Time-variation in geopolitical risk is widely considered an important determinant of fluctuations in economic activity. The financial press, international organisations, rating agencies, and the investment community all vie to assess these risks and their likely impact on the economy. Clearly, geopolitical events matter not only when they occur on rare occasions, but also when investors and consumers make decisions in anticipation of the possibility of such events.
This fact is nowhere more apparent than when it comes to geopolitical risk in energy markets. For example, many observers list risks to energy security as one of the top geopolitical risks of 2024. This assessment is driven in no small part by concerns about OPEC quota decisions, global access to Russian oil amidst Ukrainian attacks on Russian oil infrastructure and efforts to tighten the G7 price cap, dwindling strategic oil reserves, disruptions of oil shipments in the Red Sea and possibly in the Persian Gulf, and concerns about a widening conflict between Israel and Iran. Geopolitical events such as these are low probability but have potentially high impact on the economy, creating geopolitical risk.
Towards a better understanding of the role of geopolitical oil price risk
There is a deep-rooted belief in macroeconomics that higher oil price uncertainty driven by geopolitical risk lowers domestic investment and consumption and hence real GDP (e.g. Bernanke 1983). In recent research (Kilian et al. 2024), we seek to develop a better understanding of how time-variation in geopolitical risk in oil markets affects oil price uncertainty and economic fluctuations. Our analysis recognises that, while downside geopolitical risk to oil production raises oil price uncertainty, not all surges in oil price uncertainty are driven by geopolitical events.
Figure 1 shows an index of the uncertainty in the real price of oil since the 1970s constructed from the predictive variance of the real price of oil. Its evolution lines up well with that of the Chicago Board of Trade OVX index which measures the implied price volatility of oil options and is available only more recently. The figure highlights that there can be major surges in uncertainty even in the absence of shifts in geopolitical risk, as was the case during the financial crisis in 2008, for example.
Figure 1 Oil price uncertainty, 1974Q4-2023Q4
One reason that oil price uncertainty differs from geopolitical risk in oil markets is that uncertainty measures mask the distinction between two-sided risk and one-sided risk. Clearly, market participants are not concerned with the risk of an unexpected surge in global oil production driven by geopolitical events, but with the risk of an unexpected sustained drop in global oil production. Another reason is that market participants understand that uncertainty about the oil price reflects not only uncertainty about future oil production, but also uncertainty about future oil consumption driven by macroeconomic uncertainty, financial uncertainty, and policy uncertainty.
Perceptions of uncertainty in turn feed into the oil price, as market participants build or draw down oil inventories in response to shifts in uncertainty. Thus, understanding the evolution of oil price uncertainty and its role in driving economic fluctuations requires simultaneously modelling downside risk from production disasters, oil storage, and the endogenous determination of oil price and macroeconomic uncertainty in the global economy.
Our analysis of this question is based on a calibrated nonlinear dynamic stochastic general equilibrium (DSGE) model of the global economy that is designed to address the question of how geopolitical oil price risk is linked to economic fluctuations. The model includes risk averse economic agents, an oil production sector, oil storage, and limited substitutability between oil and capital. The price of oil is determined endogenously. Since the model is global, we abstract from oil imports and exports and international capital flows.
One key difference from earlier studies is that we focus on geopolitical downside risk to oil production rather than stochastic volatility in oil production or the price of oil (e.g. Gao et al. 2022). Another key difference is that our model allows for both macroeconomic uncertainty and oil price uncertainty and that uncertainty is determined endogenously.
Building on Gourio (2012), downside risk emanates from macroeconomic disasters and oil production disasters of stochastic length that occur with time-varying probabilities. Macroeconomic disasters are modelled as sharp declines in growth and may be viewed as the result of an economic crisis such as the Great Recession of 2008 or the Covid-19 recession of 2020. Oil production disasters are modelled after events such as the Iranian Revolution in 1979 or the invasion of Kuwait in 1990.
Implications for the global economy
Figure 2 shows that shocks to the probability of an oil production disaster have large and persistent effects on oil inventories, the price of oil, and oil price uncertainty, even when the disaster is not realised. These shocks can also have sizeable effects on investment and output, but they are not a major driver of fluctuations in macroeconomic aggregates because large oil disaster probability shocks are rare. Nor do they have much of an effect on macroeconomic uncertainty.
Figure 2 Responses to an oil production disaster probability shock
Shocks to the probability of a growth disaster, which increase macroeconomic uncertainty, have even larger effects on the oil market and the macroeconomy. These shocks also play a major role in the determination of oil price uncertainty, which helps explain why higher oil price uncertainty has historically been associated with lower real activity. Our analysis highlights that this association should not be interpreted as evidence of a causal link.
The calibrated model generally does an excellent job at capturing the volatility in the data. Dropping the output disaster risk from the DSGE model substantially lowers the ability of the model to explain that volatility. The resulting model not only substantially understates the standard deviation of output uncertainty and oil price uncertainty, but it also understates most other data moments. Dropping both output disaster risk and the geopolitical risk underlying oil production disasters from the model highlights the limited role geopolitical oil price risk plays in generating macroeconomic volatility.
Alternative modelling approaches
Incorporating downside oil production risk in the model is crucial for our results. While increases in oil price uncertainty may also be explained by stochastic volatility shocks to oil production, as in Gao et al. (2022), for example, the latter shocks do not have strong recessionary effects because they do not generate downside risk.
We also demonstrate that the ability to store oil plays a central role. Without storage the responses of both oil market variables and macroeconomic aggregates to higher oil production risk tend to be muted, suggesting that alternative models without storage fail to capture the full effects of shifts in oil production risk. Oil storage also matters for the responses of the global economy to growth disaster shocks, underscoring the importance of modelling oil production and macroeconomic disasters jointly.
Implications for empirical models of the effects of oil price uncertainty shocks
The model implies that changes in oil price uncertainty need not be an indication of exogenous shifts in the uncertainty about future oil supplies. Uncertainty about the oil price may reflect exogenous macroeconomic uncertainty shocks, mirroring the standard result that the real price of oil responds to shifts in the demand for oil. Perhaps less obviously, the model shows that oil price uncertainty may also reflect shocks to the level of oil market or macroeconomic variables, such as realisations of disasters.
Thus, not only are oil price level and oil price uncertainty shocks not the same, as implicitly assumed in the widely used vector autoregressive (VAR) model with generalised autoregressively conditionally heteroskedastic (GARCH) errors, but the effects of a level shock in the data-generating process are not separable from those of an uncertainty shock, as assumed in VAR models with stochastic volatility. Nor does it make sense to employ recursive linear VAR models with oil price uncertainty either ordered first or last since oil price uncertainty is simultaneously determined with macroeconomic aggregates.
These results cast doubt on the ability of these empirical models to correctly identify exogenous oil price uncertainty shocks. They also call into question a large body of empirical work that has produced seemingly robust evidence of large recessionary effects of oil price uncertainty shocks and has shaped the policy debate about geopolitical risk in recent years.
Our analysis suggests that economists and policymakers need to rethink the role of geopolitical oil price risk in the global economy and be cognizant of the interplay between oil price uncertainty, macroeconomic uncertainty, and the state of the economy. In recent years, there has been growing awareness of geopolitical risks in other commodity markets including critical minerals, natural gas, and agricultural commodities. While we focus on the market for crude oil, our modelling approach is also relevant for these other commodity markets.
Authors’ note: The views expressed in this column are those of the author and should not be attributed to the Federal Reserve Bank of Dallas or the Federal Reserve System.
References
Caldara, D and M.Iacoviello (2022), “Measuring Geopolitical Risk,” American Economic Review 112: 1194-1225.
Bernanke, B S (1983), “Irreversibility, Uncertainty and Cyclical Investment,” Quarterly Journal of Economics 98: 85-106.
Gao, L, S Hitzemann, I Shaliastovich and L Xu (2022), “Oil Volatility Risk,” Review of Financial Economics 144: 456-491.
Gourio, F (2012), “Disaster Risk and Business Cycles,” American Economic Review 102: 2734-2766.
Kilian, L, M D Plante, and A W.Richter (2024), “Geopolitical Oil Price Risk and Economic Fluctuations,” CEPR Discussion Paper 19113.
