The technological advancements over the last decade have not only reduced the costs associated with the production of unconventional oil, but also made extraction of tight oil resemble a manufacturing process in which one can adjust production in response to price changes with relative ease. This is in stark contrast to other extraction methods (e.g. offshore extraction), which require large capital expenditure and involve relatively long lead times, and more importantly, once the process is operational changing the quantity produced can be difficult. Therefore, one of the implications of the recent oil revolution is that U.S. production can play a significant role in balancing global demand and supply, and this in turn implies that the current low oil price environment could be persistent.
This paper investigates the macroeconomic consequences of the U.S. oil revolution for the global economy in general and the Middle East and North Africa (MENA) region in particular in terms of its effects on real output, oil prices and financial markets. We integrate an oil price equation, which takes account of developments in the world economy as well as the prevailing oil supply conditions, within a compact quarterly model of the global economy using a dynamic multi-country framework first advanced by Pesaran et al. (2004), known as the Global VAR (or GVAR for short). This approach enables one to analyze the international macroeconomic transmission of shocks, taking into account not only the direct exposure of countries to the shocks but also the indirect effects through secondary or tertiary channels. To discriminate the U.S. oil revolution from other supply shocks, such as disruptions caused by geopolitical tensions in the Middle-East, and oil-demand shocks in general, we employ a set of dynamic sign restrictions on the impulse responses of our GVAR-Oil model. In addition to restricting oil prices and production levels, the global dimension of the GVAR-Oil model offers an intuitive way of imposing a large number of additional cross-country sign restrictions that greatly reduces the number of admissible structural models.
Our dynamic multi-country framework consists of 38 country/region-specific models, among which is a single Euro Area region (including 8 of the 11 countries that joined Euro in 1999) as well as the countries of the Gulf Cooperation Council (GCC). These individual models are solved in a global setting where core macroeconomic variables of each economy are related to corresponding foreign variables—which have been constructed to match the international trade pattern of the country under consideration and serve as a proxy for common unobserved factors. The model has both real and financial variables: real GDP, inflation, real equity prices, real exchange rate, short and long-term interest rates, OPEC and non-OPEC oil production, and the price of oil. Our framework is able to account for various transmission channels, including not only trade relationships but also financial linkages through interest rates, equity prices, and exchange rates; see Dees et al. (2007) and Pesaran et al. (2007). We estimate the oil price equation and the 38 individual vector autoregressive models with foreign variables (VARX* models) over the period 1979Q2–2011Q2. Having combined the estimates from the oil price equation with those of the country-specific VARX* models, we solve the GVAR-Oil model and examine the effects of a U.S. oil-supply shock (while keeping the level of oil supply in Saudi Arabia constant) on the macroeconomic variables of different countries (both commodity importers and exporters), including the MENA region.
The results indicate that while oil importers typically face a long-lived rise in economic activity (ranging between 0.04% and 0.95%) in response to a U.S. supply-driven fall in oil prices, the impact is negative for energy-exporters (being on average –2.14% for the GCC, – 1.32% for other MENA oil exporters, and –0.41% for Latin America), mainly because lower oil prices weakens domestic demand as well as external and fiscal balances in these countries. To investigate the channels through which the fall in oil revenues affects oil exporters (as well as select oil importers), especially in the long run, and quantify its growth impact, we embed the long-run output relation of Esfahani et al. (2014) in individual VARX* models. Our results indicate that oil revenue shocks (such as those from the low oil price environment we are currently experiencing) have a large, long-lasting and significant impact on these economies’ growth paths operating through the capital accumulation channel.
Negative growth effects (albeit smaller) are also observed for energy-importers which have strong economic ties with oil exporters, through spillover effects. In particular, for most oil-importers in the MENA region, gains from lower oil prices are offset by a decline in external demand/financing by MENA oil-exporters given strong linkages between the two groups through trade, remittances, tourism, foreign direct investment and grants.1 These economies on average experience a fall in real output of about 0.28%. For this group, low pass-through from global oil prices to domestic fuel prices limits the impact on disposable incomes of consumers and profit-margins of firms, and thereby contains the positive effect on economic growth in these countries.
Finally, in response to a positive U.S. oil-supply disturbance, almost all countries in our sample experience long-run disinflation pressures and an increase in equity prices (apart from commodity-exporting nations). Overall, our results suggest that following the U.S. oil revolution, with oil prices falling by 51% in the first year and rebounding somewhat to 45% in year two below the pre-shock levels, global growth increases by 0.16 – 0.37 percentage points. This is mainly due to an increase in spending by oil importers which exceeds the decline in expenditure by oil exporters.
The collapse of oil prices from around $114 in June 2014 to $46 in January 2015,2 has led to a large body of literature analyzing the causes of this steep oil price drop and its macroeconomic implications. However, most of this literature is based on descriptive analysis, mainly written by international organizations (see, for instance, the IMF blog by Arezki and Blanchard, 2014), investment banks (such as Goldman Sachs Global Investment Research division’s report on “The New Oil Order”), various (energy) economists, and of course mostly internal reports by oil and gas companies (which are used to inform exploration, development, and hiring decision to name a few). There are yet only a handful of papers, which apply rigorous and quantitative analysis of the recent oil price shock. Most notably, Baumeister and Kilian (2015) argue that demand factors were most important in explaining the behavior of oil prices, while Baffes et al. (2015), Husain et al. (2015), and Mânescu and Nuño (2015) argue that supply (rather than demand) factors played the largest role.
More broadly, most papers in the literature that investigate the effects of oil shocks on macroeconomic variables have focused on a handful of industrialized/OECD countries, and in most cases they have looked at the impact of oil shocks exclusively on the United States and in isolation from the rest of the world. Moreover, the focus of those analyses has predominantly been on net oil importers—see, for example, Hamilton (2009), Kilian (2009), and Peersman and Van Robays (2012). An exception is the work of Cashin et al. (2014), who look at the differential effects of oil demand and supply shocks on the global economy, Esfahani et al. (2014), who conduct a country-by-country VARX* analysis looking at the direct effects of oil-revenue shocks on domestic output for 9 major oil exporters (six of which are OPEC members), Kilian et al. (2009), who examine the effects of different types of oil-price shocks on the external balances of net oil exporters/importers, and Mohaddes and Pesaran (2015), who examine the effects of country-specific shocks (to Iranian and Saudi Arabian oil output) on the world economy.
In this paper, we extend the literature in a number of respects. Firstly, our paper is complementary to the analysis of the effects of oil-price shocks on advanced economies, given its wide country coverage, including both major oil exporters (located in the Middle East, Africa and Latin America) as well as many developing countries. We are therefore able to analyze the macroeconomic consequences of U.S. supply-driven oil-price shocks across a wide range of developed and developing countries (including oil exporters) that are structurally very diverse with respect to the role of oil and other forms of energy in their economies. Secondly, we provide a compact model of the world economy that takes into account the economic interlinkages and spillovers (direct exposure of countries to the shocks but also the indirect effects through secondary or tertiary channels) that exist between different regions (which may also shape the responses of different macroeconomic variables to oil price shocks), rather than undertaking a descriptive analysis or a country-by-country structural VAR study of the oil market. Thirdly, we include oil production endogenously in the U.S. and Saudi Arabian models, while modelling oil prices as determined in the global oil market. This is required to answer counterfactual questions regarding the possible macroeconomic effects of the U.S. oil revolution. Finally, we demonstrate how our GVAR-Oil model, covering over 90% of world GDP, 85% of world oil consumption, and 80% of world proven oil reserves, can be used for “set-identified” impulse response analysis and to obtain a better understanding of structural shocks. In particular, we set-identify the U.S. oil supply shock3 by imposing dynamic sign restrictions on the impulse responses of oil production in the United States, Saudi Arabian oil supply, and GDP of major oil importers in our sample.4
The rest of the paper is organized as follows. Section II. describes the GVAR methodology, outlines our model specifications, and illustrates how we integrate the oil market within our framework. Section III. provides the estimates for the country-specific models, presents our identification strategy, and examines the direct and indirect effects of shocks to U.S. oil output on the world economy, on a country-by-country basis, and provide the time profile of the effects of country-specific oil shocks on real outputs, inflation, and real equity prices across countries. Section IV. investigates in greater detail the macroeconomic implications of the U.S. oil supply revolution, in terms of its real GDP effects, on individual countries in the MENA region over the short and long-term. Finally, Section V. offers some concluding remarks.
II. Modelling the Oil-Macroeconomy Relationship in a Global Context
To analyze the international macroeconomic transmission of the U.S. oil revolution we need to model the oil-macroeconomy relationship in a global context. To this end we integrate an oil price equation within a compact quarterly model of the global economy using the GVAR framework. The resulting GVAR-Oil model takes into account both the temporal and cross-sectional dimensions of the data; real and financial drivers of economic activity; interlinkages and spillovers that exist between different regions; and the effects of unobserved or observed common factors. This is crucial as the impact of the recent oil revolution cannot be reduced to just the United States (where the shock originates) but rather involves multiple regions, and may be amplified or dampened (through a number of channels) depending on the degree of openness of the countries and their trade structure. Before describing our approach in modeling individual countries and the global oil market, we provide a short exposition of the GVAR methodology below.
A. The Global VAR (GVAR) Methodology
We consider N countries in the global economy, indexed by i = 1, …, N. With the exception of the United States, all other N—1 countries are modelled as small open economies. This set of individual country-specific vector autoregressive models with foreign variables (VARX* models) is used to build the GVAR framework. Following Pesaran (2004) and Dees et al. (2007), a VARX* (pi, qi) model for the ith country relates a ki × 1 vector of domestic macroeconomic variables (treated as endogenous), xit, to a
for t = 1, 2, …, T, where ai0 and ai1 are ki × 1 vectors of fixed intercepts and coefficients on the deterministic time trends, respectively, and uit is a ki × 1 vector of country-specific shocks, which we assume are serially uncorrelated with zero mean and a non-singular covariance matrix, Σii, namely uit ∼ i.i.d. (0, Σ). For algebraic simplicity, we abstract from observed global factors in the country-specific VARX* models. Furthermore,
The country-specific foreign variables are constructed as cross-sectional averages of the domestic variables using data on bilateral trade as the weights, wij
where j = 1, 2, …N, wii = 0, and
where Tijt is the bilateral trade of country i with country j during a given year t and is calculated as the average of exports and imports of country i with j, and
Although estimation is done on a country-by-country basis, the GVAR model is solved for the world as a whole, taking account of the fact that all variables are endogenous to the system as a whole. After estimating each country VARX*(pi, qi) model separately, all the
for i = 1, … N, where
Note that given equation (2) we can write
where Wi = (Wi, Wi2, WiN), with Wii = 0, is the
where Ai (L, p) is constructed from Ai (L, pi, qi) by setting p = max (p1, p2, …,pN, q1, q2, qN) and augmenting the p – pi or p – qi additional terms in the power of the lag operator by zeros. Stacking equation (7), we obtain the Global VAR(p) model in domestic variables only
For an early illustration of the solution of the GVAR model, using a VARX*(1, 1) model, see Pesaran (2004), and for an extensive survey of the latest developments in GVAR modeling, both the theoretical foundations of the approach and its numerous empirical applications, see Chudik and Pesaran (2015). The GVAR(p) model in equation (8) can be solved recursively and used for a number of purposes, such as forecasting or impulse response analysis.
B. Country-Specific VARX* Models
We include as many major oil exporters as possible in our multi-country set up, subject to data availability, together with as many countries in the world to represent the global economy. Thus our version of the GVAR model covers 50 countries as opposed to the “standard” 33 country set-up used in the literature, see Smith and Galesi (2014), and extends the coverage both in terms of major oil exporters and also by including an important region of the world when it comes to oil supply, the MENA region.6
Of the 50 countries included in our sample, 18 are classified as major commodity exporters as primary commodities constitute more than 40 percent of their exports (these countries are denoted by * in Table 1). Moreover, 15 are net oil exporters of which 10 are current members of the OPEC (denoted by 1 in Table 1) and one is a former member (Indonesia left OPEC in January 2009). We were not able to include Angola and Iraq, the remaining two OPEC members, due to the lack of sufficiently long time series data. This was also the case for Russia, the second-largest oil exporter in the world, for which quarterly data is not available for the majority of our sample period. Our sample also includes three OECD oil exporters (Canada, Mexico, and Norway) and the UK, which remained a net oil exporter for the majority of the sample (until 2006), and therefore is treated as an oil exporter when it comes to imposing sign-restrictions (see the discussion in Section III.A.). These 50 countries together cover over 90% of world GDP, 85% of world oil consumption, and 80% of world proven oil reserves. Thus our sample is rather comprehensive.
|Systemic Countries||MENA Oil Exporters||Latin America|
|Japan||MENA Oil Importers||Malaysia|
|Rest of Advanced Economies||Morocco*|
|Australia||Syria*2||Rest of the World|
|New Zealand||Turkey||South Africa*|
For empirical applications, we create two regions; one of which comprises the six Gulf Cooperation Council (GCC) countries: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates (UAE); and the other is the Euro Area block comprising 8 of the 11 countries that initially joined the euro on January 1, 1999: Austria, Belgium, Finland, France, Germany, Italy, Netherlands, and Spain. The time series data for the GCC block and the Euro Area block are constructed as cross-sectionally weighted averages of the domestic variables (described in detail below), using Purchasing Power Parity GDP weights, averaged over the 2007-2009 period. Thus, as displayed in Table 1, our model includes 38 country/region-specific VARX* models.
Making one region out of Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates, is not without economic reasoning. The rationale is that these countries have in recent decades implemented a number of policies and initiatives to foster economic and financial integration in the region with a view to establishing a monetary union (loosely based on that of the Euro Area). Abstracting from their level of success with above objectives, the states of the GCC are relatively similar in structure, though in the short term they may face some difficulties in meeting the convergence criteria they have set for economic integration based on those of the European Union (EU). Inflation rates vary significantly across these countries and fiscal deficits, which have improved since the start of the oil boom in 2003, are about to re-emerge in some countries. However, these economies already peg their currencies to the U.S. dollar, except for Kuwait, which uses a dollar-dominated basket of currencies, and are accustomed to outsourcing their interest-rate policy. They also have relatively open capital accounts, and hence, it is reasonable to group these countries as one region.7
We specify two different sets of individual country-specific models. The first model is common across all countries, apart from the United States. These 37 VARX* models include a maximum of six domestic variables (depending on whether data on a particular variable is available), or using the same terminology as in equation (1)
where yit is the log of the real Gross Domestic Product at time t for country i, πit is inflation, eqit is the log of real equity prices,
Following the GVAR literature, the thirty-eight model (United States) is specified differently, mainly because of the dominance of the United States in the world economy. First, given the importance of U.S. financial variables in the global economy, the U.S.-specific foreign financial variables,
C. The Global Oil Market
Given that we want to consider the macroeconomic effects of the U.S. oil revolution, we also need to include nominal oil prices in U.S. dollars in the country-specific VARX* models. If we follow the literature, we would include log oil prices,
To add oil prices to the conditional country models we simply augment the VARX* models (1) by pot and its lag values
which is a standard autoregressive distributed lag, ARDL(mpo, my, mqo), model in oil prices, world real income (yt) and world oil supplies (
To take into account developments in the world economy, the oil price equation includes a measure of global output, yt, calculated as
where yjt is the log of real GDP of country j at time t,
To capture global oil supply conditions we have also included a measure for the quantity of oil produced in the world in equation (13). A key question is how should
Figure 1.Oil Production in million barrels per day, 2005M1–2015M3
Source: U.S. Energy Information Administration Monthly Energy Review.
We now turn to non-OPEC oil supply. As Figure 1 shows the increase in non-OPEC production over the last decade is more or less the result of the oil revolution which has increased U.S. production by 50% (from approximately 6 mbd to 9 mbd). The recent technological advancements has not only reduced the costs associated with the production of tight oil, but it has also made the extraction resemble a manufacturing process in which the quantity produced can be altered in response to price changes with relatively ease, which is not the case for conventional oil extraction which requires large capital expenditure and lead times. In other words, U.S. oil production can play a significant role in balancing global demand and supply. Given the developments in the last decade, we model non-OPEC oil production within the U.S. model.
III. Empirical Results
We obtain data on xit for 33 out of the 50 countries included in our sample (see Table 1) from the GVAR website: https://sites.google.com/site/gvarmodelling, see Smith and Galesi (2014) for more details. Data for the remaining 17 countries: Algeria, Bahrain, Ecuador, Egypt, Iran, Jordan, Kuwait, Libya, Mauritania, Morocco, Nigeria, Oman, Qatar, Syria, Tunisia, Venezuela, and the UAE are from Cashin et al. (2012). Oil price data are also from the GVAR website, while data on oil production are from the U.S. Energy Information Administration Monthly Energy Review.
We use quarterly observations over the period 1979Q2–2011Q2 to estimate the 38 country-specific VARX*(pi, qi) models. However, prior to estimation, we determine the lag orders of the domestic and foreign variables, pi and qi. For this purpose, we use the Akaike Information Criterion (AIC) applied to the underlying unrestricted VARX* models. Given data constraints, we set the maximum lag orders to pmax = qmax = 2. The selected VARX* orders are reported in Table 2. Moreover, for the lag order of the ARDL(mpo, my, mqo) model in oil prices, world real income, and world oil supplies AIC selects mpo = my = mqo = 2.
|Country||VARX* Order||Cointegrating relations ||Country||VARX* Order||Cointegrating relations |
Having established the lag order of the 38 VARX* models, we proceed to determine the number of long-run relations. Cointegration tests with the null hypothesis of no cointegration, one cointegrating relation, and so on are carried out using Johansen’s maximal eigenvalue and trace statistics as developed in Pesaran et al. (2000) for models with weakly exogenous I (1) regressors, unrestricted intercepts and restricted trend coefficients. We choose the number of cointegrating relations (ri) using the trace test statistics based on the 95% critical values from MacKinnon (1991). We then consider the effects of system-wide shocks on the exactly-identified cointegrating vectors using persistence profiles developed by Lee and Pesaran (1993) and Pesaran and Shin (1996). On impact the persistence profiles (PPs) are normalized to take the value of unity, but the rate at which they tend to zero provides information on the speed with which equilibrium correction takes place in response to shocks. The PPs could initially over-shoot, thus exceeding unity, but must eventually tend to zero if the vector under consideration is indeed cointegrated. In our analysis of the PPs, we noticed that the speed of convergence was very slow for Algeria, Canada, China, Iran, Korea, and Tunisia and for a few of them the system-wide shocks never really died out, so we reduced ri by one for each country, except for Korea for which we reduced
A. Identification Strategy
To discriminate oil-supply shocks due to the U.S. oil revolution from other supply shocks, such as disruptions caused by geopolitical tensions in the Middle-East, and oil-demand shocks in general, we rely on two sets of identifying restrictions within our GVAR-Oil framework: (a) dynamic sign restrictions and (b) cross-country sign restrictions arising from the global dimension of the GVAR-Oil model. Regarding these two conditions, we require the oil revolution to be associated with: (i) a decrease in oil prices; (ii) an increase in the level of U.S. oil production; (iii) a constant OPEC oil production; and (iv) an increase in the sum of real GDPs across all major oil importers in our sample. Since the effect of a positive oil-supply shock on the level of GDP of major oil and commodity exporters (for which primary commodities constitute more than 40 percent of their exports) in our sample is ambiguous, we do not impose any dynamic sign restrictions on them. Moreover, we do not impose any restriction on the GDP for Jordan as Mohaddes and Raissi (2013) show that for an oil-importing but labor-exporting small open economy which receives large (and stable) inflows of external income (the sum of FDI, remittances, and grants) from oil-rich countries, the impact of oil shocks on the economy’s macroeconomic variables can be very similar to those of the oil exporters from which it receives these large income flows. Note that other than yit we do not impose any restrictions on the remaining variables in xit, that is inflation, the real exchange rate, equity prices and both the short and long-term interest rates.
We impose these sign restrictions, (i) to (iv), to hold for one year after the shock to allow for sluggish responses of quantity measures (oil production and real GDPs). This scheme is effective in identifying oil-supply disturbances as other shocks cannot move oil prices, oil production levels, and real GDPs (across all oil-importing countries) in opposite directions. We should stress that while the quantity restrictions help with the identification of supply shocks, the global dimension of the GVAR model offer an intuitive way of imposing a large number of additional sign restrictions and can therefore greatly reduce the number of admissible models to better identify the shock. Specifically, condition (iv) imposes that the cumulated sum of the relevant individual-country outputs are positive faced with a U.S. oil-supply shock.12 Intuitively, this positive oil supply shock is perceived to be a tax reduction on oil consumers (with a high propensity to consume) at the expense of oil producers (with a lower propensity to consume) and is associated with an increase in global aggregate demand (hence the cross-country restrictions).
Given these identifying restrictions, the implementation procedure is as follows. Let vit denote the structural VARX* model innovations given by
More compactly, we construct the k × k matrix
which can be used to obtain the impulse responses of all endogenous variables in the GVAR-Oil model to shocks to the error terms
Since there are a few impulse responses that satisfy our postulated identifying restrictions, we summarize them by reporting a central tendency and the 5th and 95th percentiles as measures of the spread of responses. Although the remaining models—after imposing identifying restrictions (a) to (b)—imply qualitatively and sometimes quantitatively similar responses, the central tendency measure (i.e. median) for impulse responses of different variables (across the 38 countries/regions) may come from different impulse vectors. We therefore follow Fry and Pagan (2011) and report a single model whose impulse responses are as close to the median values of the impulse vector as possible (this is called the median target). It is important to recognize that the distribution here is across different models and it has nothing to do with sampling uncertainty.
B. The Macroeconomic Effects of the U.S. Oil Revolution
Figures 2–4 show the estimated median (blue solid) and the median target (black long-dashed) impulse responses (for up to ten years) of key macroeconomic variables of oil exporters and oil-importing countries to a supply-driven oil-price shock (emanating from the oil revolution in the United States), together with the 5th and 95th percentile error bands.14 The economic consequences of a positive oil-supply shock in the U.S., equivalent to a 51% fall in the oil prices after one year, are very different for oil-importing countries compared to energy exporters. With regard to real output, following the U.S. oil-supply shock, Euro Area and the U.S. (two major energy-importing countries) experience a long-lived boost to economic activity—0.56% and 0.60%, respectively—while similar responses are observed for the UK (a former oil exporter) and other advanced countries, being on average 0.57% and 0.42%, respectively.
Figure 2.Impact of the U.S. Oil Supply Revolution on Real Output
Notes: Figures are median (blue solid) and median target (black long-dashed) impulse responses to a one standard deviation fall in the price of oil, equivalent to an annualized drop of 51% in year 1 and 45% in year 2, together with the 5th and 95th percentile error bands. The impact is in percentage points and the horizon is quarterly.
Figure 3.Impact of the U.S. Oil Supply Revolution on Inflation
Notes: See notes to Figure 2.
Figure 4.Impact of the U.S. Oil Supply Revolution on Equity Markets
Notes: See notes to Figure 2.
Our framework takes into account not only the direct exposure of countries to the oil shock but also the indirect effects through secondary or tertiary channels. For instance, as a result of the dominance of the United States in the world economy, any increase (or decrease) in economic activity in this country can bring about positive (or negative) spillovers to other economies, as the recent global economic crisis has shown. More generally, the history of past U.S. recessions usually coincides with significant reductions in global growth. Furthermore, the continuing dominance of U.S. debt and equity markets, backed by the still-strong global role of the U.S. dollar, is also playing an important role. This is clearly illustrated in Figure 2, where what was initially an increase in domestic output due to lower oil prices, translates into a pickup in economic activity even in the medium term due to spillovers through trade and financial channels. However, these spillovers do vary greatly from country to country and depend on, for instance, a particular country’s trade exposure to the U.S. or the other advanced economies.
The GDP impact is also positive for most Asian countries (for instance, the South East Asia region experiences on average a long-run boost of 0.71%) apart from China where the median target response is negative initially, but becomes positive and around 0.04% over the long-term. However, given China’s heavy dependence on coal, as opposed to oil, for its energy consumption needs and the composition of its export basket, this result might not be that surprising after all. The United States (Euro Area) met 36% (38%) and 20% (13%) of its primary energy needs from oil and coal sources in 2014, respectively. In contrast, coal provided over 66% of China’s primary energy needs in 2014, while oil amounted to less than 18% of the total. In fact, China accounts for just over half of global coal consumption, and its coal use has almost tripled since 2000 (see British Petroleum’s Statistical Review of World Energy). Considering the dominance of coal (rather than oil) in the Chinese economy, and given that most of its coal consumption (well over 90%) is met by domestic production, oil-supply shocks will have relatively less of an impact on the Chinese economy.
Turning to the commodity exporters in our sample, it appears that an oil-supply shock in the U.S. creates a slowdown in economic activity in these regions (GCC, MENA, and Latin America) to varying degrees (given lower commodity prices)—the extent of which, at least in the short-term, depends on the size of their buffers and availability of financing. The largest effects are in the GCC and the other MENA oil exporters of −2.14% and −1.32% with the effects in Latin America being on average much smaller at −0.41%. MENA oil importers also experience an economic slowdown of 0.28% on average following a U.S. oil-supply shock given their economic ties with oil exporters in the region. For example, remittances from Jordanians working in the region are an important source of national income (equivalent to 15–20 percent of GDP); the Persian Gulf region is the primary destination for Jordanian exports, and in turn, supplies most of its energy requirements; furthermore, the country receives substantial grants and FDI from other states in the region, see Mohaddes and Raissi (2013). Given these linkages, it is no surprise that any slowdown in the GCC region (due to lower oil prices) would adversely affect the Jordanian economy, but also through similar channels other MENA oil importers in general (See Section IV.).
Looking at the GDP responses from a global perspective, our results suggest that following the U.S. oil revolution, with oil prices falling by 51% in the first year and rebounding somewhat to 45% below the pre-shock levels in year two, global growth increases by 0.16 – 0.37 percentage points. In response to lower supply-driven oil prices, we would expect the increase in spending by oil importers to exceed the decline in expenditure by oil exporters given their different marginal propensities to consume/invest. See also Arezki and Blanchard (2014), Baffes et al. (2015), and Husain et al. (2015) who obtain very similar magnitudes.
Following an oil-supply shock in the U.S., we also find strong disinflation pressures on energy-importing countries in our sample (advanced countries, China, India, South East Asia, and MENA oil importers), with the peak responses ranging between 10 and 50 basis points (see Figure 3). On impact, inflation falls in all of these oil-importing countries but the persistence of the responses changes with the magnitude of second-round effects, and the stance of monetary policy. The different responses of MENA oil-exporters and Latin American countries are probably driven by movements of the real exchange rate in these economies. The real exchange rate tends to depreciate in these countries, limiting the pass-through effect of lower international oil prices to domestic markets (and inflation).
Furthermore, in all oil-importing countries/regions, equity prices rise following a positive oil-supply shock in the United States (see Figure 4). As shown in the equity pricing model of Huang et al. (1996), the equity price equals the expected present discounted value of future cash flows. Since a lower expected inflation reduces the discount rate, a fall in oil price has a positive impact on stock market returns. The positive effect of falling oil prices on stock markets in net oil-importers has also been supported by a number of other researches, including Cheung and Ng (1998); Sadorsky (1999); and Park and Ratti (2008) among others.
IV. Focusing on the MENA Region
The results in Section III.B., and Figure 2 in particular, indicate that the U.S. oil revolution, and the resulting lower oil prices, will likely have a negative growth impact on Middle East and North Africa (MENA) region for both primary-commodity exporters and importers. This section investigates in greater detail the macroeconomic implications of the U.S. oil supply revolution, in terms of its real GDP effects, on individual countries in the MENA region over the short and long-term. Note that the resulting low oil-price environment has also political economy implications. For instance, Elbadawi (2015), argues that a negative and sustained oil price shock, by reducing oil rents per capita, could weaken the government’s effectiveness in managing the economy and maintaining civil peace (making the ruling elite more vulnerable to popular uprisings), thereby changing the developmental and sustainable political equilibrium in the GCC countries. While such political economy considerations are important, they are beyond the scope of this paper and will not be tested empirically.
The median-target impulse responses in Figure 5 show that Syria, the GCC and Iran face a long-lasting fall in their real output (more than –2% over the long-run) following a positive U.S. oil-supply shock as lower oil prices weakens the external and fiscal balances in these countries. For Algeria, an OPEC member, the median target response is negative for the first 14 quarters before stabilizing around zero over the long-run. While buffers and available financing allow most oil exporters in the region to avoid sharp cuts in government spending in the near-term (limiting the impact on short-term investment and growth), the long-term impact depends on their medium-term fiscal plans and capital spending. Table 3 shows that the fiscal break-even price for all major oil exporters, except Kuwait, is substantially above $56.25—the average Brent spot price between January and August 2015.
Figure 5.Impact of the U.S. Oil Supply Revolution on Real Output in the MENA Region
Notes: Figures are median target impulse responses to a one standard deviation fall in the price of oil, equivalent to an annualized drop of 51% in year 1 and 45% in year 2. The impact is in percentage points and the horizon is quarterly.
For most oil-importers in the MENA region, gains from lower oil prices are offset by a decline in external demand/financing by oil-exporters over the medium-term given the strong linkages between the two groups through trade, remittances, tourism, foreign direct investment and grants. The resulting estimated long-run negative growth effects on these countries, although being non-trivial, are much smaller than those on oil-exporters—about —0.5%, —0.7%, and —0.2% for Jordan, Morocco and Tunisia, respectively. For Egypt (despite having a relatively large subsidy bill) and Mauritania the median target responses are positive and about 0.2% in the medium-term. In general, low pass-through from global oil prices to domestic fuel prices limits the impact on disposable income of consumers and profit-margins of firms in MENA oil importers, and thereby reduces the direct positive impact on economic growth in these countries. The next two sub-sections discuss the main channels through which long-term growth is being affected by sustained lower commodity prices for major oil-exporters as well as select oil importers in the region.
A. A Long-Run Structural Model for Oil-Exporters
Given that oil exporters in our sample (except for Saudi Arabia) are producing at (or near) capacity, they cannot readily increase their production levels in response to lower oil prices to offset the substantial drop in oil revenues following the U.S. oil supply revolution. Even if they were able to increase production, this would only lead to an increase in global supply, which would in turn depress prices even further; at least in the short-run and until current projects from high-cost fields are completed. The question is whether the long-run growth effects of sustained lower oil revenues for major oil exporters can be modelled and empirically tested at the country level and based on a growth theory? If so this would allow one to explore the channels through which the fall in oil revenue affects these economies. Unfortunately most macroeconomic analysis of oil revenues/shocks tend to take a short-term perspective. They usually focus on the effects of oil revenues on the real exchange rate (Dutch disease) and government budget expansion, thus failing to consider their effects on long-run growth.
This approach makes sense for countries with a limited amount of oil reserves and those facing temporary shocks, but not for major oil exporting countries such as Iran, Kuwait and Saudi Arabia for which oil income should be treated more as a part of the steady state growth outcome and not as a transient state. While it is clear that the oil and gas reserves will be exhausted eventually, this is likely to take place over a relatively long period. Figure 6 shows that most OPEC members such as Algeria, Iran, Kuwait, Nigeria, Saudi Arabia, United Arab Emirates, and Venezuela, and a few countries outside OPEC such as Norway and Russia, have similar oil income to GDP ratios that have remained relatively stable.15
Figure 6.Ratio of Oil Export Revenues to Real Output Across Oil Exporters, 1980–2010
Source: British Petroleum Statistical Review of World Energy, OPEC Annual Statistical Bulletin, and IMF International Financial Statistics.
Given that there is little evidence to suggest that for the MENA oil exporting economies oil income will be diminishing any time soon and the fact that their oil income to output ratio is expected to remain high over a prolonged period, we can utilize the empirical growth model for major oil exporting countries recently developed in Esfahani et al. (2014), to empirically investigate the direct effect of a fall in oil revenue for these economies. More specifically, by extending the stochastic growth model developed in Binder and Pesaran (1999) and allowing for the possibility that a certain fraction of oil revenues is invested in the domestic economy via the capital accumulation channel, Esfahani et al. (2014) show that the long-run output equation for major oil exporters is given by
where oit is oil revenue in US dollars for country
where αi is the share of capital in output,
Rather than combining VARX*(pi, qi) models and solving the GVAR-Oil model as is done in Section III.B., we test the output equation (16) on a country-by-country basis, imposing the additional theory restrictions in (17). To this end we estimate individual country-specific models as before, using equation (12) and quarterly observations over the period 1979Q2–2011Q2, but including oil revenue (oit) rather than oil prices in xit. As predicted by the theory, we find that real output for the seven MENA oil exporters in the long run is shaped by: (i) oil revenue through its impact on capital accumulation, and (ii) technological spillovers through foreign output.
We then consider the output effects of a negative unit shock (equal to one standard error) to oil revenues using the Generalized Impulse Response Functions (GIRFs), developed in Koop et al. (1996) and Pesaran and Shin (1998).16 The associated GIRFs together with their 95% error bands are given in Figure 7. These figures clearly show that a negative oil revenue shock significantly reduces real output in all seven countries, with the full impacts of oil revenue changes showing up in these economies quite fast, and peak within 2-3 years in most cases. The equilibrium levels of these effects are between 2% and 12% (quite heterogenous across countries), with the largest real output losses occurring in three GCC countries (Kuwait, Qatar, and the U.A.E.) where the steady state value of the effect of the oil revenue shock is 11 – 12%. This difference partly reflects the much higher historical volatility of oil revenues in Kuwait (due to invasion of Kuwait by Iraq in 1990 and its aftermath) and Libya. The quarterly standard deviation of oil revenue for Kuwait and Libya is around 42.8% and 37.7% as compared to between 20.3 – 23.8% for the other countries.
Figure 7.Impact of a Negative Oil Revenue Shock for OPEC countries
Notes: Figures are median generalized impulse responses to a one standard deviation fall in oil revenue, together with 95 percent bootstrapped confidence bounds. The impact is in percentage points and the horizon is quarterly.
To contrast the results for the MENA oil exporters with the other three OPEC members in our sample, we also estimated VARX* models for Ecuador, Nigeria, and Venezuela, and found the results to be quite similar, see Figure 7. In particular the GIRFs illustrate that the equilibrium levels of these effects are between 3% and 6%, with the quarterly standard deviation of oil revenue for these economies being between 19.8 – 23.2%.
Overall, the results indicate that oil revenue shocks (such as those from the low oil price environment we are currently experiencing) have a large, long-lasting and significant impact on these economies’ growth paths, operating through the capital accumulation channel. Moreover, the theoretical model for major oil exporters outline above, and the fact that we could not reject the theory restrictions in (17), together with the impulse responses in Figure 7 indicate that these countries will be adversely affected whenever the international price of crude oil declines and will benefit whenever it rises. Therefore, macroeconomic and structural polices should be conducted in a way that the vulnerability of these countries to oil revenue (not just price) disturbances are reduced, see also Cavalcanti et al. (2015) and El-Anshasy et al. (2015).
The empirical results presented here have strong policy implications. Oil exporters in the MENA region and beyond are faced with substantial losses in government revenues as a result of a seemingly long-lasting oil price fall. With buffers eroding over the medium-term, most countries will need to re-assess and re-align their medium-term spending plans. Improvements in the conduct of macroeconomic policies, better management of resource income volatility, and export diversification can all have beneficial growth effects; as do policies which increase the return on investment, such as public infrastructure developments and human capital enhancing measures.17 Moreover, the creation of commodity stabilization funds, or Sovereign Wealth Funds in case of countries in the Persian Gulf, might be one way to offset the negative effects of commodity booms and slumps. Finally, recent academic research has placed emphasis on institutional reform. By establishing the right institutions, one can ensure the proper conduct of macroeconomic policy and better use of resource income revenues, thereby increasing the potential for growth.
B. Spillovers to MENA Oil-Importers
While it is no surprise that MENA oil exporters are affected negatively by lower oil prices, the overall long-term output effect for MENA oil importers is not clear cut (considering the direct and indirect effects of lower oil prices for these economies). While a fall in oil prices initially implies lower import costs for these economies, it also reflects a slow down in oil-exporting countries (see the discussion above), which in turn negatively impact these economies through trade, remittances, and foreign direct investment (FDI) channels. Overall, Figure 5 shows that the direct positive effect of lower oil price for all oil importers (except Egypt and Mauritania) is dominated by the indirect negative impact of spillovers from the exporters (in particular from the GCC). Below we explore the direct positive and the indirect negative channels focusing on Jordan to draw lessons for others.
Figure 8 shows the evolution of the of general government transfers, workers’ remittances, and foreign direct investment (FDI), what we refer to as external income or xit. Both remittances and external income account for a significant share of Jordan’s output, with the share of the former being around 15–20% of GDP over 1979–2009, and the latter being on average 30%. Given that the majority of Jordanian migrant workers reside in the neighboring GCC countries and that most of the official government transfers (grants) are received either from Saudi Arabia or the United States, any economic/political developments in the oil-exporting states of the region would significantly affect the flow of external income to Jordan. Therefore, even though the country is an oil importer, as long as xit from the oil-exporting economies are maintained, we expect lower oil prices to have a long-run negative growth effect on the Jordanian economy. That is, the direct positive effect of oil price booms is dominated by the indirect negative impact; see also International Monetary Fund (2010) and the detailed discussion in Mohaddes and Raissi (2013).
Figure 8.External Income and Price of Oil, in Log Level
Sources: Authors’ construction based on data from International Monetary Fund Balance of Payments Statistics and International Financial Statistics.
Figure 8 shows the relationship between log external income, xit, and log oil prices,
Given the discussion above, we augment the output gap equation (16), to include oil prices as opposed to oil revenues. Note that the inclusion of
where all variables are defined in (16).19 We estimate a VARX*(2,2) model for Jordan, imposing the restrictions in (17), and find that we cannot reject the theory derived output relation. The results therefore confirm that a fall in oil prices, by reducing external income, dampens capital accumulation and thus leads to a fall in real output. The GIRFs in Figure 9 illustrate the response of the Jordanian economy to a negative oil price shock (based on a historical quarterly standard deviation of 18.6%), where the equilibrium output effect of the shock is 4%, being similar to those in Figure 5 based on the GVAR-Oil model.
Figure 9.Impact of a Negative Oil Price Shock
Notes: Figures are median generalized impulse responses to a one standard deviation fall in oil prices, together with 95 percent bootstrapped confidence bounds. The impact is in percentage points and the horizon is quarterly.
Similar analysis can also be conducted for other oil-importing countries in the MENA region. To illustrate this, we also estimated the long-run output equation (18) for Syria, imposing the theory restrictions above, and found that we cannot reject them. The GIRFs for Syria also show that a negative oil price shock reduces real GDP by 2.4%, which is again in line with the results in Figure 5.
V. Concluding Remarks
We applied a set of dynamic sign restrictions on the impulse response of a GVAR-Oil model, estimated for 38 countries/regions over the period 1979Q2 to 2011Q2, to identify the U.S. supply-driven oil-price shock, and to study the global macroeconomic implications of the resulting fall in oil prices. We quantified the GDP impact of a 51 percent reduction in oil prices (caused by a U.S. driven supply glut) on net energy importers and net oil exporters. We found that while oil importers typically experience a long-lived rise in economic activity (between 0.04% and 0.95%) in response to a U.S. supply-driven fall in oil prices, the impact is negative for energy-exporters (–2.14% for the GCC, –1.32% for the other MENA oil exporters, and –0.41% for Latin America) and commodity-importing countries with strong economic ties with oil exporters. Specifically, we find that for most oil-importers in the MENA region, gains from lower oil prices are offset by a decline in external demand/financing by oil-exporters over the medium-term given strong linkages between the two groups through trade, remittances, tourism, foreign direct investment and grants. The resulting estimated long-run negative growth effects on these countries, although being non-trivial, are much smaller than those on oil-exporters; being on average –0.28%. Overall, our results suggest that following the U.S. oil revolution, global growth increases by 0.16 – 0.37 percentage points. Furthermore, in response to a positive U.S. oil-supply disturbance, almost all countries in our sample experience a decrease in inflation and a rise in equity prices.
The sensitivity of MENA countries (both oil exporters and importers) to oil market developments raises the question of which policies and institutions are needed in response to such shocks. While countercyclical fiscal policies (using existing buffers) are key to insulate the exporters from commodity price fluctuations, the other priority for commodity exporters should be to enhance their macroeconomic policy frameworks and institutions (such as more autonomy in conducting the monetary and exchange rate policies). Oil importers in the region should not overestimate the positive impact of the decline in oil prices on their economies given considerable uncertainty about the persistence of lower oil prices and the availability of external financing and weak demand growth in oil-exporting trade partners. For the MENA countries the current low oil-price environment provides an opportunity for further subsidy and structural reforms.
ArezkiR. and O.Blanchard (December222014). Seven Questions About The Recent Oil Price Slump. iMFdirect.
BaffesJ.M. A.KoseF.Ohnsorge and M.Stocker (2015). The Great Plunge in Oil Prices: Causes, Consequences, and Policy Responses. World Bank Policy Research Note PRS/15/01.
BaumeisterC. and L.Kilian(2015). Understanding the Decline in the Price of Oil Since June 2014. CEPR Discussion Paper 10404.
BaxterM. and M. A.Kouparitsas(2005). Determinants of Business Cycle Comovement: A Robust Analysis. Journal of Monetary Economics52(1) pp. 113–157.
BinderM. and M.Pesaran(1999). Stochastic Growth Models and Their Econometric Implications. Journal of Economic Growth 4139–183.
CashinP.K.Mohaddes and M.Raissi(2012). The Global Impact of the Systemic Economies and MENA Business Cycles. IMF Working Paper WP/12/255.
CashinP.K.Mohaddes and M.Raissi (2015). Fair Weather or Foul? The Macroeconomic Effects of El Niño. IMF Working Paper WP/15/89.
CashinP.K.MohaddesM.Raissi and M.Raissi (2014). The Differential Effects of Oil Demand and Supply Shocks on the Global Economy. Energy Economics 44113–134.
CavalcantiT. V. d. V.K.Mohaddes and M.Raissi (2011). Growth, Development and Natural Resources: New Evidence Using a Heterogeneous Panel Analysis. The Quarterly Review of Economics and Finance 51 (4) 305–318.
CavalcantiT. V. D. V.K.Mohaddes and M.Raissi (2015). Commodity Price Volatility and the Sources of Growth. Journal of Applied Econometrics30(6) 857–873.
CheungY.-W. and L. K.Ng (1998). International Evidence on the Stock Market and Aggregate Economic Activity. Journal of Empirical Finance5(3) 281–296.
ChudikA. and M.Fratzscher (2011). Identifying the Global Transmission of the 2007-2009 Financial Crisis in a GVAR Model. European Economic Review 55 (3) 325–339. Special Issue: Advances in International Macroeconomics: Lessons from the Crisis.
ChudikA. and M. H.Pesaran (2013). Econometric Analysis of High Dimensional VARs Featuring a Dominant Unit. Econometric Reviews32 (5-6) 592–649.
ChudikA. and M. H.Pesaran (2015). Theory and Practice of GVAR Modeling. Journal of Economic Surveys forthcoming.
DeesS.F. diMauroM. H.Pesaran and L. V.Smith (2007). Exploring the International Linkages of the Euro Area: A Global VAR Analysis. Journal of Applied Econometrics221–38.
El-AnshasyA.K.Mohaddes and J. B.Nugent (2015). Oil, Volatility and Institutions: Cross-Country Evidence from Major Oil Producers. Cambridge Working Papers in Economics 1523.
ElbadawiI. A. (2015). Thresholds Matter: Resource Abundance, Development and Democratic Transition in the Arab World. In I. Diwan and A. Galal (Eds.). The Middle East Economies in Times of Transition. Palgrave Macmillan.
EsfahaniH. S.K.Mohaddes and M. H.Pesaran (2013). Oil Exports and the Iranian Economy. The Quarterly Review of Economics and Finance53(3) 221–237.
EsfahaniH. S.K.Mohaddes and M. H.Pesaran (2014). An Empirical Growth Model for Major Oil Exporters. Journal of Applied Econometrics29(1) 1–21.
FryR. and A.Pagan (2011). Sign Restrictions in Structural Vector Autoregressions: A Critical Review. Journal of Economic Literature49(4) pp. 938–60.
HamiltonJ. D. (2009). Causes and Consequences of the Oil Shock of 2007-08. Brookings Papers on Economic Activity Economic Studies Program The Brookings Institution40(1) 215–283.
HuangR.R.Masulis and H.Stoll (1996). Energy Shocks and Financial Markets. Journal of Futures Markets16(1) 1–27.
HusainA. M.R.ArezkiP.BreuerV.HaksarT.HelblingP.MedasM.Sommer and an IMF Staff Team. (2015). Global Implications of Lower Oil Prices. IMF Staff Discussion Note SDN/15/15.
International Monetary Fund. (2010). Jordan – 2010 Article IV Consultation – StaffReport and Public Information Notice. Country Report No. 10/297.
KilianL. (2009). Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude Oil Market. The American Economic Review99(3) 1053–1069.
KilianL.A.Rebucci and N.Spatafora (2009). Oil Shocks and External Balances. Journal of International Economics77(2) 181–194.
KoopG.M. H.Pesaran and S. M.Potter (1996). Impulse Response Analysis in Nonlinear Multivariate Models. Journal of Econometrics74119–147.
LeeK. and M. H.Pesaran (1993). Persistence Profiles and Business Cycle Fluctuations in a Disaggregated Model of UK Output Growth. Ricerche Economiche47293–322.
MacKinnonJ. G. (1991). Critical Values for Cointegration Tests. In R. Engle and C. Granger (Eds.)Long-Run Economic Relationships: Readings in Cointegration Chapter 13 pp. 267–276. Oxford University Press.
MânescuC. B. and G.Nuño (2015). Quantitative Effects of the Shale oil Revolution. Energy Policy forthcoming.
MohaddesK. and M. H.Pesaran (2015). Country-Specific Oil Supply Shocks and the Global Economy: A Counterfactual Analysis. CESifo Working Paper No. 5367.
MohaddesK. and M.Raissi (2013). Oil Prices, External Income, and Growth: Lessons from Jordan. Review of Middle East Economics and Finance9:299–131.
MohaddesK. and O. H.Williams (2013). Inflation Differentials in the GCC: Does the Oil Cycle Matter?Middle East Development Journal5:2.
ParkJ. and R. A.Ratti (2008). Oil Price Shocks and Stock Markets in the U.S. and 13 European Countries. Energy Economics30(5) 2587–2608.
PeersmanG. and I. VanRobays (2012). Cross-country Differences in the Effects of Oil Shocks. Energy Economics34(5) pp. 1532–1547.
PesaranM. H. (2004). General Diagnostic Tests for Cross Section Dependence in Panels. IZA Discussion Paper No. 1240.
PesaranM. H.T.Schuermann and S.Weiner (2004). Modelling Regional Interdependencies using a Global Error-Correcting Macroeconometric Model. Journal of Business and Economics Statistics22129–162.
PesaranM. H. and Y.Shin (1996). Cointegration and Speed of Convergence to Equilibrium. Journal of Econometrics71117–143.
PesaranM. H. and Y.Shin (1998). Generalised Impulse Response Analysis in Linear Multivariate Models. Economics Letters5817–29.
PesaranM. H.Y.Shin and R. J.Smith (2000). Structural Analysis of Vector Error Correction Models with Exogenous I(1) Variables. Journal of Econometrics97293–343.
PesaranM. H.L. VanessaSmith and R. P.Smith (2007). What if the UK or Sweden had Joined the Euro in 1999? An Empirical Evaluation Using a Global VAR. International Journal of Finance & Economics12(1) 55–87.
SadorskyP. (1999). Oil Price Shocks and Stock Market Activity. Energy Economics21(5) 449–469.
SimsC. (1980). Macroeconomics and Reality. Econometrica481–48.