Sunday, April 19, 2009

Modeling Oil Demand

When building oil supply and demand (S&D) models, demand is especially challenging to model when there are rapid large price movements.

When oil prices are relatively stable, consumers in developing countries increase their oil demand steadily with income until they reach European (high population density) or US (low population density) levels. Once consumers reach European or US demand levels then, in a stable oil price environment, oil demand in such countries becomes much less sensitive to income growth and changes only with population size.

However, in the face of rapid price increases such as that experienced in 2008, there is little historical data available describing how any oil consumer behaves. Data for the oil shocks of 1973-1974 and 1980-1983 are useful, but the world economy and transportation vehicle ownership patterns looked quite different back then.

Comparing patterns of vehicle ownership in different parts of the world today, I find it useful to have three model touchstones : 1. the US, 2. Europe (representing the developed world outside the US) and 3. the developing world.

1. The US has a high rate of vehicle ownership at 2.28 vehicles per household according to a recent survey. From the same survey, “'The most common pairing of vehicles in American households with two to four cars is a full-sized pickup truck and a standard, mid-range vehicle.” The average efficiency of the US vehicle fleet is in the low 20 mpg range.

2. Europe has slightly smaller levels of vehicle ownership per household compared with the US. Average European fuel efficiency is much higher than the US. Public mass transit is also more readily available in Europe than in the US.

3. In the developing world there is usually less than one vehicle per household and these vehicles are very efficient (such as a motorcycle/tiny car). Public mass transit is often much less available than in Europe and the US.

When oil prices rise rapidly, households in the US have capacity for immediate efficiency gains as they can drive the more efficient of their two vehicles more and if they buy a new vehicle they can buy a much more efficient vehicle. There is a certain amount of discretionary oil spending in the US (especially vacations involving air travel) which can also be cut back rapidly.

Europeans are already driving relatively efficient vehicles and so they have less capacity for immediate oil consumption efficiency apart from using public transportation more heavily. Europeans have less discretionary oil spending (as GDP per capita is lower) compared to the US - but what discretionary oil spending exists is still oil price sensitive.

In the developing world, there is less public transportation available and vehicle owners are already driving one very efficient vehicle. This vehicle is usually essential to getting to and from a good job and there is little oil spent on vacations (jet travel, and so on). Developing country oil consumers, therefore, have the least ability to cut back on consumption when oil prices rise.

In summary, when oil prices rise, households in the US can (to a certain extent) quickly decide to buy less oil, followed closely by Europe – but for slightly different reasons. In the developing world, households cannot as easily avoid buying oil when prices rise quickly.

This demand model was clearly evident in the rapid increase in oil prices during 2008 when US oil demand collapsed first and most severely (almost 13% year/year decline at its nadir in October 2008), followed by Europe. Developing world oil demand fell only very slightly, but quickly recovered.

Follow @CommodityMD