The economic impact of bouts of severe weather is easily exaggerated

MUCH of the northern hemisphere has had a miserable winter. Snowstorms shut down many cities and airports across America’s east coast this month and last. In Britain and much of continental Europe, the worst of the weather occurred in the week before Christmas. Life there has long since returned to normal: the forecourts of terminals at Heathrow airport no longer resemble tent cities. But the economic effects of the big freeze are still becoming visible in official data. Germany’s statistical office announced on February 8th that in December the country’s industrial production shrank by 1.5%, while construction output declined by 24%. It blamed the snow for the slump in construction. January’s bad weather was blamed by many analysts for the weak jobs growth reported in the latest non-farm payrolls numbers in America. Britain’s statistical agency attributed more or less all of the 0.5% fall in GDP in the last quarter of 2010 to the bad weather in December. Some argue that the consequences of bouts of bad weather are more significant still. The Federation of Small Businesses (FSB), a British trade group, estimated that the economy lost £ 1.2 billion ($1.9 billion) for each day of the December freeze. Five severe “snow days”, say, would therefore have cost Britain around £ 6 billion. That is a substantial sum for an economy whose quarterly output is worth approximately £ 330 billion. It is true that businesses are hurt when they are forced to remain closed and people cannot get into work. Some revenue, like the money spent on coffee by commuters on their way in to work, is simply forgone. After all, people do not buy extra cups when they do eventually return to their offices. And certain industries, such as construction and transportation, are particularly vulnerable to bursts of bad weather. But numbers like those from the FSB almost certainly overestimate the true economic effect of a day of enforced winter holiday. The group based its figure on the one-fifth or so of Britain’s workers who it reckons did not make it to the office, the factory or the building site during those snowy days in late December. That may be an accurate estimate of snow-induced absenteeism, but it does not mean that a fifth of the £ 4.5 billion-5 billion of goods and services that the British economy produces on a typical working day was simply lost. Some people work from home when they cannot get into the office; others make up for lost time when they return. A few days of no production at factories might mean a bout of overtime when things return to normal: German construction output may turn out to have boomed in January because people raced to finish projects delayed by snow. Many purchases are likely to have been postponed rather than abandoned altogether. And if people expect bad weather, they may stock up on essentials in anticipation of shops being closed for a day or two, much as they might before a holiday. In America tempers are guaranteed to fray on the day before a snowstorm is expected, as supermarkets run out of supplies. So it seems reasonable to assume that a good deal of economic activity is displaced, rather than destroyed, by bad weather. Certain industries see demand for their products go up when severe weather hits. People stuck at home may consume more lighting and heating than if they had headed to work. In Britain, for example, the output of utilities grew by 1.2% in the fourth quarter, as it had done in the first three months of 2010 when the country had its last big encounter with snow. In contrast, the milder final quarter of 2009 saw utilities’ output shrink by 2.2%. SnowSports Industries America, a trade group, says that sales of equipment and accessories set a new record in December.If the net effect on economic activity is unlikely to be as large as bodies like the FSB estimate, are official calculations any closer to the truth? Britain’s official estimate of lost fourth-quarter output—half of 1% of quarterly GDP—is roughly half a day’s worth of production. Without the December snow, reckon the statisticians, output would have been “flattish”. That number is still uncertain but has a firmer basis in fact.

Statistical agencies typically have very little actual data for the last month of the quarter at the time they release their first estimates of quarterly GDP. So in the case of fourth-quarter releases, they usually estimate December’s numbers using the data for October and November, making what they hope are appropriate seasonal adjustments. This time Britain’s number-crunchers were concerned that their normal methods might lead them astray, since the unusually bad weather was concentrated in the month for which they would normally have little data. So they made a big effort to collect data sooner, to have a clearer idea than usual of what was going on in December. They found that their normal estimation methods—which can be thought of as measuring what would have happened in typical December weather—would have led to a figure 0.5% higher than the eventual, published number. This difference was their rough estimate of the impact of the effect of bad weather.