This is partly a quirk of the conventional way we measure economic activity. If someone has to spend a weekend mopping up their home, the time comes out of their leisure time.

In other words it is costed at zero, the same as if they were walking in the country or fishing. And the expense of fixing a house appears as greater economic activity . . . resulting in a higher GDP . . . a perverse result if ever there were one.

But it is also a sign of just how big a modern developed economy is. Take a �5bn figure for the cost of this disaster.

That is a bit higher than the �3bn currently quoted but I suspect the numbers will creep up. If that is right it will turn out to be a similar order of magnitude to mad cow disease.

Some numbers make this clearer. UK GDP this year will be about �1.25 trillion and will have grown by about �40bn.

So �5bn is equivalent to five or six weeks' growth. Looked at another way, it is the difference between 2.75% and 3% growth.

The greater potential danger (in economic terms) comes if there is a lasting disruption, or a disruption that damages the underlying potential of an economy or a region to recover. San Francisco recovered swiftly from the 1906 earthquake but Galveston, the port of Houston in Texas, never recovered its economic significance after its destruction by a hurricane six years earlier.

I think in this instance there will be no lasting damage to the region . . . economic activity is not going to move away from the Severn or Thames valleys.

The most recent experience of similar disasters . . . the flooding in Prague in 2002 for example . . . is that economies do bounce back. Only if there is some fundamental underlying weakness, as in New Orleans, does damage prove lasting. The economic losses to New Orleans post-Hurricane Katrina have been estimated at $125bn, which puts these floods in context.

In the coming weeks there will be a huge amount of anguish about the need for better flood defences, as well as the wisdom of building on flood plains.

Where there is an obvious and continuing danger it makes a lot of sense to spend money to avert it. The London flood barrier is designed to protect the capital from a surge of water, calculated as likely to occur only once every 100 years or more. The barrier has turned out to be useful in protecting against the more general problem of rising sea levels. Its cost, some �1.5bn in today's money, seems good value set against the damage a London flood might cost.

The country with the greatest experience of flood protection is the Netherlands, which spends nearly �2bn a year on flood defences . . . 70% of the country's GDP is generated on land below sea level.

More generally, a World Bank and US government study calculated that spending $40bn in mitigation and protection could have reduced the losses from natural disasters in the 1990s by some $280bn.

All this gives a guide as to how to proceed once the waters subside. One obvious decision will be where building should be allowed to take place.

It is fascinating looking at the aerial pictures to see how the medieval core of cities such as Tewkesbury remain above water, whereas new developments are flooded. They knew what happened to rivers when they founded the place.

On the other hand, it may be economically sensible to build on a flood plain provided you accept that your buildings will be flooded every 100 years.

Most new commercial buildings have a projected life of 60 years or less. Better to use the cheap and convenient land and take the risk of periodic losses.

The really important thing, though, is that the risks should be properly assessed and the costs properly attributed. The state has to properly fund the emergency services. It also has a responsibility for hardship relief. And it has an overriding responsibility for infrastructure, including flood defences. If it has to pay for these, it has a right to regulate where buildings are placed.

As for insurance, this will be self-policing. It may well be that insurance of certain types of property will become very expensive. So be it. People have a choice as to whether to selfinsure or whether to find property in less exposed positions.

The big point here, surely, is that we need to apply the lessons of this disaster to other potential catastrophes. We need to do this coolly and rationally. For one thing is sure: sadly, there will be much greater disasters ahead.