The effect of oil price shocks on the dynamic relationship between current account and exchange rate: evidence from D-8 Countries.
Qurat-ul-Ain, Syeda ; Tufail, Saira
The research aims to assess the dynamic relationship between
current account and exchange rate and to analyse the effect of oil price
changes on their relationship for D-8 countries. The research is based
on the time series analysis and covers the time span from 1981-2010. For
achieving the objective of the study, recursive Vector Autoregression
technique is used. Impulse response function and variance decomposition
analysis is also conducted to forecast the results for next ten years.
The results revealed that J-curve phenomenon exists in all oil importing
countries of the group. Among oil exporting countries, J-curve
phenomenon exists for Egypt and Nigeria while for Iran, Marshal Lerner
condition holds both in short and long run. The case of Malaysia is
opposite to that of Iran where depreciation could not stimulate current
account improvement even in long run. After including oil prices in the
model, J-curve phenomenon continues to exist in Bangladesh and Turkey,
though, it dampens the long run favourable effect of currency
depreciation for current account for both of the countries. For
Pakistan, in presence of oil prices exchange rate depreciation not only
deteriorates current account in short run, this deterioration
exacerbates in long run. Current account balance of Indonesia happens to
improve with depreciation of exchange rate after inclusion of oil prices
both in short and long run. For all oil exporting countries the role of
exchange rate for improving current account balance strengthens in long
run after the inclusion of oil prices. Given the results it is
recommended that oil exporting countries should diversify their exports
to overcome the recourse curse problem and oil importing countries
should consider Bangladesh as role model to reduce the vulnerability of
current account to oil price shocks.
JEL Classifications: Q43, Q48, F31
Keywords: J-Curve Phenomenon, Oil Price Shock, D-8, Vector Auto
Regression
1. INTRODUCTION
The effect of oil price shocks on global economy has been a great
concern since 1970s and has instigated a great deal of research
investigating macroeconomic consequences of oil price fluctuations.
Later on, the instability in the Middle East and recent oil price hike
confirmed the enduring significance of the issue. Though a voluminous
body of literature has evolved examining the bearings of oil prices for
internal sectors of economies [to name a few, e.g., Barsky and Kilian
(2004); Kilian (2008a,b); Hamilton (2008)], the studies analysing the
external sector response to oil price shocks are very few [see, e.g.
Kilian, et al. (2007)].
The determination of current account and exchange rate--the two
major indicators of external sector--has been studied widely in
theoretical and empirical literature but mostly the discussion of the
two variables largely remained separate [Lee and Chinn (1998)].
Similarly, investigation of simultaneous response of these two variables
to an oil price shock remained relatively less ventured avenue of
research. Initial work done on the relationship between current account
and oil price could not ascertain conclusive link between these two
variables. (1) Recent work on the issue revealed the diversity of
responses of current account of different countries to an oil price
shock. For instance, oil price increase deteriorates current account
balance of developing countries [OECD (2004); Rebucci and Spatafora
(2006); Killian, et al. (2007)] but may improve it if the country
happens to be a net oil-exporter. This implies that the relationship
depends on the number of factors among which oil dependency of country,
oil-intensity of production process (2) and responses of non-oil trade
balance (3) and sources of oil price fluctuations (4) are of particular
significance.
In this context exchange rate attains pivotal importance due to its
role for adjusting current account imbalances as advocated by both
traditional [Mundell (1962); Flemming (1962)] and advance open economy
macroeconomic approaches [Obstfeld and Rogoff (2000)] to current account
determination. However, the potency of exchange rate for smoothing
current account imbalances may be considerably affected in circumstances
where oil prices are volatile in nature. There exists a strand of
literature ascertaining the relationship between oil prices and exchange
rate for both oil importing and exporting countries. However, research
examining the effect of oil price innovations on the effectiveness of
exchange rate to lessen current account imbalances is in fact scant.
The paper bridges this gap by utilising the data for D-8 countries.
As a first step the existence of Marshal-Lerner condition and J-Curve
phenomenon is explored for each country. Following Lee and Chinn (2006)
a bivariate vector autoregressive model is employed as it minimises the
arbitrariness and helps to get several presumptions of open economy
macroeconomics validated with least possible restrictions. However,
unlike Lee and Chinn (2006) who employed reduced form model, our study
assumes identification by Cholesky factorisation considering exchange
rate is unaffected by contemporaneous innovations in current account.
This is justifiable as former is conducted for G-7 countries where the
exchange rate and current account are determined jointly, while later is
conducted for D-8 countries where assuming exchange rate relatively
exogenous seems more plausible. Given the information from the first
exercise, model is extended to allow the inclusion of oil prices to
achieve two objectives; (a) to examine the effect of oil prices on the
effectiveness of exchange rate to improve current account balance, and
(b) to examine the simultaneous response of both current account and
exchange rate to changes in oil prices. For both of these objectives
lower triangular identification scheme is followed ordering oil prices
ahead of exchange rate and current account.
The choice of countries is very critical to our objectives due to a
number of reasons. The countries not only differ as far as their trade
in oil is concerned, but also with respect to oil intensity of
production. Moreover, being the host of not only oil exporting (Iran,
Nigeria and Egypt) and importing countries (Pakistan, Turkey,
Bangladesh), but also countries transiting from being oil exporter to
importer (Indonesia and Malaysia), the group is expected to provide very
insightful and diverse outcomes for the targeted variables given the oil
price shock of same magnitude and direction. (5)
The rest of study is organised as follows. In Section 2, a review
of related literature is presented. In Section 3, descriptive analysis
of data is given. Section 4 reports the empirical results and Section 5
concludes the study.
2. RELATED LITERATURE
The relationship between current account balance and exchange rate
is explicitly established in elasticity approach to balance of payment
determination. Even the deviations from the basic model in the form of
Marshal-Lerner condition and J-Curve phenomenon could not prove the
authenticity of approach. Empirical evidence on this issue is not only
ample but also evolutionary. For instance, initial work on this issue
including Cooper (1971); Laffer (1974) and Salant (1974) provided
evidence in support of J-curve phenomenon using bivariate models of
exchange rate and trade balance.
However, according to Miles (1979) the inclusion of additional
determinants of trade balance and balance of payment nullified the
favourable contribution of exchange rate for trade balance while
Bahmani-Oaskooee (1985) reinforced the existence of J-curve phenomenon
even in a multivariate framework. Rose and Yellen (1989), Rose (1991)
conducted studies for both developed and developing countries by using
time series econometric techniques and could not find the evidence of
cointegrating relationship between exchange rate and current account.
Obstfeld and Rogoff (1995), on the other hand, assuming an infinite
horizon monetary model of monopolistically competitive world economy
showed that elasticity approach is valid if nominal prices in producer
country are rigid and exchange rate pass-through is complete. Recently,
by incorporating the standard assumptions of intertemporal macroeconomic
models in vector autoregression framework, Lee and Chinn (2006) showed
that the relationship between exchange rate and current account depends
largely on the nature of shocks. For instance, temporary shocks
depreciate the real exchange rate and improve current account balance
while permanent shocks though appreciate the exchange rate but the
effect on current account balance is not consistent.
Inclusion of oil prices in the modeling of exchange rate and
current account is not only in concordance of elasticity approach but
also consistent with both absorption and monetary approaches to balance
of payment determination. This eminence arises from the fact that oil
prices affect macroeconomy through a variety of channels most of which
either emanate from current account and exchange rate or have direct or
indirect effects on these variables. For instance, Lafer and Agmon
(1978) showed in context of monetary approach to balance of payment that
oil price shocks deteriorate trade balance markedly. This relationship
is also reported in OECD (2004); Killian, Rebucci and Spataforta (2007).
However, the size of the effect of oil price shock on trade balance is
subject to the response of non-oil trade balance to oil price shocks
[Lafer and Agmon (1978); Gruber and Kamin (2007)]. Amano and Norden
(1995), Backus and Crucini (2000) Chen and Rogoff (2003), Cashin, et al.
(2004) and Tokarick (2008) showed that effect of oil prices are
transmitted to exchange rate through changes in terms of trade.
According to Krugman (1983), Golub (1983) and Rasmussen and Roitman
(2011) this effect occurs through the transfer of wealth from oil
importing to exporting countries and is largely determined by the oil
dependence of oil importing and import patterns of oil exporting
countries. Recently, Bodenstein, et al. (2007: 2011) showed that in
order to stabilise the net foreign assets in face of positive oil price
shock exchange rate depreciates (appreciates) for oil importing
(exporting) countries. On the other hand, effect on current account
depends on the rate of depreciation of non-oil terms of trade and
adjustment of non-oil trade balance in face of oil price hike. The
magnitude of effect of oil price increase vitally depends on the level
of financial integration and efficiency of asset market.
The brief survey of literature strengthens the case for our
research as none of the study reported above takes into account the
relative effectiveness of exchange rate for adjusting current account
imbalances with and without oil price changes. Moreover, joint response
of current account and exchange rate for both oil importing and
exporting countries has not been assessed yet. The methodology to
address these issues is discussed in forthcoming sections.
3. MODEL SPECIFICATION
3.1. Model Construction
In present study two systems of equations are constructed to be
estimated by VAR. Initially two-variables-system including current
account (ca) and exchange rate (rexr) is constructed. Structural shock
includes one standard deviation positive shock to exchange rate in order
to observe the impact of current account. In the second step
three-variable system of equations is developed and oil prices (oil) are
included as third variable. The model is given as follows:
[X.sub.t] = A (L) [X.sub.t-1]+ U, ... ... ... ... ... ... (1)
Whereas for the first model X, is the 2 x 1 vector of endogenous
variables, i.e. [X.sub.t], [[rexr.sub.t], [ca.sub.t]]. A(L) is 2 x 2
matrix of lag polynomials and [U.sub.t] is the 2 x 1 vector reduced form
innovation, i.e., [U.sub.t] [equivalent to] [[u.sup.rexr.sub.t],
[u.sup.ca.sub.t]]. While for second model, [X.sub.t], is the 3 x 1
vector of endogenous variables, i.e. [X.sub.t] [oil, [rexr.sub.t],
[ca.sub.t]]. A(L) is 3 x 3 matrix of lag polynomials and [U.sub.t] is
the 3 x l vector reduced form innovation , i.e., [U.sub.t] [equivalent
to] [[[mu.sup.oil.sub.t], [u.sup.rexr.sub.t], [u.sup.ca.sub.t]. These
innovations are independently and identically distributed with variance
covariance matrix, where
E([[U.sub.t]) = 0;E([U.sub.t][U'.sub.t]) =
[summation][u.sub.t],
Amisano and Giannini (1997) suggested the following relationship
between reduced form and structural shocks in the form of AB-model:
A[U.sub.t] = B[V.sub.t] ... .... .... .... ..... .... .... (2)
[V.sub.t] are the structural shocks, whereas, A and B are 2 x 2 and
3x3 matrices for two models respectively, which show the instantaneous
relationship between variables and linear relationship between shocks
and reduced form innovation respectively. The remaining steps involved
in the construction of model are presented in Appendix A.
We employed recursive scheme of identification given the fact in
our system variables can be arranged according to degree of endogeniety.
In first system of equations including exchange rate and current
account, exchange rate is considered relatively more exogenous than
current account. This scheme of identification is considered more
appropriate for developing country due to its limited ability to affect
the value of dollar in international market. In the second system of
equation oil prices are considered most exogenous variable for both
oil-exporting and importing countries. Exchange rate is expected to be
effected by oil price and its own innovations. While current account is
considered to be affected by both exchange rate and oil prices and its
own innovations.
These identification schemes are presented as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
For three variable VAR it is given as:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Along with these short run restrictions, the same identification is
used for the long run restrictions.
3.2. Data Description
The annual data for D-8 countries; Bangladesh, Egypt, Indonesia,
Iran, Malaysia, Nigeria, Pakistan and Turkey from the year 1981 to 2011
is collected. The data set comprises of three main variables: current
account, real exchange rate and oil price. All data has been retrieved
from the World Development Indicators (2013) issued by World Bank,
except of world oil price. The data for oil price was retrieved from
International Financial Statistics issued by International Monetary
Fund. Exchange rate and oil price are taken in log form and current
account as percentage of GDP. The exchange rate is made real by
multiplying it with consumer price index (2005=100) of USA and dividing
it with consumer price index of each country.
4. ESTIMATION AND DISCUSSION OF RESULTS
4.1. Test of Stationarity
Due to the adoption of multiple exchange rate regimes and trade
reforms, it was intuitive to assume the presence of structural
instability in the exchange rate and current account balance for all D-8
countries. To affirm our assertion, the model for each country was
checked for structural stability using Chow break point test (results
are reported in Appendix B). Given the presence of significant
structural breaks for all countries, the power of conventional Augmented
Dickey Fuller test becomes dubious. In order to overcome this problem
Clemente, Montanes and Reyes (1998) test is applied that allows for two
structural breaks. By applying both innovative outlier and additive
outlier schemes, it was found that all series for each country are
integrated of order one, i.e. 1(1). The results are reported in Table 1.
4.2. Lag Order Selection
Schwartz information criterion (SIC) is used to select appropriate
lag length. The Table 2 shows appropriate lag length selected for model
with and without oil price for D8 countries.
4.3. Marshal Lerner Condition and J-Curve in D-8 Countries
Table 3 shows that J-curve phenomenon exists in all oil importing
countries of the group. Among oil exporting countries, J-curve
phenomenon exists for Egypt and Nigeria while for Iran Marshal Lerner
condition holds both in short and long run. The case of Malaysia is
opposite to that of Iran where depreciation could not stimulate current
account improvement even in long run.
After including oil prices in the model, J-curve phenomenon
continues to exist in Bangladesh and Turkey, though, it dampens the long
run favourable effect of depreciation for current account for both of
the countries. The case of Pakistan presents the extreme example of oil
price repercussions for the relationship between exchange rate and
current account. In presence of oil prices exchange rate depreciation
not only deteriorates current account in short run, this deterioration
exacerbates in long run. In contrast, for Indonesia the inclusion of oil
prices in the model makes the existence of Marshal Lemer condition
possible in both short and long run.
Many interesting results stand out when oil prices are included in
model for oil exporting countries. In short run, effectiveness of
exchange rate depreciation for current account improvement deteriorates
for Egypt, Iran and Malaysia respectively and increases for Nigeria.
However, for all exporting countries the role of exchange rate for
improving current account balance strengthens in long run after the
inclusion of oil prices. The improvement is significant for Malaysia,
which is 139 percent. These interesting results are well consistent with
the Malaysian policies of subsidising oil prices [Arshad and Shamsuddin
(2005)]. (8)
4.3.1. Impulse Response Functions and Variance Decomposition
The relationship between exchange rate and current account for both
models (with and without oil prices) is also forecasted with the help of
impulse response functions. These impulses are derived on the basis of
above specified identification scheme, in which Cholesky one-standard
deviation shocks are given to exchange rate and response of current
account balance is estimated over a period of ten years, 2012-2021,
following the initial occurrence of the shocks. The impulses for all
countries are plotted in Figure 1.
Among oil importing countries an obvious difference can be observed
in response of current account to one standard deviation positive shock
to exchange rate in models with and without oil price for Pakistan and
Turkey. This also holds for Egypt and Nigeria among oil exporting
countries. These results are also consistent with the exercise done and
results obtained in Section 4.3.
Along with derivation of impulse response function, variance
decomposition analysis is also conducted to analyse the contribution of
each shock to the variance of n-period ahead forecast error of the
variables. Table 4 presents the variance decomposition of current
account balance with and without oil prices. For all oil importing
countries, oil prices are contributing more than exchange rate in
forecasted error of current account balance. For Indonesia, Turkey and
Pakistan contribution of exchange rate reduces drastically after the
inclusion of oil prices in model. However, for Bangladesh contribution
of exchange rate after including oil prices remains almost same. This is
also the case of Nigeria among oil exporting countries. However, for
Egypt and Iran exchange rate is contributing more to the standard error
of current account balance as compared to oil prices.
[FIGURE 4.1 OMITTED]
4.4. Impact of Oil Prices on Exchange Rate and Current Account
The above exercise calls for further investigation of the issue by
analysing the response of exchange rate and current account to oil price
hike. Results are reported in Table 5. Increase in oil prices improves
current account balance for all oil importing countries in short run and
deteriorates it in long run except Bangladesh. It causes depreciation of
exchange rate for Indonesia, Pakistan and Turkey but appreciates the
exchange rate for Bangladesh in short run and other way round in long
run.
These results are supported by Wijnbergen (1984) who postulated
that oil price hike may induce recessionary pressures in oil importing
countries leading to investment cuts. This will lead to decreases in
demand of imported goods--mostly of which are energy and
capital--leading to temporary improvement in current account. These
improvements may take a permanent path depending on the availability of
alternative use of energy as in case of Bangladesh. The permanence of
improvement in current account also depends on the elasticity of
substitution between oil and other energy sources. This is also true for
Bangladesh where oil can be easily substituted with natural gas and
other non commercial sources of energy consumption. (9) Moreover,
Razzaqi and Sherbaz (2011) stated that growth of energy use is less than
growth of GDP for Bangladesh showing the less reliance of production
structure on oil and other sources of energy.
This fact is further supported by their findings that Bangladesh
has also experienced negative growth in the use of energy delineating
the highly elastic demand of energy with respect to energy prices. The
same argument can be put forward for exchange rate appreciation which is
occurring due to increase in oil price in Bangladesh. Unlike Bangladesh
current account position worsens after long run increase in oil price in
other oil importing countries. However, this worsening is insignificant
for Pakistan. This insignificance of oil price for current account
balance of Pakistan cannot be justified by the arguments posited for
Bangladesh. Contrary to Bangladesh, Pakistan has not specialised in
production of other sources of energy rather the results are pointing
toward the alarming situation in Pakistan. It is evident from results
that efforts to increase the investment or overcome the recessionary
shock of oil price hike are not enough in Pakistan leading to vicious
circle of poor investment declining demand for goods needed to encourage
investment leaving insignificant effect of oil price on current account.
On the other hand, all oil exporting countries experience
deterioration of current account in response to oil price shock both in
short and long run except Malaysia whose current account improves in
long run. For Egypt, with one percent increase in oil price current
account deteriorates by 1.67 percent and exchange rate appreciates by
0.06 percent. However, the effect of oil price on exchange rate merits
less consideration due to its insignificance. Even in long run though
insignificant yet negative effect of high oil prices prevails for both
current account and exchange rate. This means that country's oil
exports to world have not risen much as to compensate fully for rising
import bill leading to worsening of current account and appreciation of
exchange rate. However long run adverse effect of oil price is less
severe than its short run counterpart. The results are consistent with
the actual situation prevailing in the county as growth rate of oil
consumption has been more than that of production in the many years of
sample selected.
The effect of oil prices on current account in case of Iran is
similar to that of Egypt, however, for Iran growth rate of oil
production still exceeds than that of consumption. Farzanegan and
Markwardt (2009) providing more plausible reason for these results for
Iran. They showed that it's not the mounting import bill of oil as
compared to oil-export receipts which leads to current account worsening
rather these are the supply side wealth effects of increase in oil price
that stimulate real imports of variety of other goods leading to the
worsening of current account position in Iran. Morsy (2009) showed that
with the increase in oil price major oil exporting countries experience
surpluses that constitute an average of 23 percent of GDP. However,
given the increased wealth these countries spend significantly more on
imports of goods and services, amounting to an average of 37 percent of
GDP leading to the worsening of current account balance. Moreover,
appreciation of exchange rate due to long run increase in oil price is
providing strong evidence of Dutch disease phenomenon among oil
exporting countries.
4.4.1. Impulse Response
Exchange rate of all oil importing countries is depreciating
significantly in response to oil price shock except of Bangladesh whose
currency is appreciating insignificantly. As far as current account
balance is concerned, all oil importing countries are expected to
experience significant improvement in their current account balance with
one standard deviation shock to oil prices. Among oil exporting
countries, Malaysia's exchange rate is depreciating significantly,
however improvement in current account happens to be insignificant. For
all other oil exporting countries the effect of one time positive oil
price shock is appreciation of currency but insignificantly. However,
current account balance is deteriorating significantly in Iran and
Nigeria, insignificantly in Egypt. In Malaysia current account balance
is improving insignificantly in response to one standard deviation
positive shock to oil prices.
[FIGURE 4.2 OMITTED]
5. CONCLUSION AND RECOMMENDATIONS
The objective of this study is to explore the dynamic relationship
between current account and exchange rate and to analyse the effect of
oil price innovation on their relationship for D-8 countries. For
achieving this objective Vector Autoregression (VAR) approach is
employed. Impulse responses are also used to analyse the response of
current account to exchange rate shocks with and without oil price
innovations. A variance decomposition analyses is then conducted to
determine the contribution of exchange rate and oil price in the
forecasted errors of current account. The annual data for each country
is collected from 1981 to 2011 for current account, exchange rate and
oil price.
The results revealed that J-curve phenomenon exists in all oil
importing countries of the group. Among oil exporting countries, J-curve
phenomenon exists for Egypt and Nigeria while for Iran Marshal Lemer
condition holds both in short and long run. The case of Malaysia is
opposite to that of Iran where depreciation could not stimulate current
account improvement even in long run. After including oil prices in the
model, J-curve phenomenon continues to exist in Bangladesh and Turkey.
For Pakistan, in presence of oil prices exchange rate depreciation not
only deteriorates current account in short run, this deterioration
exacerbates in long run. Current account balance of Indonesia happens to
improve with depreciation of exchange rate after inclusion of oil prices
both in short and long run. For all oil exporting countries the role of
exchange rate for improving current account balance strengthens in long
run after the inclusion of oil prices.
As far as the effect of oil prices on exchange rate and current
account balance is concerned, increase in oil price improves current
account balance for all oil importing countries in short run and
deteriorates it in long run except Bangladesh. It causes depreciation of
exchange rate for Indonesia, Pakistan and Turkey but appreciates the
exchange rate for Bangladesh in short run and other way round in long
run. On the other hand, all oil exporting countries experience
deterioration of current account in response to oil price shock both in
short and long run except Malaysia whose current account improves in
long run. Moreover, appreciation of exchange rate due to long run
increase in oil price is providing strong evidence of Dutch disease
phenomenon among oil exporting countries.
The recommendations drawn from present study are that for the oil
exporting countries' exchange rate appreciates in face of oil price
hike which results in Dutch Disease phenomena. As current account
balance declines with exchange rate appreciations so these countries
should maintain stability in their exchange rates and they should
diversify their export base from oil to non-oil exports as well.
Nigeria, Iran and Egypt should reduce their dependence on oil and
natural resources and they should move towards industrial development as
well.
Bangladesh emerged as a role model for other oil importing and
developing countries through its results. Current account of Bangladesh
shows improving trend both in face of high oil price hike and with
exchange rate appreciation. It means Bangladesh has adopted alternative
resources and lowered its reliance on oil resources. Pakistan and Turkey
are oil importing countries; excessive increase in oil demand is causing
reserve depletion in these countries which in turn causes imbalance in
their current account. In order to improve current account balance,
these countries should lower their demand of crude oil by discovering
its alternatives like coal and gas reservoirs. These countries are also
in dire need of widening their export base through proper planning and
through building new infrastructure that can attract foreign investment
in these countries.
The research can be extended in number of ways. For instance, study
on exploring the transmission mechanism of oil prices to exchange rate
and analysing the relationship between exchange rate and current account
by incorporating exchange rate regimes and institutional and structural
changes taking place during the sample period can be undertaken.
Comments
The paper titled "The Effects of Oil Price Innovations on the
Dynamic Relationship between Current Account and Exchange Rate: Evidence
from D-8 Countries" is an excellent and systematic effort to
explore the relationship in the shocks in oil prices and their impact on
the current account and exchange rates in the D-8 countries. The case
study of D-8 countries is carefully selected so as to represent both the
oil exporting and oil importing countries and the relative dynamics
thereof.
However the following are some of my comments which the authors may
like to consider before the final submission of their papers:
(i) The title says oil price innovations.... in my opinion since
the authors have presented an empirical paper and it's not a pure
econometrics/statistics paper hence the term innovation which in
economics represents more of a controlled intervention sense, may be
changed to shocks.
(ii) The sample period is taken upto 1981-2010, since the paper
will be published in 2014, so if its not of a big hassle the authors may
like to increase the no of years to be taken as the sample period.
(iii) Since the authors are exploring the impact of shocks in one
of the components in the current account, a natural question arises
would the analysis change if it was some other component say for
developing countries the import of technology products, or the export of
the agro-based products. Or is it the oil specifically, then in this
case authors need to present a proper transmission mechanism of how the
shock in oil prices will lead to an impact. I am saying this because in
there results the results have some outliers on both the oil importing
and oil exporting countries. So may be we need to generalise some
results as to saying those imports which have a larger share in the
imports/exports would follow this pattern.
(iv) Since we are exploring the dynamics through the adjustments in
the exchange rates, there are a number of other institutional and
structural changes which have taken place over the sample period. Such
as (1) exchange rate regimes (authors have pointed it out in the
literature but not used it) (2) remittance from the oil exporting
countries to the oil importing countries being the origin country (3)
BOP controls such as capital account convertibility (4) active monetary
policy etc.
(v) Finally the difference in results across countries such as
Bangladesh being an outlier needs more clarification in terms of either
a theoretical justification or an administrative one.
The paper makes an interesting case and presents the results in
accordance with the theoretical understanding. Over all the paper is a
good contribution to the existing knowledge on the subject.
Mahmood Khalid
Pakistan Institute of Development Economics, Islamabad.
APPENDIX A
Recursive form of VAR can be obtained from reduced form by pre
multiplying equation 1 with A as
[AX.sub.t], = AA(L)[X.sub.t-1 + [AU.sub.t] ... ... ... ... ... (3)
Replacing A U, by BV, to get,
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
Summarised form of equation 5 can be written as:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
Equation 6 conveys autoregressive representation of the model in
which each variable is expressed as the function of the past values of
itself and of the other variables of the system. Secondly, it shows that
reduced form innovations are the linear combination of recursive
innovations.
In next step model is extended to allow for inclusion of oil prices
(oil). The above given steps are replicated and three variable system of
equation is constructed and final form is given as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
APPENDIX-B
Chow Break Point Stability Test
Log
F- Likelihood
statistics Probability Ratio Probability
Bangladesh (2003) 5.99 0.01 11.28 0.003
Egypt (1991) 9.14 0.00 15.98 0.000
Iran (1999) 5.99 0.08 5.73 0.05
Indonesia (1998) 8.31 0.0016 14.82 0.0006
Malaysia (1998) 5.52 *** 0.009 10.63 0.0049
Nigeria 4.92 0.001 9.63 0.01
Pakistan (2000) 12.64 0.000 20.38 0.000
Turkey (1989) 3.85 0.03 7.78 0.02
REFERENCES
Amano, R. A. and S. W. Norden (1998) Exchange Rates and Oil Prices.
Review of International Economics 6.
Amisano, G. and C. Giannini (1997) Topics in Structural VAR
Econometrics. Springer, Edition. 2.
Arshad, F. M. and M. N. Shamsudin (2005) Food Security in Malaysia,
http:// www.econ.upm.edu.my/~fatimah/foodsec.htm.
Bahmani-Oskooee, M. (1985) Devaluation and the J-Curve: Some
Evidence from LDCs. The Review of Economics and Statistics 67:3.
Barsky, R. B. and L. Kilian (2004) Oil and the Macroeconomy since
the 1970s. Journal of Economic Perspectives 18:4.
Buetzer, S., M. M. Habib, and L. Stracca (2012) Global Exchange
Rate Configurations: Do Oil Shocks Matter? European Central Bank.
(Working Paper).
Cashin, P., L. F. Cespedes, and R. Sahay (2004) Commodity
Currencies and the Real Exchange Rate. Journal of Development Economics
75.
Clemente, J., A. Montanes, and M. Reyes (1998) Testing for a Unit
Root in Variables with a Double Change in the Mean. Economics Letters
59.
Cooper, R. N. (1971) An Assessment of Currency Devaluation in
Developing Countries. In G. Ranis (ed.) Government and Economic
Development. New Haven: Yale University Press.
Farzanegan, M. R. and G. Markwardt (2009) The Effect of Oil Price
Shocks on the Iranian Economy. Energy Economics 31.
Golub, S. S. (1983) Oil Prices and Exchange Rates. The Economic
Journal 93.
Gruber, J. W. and S. B. Kamin (2007) Explaining the Global Pattern
of Current Account Imbalances. Journal of International Money and
Finance 26:4.
Hamilton, J. D. (2008) Oil and the Macroeconomy. In S. Durlauf and
L. Blume (eds.) The New Palgrave Dictionary of Economics Online.
Palgrave-Macmillan, London. December 31.
http://www.dictionaryofeconomics.com.
IMF (2000) Impact of Higher Oil Prices on Global Economy. Research
Department, IMF.
Kilian, L. (2008a) Exogenous oil Supply Shocks: How Big they are
and how much do they Matter for the U.S. Economy? Review of Economics
and Statistics 90.
Kilian, L. (2008b) The Economic Effects of Energy Price Shocks.
Journal of Economic Literature 46:4.
Kilian, L., A. Rebucci, and N. Spatafora (2007) Oil Shocks and
External Balances. (IMF Working Paper No. 110).
Krugman, P. (1983) Oil and the Dollar. (NBER Working Papers 0554).
Lafer, A. B. and T. Agmon (1978) Trade, Payments and Adjustment:
The Case of the Oil Price Rise. Kyklos 31.
Laffer, A. B. (1974) Exchange Rates, the Terms of Trade, and the
Trade Balance. In Effects of Exchange Rate Adjustments. Washington,
D.C.: Treasury Dept., OASIA.
Lee, J. and M. D. Chinn (1998) The Current Account and the Real
Exchange Rate: A Structural VAR Analysis of Major Currencies. (NBER
Working Paper No. 6495).
Lee, J. and M. D. Chinn (2006) Current Account and Real Exchange
Rate Dynamics in the G7 Countries. Journal of International Money and
Finance 25.
Licari, J. (1997) Economic Reform in Egypt in a Changing Global
Economy. OECD Development Centre. (Working Paper No. 129).
Marion, N. C. and L. E.O. Svensson (1984) Adjustment to Expected
and Unexpected Oil Price Changes. (NBER Working Paper 0997).
Miles, M. A (1979) The Effects uf Devaluation on the Trade Balance
and the Balance of Payments: Some New Results. Journal of Political
Economy 87.
Morsy, H. (2009) Current Account Determinants for Oil-Exporting
Countries. (IMF Working Paper).
Mundell, R. A. (1962) Capital Mobility and Stabilisation Policy
under Fixed and Flexible Exchange Rates. Canadian Journal Economics 29.
Obstfeld, M. and K. Rogoff (1995) Exchange Rate Dynamics Redux.
Journal of Political Economy 103:3.
OECD (2004) Oil Price Developments: Drivers, Economic Consequences
and Policy Responses. In OECD (ed.) OECD Economic Outlook No. 76.
Rasmussen, T. and A. Roitman (2011) Oil Shocks in a Global
Perspective: Are they Really that Bad? (IMF Working Paper WP/11/194).
Razzaqi, S. and S. Sherbaz (2012) Dynamic Relationship between
Energy and Economic Growth: Evidence from D8 Countries. The Pakistan
Development Review 50:4.
Rose, A. K. (1991) The Role of Exchange Rates in a Popular Model of
International Trade: Does the ' Marshal 1-Lemer' Condition
Hold? Journal of International Economics 30.
Rose, A. K. and J. L. Yellen (1989) Is There a J-curve? Journal of
Monetary Economics 24.
Salant, M. (1974) Devaluations Improve the Balance of Payments Even
if Not the Trade Balance. In Effects of Exchange Rates Adjustments.
Washington, D.C.: Treasury Dept., OASIA.
Tokarick, S. (2008) Commodity Currencies and the Real Exchange
Rate. Economic Letters 101, 60-62.
Wijnbergen, S. J. G. (1984) The 'Dutch Disease': A
Disease after All? Economic, Royal Economic Society 94:373,41-55.
(1) See, for instance, Lafer and Agmon (1978), Marion and Svensson
(1984).
(2) See, IMF (2000).
(3) See, Gruber and Ramin (2007).
(4) Buetzer, et al. (2012).
(5) In the analysis part Indonesia is treated as oil importing
country while Malaysia as oil exporting country.
(6) It's calculated as difference in coefficient of exchange
rate for current account with and without oil prices model as percentage
of coefficient of exchange rate for current account for model without
oil price. The exercise is done for both short and long run.
(7) A negative value is showing decrease in the effectiveness of
exchange rate for improving current account balance while a positive
sign is showing the percentage increase.
(8) The cost of oil price subsidy in Malaysia increases with the
increase in oil price which is not compensated by increased export
revenues. Subsidised oil prices also encourage oil consumption leading
to mounting oil bill and current account worsening. However, with a long
run increase in oil prices oil export revenues increase to more than
compensate initial mounted import bill leading to the existence of
J-curve phenomenon.
(9) About 66 percent of commercial energy demand is met by natural
gas and more than 50 percent of household energy demand is met by non
commercial resources.
Syeda Qurat-ul-Ain is Graduate Student, Department of Economics,
Fatima Jinnah Women University Rawalpindi. Saira Tufail
<beatingmind@gmail.com> is Lecturer, Department of Economics,
Fatima Jinnah Women University, Rawalpindi.
Table 1
Clemente-Montanes-Reyes Unit Root Test (Double Mean Shift)
Innovative Outlier
Level Difference
Country Variables (rho) Break (rho) Break
Bangladesh ca -0.83 1988, 2004 -1.79 ** 1988, 2004
lrexr -0.49 1998, 1995 -0.92 ** 1994, 2005
Egypt ca -0.7 1988, 1993 -1.74 ** 1988, 1993
lrexr -0.4 1984, 1988 -0.82 ** 1988, 1990
Iran ca -0.90 1989, 1992 -8.01 ** 1993, 1999
lrexr -0.37 1991, 2000 -0.89 ** 1992, 2000
Indonesia ca -1.0 1996, 2002 -3.0 ** 1996, 1999
lrexr -0.31 1997, 2000 -1.31 ** 1986, 1997
Malaysia ca -0.62 1986, 1996 -1.30 ** 1986, 1997
lrexr -1.0 1984, 1996 -1.54 ** 1991, 1997
Nigeria ca -1.01 1982, 2002 -1.69 ** 1992, 2004
lrexr -0.62 1984, 1997 -0.89 ** 1991, 1998
Pakistan ca -1.0 1982, 2002 -1.69 ** 1992, 2004
lrexr -0.43 1997, 2002 -1.02 ** 1994, 2000
Turkey ca -1.18 1986, 2003 -2.7 ** 2003, 2007
lrexr -0.67 1988, 2003 1.06 ** 1993, 2000
Additive Outlier
Level Difference
Country Variables (rho) Break (rho) Break
Bangladesh ca -1.0 1987, 2003 -1.96 ** 1984, 1993
lrexr -0.43 1995, 2003 -0.83 ** 1995, 2004
Egypt ca -0.6 1987, 1994 -1.43 ** 1992, 2001
lrexr -0.5 1986, 1992 -0.82 ** 1988, 1990
Iran ca -0.61 1989, 1995 -2.89 ** 1992, 1998
lrexr -1.0 1994, 2003 -1.8 ** 1994, 2004
Indonesia ca -0.91 1997, 2003 -1.87 ** 1996, 2000
lrexr -0.42 1995, 2003 -1.60 ** 1988, 1996
Malaysia ca -0.50 1995, 2000 -1.29 ** 1985, 2006
lrexr -0.76 1987, 1999 -1.05 ** 1990, 1996
Nigeria ca -1.0 1990, 2001 -1.70 ** 1991, 2007
lrexr -0.22 1988, 1998 -0.66 ** 1992, 1997
Pakistan ca -0.78 1990, 2001 -1.79 1991, 2007
lrexr -0.60 1986, 1998 1.67 ** 1997, 1999
Turkey ca -1.06 1986, 2003 -2.05 ** 2002, 2006
lrexr -0.79 1987, 2004 -1.2 ** 1992, 1999
** Denotes rejection of null hypothesis at 5 percent level of
significance.
Table 2
Lag Length Selection *
Without Oil With Oil
Countries Price Lags Price Lags
Bangladesh 1 1
Egypt 2 1
Iran 1 1
Indonesia 3 1
Malaysia 1 1
Nigeria 2 1
Pakistan 3 1
Turkey 2 1
* Selection is based on the minimum value of SIC.
Table 3
Marshal Lerner Condition and J-Curve in D-8 Countries
Bangladesh Without Oil Prices With Oil Price
Short Run Long Run Short Run Long Run
Ca Ca Ca Ca
Rexr -4.02 *** 6.09 *** -4.14 *** 4.79 ***
(-19.17) (29.83) (-20.27) (23.44)
Egypt
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -10.76 *** 4.38 *** -12.60 *** 9.39 ***
(56.94) (23.18) (-67.86) (50.57)
Iran
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr 0.3 * 1.49 *** -0.07 2.54 ***
(1.69) (8.07) (-0.42) 13.4
Indonesia
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -4.88 *** 3.63 *** 0.39 ** 2.54 ***
(-.25.36) (18.86) 2.11 13.4
Malaysia
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -18.96 *** -14.5 *** -19.75 ** 5.59 ***
(-102.08) (-78.18) (-106.36) (30.11)
Nigeria
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -8.609 *** 6.211 *** -1.905 ** 7.932 ***
(-45.55) (32.866) (-10.258) (42.715)
Pakistan
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -8.13 *** 6.82 *** -10.93 ** -18.28 **
(-42.29) (35.44) (-58.89) (-98.48)
Turkey
Short run Long run Short run Long run
Ca Ca Ca Ca
Rexr -14.69 *** 9.47 *** -7.88 ** 1.759 ***
(-77.74) (50.126) (-42.47) (9.47)
Bangladesh Percentage Change (6,7)
Short Run Long Run
Rexr -0.03 -0.21
Egypt
Rexr -0.17 114.38
Iran
Rexr -123.3 70.47
Indonesia
Rexr 107.99 -30.02
Malaysia
Rexr -4.17 138.55
Nigeria
Rexr 77.87 27.70
Pakistan
Rexr -34.4 -368.03
Turkey
Rexr 46.35 -81.42
*** Denote significance at 1 percent level, ** Denotes significance
at 5 percent level.
Table 4
Percentage Contribution of Exchange Rate in Standard Error of
Current Account Balance
Bangladesh
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.03 1.24 98.75
2 0.05 2.73 97.26
9 0.14 16.68 83.31
10 0.16 18.83 81.16
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.21 9.24 1.93 88.82
2 0.25 21.97 3.02 75.00
9 0.27 29.88 16.71 53.40
10 0.27 28.96 19.164 51.86
Egypt
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.12 33.77 66.22
2 0.19 58.88 41.11
9 0.23 69.31 30.68
10 0.23 69.63 30.36
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.26 5.23 48.28 46.48
2 0.35 4.44 46.12 49.42
9 0.56 5.59 42.25 52.15
10 0.57 5.86 42.15 51.98
Iran
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.61 0.23 99.76
2 0.85 3.76 96.23
9 1.40 21.96 78.03
10 1.43 22.87 77.12
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.22 20.52 0.014 79.46
2 0.28 18.91 4.046 77.03
9 0.35 18.96 21.63 59.40
10 0.35 18.95 22.46 58.57
Indonesia
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.16 15.78 84.21
2 0 18 27.06 72.93
9 0.20 36.67 63.32
10 0.20 36.67 63.32
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.25 15.03 0.13 84.82
2 0.29 18.35 1.68 80.00
9 0.41 31.44 11.20 57.35
10 0.41 31.84 11.22 56.925
Malaysia
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.06 9.37 90.62
2 0.07 7.56 92.43
9 0.08 10.90 89.09
10 0.08 10.92 89.07
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.22 0.11 5.34 94.53
2 0.29 0.12 4.27 95.56
9 0.36 1.69 7.78 90.52
10 0.36 1.76 7.78 90.44
Nigeria
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.33 14.21 85.78
2 0.47 20.03 79.96
9 0.58 19.87 80.12
10 0.58 19.87 80.12
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.23 39.06 0.55 60.37
2 0.28 34.96 0.51 64.51
9 0.33 34.06 1.05 64.87
10 0.33 34.08 1.09 64.81
Pakistan
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.04 14.80 85.19
2 0.07 50.45 49.54
9 0.12 60.33 39.60
10 0.12 60.41 39.58
Percentage Contribution to Standard Error of Current Account With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.22 28.04 3.07 68.88
2 0.27 35.43 2.34 62.22
9 0.33 40.90 1.96 57.12
10 0.33 40.91 1.96 57.12
Turkey
Percentage Contribution to Standard Error of Current Account Without
Oil Price
Forecasted Exchange Current
Period Standard Error Rate Account
1 0.11 63.56 36.43
2 0.15 65.29 34.70
9 0.24 76.14 23.85
10 0.24 76.33 23.66
Percentage Contribution to Standard Error of Current Accounts With
Oil Price
Forecasted Exchange Current
Period Standard Error Oil Price Rate Account
1 0.25 38.60 15.60 45.79
2 0.33 41.78 13.86 44.35
9 0.58 60.33 11.02 28.63
10 0.60 61.46 10.84 27.68
Source: Author's Calculations.
Table 5
Impact of Oil Prices on Current Account Balance and Exchange Rate
Bangladesh
Short-run Long-run
Ca Rexr Ca Rexr
Oil 1.684 *** -0.01 6.48 *** 0.97 ***
(8.25) (-0.04) (6.49) (4.76)
Egypt
Short-run Long-run
Ca Rexr Ca Rexr
Oil -1.67 *** -0.06 -0.81 -0.02
(-9.01) (-0.33) (-0.46) (-0-12)
Iran
Short-run Long-run
Ca Rexr Ca Rexr
Oil -8.07 *** 0.44 *** -0.69 ** -0.12
(-39.84) (2.34) (-1.96) (-0.64)
Indonesia
Short-run Long-run
Ca Rexr Ca Rexr
Oil 1.75 *** 0.27 -2.32 *** -0.39
8.91 1.44 -1.5 -2.09
Malaysia
Short-run Long-run
Ca Rexr Ca Rexr
Oil -2.12 *** 0.14 3.84 *** -0.16
(-11.31) (0.46) (3.64) (-0.84)
Nigeria
Short-run Long-run
Ca Rexr Ca Rexr
Oil -22.87 *** -0.153 -1.293 *** -0.79
(-121.00) (-0.824) (-0.871) -4.256
Pakistan
Short-run Long-run
Ca Rexr Ca Rexr
Oil 4.96 *** 0.07 -5.30 -0.28
(26.66) (038) (-1.56) (-1.50)
Turkey
Short-run Long-run
Ca Rexr Ca Rexr
Oil 2.98 *** 0.197 -3.267 *** -0.36 ***
(15.79) (1.06) (-8.69) (-1.977)
*** Denote significance at 1 percent level.
