The cost of calamity

THE full extent of the damage from the tsunami that hit Japan's north-eastern coast on March 11th is not yet known, but early estimates of the cost are big. Rebuilding homes, factories, roads and bridges could cost as much as $200 billion, some reckon. Quite apart from these direct costs, is the disaster likely to do lasting harm to Japan's economy?

Much will depend on the success of efforts to prevent a nuclear catastrophe. Assuming the situation at the Fukushima Dai-ichi plant stabilises, the contours of the economic impact of the tsunami itself can already be discerned. Natural disasters disrupt production, much as less destructive episodes of bad weather do. In Japan the interruption to electricity supply means that output has been affected even in areas the tsunami did not directly inundate. Toyota, for example, halted production because of problems with parts and supplies. Operations were suspended in six of Sony's factories, only one of which was flooded.

But such disruption is unlikely to persist. On March 16th Toyota announced that it was restarting the production of spare parts. As with bad weather, disasters cause some output to be postponed rather than lost. When production resumes, it is likely to be at a faster clip than usual. Studies of the economic effects of past natural disasters, as well as Japan's own experience after the 1995 earthquake in Kobe, provide further reassurance. They suggest that the macroeconomic effects of the tsunami, though hardly negligible, will not be devastating and will not last very long.

How much a natural disaster reduces output over the medium term depends on a number of factors. Location matters: disasters that strike an industrial belt will be more economically crippling than ones that hit an area where little is produced; the economic effects of the tsunami would have been much worse if it had struck Japan's industrial heartland. And different kinds of natural disasters have different consequences for growth. In the medium term, not all effects are negative. A 2009 World Bank study found that by increasing soil fertility, a typical flood increases agricultural output in the year after it strikes (though output falls in the year it occurs). The benefits from higher agricultural production spill over to other sectors, and in developing countries where the farm sector is a bigger part of the economy this may be enough to lead to faster growth in manufacturing and services in subsequent years.

Earthquakes, on the other hand, have small but consistently negative effects on economic growth. This is because earthquakes do not just shut production down for a while. They also destroy factories, roads, electricity lines and offices. This destruction does not directly reduce a country's GDP, which measures the value of the flow of goods and services that an economy produces. But it does affect an economy's underlying productive capacity. The Japanese tsunami fits this template.

As long as these assets remain out of commission, the output they would have produced is, in theory, lost. In practice, this negative effect can partly be made up by using plant and machinery in areas unaffected by the disaster. Most factories do not run at full steam all the time; output from plants that are still working can be increased to make up for lost production elsewhere.

An analysis of the effect of the Kobe earthquake by George Horwich of Purdue University provides some reason to hope that this might happen in Japan. The quake ravaged many of the facilities of what was then the world's sixth-largest container port and the source of nearly 40% of Kobe's industrial output. Over 100,000 buildings were completely destroyed, and many more badly damaged; 300,000 people were rendered homeless; over 6,000 died. Yet despite this devastation in a big production centre, the local economy recovered very fast. Even though less than half the port facilities had been rebuilt by that stage, within a year import volumes through the port had recovered fully and export volumes were nearly back to where they would have been without the disaster. Less than 15 months after the earthquake, in March 1996, manufacturing activity in greater Kobe was at 98% of its projected pre-quake level.

Mr Horwich reckons that the likely reason for this rebound in output and economic activity, even as swathes of infrastructure still lay in ruins, is that output can be produced using different combinations of labour and capital. Although a disaster may destroy physical capital, things can be made using more labour or using it more intensively than before. In addition, rebuilding is easier than building up capital the first time around, because it mainly aims to replicate a pattern of investment rather than figure out what to invest in. And productivity growth may accelerate when new, and often superior, machinery is installed.

A glimmer, at least

Reconstruction itself, of course, also helps to offset the negative impact of a drop in output in the aftermath of a disaster. Business booms for builders and producers of capital goods. Disasters probably do not actually stimulate the economy because additional production in some sectors may be displacing spending elsewhere, though this is less of a worry in an economy with a lot of spare capacity. Certainly, the year of the Kobe quake was not a bad one for the Japanese economy, which grew by 1.9% in 1995 compared with 0.9% growth in 1994.

There are grounds to hope, then, that this month's terrible events will not cause lasting damage to Japan's economy. But there are worries, too. The nuclear crisis adds greatly to uncertainty. Consumer and business confidence is fragile. With interest rates already at zero, policymakers have little wiggle room. Japan's manufacturing sector is running closer to full capacity now than in the mid-1990s, making it harder to make up for lost output. When disasters occur can matter as much to the economy as how bad they are.