Policymakers monitor financial markets and macroeconomic data releases to obtain cues about changes in the state of the global economy and the nature of the underlying drivers (Draghi 2016). But macroeconomic data are only available with a significant lag and are subject to non-negligible revisions. Asset prices, on the other hand, are available in real time but are only indirectly related to macroeconomic variables like output and inflation. The tension between using information that is timely but potentially noisy, and information that is accurate but lagging, is a key challenge for real-time decision making.
The oil market plays a special role in assessing the state of the global economy in real time. First, the price of oil is closely related to economic activity and global inflation. Second, oil prices can have a lasting impact on inflation expectations and therefore affect monetary policy decisions. Third, oil is traded in liquid financial markets and its price is available at high frequency.
However, interpreting the drivers of daily fluctuations in oil prices is not straightforward. To date, these structural drivers have typically been identified via structural vector autoregressions (SVAR) (Kilian 2009, Caldara et al. 2019, Baumeister and Hamilton 2019). Unfortunately, such SVAR models, while interesting ex post, are of little use to policymakers in real time as they rely on monthly data which are plagued by publication lags. Hence, policymakers are left in the dark about the sources of shocks precisely when a structural narrative would be valuable to inform their decisions.
In a recent paper (Gazzani et al. forthcoming), we show how to exploit information from oil market developments to resolve this tension. We propose a credible structural identification strategy that connects the oil market, financial markets and the macroeconomy and that can be used to build a macroeconomic narrative that is timely and hence valuable for monetary policymakers.
We set out to fill this void. To this end, our model is a daily Bayesian SVAR that provides a real-time decomposition of the price of oil in three structural shocks, through a combination of narrative and sign restrictions. Two of these shocks are related to, respectively, future and current global demand. We treat current and future demand separately since the price of oil, much like financial asset prices, is sensitive to news on future economic prospects. The third is a shock to the supply of oil.
To further validate our estimated shocks, we document how they satisfy a number of appealing properties. Their volatility is higher upon days with specific news hitting the market: oil supply shocks tend to be larger around OPEC announcements, while forward-looking and current demand shocks relate to macroeconomic data releases and/or announcements of the US Federal Reserve Open Market Committee (FOMC).
The model helps real-time policymaking
We demonstrate the usefulness of our model in the context of real-time policymaking. Focusing on specific US Federal Reserve meetings between 2011 and 2022, we show how FOMC members not only discussed oil price developments but also aimed to disentangle the source of their movement into underlying demand/supply factors. Our model could have dispelled those doubts by promptly quantifying these drivers.
We describe here three more recent episodes, outside the sample of our paper, when the oil market was buffeted by particularly relevant shocks. First, we consider the aftermath of a disappointing macroeconomic data release from China, when PMI manufacturing and services turned out to be well below expectations. The news was released on Sunday 1 October 2023, and oil markets responded with a fall in prices on the following Monday, as China’s economic prolonged slump weighed on the crude demand outlook. The model rightly attributes most of the fall (by 5% in the day) to a negative demand shock (the falling contribution from the orange bar).
Second, we consider the outburst of the conflict in the Middle East (on 7 October 2023) which roiled global markets on Monday 8 October, as sudden fears of a larger conflagration boosted oil prices. On that day Brent prices gained 4%, and our model pins down most of the price increase to a supply shock (shown by the greater contribution from the yellow bar). Throughout the sample oil prices were compressed by negative shocks to the appetite for risk. This is connected to the risk of higher-for-longer interest rates due to relatively hawkish FED communication in September and October.
Third, we look at the OPEC announcement of supply cuts at the end of November. Crude oil prices fell upon the announcement, as markets deemed the announcements to be rather underwhelming compared to original expectations of deeper cuts by OPEC members.
Figure 1 Decomposing the price of oil in real time
A set of simplified replication codes that allow the user to decompose the price of oil in real time, as in Figure 1, is available online. Updated shocks from the daily VAR can be found here.
Conclusions and policy implications
Oil prices provide timely information on global shocks that is of key relevance for monetary policy decisions. Workhorse monthly VAR models of the oil market are, however, of limited use for monetary policy in real time, as they rely on data that are released with considerable delay. We propose a novel approach to identify shocks to the price of oil at a daily frequency and in real time, through a combination of narrative and sign restrictions. Unlike existing monthly VARs, our model can inform policymakers in a timely fashion. We also make a broader methodological contribution, to circumvent publication lags, by identifying shocks on daily data and using their lower frequency average as instruments in lower frequency models to validate them based on their macroeconomic impact, a feature which could be applicable and useful in other settings.
Authors’ note: This column does not necessarily reflect the view of the Bank of Italy or ESCB.
References
Baumeister, C and J D Hamilton (2019), “Structural interpretation of vector autoregressions with incomplete identification: Revisiting the role of oil supply and demand shocks”, American Economic Review 109: 1873–1910.
Caldara, D, M Cavallo and M Iacoviello (2019), “Oil price elasticities and oil price fluctuations”, Journal of Monetary Economics 103: 1–20.
Draghi, M (2016), “On the importance of policy alignment to fulfil our economic potential”, Speech given for the 5th Tommaso Padoa-Schioppa Lecture at the at the Brussels Economic Forum 2016, Brussels.
Gazzani, A, F Venditti and G Veronese (forthcoming), “Oil price shocks in real-time”, Journal of Monetary Economics.
Kilian, L (2009), “Not all Oil Price Shocks are Alike: Disentangling Demand and Supply Shocks in the Crude Oil Market”, American Economic Review 99: 1053–69.
