Putting Home Economics Into Macroeconomics: Replication of Greenward, Rogerson and Wright's Work (1993)
Putting Home Economics Into Macroeconomics: Replication of Greenward, Rogerson and Wright's Work (1993)
Putting Home Economics Into Macroeconomics: Replication of Greenward, Rogerson and Wright's Work (1993)
University of Bern
Switzerland, June 2018
ABSTRACT
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
……………………………………………………………………………………………………………………………………
…………………
Keywords:
Contents
Introduction ............................................................................................................................................ 2
The Basic Model ..................................................................................................................................... 4
Assumptions ....................................................................................................................................... 5
Income sources: ................................................................................................................................. 5
Instantaneous utility function, market and home production function: ..................... 6
Firm problem ...................................................................................................................................... 7
Home Problem: .................................................................................................................................. 0
Optimization Problem Definition:................................................................................................... 0
First Order Conditions (FOCs): ........................................................................................................ 1
General Equilibrium Equations ....................................................................................................... 0
Calibrtion .................................................................................................................................................. 0
Introduction
Economic activity is generally classified into business sector (named as market) and
household sector (named as home economic). Efforts have been made considering
business sector more of the time.
On the factual grounds, home economics has a share that cannot be neglected. U.S.
national income and product accounts indicate that investment in household capital
actually exceeds investment in business capital by about 15%. Couples spend around
25% of their discretionary time on unpaid work in their homes, while 33% time doing
work in the market. This research aims to shape the understanding of the interaction
between activities in home and in the market in form of time distribution, production,
capital and investments, and methods to incorporate home activities into standard real
business cycle.
These are the steps which are followed to develop the model:
Step I: Taking a utility function which is dependent upon consumption and time (leisure
and working time)
Here in our case of home consumption is of two types.
c. Leisure time, remaining time after both home and market work (lt )
Step II: In this phase I provide relations of four arguments of utility function by
considering income distribution (budget allocation) and argument of utility function
will appear in these relations. Firstly, I talk about income from three sources and see
that hMt appears in income from market work. Here further variables have appeared
like k Mt , k , h , taxes and depreciation M , H etc. After knowing about income from
all sources, I provide its distribution about how this income is consumed. This forms a
budget constraint.
Step III: In this phase, utility function is presented explicitly. Consulting the relevant
literature on neoclassical growth model, utility function is expressed as related to
consumption and time.
u log Ct b lt1b
b log(Ct ) (1 b)log(lt )
And using the expression of Ct we explicitly formulated utility function.
Step VI: General Equilibrium equations were summarized from the developed model.
In next phase, keeping in view the formulated model, variables will be classified in form of
endogenous and exogenous variables. While model parameters will be evaluated on the
basis of present information or balanced growth path will be considered to evaluate
parameters. In this way, Dynare code input will be established.
Consider large number of infinitely lived households which have preferences described
by utility function as follows:
Assumptions
u
u1 Derivative of instantaneous utility w.r.t 1st argument ( cMt ) = 0
cMt
u
u2 Derivative of instantaneous utility w.r.t 2nd argument ( cHt ) = 0
cHt
u
u3 Derivative of instantaneous utility w.r.t 3rd argument ( hMt ) = 0
hMt
u
u4 Derivative of instantaneous utility w.r.t 3rd argument ( hHt ) = 0
hHt
Individual has a budget and income which is allocated according to this budget (a
constraint) in 3 usages:
This usage is done in a way that capital goods purchased in one period in time are brought
forward and become available for usage in the next period in time.
Income sources:
It is supposed that all taxes are rebated so “G” becomes equal to zero.
T wt hMt h (rt kMt M kMt ) k Eq. 6
Or we can write it as
b
u (cMt , cHt , hMt , hht ) log acMt
e
(1 a)cHt
e
(1 b)log(1 hMt hht ) Eq. 9
e
Firm problem
There is representative firm which has market production with constant return to
scale technology:
zMt hMt
1
yt f (hMt , kMt , zMt ) kMt Eq. 10
While capital which can be invested in market in the future is the sum of capital after
depreciation and investment in time t.
So, we can write total capital by putting Eq. 14 and Eq. 15 in Eq. 13
k(t 1) k Ht (1 H ) xHt kMt (1 M ) xMt
or
k(t 1) k Ht (1 H ) kMt (1 M ) xMt xHt
t
0 , t 0
hMt kMt
t
f (hMt , kMt , zMt ) wt hMt rt kMt 0 Eq. 18
hMt hMt hMt hMt
t
f (hMt , kMt , zMt ) wt hMt rt kMt 0 Eq. 19
kMt kMt kMt kMt
f (hMt , kMt , zMt ) rt kMt
kMt kMt
f 2 (hMt , kMt , zMt ) rt kMt
kMt
f (hMt , kMt , zMt ) wt hMt
hMt hMt
f1 (hMt , kMt , zMt ) wt hMt
hMt
We will use results from Eq. 20 and Eq. 21in developing FOC in the following section.
Home Problem:
function
g (hHt , kHt , zHt ) as it is function of home capital stock k Ht at period t, the time
We assuming here that zMt zMt and zHt zHt . Here t is deterministic and
t t
max
t 0
t u (cMt , cHt , hMt , hht ) i e;
b
max t 0 t log acMt
e
(1 a)cHt
e
(1 b)log(1 hMt hht )
e
Eq. 9
Subject to:
Market budget constraint
u b
log acMt (1 a)cHt (1 b)log(1 hMt hHt )
e e
Eq. 26
hMt hMt e
u t b
kMt kMt
t 0
log acMt
e
e
(1 a)cHt
e
(1 b)log(1 hMt hht ) Eq. 27
u b
log acMt (1 a)cHt (1 b)log(1 hMt hHt )
e e
Eq. 28
hHt hHt e
u t b
kHt kHt t 0 e
log
ac e
Mt (1 a ) c e
Ht
(1 b )log(1 hMt hht ) Eq. 29
These derivatives from Eq. 26 - Eq. 29 yield the following equations:1
u1 (t ) wt (1 h ) u3 (t ) Eq. 30
1
u1 (t ) (1 M ) rt (1 k ) M k (1)u1 (t 1) Eq. 31
1
u2 (t ) g1 (t ) u4 (t ) Eq. 32
1
u1 (t )(1 H ) u2 (t ) g 2 (t ) u1 (t 1) Eq. 33
These has been simplified into
(1 b)hMt
ab Ct e cMt
e1
y (1 )(1 h ) Eq. 34
lt
yt
(1 k )
1 M (1 k ) Eq. 35
kMt
cMt
e 1e
b(1 a)cHt
1 H Eq. 36
akHt
(1 b)hHt
b(1 a)Ct ecHt
e
(1 ) Eq. 37
lt
The details of derivation of Eq. 30 - Eq. 37, is provided in Appendix-A in the end of this
report.
1 Here u1 (t ) means derivative of u(cMt , cHt , hMt , hht ) with respect to first argument that is cMt .
For steady state, we can write Eq. 34 - Eq. 37 as follows:
(1 b)hM
ab C e cMe1 y (1 )(1 ) Eq. 38
l
y
(1 k ) 1 M (1 k ) Eq. 39
kM
b(1 a)cHe c1Me
1 H Eq. 40
ak H
(1 b)hH
b(1 a)C ecHe (1 ) Eq. 41
l
Equilibrium path to which the economy converges when zMt = zHt = t for all t. In this
case, the economy converges to a path on which hMt = hM and hHt =hH are constant while
all other grow at rate X, so that yt = y for some constant while y, cMt =cM t for some
t
constant cM. Now, the model developed is being summarized in form of General
Equilibrium Equations in the section below.
General Equilibrium Equations
2 This is the Dynare® code latex output for the model generated for considering home production.
Calibration
Calibration is meant for finding the values of parameters within the model using some
information/moments present in hand like consumption to output ratio, capital to
output ratio, average time allocation for work in the home and work in the market etc.
When on balanced growth path or in a steady state, we substitute these moments into
model equations to find the values of parameters.
In this section, calibration of model has been discussed and explained. This is done by
taking balanced growth path and prior information about growth rate. Magnitudes of
endogenous variables are taken as an average after WWII period. Some characteristics
of balanced growth path first evaluated. Economy goes to the path where hM,t = hM and
hH,t = hH and other endogenous variables get growing steadily with a growth rate of X.
Such that
Yt = (X^t) * y
Where “y” is some constant value of market output. To explain this in detail, market
profit maximization and household maximization problem is solved as previously
obtained FOC (First order conditions). Eq. 30 and 31 are called as efficiency conditions.
For example, in Eq. 30, right hand side that is –u3(t) is known as disutility that have to
be compensated by household extra working in market. While u1(t)*wt*(1-tauk) is the
increase in utility by working an extra unit of time in the market which balances the
disutility. Maximization of household utility emphasizes that costs and production
benefits must be balanced.
Eq. 32 and 33 are related to capital accumulation in home and market. Eq. 33 describes
that whenever household comes to purchase house hold capital in previous time period
which is done by consuming unit of consumption of good in market. This results into a
loss of utility at that period by a factor of 1/beta.
There are 7 parameters m , h , , , a, b, e,,where first 4 are technological parameters
and remaining 3 are preferences parameters. Their values are found by putting values
of endogenous variables based upon information already available and putting other
values based upon balanced growth path. Here, period of time t is one quarter of year.
So, X value is taken to be 1.005 according to output growth rate of U.S economy after
World War II.
By taking annual rate of return that is 6%. So the value of is found to be 0.9852.
h = 0.25 (Labor income tax rate)
This is an average value as reported by [reference McGrattan, Rogerson, and Wright 1992].
There is much debate in literature for the choice of tax on the capital income. [reference
McGrattan, Rogerson, and Wright 199]
Here we set k = 0.70 that is a very high tax rate to capture all the regulations and duties
direct and indirect taxations imposed by government on market and non-market
investments.
Eq. 38 – 41 has been used for matching six observations as follows:
1. xm/y = 0.118 ( Market investment to market output ratio)
2. xh/y = 0.135 ( Market investment to market output ratio)
3. hH = 0.25
4. hM = 0.33
5. kM/y = 4 ( Market capital to market output ratio)
6. kH /y =5 ( Home investment to market output ratio)
These observations results into value of M = 0.0247, H =0.0235, and
Now, a, b, e the preference parameters are left while there are two
observations remaining. By taking the different values of “e”, values of a and b are found
out. For each value of e, value of a and b are different however values of other
parameters remain unaffected.
The stochastic part of model must be defined to proceed for stochastic simulation of the
model. The values of m = h = 0.95 with m = h = 0.007, in this way the market
technology shock as well as home technology shock, are similar and mimic each other.
As all the parameters have been calibrated based upon a specific value of e. This
parameters dictates the substitution between market and home consumption. Its value
lies between 0 and 1 such that a value of 0 means perfect substitute to each other. A
large value means that substitution of consumption in market sector output in other
sector willingly.
3 Excel file named “Calibration.xls” has been provided for the calculation of Calibration section.
Simulation
The model developed has been simulated using Dynare® 4.5.4 4 for its assessment to
capture U.S economy behavior. It comprises of comparison of volatility of market
output, relative volatility of investment with respect to market output as well as
correlation of market hours and market output. The result has been tabulated in Table
1.
In the table there are three different kind of model behavior which differs from the
value of “e”. The statistics which have been compared are standard deviation in percent
of y, standard deviation of total investment relative to y, market consumption relative
to y, market hours relative to y, the correlation between market hours hM and market
wage and correlatin between market investment to household investment.
The real wage w can be interpreted as an average hours worked outcome in the market
(productivity). Hours worked in market are taken from survey.
1. e = 0, this means that market consumption cM and cH are perfect substitute and
by putting e =0, total consumption becomes Cobb-Douglass function of cM and cH.
This is the base model as discussed in [reference : Greenwood and Hercowitz
1991] and it mimics the same behavior as standard real business cycle without
involving home production. There exists differences in statistics provided by the
model with e=0 and the statistics provided by real data of U.S economy. We can
see that standard deviation of y in percent is too low given by model that is
market output is comparatively far less volatile 0.21 as compared to 1.96. Higher
value of standard deviation of investment with respect to market output that is
2.67 from the model and 2.61 from the real U.S economy data. While volatility of
market consumption w.r.t market output is also very low 0.1 as compared to
0.54 in the real U.S economy. While market wage w.r.t market hours fluctuates
too much un realistically in the model results with a high relative volatility of
2.65. The model results show that hours worked in market and wage are
positively correlated in contrast with the negative correlation existed in the real
data.
U.S. Time
1.96 2.61 0.54 0.78 0.73 1.06 -0.12 0.30
Series
1. Standard
: Home
production 0.21 2.67 0.1 0.62 2.65 0.72 0.99 1
minimized
(e=0)
2. Increased
Willingness
to
Substitute
0.4 2.48 1.0 0.63 0.75 0.83 0.69 0.42
Between
Home and
Market
(e=2/3)
3. Increased
Incentive to
Substitute
Between 0.27 2.63 0.70 0.63 0.81 0.77 0.93 0.87
Home and
Market
(e=0.4)
Appendix:
/*
Putting Home economics into Macroeconomics
by
*/
close all;
// Variables
// Endogeneous variables
var Cm, Ch, y, x, xM, xH, K, Km, Kh, L, Hm, Hh, r, w, Zm, Zh, log_Cm, log_Ch, log_y, log_x, log_xM, log_xH, log_w, log_Hm;
//14 variables
// Exogenous variables
varexo e_m, e_h;
// Parameters
parameters a, b, e, beta, theta, eta, delta_M ,delta_H , tau_k, tau_h, rho_M , rho_H, sigmae_M, sigmae_H, gamm;
// Calibration present in Excel file.
beta=0.9898;
e=0.66; a= 0.611446815;b= 0.799910372;
theta=0.365666667;eta= 0.368567049;delta_M=0.0235;delta_H=0.0235;rho_M=0.95;rho_H=0.95;
tau_k=0.70;tau_h=0.25;sigmae_M=0.007;sigmae_H=0.007;gamm=0.66;
log_Cm=log(Cm);
log_Ch=log(Ch);
log_y=log(y);
log_x=log(x);
log_xM=log(xM);
log_xH=log(xH);
log_w=log(w);
log_Hm=log(Hm);
end;
// Initial values
initval;
y = 2;
Cm = 0.747;
Ch = 0.42;
L=0.4;
Hm = 0.33;
Hh = 0.25;
K = 12;
Km= 5;
Kh= 2;
x = 0.25;
xM=0.1;
xH=0.2;
w = (1-theta)*y/Hm;
r = theta*y/Km;
Zh = 1;
Zm = 1;
e_h = 0;
e_m = 0;
end;
// Steady state
steady;
// Blanchard-Kahn conditions
check;
// Disturbance analysis
shocks;
var e_m=sigmae_M^2;
var e_h=sigmae_H^2;
end;
// Stochastic simulation
stoch_simul (hp_filter=1600,irf=50,order=1) ;
//Write latex file
write_latex_dynamic_model;