Forecasting oil prices: What works best and when

Oil price forecasts help central banks assess risks to inflation and economic activity. Forecasters often rely on the latest price of oil, on futures markets or on economic models. The best approach depends on how far ahead forecasters are looking and whether they need only a forecast or also an explanation of what’s behind it.

Oil is a key input when producing and transporting everyday goods, and its price affects costs across the economy, including the price consumers pay for gasoline. Oil prices can change quickly as new information reaches the market about supply, world events or the global economic outlook. The closure of the Strait of Hormuz early in 2026 provides a good example of this.

Because oil prices have such widespread effects, policy‑makers, businesses, investors and households all want to know where they are headed. This raises a natural question: what is the best way to forecast oil prices?

I’ve spent much of my career studying the oil market, with a particular focus on understanding how to forecast oil prices. In this article, I share some of the lessons I've learned about what works and what doesn't. The main takeaway is simple: the best forecasting approach depends on how far into the future you’re looking and whether understanding what is driving oil prices matters to you.

The latest price of oil is a useful starting point for building near‑term forecasts

Every day, traders form views about future oil production, consumption and inventories, as well as geopolitical events and the global economy. As these traders buy and sell oil, their expectations get incorporated into market prices. Short‑term changes in prices therefore mainly reflect new, unexpected information. This makes the latest price a natural benchmark for evaluating forecasts produced using oil futures or economic models.

Empirically, other forecasting approaches have difficulty improving on the latest‑price forecast at short horizons. Like stocks and exchange rates, short‑term movements in oil prices are driven largely by new information and are therefore difficult to predict using tools such as oil future prices or models.

Futures are most informative at longer horizons

Futures contracts offer a market‑based view on where oil prices may be heading. In these contracts, a seller agrees to deliver oil to a buyer at a specified future date and price. Futures prices therefore reflect market participants’ expectations and compensation for the risk they are taking.

The usefulness of oil price futures depends on the forecast horizon. Futures‑price forecasts provide little improvement over latest‑price forecasts in the short run, but they become more informative further ahead (Chart 1). For example, at a one‑year horizon, they are 13% more accurate than the latest‑price forecasts. At three years, they are 42% more accurate.

But getting the price exactly right is not the only metric that matters. For policy‑makers, knowing whether oil prices are likely to move up or down can be just as important as knowing by how much, because the direction of the movement often shapes the implications for inflation and economic activity. Oil price futures correctly predict the direction of changes 62% of the time at a one‑year horizon and 87% of the time at a three‑year horizon.

Over longer horizons, financing, storage and other costs accumulate, making it more challenging for traders to profit from differences between current and future oil prices. As a result, the longer‑term outlook can differ significantly from today’s price.

This predictive power is a relatively recent development. Although earlier research found little evidence that oil price futures improved forecast accuracy at longer horizons, since the late 2000s, accuracy has been found to have increased (Chart 2).

Several factors may have contributed to this improvement, including changes in risk premiums. Oil price futures markets have grown substantially over the past two decades, attracting a broader set of participants and becoming more liquid. These deeper markets may incorporate new information into prices more effectively, improving accuracy at longer horizons.

Models can improve forecasts and explain price movements

Economic models take a different approach. Rather than relying solely on market prices, they also use information on economic variables, such as production, consumption, inventories and economic activity.

Like futures‑price forecasts, models’ usefulness depends on the forecast horizon and the information used. Models based on monthly average prices add little to short‑run forecasts because averaging obscures the latest information reflected in prices (Chart 3). Models that incorporate daily or end‑of‑month prices perform better and can sometimes improve on the latest‑price benchmark. However, these gains are generally modest. For example, the best vector autoregressive model correctly predicts the direction of change one and three months ahead only 53% of the time—little better than a coin toss.

For longer‑term forecasts, the choice between average and latest prices matters less, while other features of how oil prices behave become more important. For example, models that allow oil price volatility to change over time have been found to perform better at longer horizons.

Models can also do something that latest‑price and futures‑price forecasts cannot do on their own: help explain why oil prices change. With additional assumptions, models can distinguish between changes in oil supply, demand and other forces. This allows economists to explain revisions to the outlook and examine various alternative scenarios, such as a disruption to oil production or an unexpected change in global activity.

Putting it all together

Oil prices can have important implications not only for inflation but also for the outlook for activity in the oil sector and the broader economy. For policy‑makers, forecasting oil prices is about more than predicting where prices are headed. It is also about understanding the risks around the outlook and the forces driving changes in prices.

No forecasting approach works best in every situation. The latest price of oil is a strong short‑term forecast and a natural benchmark for evaluating other forecasts. Futures can offer signals about where prices are headed beyond one year. Economic models can sometimes improve accuracy while also helping assess whether changes in oil prices reflect shifts in supply, demand or other factors. Used together, these tools can help policy‑makers better assess the outlooks for inflation and the economy.

Find out more