Alpha-1))

which is the one corresponding to the value of the steady-state production.

Similarly, in cell "B12" we have introduced the expression corresponding to
the steady-state value of the stock of capital, so the expression that appears
in said cell is:

=((1-Beta+Delta*Beta)/(Alpha*PTF*Beta))^(1/(Alpha-1))

Similarly, in cell "B13" we calculate the steady-state value of the con-

sumption, in cell "B14" the steady-state value of the investment and in cell
"B15" the steady-state value of the interest rate. Finally, in cell "B18", we

7
assign the value of the technological change that we assume occurs in period
1, taking a zero value initially.
The variables of the model are de…ned in the columns "D-K", where the
values corresponding to the initial steady state appear in the period 0. The
column "D" is the time index, the column "E" is the TFP, while the column
"F" gives us the optimal path of consumption, which is the variable that
we have to calculate. Column "G" is the investment, which is simply the
di¤erence between what is produced and what is consumed, column "H" is
production, column "I" is the stock of capital, column "J" is the return on
capital and …nally the "K" column is the discounted utility. In cell "E4", the
expression

=E3^Rho*Epsilon

appears with the objective of simulating a productivity shock in period 1. In

cell "E5" the introduced expression is "=E4^Rho", since we assume that the
shock takes a positive or negative value at time 1 and zero in the following
periods. This expression is copied to the following cells in the column.
In cell "I3", the initial capital stock appears. For its part, in cell "I4"
appears the expression:

=(1-Delta)*I3+G3

where the stock of capital in each period of time is the stock of capital of the
previous period, discounting the depreciation, plus the new capital that is
incorporated, which is determined by savings. Finally, column "K" presents
the value of the utility in discounted terms.
The sum of the discounted utilities is calculated in cell "K34", which will
be the target cell to be maximized in the "Solver" tool. The solution to
the problem is obtained by executing the "Solver", once we have de…ned the
target cell to be maximized (the "K34"), the …nal condition ("$I$34=K0"),
and the cells to change with the solution ("$F$4:$F$33").
As we can now see, the optimal path of consumption that we obtain is
completely horizontal, indicating that the consumption is the same period
to period. This is because we are calculating its steady-state value, and in
steady state the variables are constant period to period. In the previous
exercises in which the optimal consumption path was calculated, the slope
of the same depended on the relationship between the discount factor and

8
the interest rate, which was assumed exogenous. However, in this general
equilibrium model, the interest rate is an endogenous variable, and its equi-
librium value is such that, given a discount factor, it makes the rest of the
variables constant, so the optimal path of the resulting consumption is hori-
zontal. In fact, we can verify that in a steady state, given a value of = 0:96,
corresponding to an intertemporal subjective rate of = 0:042, the steady-
state value of the interest rate is 10.2% per period. Discounting the physical
depreciation rate of capital, which is 6% per period, it turns out that the net
return of capital is 0.102-0.06=0.042, which is exactly equal to the value of
the subjective rate of intertemporal preference.
Second, the model can be solved directly in Excel using the log-linearized
system. The numerical resolution of the log-linearized model corresponds
to the spreadsheet "RBC-2.xlsx". First, we de…ne the parameters of the
model. In this exercise, we will use the same parameters as in the previous
exercise –the intertemporal discount factor, the elasticity of production with
respect to the stock of capital and the rate of depreciation of capital. We
also calculate two parameters that are a combination of the previous ones
to simplify the used expressions. The corresponding values appear in cells
"B14" to "B18". In column "C", these values are reproduced, in order to
perform a sensitivity analysis and study the implications of changes in the
values of these parameters. Below, we present the steady-state values, rows
21 to 25, which are the same as those obtained in the previous exercise. In
column "B", the steady-state values are presented with the initial values,
while in column "C", these values are presented with the …nal values. Given
that in this exercise we have considered TFP as an exogenous variable, we
have introduced its initial value in cell "B25". If we want to introduce a new
value to simulate a permanent technological disturbance, we would do so by
changing the corresponding value in cell "C25".
Rows 28 and 29 calculate the eigenvalues associated with the dynamic
system, in column "B" for the initial steady state and in column "C" for the
…nal steady state. Given the restrictions on the parameters, the roots are
going to be real, so the calculation of the imaginary part is not necessary. If
we place the cursor in cell "B28", the expression that appears is:

=((-((Alpha_0-1)*OMEGA_0*PHI_0-Alpha_0*(1-Beta_0)))/(Alpha_0*Beta_0)
-ROOT((((Alpha_0-1)*OMEGA_0*PHI_0-Alpha_0*(1-Beta_0))/(Alpha_0*Beta_0))^2
-4*((Alpha_0-1)*OMEGA_0*PHI_0)/(Alpha_0*Beta_0)))/2

9
which corresponds to the …rst root, while in cell "B29", the equivalent ex-
pression for the other root appears. In rows 32 and 33, the module of each
root plus the unit is calculated.
The information that results from numerically simulating this model ap-
pears in the "G-U" columns. Column "G" is the time index. The variables of
the model are de…ned in the columns "H", the consumption, "I" corresponds
to the investment, "J" the level of production, "K" the stock of capital, while
the columns "L", "M", "N", and "O", present the previous variables in the
same order, but in logarithms. Next, the column "P" corresponds to the log-
arithmic deviation of the consumption with respect to its steady-state value,
the "Q" is the logarithmic deviation of the investment with respect to its
steady-state value, which is simply the di¤erence between what is produced
and what is consumed, in terms of deviations. Column "R" is the logarithmic
deviation of production, which depends on the deviation of the stock of cap-
ital from its steady-state value, and column "S" is the logarithmic deviation
of the capital stock with respect to its steady-state value.
To determine the initial consumption, column "H", we will start from its
steady-state value. To determine the consumption in period 1 (i.e., "H4"),
we use the following expression:
=EXP(P4+LN(Css_1))
This same expression appears in the following cells. A similar expression
appears in the "K" column to calculate the stock of capital from its logarith-
mic deviation. To determine the values corresponding to column "J", we use
the expression, "=PTF_0*K_0^Alpha_0", corresponding to the initial period,
while the initial expression entered in the column "I" is "=K_0*Delta_0".
For the following periods, it is determined using the following expression,
"=Alpha1*Q4". The only value that would change is "Q" that corresponds
to the logarithmic variations of the stock of capital in each period of time
with respect to the steady state.
Columns "P" to "S" show the deviations of each variable with respect to
its steady state, where the key cells are "P4" and "S4". The "P" column
corresponds to the logarithmic variation in consumption with respect to its
steady-state value. For the initial period (zero), cell "P3", is the di¤erence
between the logarithm of the steady-state consumption and the logarithm
of the same, whose result is zero. Cell "P4" contains the new value of the
deviation of the consumption before a disturbance that places said variable
in the new stable path. Thus, b
c1 is determined with the following expression:

10
 =Alpha_1*(Alpha_1-1)*OMEGA_1/((Alpha_1-1)*OMEGA_1*PHI_1
+Alpha_1*Beta_1*Lambda1_1)*S4
which corresponds to jump value. For the successive periods, the consump-
tion deviation is determined using the expression, "=N4+R4", that is, the
consumption in the previous period plus the variation in consumption, which
is the value corresponding to cell "R4". This expression is copied into the
remaining rows of that column.
Column "S" contains the di¤erences of the logarithm of the stock of
capital with respect to the steady state. For the period zero it would be
determined using the following expression, "=LN(K3)-LN($K$3)", it is the
di¤erence between the logarithm of the stock of capital in steady state and
the same, therefore it is zero. On the other hand, to calculate this deviation
in period 1, corresponding to cell "S4", it would correspond to the di¤erence
between the logarithm of the stock of capital in the steady state and the
logarithm of the stock of capital in the …nal steady state, the expression that
we use is:
=LN(K_0)-LN(Kss_1)
On the other hand, the column "R" calculates the deviations of the pro-
duction level, using the expression "=Alpha_0*S3". In the following periods
and until the end of the column, the expression used is, "=Alpha_1*S4".
Finally, columns "T" and "U" show the variations in the deviations of
consumption and capital stock, respectively. Column "T" contains the vari-
ations of the logarithmic deviations of the consumption with respect to the
steady state. If we place the courses in cell "T3", the expression that appears
is:
=-(Alpha_0-1)*OMEGA_0*PHI_0/(Alpha_0*Beta_0)*P3
+(Alpha_0-1)*OMEGA_0/Beta_0*S3
This same expression appears in the following cells of this column but
referred to the values of the …nal parameters. Finally, column "U" presents
the value of the variations of the deviations of the stock of capital with respect
to the initial steady state. In this case, the expression that appears in cell
"U3" is
=-PHI_0/(Alpha_0*Beta_0)*P3+(1-Beta_0)/Beta_0*S3
This same expression appears in the following cells of the column but
refers to the values of the parameters in the …nal steady state.

11
3 Exercises
1. Study the e¤ects of a change in the elasticity of output with respect to
capital, (for example, the value changes from 0.35 to 0.3). Analyze
the consequences of this change on the steady state and on the e¤ect
of a technological shock. Repeat this exercise using the spreadsheets
"RBC-1.xlsx" and "RBC-2.xlsx".

2. Suppose that an earthquake decreases by 20% the capital stock of the

economy. Using the spreadsheet "RBC-2.xlsx", The Model tab, study
what the e¤ects of this shock. To perform this experiment, simply
enter in the spreadsheet that the initial value of the capital stock in
cell "K3" is 20% lower than the corresponding steady-state value, i.e.,
"K_0*0.8".

3. Solve the dynamic general equilibrium model assuming that house-

holds’utility function is:

U (Ct ; Lt ) = ln Ct + (1 ) ln(1 Lt )

where 0 < < 1. Build a spreadsheet similar to the "RBC-1.xls",

which calculates both the optimal path of consumption and the optimal
labor supply. What e¤ects does a positive technological disruption have
on the labor supply?

12