Spiralling oil prices: not such a shock?

Paul Edelstein, Lutz Kilian, 06 June 2007



Oil shocks have a chilling reputation for wrecking economies: memories of fuel rationing and recession in the 1970s mean a sudden jump in the price of a barrel of crude is assumed to have a singularly disastrous effect on GDP growth today. Yet while the cost of energy in the US rose an extraordinary 68% in real terms between January 2002 and July 2006, the US was booming, and the global economy achieved one of its strongest periods of sustained growth for three decades.

CEPR Research Fellow Lutz Kilian and his co-author Paul Edelstein use a meticulous analysis of energy prices and consumer spending in the US from 1970 to 2006, to unpack exactly how changes in the oil price affect the economy, and discover that the declining size of America’s once-mighty auto sector has actually helped to contain the effects of an oil shock on the wider economy.

After a year, a 1.5% increase in energy prices causes total consumption to fall by an average of 2.3% in real terms. Given the small share of energy in total consumption, that’s a bigger change than could be explained simply by the ‘discretionary income effect’ – the impact of consumers finding that more expensive energy leaves them less money to spend on other things.

Edelstein and Kilian’s analysis shows that as energy prices rise, two more effects are at work: a ‘precautionary savings effect’ and an ‘operating cost effect’. They use evidence from the Michigan University consumer confidence survey to show that shocks to spending power tend to leave consumers feeling less confident, and expecting higher unemployment, and lower income: in other words, they tend not to feel it’s a great time to go shopping. Instead, they squirrel extra money away, in case things are about to get worse.

At the same time, consumers begin to change their spending patterns to adjust to the higher energy price. In what the authors call the ‘operating cost effect,’ they may opt for more energy efficient fridges or washing machines, for example, and buy a smaller, energy efficient car instead of a gas-guzzler.

Predictions of widespread economic damage following an oil shock are often based on the assumption that changes in spending patterns like these lead to an indirect, ‘reallocation effect’: as households switch away from energy-intensive goods – critically, cars – US industry is forced to adjust, causing a disruptive shake-up, with potential for further knock-on effects in lost jobs and income.

The operating cost effect on the car market is large: the authors calculate that an oil shock of the size that followed Hurricane Katrina could knock out 10% of demand for domestic cars, as consumers switch to foreign models, which tend to be more efficient.

Comparing the sales of different types of vehicle confirm that it’s energy efficiency – operating costs – driving the switches in spending. While the total number of vehicles sold is roughly the same, demand for trucks, including gas-guzzling SUVs, falls by as much as 11.2% after a Hurricane Katrina-type price rise. And since trucks account for a large chunk of the output of Ford, Daimler-Chrysler and GM, the impact on the industry would be considerable.

In the 1970s, when the US auto industry was much larger, and produced few energy-efficient models, leaving that end of the market to the Japanese, the economic effects of such a sharp switch in spending would have been much larger; but today, the car manufacturers account for just 1% of US employment, and 1% of output. Their relatively small size helps to contain the reallocation effect, and thus to limit the damage done by oil shocks today.

As well as thinking about the energy efficiency of their car, consumers respond with a myriad of tiny economies, as they trim their day-to-day spending in response to rising prices. Using the detailed spending data, Edelstein and Kilian are able to detail these smaller changes. Families spend less on eating out, for example, and more on eating at home; less on private travel and more on public transport. They tend not to let their insurance policies lapse as their fuel bills go up - but they do cut back on car repairs.

Dissecting the impact of an energy price shock in this way also helps to solve a puzzle. When Opec, the oil producers’ cartel, collapsed in 1986, and oil prices plunged, there was no resulting economic boom, despite the extra cash in consumers’ pockets from falling energy bills. This is often regarded as evidence that the response to oil shocks is asymmetric, with price rises being unambigously bad, while falls have more complex effects. However, none of the mechanisms uncovered by the analysis – the precautionary savings effect, the discretionary income effect, and the operating costs effect – are asymmetric. Just as consumers increase their savings when prices are rising, they feel safe to spend more when prices are falling.

By re-examining the 1980s episode using the detailed consumption data, the authors find that in fact spending did pick up sharply, but there was no boom, because rising household consumption was offset by a sharp decline in business investment, perhaps caused by a completely unrelated piece of legislation, the 1986 Tax Reform Act. Just as oil price rises are damaging, falling costs can tempt consumers back to the shops. As the experience of the past five years suggests, the idea that a sharp increase in energy prices – an ‘oil shock’ – has a unique, devastating power over the whole economy is simply a myth, based on a misreading of the past.

CEPR DP 6255 Retail Energy Prices and Consumer Expenditures

URL:  to come

Topics:  Energy Global economy

Tags:  oil shocks, consumer spending, energy prices


CEPR Policy Research