The Shock Reactions in the Closed Digital Economy
ADIREK VAJRAPATKUL, ATHAKRIT THEPMONGKOL
School of Development Economics
National Institute of Development Administration
148 Serithai Road, Klong-Chan, Bangkapi, Bangkok 10240
THAILAND
Abstract: - Economic variability can affect economic agents’ risk perception and behavior, which in turn affects
negatively economic activities and prosperity. The government, therefore, tries to raise their confidence by
designing proper policies to stabilize the economy. To learn the effects of the policies, several models are
utilized, and the Dynamic Stochastic General Equilibrium (DSGE) model is recognized as a potential choice for
discovering such effects. Also, this work applies the DSGE model to extend its application and contributes to
this research area in terms of model construction technique by learning the policy effects in the Thai context
through the closed economy models. In this study, Thailand's quarterly detrended data from 2001:Q1 to
2019:Q2 and the Bayesian estimation method were used. The results showed that the positive effect of
technological evolution on economic growth occurred in both economies, but the effect in the two-sector
economy was less than what occurred in the one-sector economy. Additionally, it was demonstrated that
monetary policy was more effective than fiscal policy. Hence, the recommendations were that policy designers
had to design policies to improve technology in all sectors simultaneously, and the fiscal authority had to
recognize the effect of the number of related agents on the effectiveness of its policies. Also, the monetary
authority had to design a boundary for interest rate volatility to stabilize the economy.
Key-Words: - Business cycle, DSGE, Bayesian estimation, Digital economy, Technology shock, Thailand
Received: July 25, 2021. Revised: February 27, 2022. Accepted: March 13, 2022. Published: March 24, 2022.
1 Introduction
The risks emerging from an unpredictable economic
situation can affect the decisions of related agents
within the economy. When those agents, such as
consumers, producers, and investors, are unsure
about the future, they often delay or change their
actions, which can have a negative impact on
economic activities and transactions as well as
economic growth. To instill confidence in these
agents, the government attempts to establish policies
that promote economic unpredictability within an
acceptable range. For these purposes, the effects of
two major types of policies, namely fiscal and
monetary policy, on the economy are thoroughly
investigated. To understand the potential effects of
those policies, economists have utilized several
econometric methods and models to investigate the
reactions of the target macroeconomic variables to
such policies. Dynamic Stochastic General
Equilibrium (DSGE) is a candidate that is often used
to meet that objective. This model has been
recognized as a useful tool because it enables
analysts to work with various assumptions in the
analysis. Although this model is primarily used to
analyze the economic consequences of monetary
policy, fiscal policy, and technological
advancement, it has been extended to examine the
economic implications of numerous anomalous
phenomena, such as preferences and risks [1, 2].
Concerning monetary policy's effects on the
economy, earlier research established that certain
macroeconomic indicators, e.g., the marginal cost
[3], work hours, investment [4, 5], consumption,
inflation, and output [6 - 8], reacted unfavorably to
a positive monetary policy shock. However, it has
been recognized that there are many factors that can
influence the pattern of reactions [9]. Also, when the
analysts change their model assumptions, some
variables may react in the reversed direction. Fiscal
policy is likewise inconclusive in terms of the
implications of a government expenditure shock on
the economy [10], i.e., it may produce crowding-in,
crowding-out, or a neutral outcome, which depends
on how the government finances its spending [11].
For technological evolution, it is a critical variable
in explaining economic growth because it has the
potential to affect productivity improvement. In
recent years, it has been asserted that technological
evolution is responsible for changes in production
and consumption patterns [12]. It is referred to as
"disruptive technology" in the so-called "digital
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
905
Volume 19, 2022
economy" [13]. The importance of technology and
the digital economy in economic growth and
volatility is discussed in detail in the works of, for
example, [14–18]. In terms of technological
evolution's effects, it has been discovered that
advancement can enhance employment and output
while simultaneously lowering prices and interest
rates [19]. What is learned from the preceding
discussion is valuable, particularly for government
officials, as they can utilize it to make policy
decisions that will help stabilize the economy. As
stated previously, the pattern of reaction might
change depending on the model specification,
assumptions, and context of analysis; hence, model
modification and assumptions that are appropriate
for a particular context, i.e., time and location,
become critical.
The importance of economic stability and the
volatility caused by numerous shocks motivate this
effort. This study use a dynamic stochastic general
equilibrium model and Bayesian estimate
techniques to examine the impact of shocks on the
Thai economy in a closed economy context in order
to identify appropriate policies to promote stability
and growth. In the process of investigating the
results, two models are constructed, i.e., a one-
sector model and a two-sector model, which
separate the digital sector from the non-digital ones,
and then the results are compared to highlight the
dynamic properties of the economy when it is
integrated by digital production activity. This effort
is organized in the following manner. The following
section will describe how the models are
constructed. Section 3 will summarize the findings
and make policy recommendations.
2 Problem Formulation
2.1 Model Construction
In the following, how the models are constructed
will be explained, and the details are as follows.
Let's begin with a one-sector model that acts as
the baseline model for comparison. In this model,
the representative household is supposed to
maximize its lifetime utility by consuming at the
optimal amount and working the optimal number of
hours. Their utility [20] is represented by the
following function:
11
β,1 1
0
CN
ttt
EA
t C t
t
, (1)
where the constant
,
, and
, respectively,
represent the intertemporal discount factor, the
inverse elasticity of consumption, and the inverse
elasticity of labor supply.
Ct
and
Nt
stand for
consumption and working hours.
ACt
denotes
consumption evolution that takes into account
anomaly changes in consumption [21] which is
evolved according to a first-order autoregressive
process. The constraint of household’s budget is
represented by:
1 1 , 1 1 1
1
W N R K R B
NK
t t K t t t t
P C I B
C t t
tt
, (2)
where the constant
N
,
K
, and
C
are ,
respectively, labor tax, capital tax, and consumption
tax. The variables
Kt
,
Bt
, and
It
are capital, riskless
one-period bonds, and investment, respectively.
Wt
,
,
RKt
,
Rt
, and
P
t
stand for, respectively, wage,
capital rental rate, policy rate, and general price
index. The capital is assumed to be evolved
according to the following law:
1
1
K K I
t t t
, (3)
where
is capital depreciation rate.
For the representative firm, it is assumed to
process its final output by using Cobb-Douglas
technology which represented by:
α 1 α
,1
Y A K N
t T t t t
, (4)
where
is the share of capital in the production and
,
ATt
is a technology evolution which is also assumed
to followed a first-order autoregressive process.
The rules that government authority impose in
the economy include fiscal and monetary rules [22 -
23]are represented respectively as follows.
1
1 1 1 ,
R R R
R R Y P
t t t t AMt
R R Y P
ss ss ss ss
, (5)
1
11 ,
1
G G G
G G Y B
t t ss t AGt
G G Y B
ss ss t ss
. (6)
where
,Gt
A
and
,Mt
A
denote fiscal and monetary
policy evolutions which also follows the first-order
autoregressive process. The government budget
constraint is defined by:
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
906
Volume 19, 2022
, 1 1
P C I R K W N B PG R B
C t t t K K t t N t t t t t t t
, (7)
Finally, the market clearing condition can be
expressed by
Y C G I
t t t t
.
By following the regular process, we get the
following key log-linear equations.
,
W C N A P
t t t C t t
, (8)
1 1 1
, 1 1 1
1, , 1
R P P
K t t t
C C A A
ttC t C t
, (9)
()
1 , 1 , 1
C C A A P P R
t t C t C t t t t
, (10)
W N P Y
t t t t
, (11)
,1
R K P Y
K t t t t
, (12)
1,,
P W A R
t t T t K t
, (13)
1
1 1 1 ,
G G Y B A
t G t G G t t G t
, (14)
1
1 1 1 ,
R R P Y A
t R t R R t t R t
. (15)
In a two-sector model, the representative
household's utility function is identical to that in a
one-sector economy. However let define:
1
11
1
,,
N N N
t N t D t
, (16)
where
0
controls the willingness of households
to substitute labour between sectors.
,
NNt
and
,
NDt
are the working hours in non-digital and digital
sector, respectively. The household’s budget
constraint as follow:
1 1
, , , , 1
,,
1
,
11 ,
W N R K
N i t i t K Ki t i t
i N D i N D
R B P C I B
C t i t
t t t t
i N D
, (17)
As N denoted the non-digital sector and D is
for the digital sector, the further variables
augmented by these letters will be assigned to such
sectors accordingly, e.g.,
P
is the aggregated price,
PN
is the non-digital product’s price, and
PD
is the
digital product’s price.
In this model firm is separated into two groups,
i.e., non-digital and digital firms, with the following
explanations.
The representative non-digital firm uses the
Cobb-Douglas technology of the form:
α 1 α
, , , 1 ,
NN
Y A K N
N t TN t N t N t
, (18)
where
,
A
TN t
is a technological evolution in non-
digital sector which usually assumed to follow a
first-order autoregressive process.
Also the representative digital firm uses the
same technology and hence its production function
is expressed by:
α 1 α
, , , 1 ,
DD
Y A K N
D t TD t D t D t
, (19)
For the final good firm, it assembles the
outputs by using the following production function:
1 1 1 1
1
1
,,
Y Y Y
t N t D t
, (20)
where
is the share of non-digital input in the
production of the final good and
is the
intratemporal elasticity of substitution between the
non-digital and digital good.
Similarly, this model assumes the role of fiscal
and monetary authority as the same in the one-sector
model. However the market clearing condition is
defined by
,,
Y C I I G
t t N t D t t
.
The key log-linear form of the two-sector
model can be written as follows.
, , ,
W N A C N N P
N t t C t t N t t t
, (21)
, , ,
W N A C N N P
D t t C t t D t t t
, (22)
1 1 1
, 1 1 1
1, , 1
R P P
KN t t t
C C A A
ttC t C t
, (23)
1 1 1
, 1 1 1
1, , 1
R P P
KD t t t
C C A A
ttC t C t
, (24)
, , , ,
W N P Y
N t N t N t N t
, (25)
, , , ,
W N P Y
D t D t D t D t
, (26)
, , 1 , ,
R K P Y
KN t N t N t N t
, (27)
, , 1 , ,
R K P Y
KD t D t D t D t
, (28)
1
, , , ,
P W A R
N t N N t N TN t N KN t
, (29)
1
, , , ,
P W A R
D t D D t D TD t D KD t
, (30)
,,
Y P P Y
N t t N t t
, (31)
,,
Y P P Y
D t t D t t
, (32)
1 1 1
, , ,
1 1 1
P P P P P P
N t Nss D t Dss D t Dss
P
tP P P
Nss Dss Dss
,(33)
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
907
Volume 19, 2022
The Bayesian estimation technique is then used
to estimate the models using Thailand's quarterly
data acquired from the World Bank database. These
data included three detrended series [24]: GDP,
policy rate, and employment, which covered the
period 2001:Q1 to 2019:Q2.
2.2 Estimation Technique
This section will highlight the methods used for
determining parameters and the basic idea behind
Bayesian estimation. For the method of determining
the parameters, it could be separated into two main
groups of approaches [25]. In the first approach, a
DSGE model duplicates the world in a certain set of
dimensions using a calibration procedure that
assigns values to parameters based on data from past
research and scholarly knowledge. Another
technique in this approach is the Generalised
Method of Moments (GMM), which involves the
selection of parameters for a specific equilibrium. In
contrast, the second approach attempts to obtain an
estimation that provides a full characteristic of the
observed data. This category includes two methods:
classical and Bayesian maximum likelihood
estimation (MLE). These techniques are based on
the likelihood function, which represents the chance
of observing a particular dataset as a function of the
model's parameters. This likelihood can be
computed for various parameter combinations in
order to make the data set "more likely". Parameter
estimations are obtained directly from this method
in classical MLE. However, in Bayesian MLE, an
additional function, namely the prior function, is
considered before observing the data. The prior is
then mixed with the likelihood, and the resulting
function can be maximized in terms of the
parameters until the objective function is
maximized. With the whole information provided
by the Bayesian MLE technique, it is possible to
characterize the data production process more
consistently. Another significant advantage of
Bayesian approaches is that they generate
probability distributions for model parameters,
IRFs, and forecasts, etc., and so explicitly quantify
the uncertainty associated with model-based
analysis and forecasting. Basically, Bayesian
estimation is a technique that combines calibration
with maximum likelihood estimation. The model is
calibrated by specifying priors, whereas the
maximum likelihood is derived from the data. These
priors and the likelihood function are linked
according to the Bayes rule. The first step in
applying Bayesian MLE is to determine the
likelihood function that corresponds to the joint
density of all variables in the data sample,
depending on the model's structure and parameters.
The following step is to specify the prior
distributions. Each prior is a probability density
function for a parameter, providing a formal manner
of expressing the probabilities associated with the
values that parameters can adopt based on previous
research. It is a representation of belief within the
model's context, set independently of the data and
serving as an additional source of information. After
deriving the likelihood and specifying the priors, the
posterior distribution is computed, which indicates
the probability given to various parameter values
after observing the data. It is essentially an update of
the prior probability based on the additional
information provided by the sample variables.
Consider this concept further by applying the Bayes
theorem to the following two random events [26,
27].
Let
()P
is the probability of
,
/PY
is the
conditional probability, and
PY
is the marginal
probability. Also, define the joint probability of
obtaining such
on data
Y
by
/P Y P Y P Y
and vice versa. Hence we have:
/
/P Y P
PY PY
, (34)
where
P
and
/PY
are, respectively, the prior
and the posterior distribution of
, given the
observed data
Y
.
/PY
is the density of the data
that is conditional on the parameters (the
likelihood). Note that
PY
does not depend on 𝜃
and therefore can be treated as a constant for the
estimation, producing:
/ / /P Y P Y P Y
, (35)
where ∝ denotes proportional to.
/Y
is the
posterior kernel. Using this fundamental equation, it
is possible to acquire all the posteriors of interest. In
order to obtain the likelihood, one must use the
Kalman Filter and then simulate the posterior kernel
by a Monte Carlo method with the help of Matlab
and a toolbox called Dynare.
3 Problem Solution
To perform the estimation by using the Bayesian
estimation technique, the first step is to set
particular values of parameters [28, 29] as shown in
Table 1.
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
908
Volume 19, 2022
Table 1. Predefined parameters
Parameters
Value
0.300
3.200
2.000
0.997
G
0.950
G
0.100
0.900
Then the rest of the parameters will be
estimated based on the available information of
priors obtained from previous studies and the results
from the one-sector economy are shown in Table 2
Table 2. One-sector estimated parameters.
Par.
Prior
Posterior
Distr.
Mean
Mean
HPD
inf
HPD
sup
gamma
0.050
0.057
0.039
0.069
M
gamma
0.510
0.789
0.718
0.862
R
gamma
0.280
0.274
0.184
0.367
AT
beta
0.500
0.335
0.212
0.485
AC
beta
0.500
0.487
0.190
0.822
AG
beta
0.500
0.953
0.919
0.987
AM
beta
0.500
0.632
0.522
0.749
Source: Authors' calculation
For the reactions of economic variables to
shocks in the one-sector economy, they are
presented in Fig. 1. In the case of the two-sector
economy, the results are presented in Table 3
and Fig. 2 to Fig. 5.
Table 3. Two-sector estimated parameters.
Par.
Prior
Posterior
Distr.
Mean
Mean
HPD
inf
HPD
sup
gamma
1.000
1.007
0.935
1.115
gamma
0.050
0.060
0.049
0.071
gamma
2.000
2.160
1.972
2.353
gamma
0.500
0.569
0.453
0.700
G
gamma
0.510
0.763
0.675
0.845
R
gamma
0.280
0.208
0.156
0.262
ATN
beta
0.500
0.288
0.117
0.417
ATD
beta
0.500
0.519
0.170
0.849
AC
beta
0.500
0.493
0.192
0.817
AG
beta
0.500
0.974
0.966
0.982
AM
beta
0.500
0.704
0.577
0.825
Source: Authors' calculation
(a)
(b)
(c)
Fig. 1: Reactions in the one-sector economy:
(a) Tech. shock, (b) Gov. shock, (c) Mon. shock
Source: Authors’ presentation
Fig. 2: Reactions to Tech. shock in non-digital
sector
Source: Authors’ presentation
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
909
Volume 19, 2022
Fig. 3: Reactions to Tech. shock in digital sector.
Source: Authors’ presentation
Fig. 4: Reactions to Gov. shock in two sector
economy.
Source: Authors’ presentation
Fig. 5: Reactions to Mon. shock in two sector
economy.
Source: Authors’ presentation
Three shocks that are frequently mentioned in
DSGE analysis will be explored in the following,
namely the technology shock, the government
spending shock, and the monetary policy shock.
Let's begin with the results obtained from a
one-sector economy. According to Fig. 1(a), which
represents the reactions to the technology shock, it
is discovered that when there is an improvement in
technology, there is an increase in demand for
inputs of production, i.e., demand for labor and
capital, and investment. The price of these inputs,
i.e., wages and capital rental rates, has also
increased. However, the fall in the cost of
production caused by technological progress
supports the reduction of the aggregated price. The
increase in demand for labor and capital, as well as
wage and capital rental rates, supports the rise in
output, income, and consumption, which in turn
encourages national income. To stabilize the
economy, the central bank raised the interest rate.
Fig. 1(b), which presents the reactions to a
government spending shock, shows that when the
government increases its spending, it incurs an
increase in demand for labor and capital as well as
investment. Although wages are decreased, the
capital rental rate is increased. An increase in this
spending suppresses the aggregated price. Although
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
910
Volume 19, 2022
there is an increase in demand for labor, demand for
capital, and capital rental rates, consumption has
decreased. It is possible that the increase in
government spending and investment will overcome
the decrease in consumption, thereby increasing
national income. This situation may be interpreted
as a crowding-out effect, as the increase in
government spending causes a reduction in private
consumption and investment thereafter. Regarding
Fig. 1(c), which presents the reactions to a monetary
policy shock, it shows that when the central bank
raises the interest rate, there is a decrease in demand
for labor and capital as well as investment. Also,
wages and capital rental rates have decreased. An
increase in interest rates suppresses the aggregated
price, which in turn encourages consumption. As
labor and capital demands, capital rental rates, and
wages are decreased, national income is also
decreased.
The results obtained from the two-sector
economy will now be discussed. According to Fig.
2, which represents the effects of a technology
shock in the non-digital sector, it is discovered that
when there is an improvement in technology in this
sector, there is an increase in demand for non-digital
and digital labor, non-digital capital, as well as non-
digital investment. Also, non-digital and digital
wages, as well as non-digital and digital capital
rental rates, are increased. However, the fall in the
cost of production in the non-digital sector caused
by technological progress supports the reduction of
the non-digital price and the aggregated price. The
increase in demand for labor and capital, along with
wage and capital rental rates, supports the rise in
output, income, and consumption, and hence
national income. To stabilize the economy, the
central bank raised the interest rate. In Fig. 3, which
represents the effects of a technology shock on the
digital sector, it is discovered that when there is an
improvement in technology in this sector, there is an
increase in demand only for digital labor and digital
capital. But both non-digital and digital investments
have increased. Also, non-digital and digital wages,
as well as non-digital and digital capital rental rates,
are increased. However, the fall in the cost of
production in the digital sector caused by
technological progress supports the reduction of the
digital price and the aggregated price. The increase
in demand for labor and capital, as well as wage and
capital rental rates, supports the rise in output,
income, and consumption, and hence national
income. To stabilize the economy, the central bank
raised the interest rate. Figure 4, which depicts the
reactions to a government spending shock, shows
that when the government increases its spending,
demand for both non-digital and digital labor rises,
as does demand for capital and investment. Here,
the wage is decreased but the capital rental rate is
increased. An increase in this spending makes the
aggregated price decline. Although there is an
increase in demand for labor and the capital rental
rate, consumption has decreased. It is possible that
the effect of an increase in government spending
and investment will overcome the decrease in
consumption, thereby increasing national income.
To stabilize the economy, the central bank raised the
interest rate. Referring to Fig. 6, which represents
the reactions to a monetary policy shock, it shows
that when the central bank increases the interest
rate, there is a decrease in demand for both non-
digital and digital labor, investment in both non-
digital and digital sectors, and a slight decrease in
capital in both non-digital and digital sectors. In this
case, both the wage rate and the capital rental rate
are decreased. An increase in interest rates makes
the aggregate price decline, which in turn
encourages domestic consumption. However, it is
possible that the increase in aggregated consumption
is overcome by the decline in returns from inputs,
investment, and government spending, and hence
national income is decreased.
The implications of these findings are as
follows. In the case of the technology shock, when
comparing the results obtained from a one-sector
economy and a two-sector economy, it is found that
although the reaction of each variable is in the same
direction, some variables, i.e., aggregate price,
aggregate consumption, and national income, are
less sensitive to partial shocks from each sector, i.e.,
in a two-sector economy, aggregate price decreases
more and aggregate consumption, as well as
national income, increase less than in a one-sector
economy. When comparing between non-digital and
digital sectors, it is found that the effect of a
technology shock in the non-digital sector on
national income is larger. This is because the
proportion of non-digital output in the final goods is
larger than the digital ones. Hence, it reflects that
the effect of digital goods on economic growth
depends on how much it is integrated into the final
goods. In the case of the government spending
shock, when comparing results obtained from a one-
sector economy and a two-sector economy, it is
found that some variables react in the same
direction, but all of them in the two-sector economy
are less volatile to this shock, i.e., the aggregated
price and aggregated consumption are less
decreased, while national income is more increased
than what happens in the one-sector economy.
Therefore, it can be concluded that the positive
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
911
Volume 19, 2022
effect of government spending on economic growth
is smaller in the two-sector economy, i.e., this
policy is less effective when the number of sectors
is increased in the economy. For monetary policy
shock, when comparing results obtained from a one-
sector economy and a two-sector economy, it is
discovered that some variables react in the same
direction but become more volatile, i.e., the
aggregated price and national income decrease
more, while aggregated consumption increases more
than what happens in the one-sector economy.
Therefore, it can be concluded that the negative
effect of monetary policy contraction on economic
growth is larger in a two-sector economy, i.e., this
policy is more effective when the number of sectors
is increased in an economy.
Therefore, the interpretation of these
discoveries is that in the two-sector economy,
technology development in a particular sector can
produce a partially positive effect on economic
growth. This effect is, however, less than what
happens when the overall technology in one sector
is improved. Thus, a valuable strategy is to
encourage technology development in all sectors,
instead of focusing on a particular sector. In the case
of policies for stabilizing the economy, monetary
policy is perceived as an effective tool for
governments.
The recommendations based on the results of
the analysis include: 1) the policy designers should
take into consideration the importance and benefits
of technology development in all sectors for
economic growth, i.e., only paying attention to
technology development in a particular sector will
produce less benefit than supporting all technology
development in all sectors; 2) the fiscal authority
should be aware of how to allocate its budget as
when there is an increase in the number of economic
agents, the effectiveness of this policy declines; and
3) the monetary authority should define an
appropriate boundary of interest rate volatility as it
can effectively impact on economic growth and
hence stability.
4 Conclusion
Unpredictable economic situations can create risk
and affect the decisions of economic agents within
the economy. They intend to delay or change their
actions when they feel less confident in the
economy. This situation can have a potentially
negative effect on economic activity and prosperity.
Hence, to raise their confidence, the government
tries to design policies to stabilize the economy.
Here, two major types of policy, i.e., fiscal policy
and monetary policy, with the application of some
economic models, are utilized to meet that
objective. To learn the effects of these policies on
the variability of the economy, the DSGE model is
often employed for this purpose. Also, this work
tries to apply this model to understand such a
situation by using two models of the closed
economy, i.e., one-sector and two-sector models,
and comparing their results. The analysis is
conducted by using the quarterly detrended data of
Thailand, 2001:Q1 to 2019:Q2, and the Bayesian
estimation technique. The results show that although
the positive effect of technology evolution on
economic growth occurs in both economies, the
effect in a two-sector economy is less than what
occurred in a one-sector economy. In the case of
policies for stabilizing the economy, monetary
policy is perceived as an effective tool for
governments. Hence, this work recommends that
policy designers should place an importance on
improving technology in all sectors. While fiscal
authority should allocate its budget by considering
the number of related agents within the economy.
Also, the monetary authority should define an
appropriate boundary for interest rate volatility to
stabilize the economy.
Acknowledgement:
This research is supported by the School of
Development Economics, National Institute of
Development Administration, Thailand.
References:
[1] M. Adolfson, S. Laséen, J. Lindé, and M.
Villani, Bayesian Estimation of An Open
Economy DSGE Model with Incomplete Pass-
Through, J. Int. Econ., Vol. 72, No. 2, 2007,
pp. 481–511.
[2] F. Breuss and K. Rabitsch, An Estimated Two-
Country DSGE Model of Austria And The
Euro Area, Empirica, Vol. 36, No. 1, 2009, pp.
123–158.
[3] M. Rubio and S. Schuh, Monetary Policy And
The Role of Inventory Investment, Appl. Econ.
Lett., Vol. 24, No. 21, 2017, pp. 1605–1612.
[4] S. Mouabbi and J. Sahuc, Evaluating The
Macroeconomic Effects of The ECB’s
Unconventional Monetary Policies, J. Money
Credit Bank., Vol. 51, No. 4, 2019, pp. 831–
858.
[5] R. Wang, Unconventional Monetary Policy in
Japan: Empirical Evidence from Estimated
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
912
Volume 19, 2022
Shadow Rate DSGE Model, J. Int. Commer.
Econ. Policy, Vol. 10, No. 02, 2019, pp. 1-29.
[6] G.-A. Ionita, Monetary Policy in A DSGE
New Keyesian Model–Case Study for
Romania, Acta Univ. Danub. Econ., Vol. 13,
No. 1, 2017, pp. 5–16.
[7] W. E. Melesse, Business Cycles in Ethiopia
Under Alternative Monetary Policy Rules, Afr.
J. Econ. Manag. Stud., Vol. 10, No. 3, 2019,
pp. 299-313.
[8] A. Goyal and A. Kumar, A DSGE Model-
Based Analysis of The Indian Slowdown, J.
Int. Commer. Econ. Policy, Vol. 11, No. 01,
2020, pp. 1-57.
[9] B. Perry, K. Phillips, and D. E. Spencer, Real
Wages and Monetary Policy: A DSGE
Approach, Journal of Economic Studies, Vol.
42, No 5, 2015, pp. 734-752.
[10] A. Nandi, Indian Fiscal Policy: A DSGE
Primer, J. Dev. Areas, Vol. 54, No. 2, pp. 179-
200, 2020.
[11] C. Jooste and R. Naraidoo, The
Macroeconomics Effects of Government
Spending Under Fiscal Foresight, South Afr. J.
Econ., Vol. 85, No. 1, 2017, pp. 68–85.
[12] N. Ahmad and J. Ribarsky, Towards A
Framework for Measuring The Digital
Economy, presented at the 16th Conference of
the International Association of Official
Statisticians, 2018, pp. 1-33.
[13] P. Schreyer and I. Mexico, Measuring GDP in
a Digitalised Economy, Paris, OECD
Conference Centre, 2016, pp. 1-24.
[14] M. Afonasova, E. Panfilova, and M.
Galichkina, Social and economic background
of digital economy: conditions for transition,
European Research Studies Journal, Vol.
XXI, Special Issue 3, 2018, pp. 292-302.
[15] K. B. Murthy, A. Kalsie, and R. Shankar,
Digital Economy In A Global Perspective: Is
There A Digital Divide?, Transnatl. Corp.
Rev., Vol. 13, No. 1, 2021, pp. 1–15.
[16] S. Abendin and P. Duan, International Trade
and Economic Growth in Africa: The Role of
The Digital Economy, Cogent Econ. Finance,
Vol. 9, No. 1, 2021, pp. 1-25.
[17] I. Irtyshcheva et al., The Effect Of Digital
Technology Development on Economic
Growth, Int. J. Data Netw. Sci. , 2021, Vol. 5,
No. 1, pp. 25–36.
[18] X. Jiang, X. Wang, J. Ren, and Z. Xie, The
Nexus Between Digital Finance and Economic
Development: Evidence from China,
Sustainability, 2021, Vol. 13, No. 13, pp. 2-17.
[19] I. W. Bache, L. Brubakk, and J. Maith, Simple
Rules Versus Optimal Policy: What Fits?, fra
Norges Bank Working papers , 2010, pp. 1-34.
[20] J. Jääskelä and R. McKibbin, Learning in An
Estimated Small Open Economy Model.
Economic Group, Reserve Bank of Australia,
2009, pp. 1-46.
[21] M. H. Alp and S. Elekdag, Shock Therapy!
What Role for Thai Monetary Policy?, IMF
Working Paper, 2012, pp.1-48.
[22] Y. Iwata, Fiscal Policy in an Estimated DSGE
Model of the Japanese Economy: Do Non-
Ricardian Households Explain All?, Economic
and Social Research Institute, Cabinet Office ,
2009, pp. 1-64.
[23] K. Bhattarai and D. Trzeciakiewicz,
Macroeconomic Impacts of Fiscal Policy
Shocks in The UK: A DSGE Analysis, Econ.
Model., 2017, Vol. 61, pp. 321–338.
[24] S. Xiaojin and T. K. Ping, What Cycles? Data
Detrending in DSGE Models, Studies in
Nonlinear Dynamics & Econometrics, 2019,
Vol. 23, No. 3, pp. 1-23.
[25] V. Almeida, Bayesian Estimation of A DSGE
Model for The Portuguese Economy, Estud. E
Doc. Trab. Work. Pap., 2009, pp. 1–65.
[26] A. Carriero, T. E. Clark, and M. Marcellino,
Bayesian VAR: Specification Choices And
Forecast Accuracy, J. Appl. Econom., 2015,
Vol. 30, No. 1, pp. 46–73.
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
913
Volume 19, 2022
[27] C. Lidiema, A. G. Waititu, and A. Ngunyi,
Forecasting Household Credit in Kenya Using
Bayesian Vector Autoregressive (BVAR)
Model, Am. J. Appl. Math. Stat., Vol. 6, No. 1,
2018, pp. 17–24.
[28] S. Tanboon, The Bank of Thailand Structural
Model for Policy Analysis, Bank of Thailand.
Work. Paper, 2008, pp. 1–41.
[29] K. Phrommin, Monetary Policy Analysis
Under Headline and Core Inflation Targeting
In Thailand, Thailand and The World
Economy, 2018, Vol. 36, No. 2, pp. 1–31.
Author Contributions:
Adirek Vajrapatkul carried out the simulation and
discussion on the result.
Athakrit Thepmongkol organized and executed the
literature review and experiments.
Sources of funding:
School of Development Economics, National
Institute of Development Administration, Thailand.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on BUSINESS and ECONOMICS
DOI: 10.37394/23207.2022.19.79
Adirek Vajrapatkul, Athakrit Thepmongkol
E-ISSN: 2224-2899
914
Volume 19, 2022
