Impact of oil fluctuations

Impact of oil fluctuations

Research Question

1. An external reserve is a self-insurance against capital flow volatility. The financial crises of the 1990s bring it to the fore, the need for countries to hold sufficient amount of reserves (Jeanne and Ranciere 2011). External reserves include foreign currencies gold, silver, Special Drawing Rights (SDRs) and International Monetary Fund (IMF). The behavior and what factors affect the behavior of external reserves have been discussed in literature. Some authors have attributed fluctuation in external reserves to change in exchange rate regime. However, Bahmani-Oskooee (1988) showed that oil price shock is a major reason for instability in external reserves. The reason for this is that increase in oil prices may result in jump in imports and the variability of BOP altering the reserve holding of Oil importing countries. Moreover, for many oil exporting countries, their source of external reserve is from oil revenue which means a fluctuation in oil price will affect their external reserves. It is against this background we set out to determine if crude oil price fluctuation affects external reserve and if the effect differs depending on whether the country is a net oil exporter. Specifically, this study seeks to address two questions: (i) whether and to what extent oil price fluctuations affect countries’ external reserves; and (ii) whether this effect depends on a country’s net oil exporting status

Abstract

This paper sought to present a critical discussion of the impact of oil fluctuations and how they affect the countries external reserve. A decline in oil prices has a positive and negative impact. Oil is the most traded material and it has a significant impact on the transport expenses, so it might lead to inflation which can impact on the rising rates of economic growth. The oils shocks do play a big role in determining the exchange market pressures in the case of exporting oil, which is the main reason for the floating currencies and is responsible for the rising and falling of the external reserve holdings. The research paper also looks at how the fluctuation of prices affects the net oil-exporting status, a decline in the oil prices is a joy to the nations that are importing oil while it hurts the oil exporters. If a nation experiences low oil prices for a long period then it might result in long term reductions of OPEC.

Literature Review

Studies have been carried out on the effect of oil price. One of such studies is by Bahmani-Oskooee (1988) who studied the effect of oil price shock on stability of demand for external reserves. His main research question was if a sharp and sudden increase in oil price could lead to a structural shift in the reserve demand function. The data for the study is time series data for 19 developed and 17 less developed countries for the period of 1973-1985. He used GLS estimation techniques to estimate a pooled regression model and them used Quandt’s log-likelihood ratio test and chow test to determine if there is structural change in demand pre and post crisis. His result shows that reserve demand function experienced structural shifts as a result of oil price shock. Another study was carried out on Nigeria, a less developed, African oil exporting country, using monthly data from 1995 to 2013 (Imarhiagbe 2015). The question the research is interested in answering is if there is any impact of crude oil prices on Nigeria’s external reserve. He used GARCH-M and EGARCH-M to estimate the model and found that oil price variability (volatility) has a positive impact on the volatility of external reserves. Bahmani-Oskooee and Brown (2004) used the Kalman filter approach to estimate demand for international reserves. They observed that due to fluctuations in exchange rates and oil prices the estimated elasticities are time-dependent. Therefore, there is need to use a technique that incorporates time-varying properties of coefficients into estimation procedure such as the Kalman filter. Using quarterly data  over 1959 to 1994 for 19 industrialized countries and a rolling window regression method to examine time-varying properties of elasticities, he found that elasticity of reserve demand have been unstable over time.

Data and Methodology

The data consists of time series data from 1995 to 2017 for 10 countries. The countries are China, UAE, Norway, Singapore, Saudi Arabia, South Africa, Australia, France, Canada and Nigeria. 5 of the countries are net oil exporter while five are not oil exporter. The data set include data on External Reserves, GDP at constant price, Import, Current account sourced from World Bank (data.worldbank.org) while the oil price is proxied by the Western Intermediate Select (WTI) petroleum price sourced from the LCDMA.

Model

Previous studies have modeled the demand for foreign reserves as follows

Where R is the real reserves held by a country at time t;M is the real imports; VR is a variability measure of balanceof payments and m is average propensity to import.It is usually expected that the higher the imports, the morereserves a country needs. Thus, estimate of  is expectedto be positive. The choice of variability measure of balanceof payments (VR) is the reserve demand function stemsdirectly from the role of reserves in serving as a bufferstock accommodating fluctuations in external transactions.Thus, the higher the VR, the more reserves will bedemanded yielding a positive estimate of . The rationalefor including the marginal or average propensity to import(m) stems from the Keynesian foreign trade multiplier.The higher the m, the smaller the multiplier which impliesa lower adjustment cost of an external imbalance, thusa lower demand for reserves. Thus an estimate of  isexpected to be negative. However, it is argued that if mstands for degree of openness of a country, an estimate of  could be positive (Bahmani-Oskooee and Brown 2014).

However, in line with the purpose of this study and given that the data is panel, I modify this reserve demand model by including oil price, an indicator whether a country is an net exporter of oil or not and the interaction of both variables. The model is re-specified as below.

Where OP represent oil price, oil is an indicator variable which equals 1 if the country is net oil exporter and 0 of the country is not a net oil exporter. VR is measured as change in current account balance and a constant is added to all observation in order for the logarithm to be possible in case of negative values.

Results

Table 1 presents the summary statistics of external reserves, import oil price and current account balance. The average reserves for all the country is $3146.22 (in billions) with a variation of $4448.98 (in billions).Within country variation in reserves is $3231.806 (in billions) while between country variation is $3216.779 (in billions). Average import is 281.87 (in billions) and the overall variation is $406.049 (in billions). Variation between countries is $293.499 (in millions) while variation within country is 295.04 (in millions). Average oil price over the year is $50.82 and the variation is $26.502. Between variation is 0 because the price of crude oil is the same for all countries and the within variation is $26.502. Average change in current account balance is $19.08 (in billions) and the overall variation is $62.939 (in billions). The between country change in current account balance is $45.679 (in billions) while the within country variation is $45.565 (in billions)

Table 1: Summary Statistics

Regression result

The first decision to make in a static panel data analysis is to decide between pooled regression, fixed effect regression and random effect regression. Pooled regression treats all units as the same and does not cater unit specific effect. The fixed effect model considers the unit specific effect but assume the effect is fixed. The random effect model treats this effect as random. Therefore to decide between these models, the first step is to estimate the fixed effect model and test for the presence of unit specific effect. The result shows that F(7, 184)=19.38, p<0.001 which means rejection of the null hypothesis that individual specific effect is 0. This means we cannot use pooled regression model. The next step after confirming the fixed effect is to determine if they are random. The Breusch-Pagan test LM was used for the result in table 2 shows that chi2(1)=275.91, p<0.001. This means that we reject the null hypothesis that random effect is not present. Now that we have confirmed the presence of fixed and random effect, we turn to Hausman test to decide between the two models. The Hausman test result presented in table 2 shows that chisquare(5) =4.05, p=0.5428. This means that, we cannot reject the null hypothesis that effect is not correlated with other regressors which connotes that fixed effect model is better than a random effect model.

Table 2: Model Specification Test

The regression result is presented in table 3. From the result, the coefficient of the dummy variable has omitted from the analysis due to the data. This connotes that demand for foreign reserve does not depend on whether a country is net exporter of oil or not. We also observe that log oil price have positive significant effect on reserves. 1% increase in oil price increases reserves by 114.66% (in billions). But this is not significant relationship (p = 0.283 > .05).This is in line with the findings of Imarhiagbe (2015) who also found positive effect of oil price on external reserves. Moreover, the coefficient of the interaction is also not significant (p = 0.572 > .05) and is negative. This connotes that the effect of oil price on external reserve is less for countries that are net oil exporter than countries that are not net oil exporter. The sign of coefficients of import and change in current account balance are consistent to what has been proposed in literature (Bahmani-Oskooee and Brown 2014). A 1% increase in import increases reserves by 0.95% (in billions) (p<0.001) while a 1% increase in balance of payment variation increases reserve by 110.78% (in billions) and a 1% increase in change in log of current account balance reserve by 159.05% (in billions). Time variable is not significant in the model (p < .05). Therefore there is no significant increase or decrease in reserve with time.

Table 3: Regression result using fixed effect model

Conclusion

This study examined the effect of oil price on external reserve and whether the effect depends on the country being net oil exporter or not. Using a panel data of 10 countries between 1995 and 2017, we showed that the effect of oil price on reserves depends on whether a country is net oil exporter or not and oil price have more effect on reserve for net oil exporter than non-net oil exporter. The parameter estimate of the traditional variables that have been included in reserve demand function was also found to be consistent with theory. This work is robust because individual country heterogeneity was considered in the model to cater for the bias that would have arisen from pooling together less developed and developed countries.

References

Bahmani-Oskooee, M. (1988). “Oil Price Shocks and Stability of theDemand for International Reserves” Journal of Macroeconomics, Fall 1988, Vol. 10, No. 4, pp. 633-641

Bahmani-Oskooee, M. and Brown, F. (2014).  “Kalman filter approach to estimate the demand for international reserves” Applied Economics, 36:15, 1655-1668, DOI: 10.1080/0003684042000218543.

Imarhiagbe, S. (2015). “Examining the Impact of Crude Oil Price on External Reserves:Evidence from Nigeria” International Journal of Economics and Finance; Vol. 7, No. 5; 2015

Jeanne, O.  andRanciere, R. (2011). “The Optimal Level of International Reserves forEmerging Market Countries: A New Formula andSome Applications” The Economic Journal, 121 (September), 905–930. Doi: 10.1111/j.1468-0297.2011.02435.x.