Corrected: Correction

ARTICLE
https://doi.org/10.1057/s41599-019-0304-z OPEN

Twin deficit hypothesis and reverse causality: a

case study of China
Umer Jeelanie Banday1 & Ranjan Aneja1

ABSTRACT This paper analyses the causal relationship between budget deficit and current
account deficit for the Chinese economy using time series data over the period of 1985–2016.
1234567890():,;

We initially analyzed the theoretical framework obtained from the Keynesian spending
equation and empirically test the hypothesis using autoregressive distributed lag (ARDL)
bounds testing and the Zivot and Andrew (ZA) structural break for testing the twin deficits
hypothesis. The results of ARDL bound testing approach gives evidence in support of long-
run relationship among the variables, validating the Keynesian hypothesis for the Chinese
economy. The result of Granger causality test accepts the twin deficit hypothesis. Our results
suggest that the negative shock to the budget deficit reduces current account balance and
positive shock to the budget deficit increases current account balance. However, higher effect
growth shocks and extensive fluctuation in interest rate and exchange rate lead to divergence
of the deficits. The interest rate and inflation stability should, therefore, be the target variable
for policy makers.

1 Central University of Haryana, Haryana, India. Correspondence and requests for materials should be addressed to U.J.B. (email: [email protected])

PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms 1

ARTICLE PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z

F
Introduction
iscal and monetary strategies, when executed lucidly, assume Fleming model (Fleming, 1962; Mundell, 1963), which asserts
a conclusive part in general macroeconomic stability. The that an increase in budget deficit will cause an upward shift in
macroeconomic theory which assumes an ideal connection interest rate and exchange rate. The increase in interest rates
between budget (or fiscal) deficit and trade balance is known as makes it attractive for foreign investors to invest in the domestic
twin deficit hypothesis. The growing literature on twin deficit market. This increases the domestic demand and leads to an
hypothesis (TDH) has been theoretically and empirically resear- appreciation of the currency, which in turn causes imports to be
ched by researchers like Kim and Roubini (2008), Darrat (1998), cheaper and exports costlier. However, the appreciation of
Miller and Russek (1989), Lau and Tang (2009), Abbas et al. domestic currency will increase imports and lead to a current
(2011), Bernheim and Bagwell (1988), Lee et al. (2008), Corestti account deficit (Leachman and Francis, 2002; Salvatore, 2006).
and Muller (2006), Altintas and Taban (2011), and Banday and Conversely, the Ricardian Equivalence Hypothesis (REH) dis-
Aneja (2016). agrees with the Keynesian approach. It states that, in a setting of
A study by Tang and Lau (2011) found that a 1% increase in an open economy, there is no correlation between the budget and
budget deficit causes 0.43% increase in current account deficit in current account deficits and hence the former would not cause
the US from 1973 to 2008. After the currency crises and Asian the latter. In other words, a change in governmental tax structure
financial crises, various countries simultaneously experienced will not have any impact on real interest rate, investments or
budget deficit and current account deficit. Lau and Tang (2009) consumption (Barro, 1989; and Neaime, 2008). The assumption
confirmed the twin deficits hypothesis for Cambodia based on here is that consumption patterns of consumers will be based on
cointegration and Granger causality testing. Banday and Aneja the life cycle model, formulated by Modigliani and Ando in 1957,
(2016) and Basu and Datta (2005) confirmed the twin deficits which suggests that current consumption depends on the
hypothesis for India by applying cointegration and Granger expected lifetime income, rather than on the current income as
causality testing. Kulkarni and Erickson (2011) found that, in proposed by the Keynesian model. Furthermore, the permanent
India and Pakistan trade deficit was driven by the budget deficit. income hypothesis developed by Milton Friedman in 1957, states
Thomas Laubach (2003) found that for every one percentage that private consumption will increase only with a permanent
increase in the deficit to GDP ratio, long-term interest rates increase in income. This means that a temporary rise in income
increase by roughly 25 basis points. The recent study by Engen fueled by tax cuts or deficit-financed public spending will increase
and Hubbard (2004) found that when debt increased by 1% of private savings rather than spending (Barro, 1989, p. 39;
GDP, interest rate would go up by two basis points. China needs Hashemzadeh and Wilson, 2006). As private savings rise, the
to take care of lower interest rate rather than higher interest rate need for a foreign capital inflow declines. In this situation a
especially on home loans which is very low in real estate market current account deficit will not occur (Khalid and Guan, 1999,
and has created a serious trouble in the economy. p. 390).
The literature identifies that the FDI “crowding in effect” as an This paper attempts to find out how strong is the relationship
important source of China’s current account surpluses. In fact, between budget and current account deficits in China? Addres-
the World Developmental Report (1985) demonstrated that a sing this question is relevant for both policy and academic per-
country can promote growth using FDI and technology transfers. spectives. Policymakers would like to know to what extent fiscal
China employed such strategies to a remarkable effect on foreign adjustment contributes to addressing external disequilibria,
exchange reserves which increased from US$ 155 billion in 1999 especially in the case of increasing external and fiscal imbalances.
to US$ 3.8 trillion by the end of 2013. Since then, there has been a Secondly, we examined the effects of budget deficit on the current
sharp decline in the capital account surplus. The data from the account in the multivariate framework, and also have derived the
State Administration of Foreign Exchange (SAFE) showed that in general relationship among those variables. The study is based on
2015, China recorded a deficit of US$142.4 billion on the capital ARDL model to investigate twin deficit hypothesis for china both
and financial account. However, it posted a current account in short-run and long-run. Further, we analyze the direction of
surplus of US$ 330.6 billion. causality if such a relationship exists. We utilize econometric
However, majority of the countries are facing both budget methods for example, ARDL bound testing approach and
deficit and current account deficit. The increasing budget deficit Granger causality test to achieve these goals. In section “Theo-
alongside with current account deficit has been an imperative retical Background and Literature Review” we explore the theo-
issue for policy makers in China. Besides, given the importance retical foundation of the twin deficits hypothesis. Section
on free trade, decentralization and development there is a need to “Macroeconomics Aspects of the Chinese Economy” will give
understand the association between budget deficit and trade Macroeconomics Aspects of the Chinese Economy. Section “Data
imbalance in Chinese economy. Few researchers like Burney et al. and Empirical results” describes data and empirical results. Sec-
(1992), Banday and Aneja (2015), Lau and Tang (2009), Corsetti tion “Results and Discussion” provides empirical results and
and Müller (2006) and Ghatak and Ghatak (1996) have empha- section “Conclusions”, provides conclusion to the study.
sized the issue emerged because of growing budget deficit and its
relationship with macroeconomic factors likes interest rate, Theoretical background and literature review
exchange rate, inflation, and consumption. The theoretical relationship between budget deficit and current
In the economic literature there are two distinct theories that account deficit can be represented by the national income
show the linkage between budget deficit and current account accounting identity:
deficit. The first theory is based on the traditional Keynesian

approach, which postulates that current account deficit has a Sp
I þ ðIM
EXÞ ¼ ðG þ TR
TÞ ð1Þ
positive relationship with budget deficit. This means that an
increase in budget deficit will lead to a current account deficit and where IM stand for imports, EX for exports, Sp for private
a budget surplus will have a positive impact on the current savings, I for real investments, G for government expenditure, T
account deficit. The increase in budget deficit will lead to for taxes, and TR for transfer payments. When IM is greater than
domestic absorption and, thus, increase domestic income, which EX, the country has current account deficit. From the right-hand
will lead to an increase in imports and widen the current account side of the equation, when (G + TR − T) is greater than 0, the
deficit. The twin deficits hypothesis is based on the Mundell- country is running a budget deficit. The difference between

2 PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms

PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z ARTICLE

budget deficit and current account deficit must equal the private Nazier and Essam (2012) studied the Egyptian economic data
savings and investments which are shown below. from 1992 to 2010. The data included five variables: GDP, gov-

 ernment budget deficit (as primary deficit), current account
Sp
I ¼ ðG þ TR
TÞ
ðIM
EXÞ ð2Þ deficit, real lending interest rate (RIR) and real exchange rate
(RER). The study used structural vector autoregression (SVAR)
In the literature, various studies attempt to find the relation- analysis, which also gave an impulse response function (IRF) to
ship between budget deficit and current account deficit for dif- capture the impact of budget deficit on current account deficit
ferent countries or groups of countries using various methods and real exchange rate. The findings support the twin divergence
depending upon the sample size and pre-testing of the data. The hypothesis. This contradicts the theoretical framework, which
researchers have generally focused on the U.S economy because finds that a shock given to a budget deficit leads to an
of its simultaneous budget and current account deficit since 1980s improvement in the current account deficit and exchange rate.
(Miller and Russek, 1989; Darrat, 1998; Tallman and Rosensweig, Sobrino (2013) investigated the causality between budget deficit
1991; Bahmani-Oskooee, 1992). However, researchers from dif- and current account deficit for the small open economy of Peru
ferent countries have studied twin deficit hypothesis and obtained for the period 1980–2012 using the Granger causality and Wald
different results for different countries (Like India, U.S and tests, generalized variance decomposition and the generalized
Brazil) ((Bernheim, 1988; Holmes, 2011; Salvatore, 2006; Mukh- impulse-response function. The study found no evidence of a
tar et al. 2007; Kulkarni and Erickson, 2011; Banday and Aneja, causal relationship between budget deficit and current account in
2016; Ganchev, 2010; Lau and Tang, 2009; Rosensweig and the short run.
Tallman, 1993; Fidrmuc, 2002; Khalid and Guan, 1999)). These Goyal and Kumar (2018) investigates the connection between
studies do not support the REH but rather accept the Keynesian the current account and budget deficit, and the exchange rate, in a
traditional theory, finding that budget deficit does have an impact structural vector autoregression for India over the period 1996Q2
on current account deficit in the long run. There are studies that to 2015 Q4. The impact of oil stuns and the differential effect of
support the Ricardian equivalence theorem which denies any consumption and venture propose compositional impacts and
correlation between budget deficit and current account deficit supply stuns rule the conduct of India's CAD, directing the total
(Feldstein, 1992; Abell, 1990; Kaufmann et al. 2002; Enders and request channel. Ricardian hypothesis is not supported.
Lee, 1990; Kim, 1995; Boucher, 1991; Nazier and Essam, 2012; Afonso et al. (2018) studied 193 countries over the period of
Khalid, 1996; Modigliani and Sterling, 1986; Ratha, 2012; Kim 1980–2016 using fixed effect model and system GMM model. The
and Roubini, 2008; Algieri, 2013; Rafiq, 2010). These contra- results find the existence of fiscal policy reduces the effect of
dictory results may be due to differences in the sample period and budget deficit on current account deficit. When there is an
the methods of measuring variables, which use different econo- absence of fiscal policy rule twin deficit hypothesis exists.
metric techniques. Bhat and Sharma (2018) examines the association between
Khalid (1996) researched 21 developing countries from the current account deficit and budget deficit for India over the
period of 1960–1988, taking three variables into consideration: period of 1970–1971 to 2015–2016 using ARDL model. The
real private consumption, real per capita gross domestic product results accepts Keynesian proposition and rejects Ricardian
(GDP) as a proxy of real disposable income and real per capita equivalence theorem. The results find long-run relationship
government consumption expenditure as a proxy of real public between current account deficit and budget deficit. Rajasekar and
consumption. This was done using Johansen cointegration (1988) Deo (2016) find the long-run relationship and bidirectional
and full information maximum likelihood (FIML) for parameter causality between the two deficits in India. Garg and Prabheesh
estimates. The model gives us restricted and unrestricted para- (2017) investigate the twin deficit for India by using ARDL model
meter estimations when restricted parameters are used, meaning and confirm the twin deficit hypothesis. Badinger et al. (2017)
that parameter estimation is non-linear when testing the REH for investigated the role of fiscal rules in the relationship between
the sample countries. The results did not reject the REH for fiscal and external balances for 73 countries over the period of
twelve of the countries, but the remaining five countries do 1985–2012. Their results confirm the twin deficits hypothesis.
diverge from the REH. The rejection of Ricardian equivalence in Litsios and Pilbeam (2017) investigates Greece, Portugal and
the latter group of countries demonstrates lack of substitutability Spain using ARDL model. The empirical results suggest negative
between government spending and private consumption. Ghatak relationship between saving and current account deficit in all
and Ghatak (1996) studied variables such as private consumption, three countries.
government expenditure, income, taxes, private wealth, govern- It is clear that the research so far has not yielded any concrete
ment bonds, government deficits, investments, government evidence regarding the causal relationship between budget deficit
spending and interest on bonds to test the Ricardian Equivalence and current account deficit. As such there is a disagreement on
hypothesis for India for the period from 1950 to 1986. The study the causality between the two deficits. In this paper we test the
employed multi-cointegration analysis and rational expectation theory with the support of autoregressive distributed lag (ARDL)
estimation, and both the tests rejected the REH, finding evidence bounds testing approach using data for the emerging
that tax cuts induce consumption. Thus, the results invalidate the country China.
REH for India.
Ganchev (2010) rejected the Ricardian equivalence theorem for
the Bulgarian economy using monthly time series data for the Macroeconomics aspects of the Chinese economy
period 2000–2010. The long-run results of the vector error cor- The Chinese economy has experienced an unparalleled growth
relation model (VECM) showed evidence of the structural gap rate over the past few decades, with increases in exports, invest-
theory, which states that fiscal deficit influences current account ments and free market reforms from 1979 and an annual GDP
deficit. However, vector autoregression (VAR) results did not growth rate of 10%. The Chinese economy has emerged as the
show any evidence of a short-run relationship between budget world’s largest economy in terms of purchasing power parity,
deficit and current account deficit. The study, therefore, found manufacturing and foreign exchange reserves.
that fiscal policy should not be used, in the Bulgarian economy, as In 2008, the global economic crisis badly affected the Chinese
a substitute for monetary policies in maintaining the internal economy, resulting in a decline in exports, imports and foreign
equilibrium. direct investments (FDI) inflow and millions of workers losing

PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms 3

ARTICLE PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z

Budget Deficit (BD) Current Account Balance (CAB)

1 12

0
8

-1
4
-2

0
-3

-4 -4
86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16

Interest Rate (INT) Inflation (INF)

12 25

10 20

8 15

6 10

4 5

2 0

0 -5
86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16

Money Supply (MS) Real Effective Exchange Rate (REER)

50 180

160
40
140
30
120
20
100

10 80

0 60
86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16

Fig. 1 Behavior of Macroeconomic variables. Overview of Budget Deficit (BD), Current Account Balance (CAB), Interest Rate (INT), Inflation (INF), Money
Supply (MS) and Real effective Exchange Rate (REER) from 1985 to 2017. Source: Compiled based on data taken from the World Bank. BD and CAB are
expressed as percentage of GDP

their jobs. It is visible from the Fig. 1 that China’s current and finally enhances current account surplus. However, the
account surplus has dropped significantly from its boom dur- negative shift in the current account is due to structural and
ing 2007–2008 financial crises. After the financial crisis, Chi- cyclical forces. The cyclic forces can be seen from the trade
na’s exports fell by 25.7% in February 2009. Further, the prospective: the increasing prices of Chinese imports like oil and
exclusive demand for Chinese goods in the international semiconductors, pulls the current account balance downwards.
market pushed up current account surplus to 11% of the GDP However, the structural shift can be seen from the financial side,
in 2007, in a global economy (Schmidt and Heilmann, 2010). It which affects Chinese saving and investments. The investments
is said that China responded to the 2008 crisis with a greater got reduced to 40% and household savings has reduced from 50%
fiscal stimulus in terms of tax reduction, infrastructure and to 40% of GDP.
subsidies when compared to the Organization for Economic
Co-operation and Development (OECD) countries (Herd et al., Data and empirical results
2011; Morrison, 2011). For undertaking the empirical analysis, we use the World Bank
However, we draw an important inference from the Fig. 1, data for the following variables. The study covers the period from
when the budget deficit was lowest (−0.41% of GDP in 2008), the 1985–2016 and is based secondary data.
current account surplus increased to 9.23% of the GDP. The
negative shock to the budget deficit reduces current account a. Current account deficit (CAD) indicates the value of goods,
surplus and positive shock to the budget deficit increases current services and investments imported in comparison with
account surplus. exports on the basis of percentage of the GDP
The deprecation of the exchange rate is directly related to the b. Budget deficit (BD) indicates the financial health in which
elasticity of demand which improves the exports of the country expenditure exceeds revenue as a percentage of the GDP

4 PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms

PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z ARTICLE

c. Deposit interest rate (DIR) as a proxy of interest rate (INT) Empirical results
the amount charged by lender to a borrower on the basis of Unit root test. As we are using time series data, it is important to
percentage of principals check the properties of the data; otherwise, the results of non-
d. Inflation (INF) is measured on the basis of consumer price stationary variables may be spurious (Granger and Newbold,
index which reflects annual percentage change in the cost of 1974). To assess the integration and unit root among the vari-
goods and services ables, numerous unit root tests were performed. Apart from
e. Broad money (MS) a measure of the money supply that applying the unit root test with one structural break (Zivot and
indicates the amount of liquidity in the economy. It includes Andrews, 1992), we also employed a traditional Augmented
currency, coins, institutional money market funds and other Dickey-Fuller (ADF) unit root test (1981) and Phillips-Perron test
liquid assets based on annual growth rate and real effective (PP 1988).
exchange rate (REER) To ascertain the order of integration, we first applied the ADF
and PP unit root tests. The results of the unit root test suggest
that the RER and INF series is integrated of order zero I(0), and
Model. The basic model to find out the relationship among BD, other variables are integrated of order I(1). The Zivot and
CAD, INF, INT, REER, and MS is as follows:
Andrews (ZA) unit root test begins with three models based on
CAD ¼ f ðBD; REER; INF; INT; MSÞ ð3Þ Perron (1989). The ZA model is as follows:
Xk
where INF is inflation and DIR is direct interest rate. Yt ¼ μA þ θA DUt ðλÞ þ βAt þ aA Yt
1 þ cA ΔYt
j
j¼1 j
þ εt ð4Þ
Based on Fig. 2 we will estimate the two models in which CAD
and BD is dependent variables and others are independent Xk
variables which is as below: Yt ¼ μB þ βBt þ γB DTt ðλÞ þ aA Yt
1 þ cB ΔYt
j
j¼1 j
þ εt ð5Þ
To estimate the above models we have employed various
Xk
econometric techniques to achieve the objective of the study are Yt ¼ μC þ θCDU ðλÞ þ βCt þ γCDT ðλÞ þ aC Yt
1 þ cC ΔYt
j þ εt
t t j¼1 j
as below:
ð6Þ
● Unit root test to check stationarity of the variables.
● ARDL bound testing for long-run and short-run relationship From the above equations, DUt(λ) = 1, if t > Tλ, 0 otherwise:
among the variables. DTt(λ) = t − Tλ if t > Tλ, 0 otherwise. The null hypothesis for
● Granger causality to check the causal relationship among the Eqs. 4–6 is α = 0, which implies that (Yt) contains a unit root
variables. with drift and excludes any structural breakpoints. When α < 0 it
simply means that the series having trend-stationary process with
a structural break that happens at an unknown point of time. DTt
is a dummy variable which implies that shift appears at time TB,
where DTt = 1 and DTt = t − TB if t > TB; 0 otherwise. The ZA
REER REER test (1992) suggests that small sample size distribution can
deviate, eventually forming asymptotic distribution. The results of
the unit root tests are given below.
INF INT INF INT The first step in ascertaining the order of integration is to
CAD BD perform ADF and PP unit root tests. Table 1, above, provides the
results of the ADF and PP tests, which suggest that four out of six
variables are non-stationary at level and two variable is stationary
at level. However, it is important to check the structural breaks
BD MS CAD MS and their implications, and the ADF and PP tests are unable to
find the structural break in the data series. To avoid this obstacle,
Fig. 2 Relationship of the dependent variables (CAD and BD) with other we applied a ZA (1992) test with one break, the results of which
macroeconomic variables are given in Table 2. The structural break test reveals four breaks

Table 1 Augmented dickey fuller test (ADF) and Phillips-Perron (PP) test for unit

ADF PP
Intercept Intercept-trend Intercept Intercept-trend
CAD(I0) −2.48615 (0.1214) −2.45463 (0.3466) −2.52978 (0.1184) −2.5654 (0.2973)
CAD(I1) −5.1057 (0.0003)a −5.07512 (0.0015)a −5.165446 (0.0002)a −5.3379 (0.0008)a
BD(I0) −1.98887 (0.2895) −2.0139 (0.5669) −2.51257 (0.1223) −2.56240 (0.2986)
BD(I1) −3.70742 ((0.0098)a −3.61588 (0.0472) −7.02441 (0.0000)a −7.6921 (0.0000)a
REER(I0) −4.06142 (0.0037)a −4.07804 (0.0162) −3.95265 (0.0049)a −4.38008 (0.0080)a
REER(I1) −6.09864 (0.0000)a −6.142736 (0.0001)a −6.54687 (0.0000)a −14.7227 (0.0000)a
MS(I0) −2.400962 (0.1497) −2.963096 (0.1581) −2.491013 (0.127) −3.12720 (0.1179)
MS(I1) −6.748769 (0.000)a −6.621696 (0.000)a −6.826333 (0.000)a −6.690086 (0.000)a
INF(I0) −3.039361 (0.0426)a −4.143720 (0.0142)a −2.25412 (0.1925) −2.37494 (0.3845)
INF(I1) −4.936068 (0.0005)a −4.92759 (0.0026)a −7.70762 (0.000)a −7.358584 (0.000)a
INT(I0) −0.96954 (0.751) −1.99693 (0.580) −1.03964 (0.726) −2.16663 (0.4905)
INT(I1) −4.43685 (0.001)a −4.36859 (0.008)a −4.54219 (0.001)a −4.31192 (0.009)a
Source: Computed by authors
Note: Critical value at the 1% significance level denoted by superscript ’a‘, with intercept and intercept-trend

PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms 5

ARTICLE PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z

Table 2 Zivot and Andrews (ZA) test for unit roots with one Table 4 Cointegration test results
structural break
Calculated 90% LB 90% UB 95% LB 95% UB
Variable CAD BD REER MS INT INF F-statistic
Test- −3.178 −2.548 −3.159 −3.076 −3.221 −4.923 F = 9.439 2.05 3.33 2.51 3.99
statistic (α)
Note: When the F-value is greater than the lower and upper bounds value, we can say the
Time of 2008 2008 1992 1994 2003 2003 variables are co-integrated
Breakdown
Lags (k) 0 2 2 0 0 0
greater than the critical value and reject the null hypothesis when
Note: Structural breaks are based on breaks in trend; critical values are obtained from ZA (1992)
with one structural break and an optimal lag structure by AIC t is less than the critical value.
As seen in Table 4, the calculated F-statistic F = 9.439 is higher
than the upper bound critical value 3.99 at the 5% level. The
results over the period of 1985 to 2016 suggest that there is a long
Table 3 ARDL model (2,0,2,3,0,1) dependent variable (BD) run relationship among the variables and the null-hypothesis of
no cointegration is rejected meaning acceptance of traditional
Variables Coefficient t-value Prob Keynesian approach for China. The bounds test results conclude
BD(−1) .52504 4.3020 .001 that there is strong cointegration relationship among budget
BD(−2) −.47852 4.4321 .000 deficit, current account deficit, interest rate, exchange rate,
CAD .13298 4.6230 .000 inflation and money supply. Since the F-statistic was greater
DIR .22922 3.0986 .007 than the upper bound critical value, we performed a diagnostic
INT(−1) −.14730 1.5119 .150 testing based on auto-correlation, normality and heteroskedasti-
INT(`2) −.38099 4.6848 .000 city, the results of which were insignificant based on the
INF .009286 .4124 .685 respective P-values. The Fig. 3 gives the results of CUSUM and
INF(−1) .041560 1.6862 .111 CUSUMSQ tests which check the stability of the coefficient of the
INF(−2) .057117 2.5665 .021
regression model. The CUSUM test is based on the sum of
INF(−3) .084059 3.5271 .003
MS −.010333 .78506 .444
recursive residuals, and the CUSUMSQ test is based on the sum
REER .0013332 .18609 .855 of the squared recursive residual. Both the graphs are stable, and
REER(−1) −.017791 2.1974 .043 the sum does not touch the red lines. Hence there were no issue of
serial corelation, Heteroscedasticity and normaility in this model
Note: Lag selection is based on Schwarz Bayesian Criteria (SBC) see Table 5.

in Model A. The first, in 1992, may be due to inflation caused by Long-run and short-run relationship. After discovering the
privatization; the second, in 1994 may be due to higher inflation cointegrating relationship among the variables, it is important to
which caused the consumer price index to shot up by 27.5% and determine the long-run and short-run relationships using the
imposition of a 17% value-added tax on goods; the third, in 2003, ARDL model.
may be due to a 48% decline in state owned enterprises at the Pq Pq Pq
BDi ¼ θ0 þ i¼1 θ1i CADt
i þ i¼1 θ2i BDt
i þ i¼1 θ3i INFt
i
same time as a reduction on trade barriers, tariffs and regulations Pq Pq Pq
was put in place and the banking system was reformed; the þ i¼1 θ4i REERt
i þ i¼1 θ5i MSt
i þ i¼1 θ6i INTt
i þ ut
fourth, in 2008, is probably due to the global financial crisis ð7Þ
The ZA test with one structural break gives different results, as Pq Pq
all six variables are non-stationary at a 1% level. ΔBDi ¼ θ0 þ θ1 ΔECMt
1 þ i¼1 θ2i ΔCADt
i þ i¼1 θ3i ΔBDt
i
Pq Pq Pq
þ i¼1 θ4i ΔINFt
i þ i¼1 θ5i ΔREERt
i þ i¼1 θ6i ΔMSt
i
Pq
ARDL bounds testing approach. We applied the ARDL bounds þ i¼1 θ7i ΔINTt
i þ ust
testing approach to check the cointegration long-run relationship
by comparing F-statistics against the critical values for the sample ð8Þ
size from 1985 to 2016. The bounds testing framework has an The above Eqs. 7, 8 of the ARDL model capture the short and
advantage over the cointegration test developed by Pesaran et al. long-run relationship among the variables. The model is based on
(2001), in that the bounds testing approach can be applied to Schwarz Bayesian Criteria (SBC) optimized over 20,000 replica-
variables when that have different orders of integration. Thus, it is tions. The lagged error correction term (ECM) is estimated from
inappropriate to apply a Johansen test of cointegration, and we the ARDL model. The coefficient ECMt−1 should be negative and
applied ARDL bounds testing to determine the long-run and significant to yield the evidence of a long run relationship and
short-run relationships. The F-statistics are compared with the speed of equilibrium (Banerjee et al. 1993). The results of Eqs. 7, 8
top and bottom critical values, and if the F-statistics are greater are provided in Table 6. The results find a strong long-run
than the top critical values, it means there is a cointegrating relationship among the variables for the period from 1985 to
relationship among the variables Table 3. 2016.
All values are calculated by using Microfit which defines At 5% level of significance the long-term estimates of the
bounds test critical values as “k” which denotes the number of ARDL model find evidence of the twin deficits hypothesis for
non-deterministic regressors in the long-run relationship, while China, as all the variables are found to be significant at this level.
critical values are taken from Pesaran et al. (2001). The critical Thus, our study upholds the empirical validity of the Keynesian
value changes as ‘k’ changes. The null hypothesis of the ARDL proposition for China, while rejecting the Ricardian equivalence
bound test is Ho: no relationship. We accept the null hypothesis hypothesis.
when F is less than the critical value for I(0) regressors and we The short-run results are significant; the coefficients of BD,
reject null hypothesis when F is greater than critical value for I(1) CAD, INF, and INT are significant at a 5% level. This shows that
regressors. For t-statistics we accept the null hypothesis when t is a small change in the budget deficit has a significant impact on

6 PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms

PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z ARTICLE

1.6 10.0

7.5

1.2
5.0

2.5
0.8

0.0

0.4 -2.5

-5.0
0.0
-7.5

-0.4 -10.0
06 07 08 09 10 11 12 13 14 15 16 06 07 08 09 10 11 12 13 14 15 16

CUSUM of Squares 5% Significance CUSUM 5% Significance

Fig. 3 CUSUM and CUSUMSQ stability tests. Plot of cumulative sum of recursive residuals and plot of cumulative sum of squares of recursive residuals

economy. The pace of growth in china’s economy has accelerated

Table 5 Diagnostic testing since the decades in which there was a higher export promotion
due to market liquidity and flexible governmental policies. In the
Test statistic LM version F version early 1990s, it was reported in the Chinese print media that
(A) Serial correlation .01999 (.964) 0.1034 (.975) although the Chinese economy was expanding rapidly, the deficits
(B) Normality .94082 (.625) still existed. These were called ‘hard deficits’ primarily because
(C) Heteroscedasticity .58695 (.444) .55776 (.462) they were financed by expanding money supply and hence
A: Based on Lagrange multiplier test of residual serial correlation
exerted inflationary pressure on the economy (People’s Daily,
B: Based on a test of skewness and kurtosis of residual March 28, 1993). Deficit financing may also increase interest rates
C: Based on the regression of squared residuals on squared fitted values
because once the monetary accommodation of the deficit is ruled
out, the government has to incentivize consumers and firms to
buy more government bonds. If the purchase of government
Table 6 Long-run coefficient using ARDL model based on bonds does not increase in direct proportion to the rise in deficit,
(2,0,2,3,0,1) dependent variable (ΔBD) government must borrow more money. This, in turn, crowds out
private investment. Such a reduction in the private sector's
demand for capital has been summarized by Douglas
Coefficient t-statistic Prob
Holtz–Eakin, the director of the U.S Congressional Budget Office,
ADJ(ECM) BD(−1) −0.95348 9.4018 0.000 as a “modestly negative” effect of the tax cut or budget deficits on
Long Run CAD 0.13946 4.905 0.000
INT −0.31366 2.924 0.010
long-term economic growth.
INF 0.20139 3.6196 0.002 This study finds that a higher interest rate significantly affects
MS −0.010837 .77795 0.448 the budget deficit and current account deficit in both the short-
REER −0.017262 9.9136 0.000 run and long-run. Chinese banks are increasing interest rates on
Short Run ΔBD (−1) 0.47852 4.432 0.000 home loans that had previously been very low; however, because
Short Run ΔCAD 0.13298 4.6230 0.000 of its economic bubble, especially in the real estate market, this is
Short Run ΔINT 0.22922 3.0986 0.006 creating serious trouble in the economy. This is similar to the
ΔINT(−1) 0.38099 4.6848 0.000 American bubble, in which most people were unable to repay
Short Run ΔINF 0.00928 −0.4125 0.685 loans, creating a financial crisis. Still, China must work within the
ΔINF(−1) −0.14118 4.2123 0.000
existing bubble, even though it creates a lack of investments and
ΔINF(−2) −0.084059 3.5271 0.002
Short Run ΔMS −0.010333 0.7850 0.442
can cause trouble for economies worldwide. The debt bubble is
Short Run ΔREER 0.001332 0.1860 0.854 due to the elimination of loan quotas for banks in an attempt to
increase small business. These companies are still struggling to
R-squared = 0 .8326 F-Stat = 16.815 (.000) DW Stat = 1.9049 repay that debt, which is almost half the amount of GDP of both
Note: The table reports the ECM value to be −0.95348 negative and statistically significant. The
results ensure there is convergence among the BD, CAD, INT, INF, MS, and REER, meaningthere private and public debt (The Economist, 2015).
is significant long-run relationship While empirical research shows a statistically significant
relationship between higher budget deficits and long-term interest
the current account deficit; similarly most of the macroeconomic rates, there are differences on the magnitude of the impact.
variableshave a significant impact on the current account deficit. Chinese needs to take care of lower interest rate rather than
As we found evidence of a long-run relationship, we calculated higher interest rate especially on home loans which was very low;
the lagged ECM from the long-run equations. The negative in real estate market which has created a serious trouble in the
(−0.95348) ECM coefficient and the high speed of adjustment economy.
brought equilibrium to the economy, with the exogenous shocks It is also important to note that the concept of the deficit
and endogenous shocks restoring it after a long period. includes both hard and soft deficits. The “hard” part of the deficit,
The Chinese economy is one of the most integrated economies being financed by printing money, is inflationary. The “soft” part,
in the world and has emerged as a powerful actor in the global being financed by the government's domestic and foreign debts, is

PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms 7

ARTICLE PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z

non-inflationary. However, they are not reported by the Chinese reverse causality from budget deficit to macro variables and
government. current account deficit to macro variables holds at 5% level of
Moreover, hard deficits and consequent inflation increases significance. Thus, the results of Granger causality gives us more
capital inflows (to prevent interest rates from rising) and causes evidences in support of the Keynesian proposition for China in
current account deficit. In China, deficits occurred as a result of the light of above data. The reverse causality was not apparent
overestimating revenues or underestimating expenditures (Shen because Chinese economy is one of the most integrated econo-
and Chen, 1981) and contributed to the development of bottle- mies with higher capital outflows, export-led growth and export
necks in sectors such as energy, transportation, and communica- promotion due to market liquidity and flexible governmental
tions sectors that are of vital importance for the long-term growth policies. While the capital inflow determines a tight fiscal policy
of China. However, given the reluctance of the non-government to avoid overheating of economy (see author Castillo and Barco
investors to venture into such low pay-back sectors, the Chinese (2008) and Rossini et al. (2008).
government’s policy to balance budgets by cutting its capital
expenditure can severely restrict the development of these sectors
Results and discussion
(Luo, 1990, and Colm and Young, 1968).
The ambiguous verdict on the twin deficits hypothesis based on
However, the natural explanation for the external surplus is
two competing theories: the Keynesian preposition and the
that the economy follows an export-driven growth model by
Ricardian Equivalence theorem. In this paper we investigated the
increasing FDI inflows. In the rising tide of economies, China’s
link between budget deficit and current account deficit with an
demographic boom has created an economic miracle, allowing
emphasis on the impact of macro-variables shock on current
them to invest in education and skilled workers that will
account balance and budget deficit in the Chinese economy over
accordingly benefit the economy.
the period 1985–2016.
During the global financial crisis of 2008, the Chinese
The conclusion of this study is based on ARDL bound testing
government combined active fiscal policy and loose monetary
approach. The model is based on Schwarz Bayesian Criteria
policy and introduced a stimulus package of $580 billion (around
(SBC) optimized over 20,000 replications. With this data and
20% of China’s GDP) for 2009 and 2010. The package partially
analysis, we conclude that there is a long-run and short-run
offset the drop-in exports and its effect on the economy, so that
relationship among the variables. We accept the Keynesian
the economic growth of China remained well above international
hypothesis, which means we did not find evidence in support of
averages. During this period (2008–2009) the current account
the Ricardian Equivalence theorem in the Chinese economy. This
surplus, as a percentage of GDP was at its highest and the budget
was surprising because higher governmental expenditure and
deficit was at its lowest point of the entire decade. Our findings
flexible governmental policies could have pushed up the current
suggest that a strong long-run and short-run relationship exists
account deficit. We further applied Granger causality test the
among the variables. The empirical findings support the
results concluded that there is a strong inverse causal relationship
Keynesian proposition for Chinese economy.
between budget deficit and current account deficit meaning that
Keynesian preposition is more plausible than Ricardian Equiva-
Granger causality lence theorem in the light of above data. Our results are con-
In this section we attempt to estimate the causality from Xt to Yt sistent with Banday and Aneja, (2019); Banday and Aneja, (2016);
and vice versa. We apply Granger causality to check the robust- Ganchev, 2010; Bhat and Sharma (2018); Badinger et al. (2017),
ness of our results and to detect the nature of the causal rela- and Afonso et al. (2018).
tionship among the variables based on Eqs. 9, 10. The findings of the Granger causality test revealed that there
P P P
ΔBDt ¼ α1 þ β1 ΔCADt
i þ θ1 ΔINFt
i þ γ1 ΔREERt
i exists bidirectional causality running from budget deficit to cur-
P P rent account balance and vice versa in China. A similar move-
þ δ1 ΔINTt
i þ λ1 ΔMSt
i þ εt
ment of the budget deficit and the current account balance is the
ð9Þ thing that one would expect when there are cyclic shocks to
P P P output. Detailed empirical findings highlight that the interest rate
ΔCADt ¼ α2 þ β2 ΔBDt
i þ θ2 ΔINFt
i þ γ2 ΔREERt
i
P P bubble and higher inflated housing prices are becoming a chal-
þ δ2 ΔINTt
i þ λ2 ΔMSt
i þ εt lenge for the Chinese economy, as they are now nearing the prices
ð10Þ of the US bubble before the financial crisis popped it. If the US
increases interest rates, the money will flow out of the Chinese
Table 7 describes the results of Granger causality test. The market, which could cause a similar crisis in China. It will be a
reverse causality holds in light of the fact that budget deficit cause challenge for the monetary authority to bring stability in China
Granger to the current account deficit. For all the variables, the where inflation, interest rate bubble and exchange rate volatility
are of primary concern. Expenditure reduction for low payback
Table 7 Granger causality sectors, such as energy and communication, could also be a worry
in the future. Rising inflation can significantly increase the inflow
Equation Chi2 Prob of capital due to an increase in domestic demand; this can lead to
BD to CAD 16.742 0.001 a current account deficit, as it is now more than half of the GDP.
BD to RER 10.949 0.012 The indebtedness in the economy is at its peak, and this can crush
BD to INF 15.648 0.001 the financial cycle and cause financial crisis.
BD to INT 35.239 0.000
BD to MS 12.199 0.007
CAD to BD 13.795 0.003 Conclusions
CAD to RER 6.08 0.108 The cointegration investigation confirms the presence of a long-
CAD to INF 15.636 0.001 run relationship between the variables. The results propose that
CAD to INT 9.4748 0.024 there is a positive connection between the current account deficit
CAD to MS 26.102 0.000 and fiscal deficit. Our outcomes show that increase in money
supply and the devaluation of the exchange rate increases current
Note: Computed by authors
account balance. The effects of the current account deficit and the

8 PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms

PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z ARTICLE

exchange rate have been taken as exogenous. A long-run sta- Darrat AF (1998) Tax and spend, or spend and tax? An inquiry into the Turkish
tionary association between the fiscal deficit, current account budgetary process. South Econ J 940–956
deficit, interest rate, inflation, money supply, and the exchange Dickey DA, Fuller WA (1981) Likelihood ratio statistics for autoregressive time
series with a unit root. Econometrica: J Econometric Soc 1057–1072
rate has been found for China. Granger causality tests reveal Enders W, Lee BS (1990) Current account and budget deficits: twins or distant
bidirectional causality running from budget deficit to current cousins? The Review of economics and Statistics 72(3):373–381
account balance and macroeconomic variables to budget deficit Engen EM, Hubbard RG (2004) Federal government debt and interest rates. NBER
and current account balance. Macroecon Annu 19:83–138
A genuine answer for the issue of budget deficit and current Feldstein M (1992) Analysis: the budget and trade deficits aren’t really twins.
Challenge 35(2):60–63
account deficit lies with a coherent package comprising of both Fidrmuc J (2002) Twin deficits: Implications of current account and fiscal imbal-
fiscal and monetary approaches. It must concentrate on stable ances for the accession countries. Focus Transit 2(2002):72–83
interest rate, inflation target and monetary position will supple- Fleming JM (1962) Domestic financial policies under fixed and under floating
ment the budget-cut policies. The findings of the paper will be exchange rate. Staff Pap Int Monet Fund 10:369–380
helpful for future empirical models, which attempt to further Ganchev GT (2010) The twin deficit hypothesis: the case of Bulgaria. Financ theory
Pract 34(4):357–377
highlight the transmission mechanism. Garg B, Prabheesh KP (2017) Drivers of India’s current account deficits, with
implications for ameliorating them. J Asian Econ 51:23–32
Data availability Ghatak A, Ghatak S (1996) Budgetary deficits and Ricardian equivalence: the case
For undertaking the empirical analysis, we use the World Bank of India, 1950–1986. J Public Econ 60(2):267–282
Goyal A, Kumar A (2018) The effect of oil shocks and cyclicality in hiding Indian
data for the following variables. The study covers the period from twin deficits. J Econ Stud 45(no. 1):27–45
1985 to 2016 and is based on secondary data which are available Granger CW, Newbold P (1974) Spurious regressions in econometrics. J Econ 2
at: https://data.worldbank.org/country/china. (2):111–120
Hashemzadeh N, Wilson L (2006) The dynamics of current account and budget
deficits in selected countries if the Middle East and North. Afr Int Res J
Received: 21 May 2019 Accepted: 23 July 2019 Financ Econ 5:111–129
Herd R, Conway P, Hill S, Koen V, Chalaux T (2011) Can India achieve double-
digit growth?. OECD Economic Department Working Papers, (883), 0_1
Holmes MJ (2011) Threshold cointegration and the short-run dynamics of twin
deficit behaviour. Res Econ 65(3):271–277
Kaufmann S, Scharler J, Winckler G (2002) The Austrian current account deficit:
References driven by twin deficits or by intertemporal expenditure allocation? Empir
Abell JD (1990) Twin deficits during the 1980s: an empirical investigation. J Econ 27(3):529–542
Macroecon 12(1):81–96 Khalid AM (1996) Ricardian equivalence: empirical evidence from developing
Abbas SA, Bouhga-Hagbe J, Fatás A, Mauro P, Velloso RC (2011) Fiscal policy and economies. J Dev Econ 51(2):413–432
the current account. IMF Econ Rev 59(4):603–629 Khalid AM, Guan TW (1999) Causality tests of budget and current account def-
Afonso A, Huart F, Jalles JT, Stanek, PL (2018) Twin deficits revisited: a role for icits: cross-country comparisons. Empir Econ 24(3):389–402
fiscal institutions? REM working paper, 031-2018 Kim KH (1995) On the long-run determinants of the US trade balance: a comment.
Algieri B (2013) An empirical analysis of the nexus between external balance and J Post Keynes Econ 17(3):447–455
government budget balance: the case of the GIIPS countries. Econ Syst 37 Kim S, Roubini N (2008) Twin deficit or twin divergence? Fiscal policy, current
(2):233–253 account, and real exchange rate in the US. J Int Econ 74(2):362–383
Altintas H, Taban S (2011) Twin deficit problem and Feldstein-Horioka hypothesis Kulkarni KG, Erickson EL (2001) Twin deficit revisited: evidence from India, Pakistan
in Turkey: ARDL bound testing approach and investigation of causality. Int and Mexico. Journal of Applied Business Research (JABR) 17(2):(90–104)
Res J Financ Econ 74:30–45 Lau LJ (1990) Chinese economic reform: how far, how fast? In: Bruce J. Reynolds
Badinger H, Fichet de Clairfontaine A, Reuter WH (2017) Fiscal rules and twin (ed) Academic Press, New York, 1988. pp. viii+ 233
deficits: the link between fiscal and external balances. World Econ 40 Lau E, Tang TC (2009) Twin deficits in Cambodia: are there reasons for concern?
(1):21–35 An empirical study. Monash University, Deartment Economics, Disscussion
Bahmani-Oskooee M (1992) What are the long-run determinants of the US trade Papers 11(09):1–9
balance? J Post Keynes Econ 15(1):85–97 Laubach T (2003) New evidence on the interest rate effects of budget deficits and
Banday UJ, Aneja R (2015) The link between budget deficit and current account debt. Finance and Economics Discussion Series 2003-12, Board of Governors
deficit in Indian economy. Jindal J Bus Res 4(1-2):1–10 of the Federal Reserve System, May
Banday UJ, Aneja R (2016) How budget deficit and current account deficit are Leachman LL, Francis B (2002) Twin deficits: apparition or reality? Appl Econ
interrelated in Indian economy. Theor Appl Econ 23(1 (606):237–246 34:1121–1132
Banday UJ, Aneja R (2019) Ricardian equivalence: empirical evidences from China. Lee MJ, Ostry MJD, Prati MA, Ricci MLA, Milesi-Ferretti MGM (2008) Exchange rate
Asian Aff Am Rev 0(1–18):2019 assessments: CGER methodologies (No. 261). International Monetary Fund
Barro RJ (1989) The Ricardian approach to budget deficits. J Econ Perspect 3 Litsios I, Pilbeam K (2017) An empirical analysis of the nexus between investment,
(2):37–54 fiscal balances and current account balances in Greece, Portugal and Spain.
Basu S, Datta D (2005) Does fiscal deficit influence trade deficit?: an econometric Econ Model 63:143–152
enquiry. Economic and Political Weekly, 3311–3318 Miller SM, Russek FS (1989) Are the twin deficits really related? Contemp Econ
Bernheim BD (1988) Budget deficits and the balance of trade. Tax policy Econ Policy 7(4):91–115
2:1–31 Mukhtar T, Zakaria M, Ahmed M (2007) An empirical investigation for the twin
Bernheim BD, Bagwell K (1988) Is everything neutral? J Political Econ 96 deficits hypothesis in Pakistan. J Econ Coop 28(4):63–80
(2):308–338 Mundell RA (1963) Capital mobility and stabilization policy under fixed and
Bhat JA, Sharma NK (2018) The twin-deficit hypothesis: revisiting Indian economy flexible exchange rate. Can J Econ Political Sci 29(4):475–85
in a nonlinear framework. J Financ Econ Policy 10(3):386–405 Modigliani F, Ando AK (1957) Tests of the life cycle hypothesis of savings: com-
Boucher JL (1991) The US current account: a long and short run empirical per- ments and suggestions. Oxf Bull Econ Stat 19(2):99–124
spective. South Econ J 93–111 Modigliani F, Sterling A (1986) Government debt, government spending and
Burney NA, Akhtar N, Qadir G (1992) Government budget deficits and exchange private sector behavior: comment. Am Econ Rev 76(5):1168–1179
rate determination: evidence from Pakistan [with Comments]. Pak Dev Rev Morrison WM (2011) China-US trade issues. Curr Polit Econ North West Asia 20
31(4):871–882 (3):409
Castillo P, Barco D (2008) Facing up a sudden stop of capital flows: Policy lessons Nazier H, Essam M (2012) Empirical investigation of twin deficits hypothesis in
from the 90’s Peruvian experience (No. 2008-002) Egypt (1992–2010). Middle East Financ Econ J 17:45–58
Colm G, Young M (1968) In search of a new budget rule. Public Budgeting and Neaime S (2008) Twin Deficits in Lebanon: A Time Series Analysis, American
Finance: Readings in Theory and Practice. FE Peacock Publishers, Itasca, IL, University of Beirut Institute of Financial Economics, Lecture and Working
p 186–202 Paper Series, No. 2
Corsetti G, Müller GJ (2006) Twin deficits: squaring theory, evidence and common Perron P (1989) The great crash, the oil price shock, and the unit root hypothesis.
sense. Econ Policy 21(48):598–638 Econometrica: Journal of the Econometric Society 57(6):1361–1401

PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms 9

ARTICLE PALGRAVE COMMUNICATIONS | https://doi.org/10.1057/s41599-019-0304-z

Pesaran MH, Shin Y, Smith RJ (2001) Bounds testing approaches to the analysis of Author contributions
level relationships. J Appl Econ 16(3):289–326 UJB did the write-up and then analyzed the results. RA read the paper and made
Phillips PC, Perron P (1988) Testing for a unit root in time series regression. necessary corrections prior to submission. Both authors read and approved the
Biometrika 75:335–346 final paper.
Rafiq S (2010) Fiscal stance, the current account and the real exchange rate: some
empirical estimates from a time-varying framework. Struct Change Econ Dyn
21(4):276–290
Additional information
Competing interests: The authors declare no competing interests.
Rajasekar T, Deo M (2016) The Relationship Between Fiscal Deficit and Trade
Deficit in India: An Empirical Enquiry Using Time Series Data. The IUP
Reprints and permission information is available online at http://www.nature.com/
Journal of Applied 28 Economics 15(1)
reprints
Ratha A (2012) Twin deficits or distant cousins? Evidence from India1. South Asia
Econ J 13(1):51–68
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in
Rosensweig JA, Tallman EW (1993) Fiscal policy and trade adjustment: are the
published maps and institutional affiliations.
deficits really twins? Econ Inq 31(4):580–594
Rossini R, Quispe Z, Gondo R (2008) Macroeconomic implications of capital inflows:
Peru 1991–2007. In Participants in the meeting (p. 363–387), BIS Papers No 44
Salvatore D (2006) Twin deficits in the G-7 countries and global structural
imbalances. J Policy Model 28(6):701–712 Open Access This article is licensed under a Creative Commons
Schmidt D, Heilmann S (2010) Dealing with economic crisis in 2008-09: The Attribution 4.0 International License, which permits use, sharing,
Chinese Government’s crisis management in comparative perspective. China adaptation, distribution and reproduction in any medium or format, as long as you give
Anal 77:1–24 appropriate credit to the original author(s) and the source, provide a link to the Creative
Shen J, Chen B (1981) China’s Fiscal System. Economic Research Center, the State Commons license, and indicate if changes were made. The images or other third party
Council of the People’s Republic of China (Ed.), Almanac of China’s Econ- material in this article are included in the article’s Creative Commons license, unless
omy. Modern Cultural Company Limited, Hong Kong, p 634–654 indicated otherwise in a credit line to the material. If material is not included in the
Sobrino CR (2013) The twin deficits hypothesis and reverse causality: a short-run article’s Creative Commons license and your intended use is not permitted by statutory
analysis of Peru. J Econ Financ Adm Sci 18(34):9–15 regulation or exceeds the permitted use, you will need to obtain permission directly from
Tallman EW, Rosensweig JA (1991) Investigating US government and trade defi- the copyright holder. To view a copy of this license, visit http://creativecommons.org/
cits. Econ Rev-Fed Reserve Bank Atlanta 76(3):1 licenses/by/4.0/.
Tang TC, Lau E (2011) General equilibrium perspective on the twin deficits
hypothesis for the USA. Empir Econ Lett 10(3):245–251
Zivot E, Andrews DWK (1992) Further evidence on the Great Crash, the Oil Price © The Author(s) 2019
Shock, and the unit root hypothesis. J Bus Econ Stat 10:251–70

10 PALGRAVE COMMUNICATIONS | (2019)5:93 | https://doi.org/10.1057/s41599-019-0304-z | www.nature.com/palcomms