The Time to Repair the Roof Is When the Sun Is Shining

On 29 October 1999 a super cyclone made landfall on the coast of Odisha with winds of over two hundred and fifty kilometres an hour, and killed around ten thousand people. The state had no effective early warning system and few places to shelter. After the disaster Odisha created a State Disaster Management Authority, built hundreds of cyclone shelters and trained local volunteers. In October 2013, when Cyclone Phailin of similar strength approached, the state moved about a million people out of the path of the storm in the largest evacuation in its history, and fewer than fifty people died. President Kennedy had put the principle in his State of the Union address of January 1962, when he said that the time to repair the roof is when the sun is shining.

The logic of the saying is simple. Repairing a roof in the rain is difficult, expensive and dangerous, since the work is hurried, materials are scarce and the damage is already done. Repairing it in fair weather is cheap and leisurely. The same asymmetry applies to many things. A bridge inspected in the dry season, a vaccine stockpiled before the epidemic and a savings account built in good years each cost little compared with the cost of fixing the problem after it has arrived. Preparation converts a future emergency into an ordinary expense.

Economists have built institutions on the idea. In 1990 Norway passed a law to place its oil revenues into a fund that invests abroad, and it received its first deposit in 1996. The fund is now worth more than a trillion dollars, and it allows the country to spend the proceeds of a finite resource over generations, without being whipsawed by the oil price. Chile created a stabilisation fund for its copper revenues, and many states in India now maintain budget stabilisation funds. Savings made when the sun shines are meant to be spent when the rain comes.

Public health shows the stakes. Epidemiologists had warned for years that a respiratory pandemic was likely, and some countries acted on the warning. South Korea, which had suffered the MERS outbreak of 2015, revised its laws to allow rapid testing and tracing, and was able to use them when Covid arrived in 2020. Countries without that experience had to improvise. Preparedness had been seen as a cost, but it proved to be a very small one beside the trillions lost in the pandemic.

Preparedness had been seen as a cost, but it proved to be a very small one beside the trillions lost in the pandemic.

Japan's experience offers a harder lesson. The Fukushima Daiichi plant was struck in March 2011 by a tsunami that overwhelmed its defences. Investigations found that the possibility of a much higher wave had been raised inside the operator before the disaster, and that the improvements which could have been made had not been, partly because they would have been costly and because the existing standards were thought sufficient. The sun had been shining for decades, and the roof had gone unrepaired because no one could show that the storm would come.

This points to the difficulty at the heart of the saying. Prevention is a thankless business, because when it succeeds nothing happens. A politician who spends money on flood defences gets no credit for the flood that does not occur, whereas the one who opens a new stadium is admired immediately. The Y2K computer problem is a case. Governments and companies spent hundreds of billions of dollars to fix software before 2000, the new year passed without disaster, and many people concluded that the danger had been exaggerated. Success looked exactly like waste. Insurance markets, building codes and fire brigades are all ways of paying in advance for things that one hopes will never happen, and the reward for foresight is a story that nobody tells.

Politics compounds the problem. Electoral cycles are short and the benefits of preparation arrive late, so leaders are rewarded for spending in good times and not for saving. Tax cuts are popular when the economy booms, and stockpiles are cut when they are not needed. The consequence is that societies tend to repair roofs in the storm. Michele Wucker calls the obvious threats that are ignored gray rhinos, in contrast with the black swans that no one could foresee, and her point is that many disasters are neither surprising nor unpredictable but merely uncomfortable to think about.

There is, however, a case on the other side. Resources are limited, and a roof can be repaired too much. France built the Maginot Line along its border with Germany in the 1930s at enormous cost, in the belief that the next war would resemble the last, and in 1940 the German army went around it through the Ardennes. A country that prepares against the wrong threat spends its money and loses its safety. Preparation also has opportunity costs, because money spent on a seawall is not spent on a school, and the poor cannot always afford to save for a storm that may never arrive.

How should a prudent society proceed? One approach is to prepare for the consequences and not the causes, building general resilience such as reserves, redundancy, flexible institutions and trained people, which help against many kinds of shock. Another is to give the work of foresight to bodies insulated from the electoral calendar, such as a fiscal council, a disaster authority or an independent statistical office. Odisha's success rested on exactly such a body, which was given a permanent mandate and a stable budget. The sun does not need to be shining politically for the roof to be repaired.

Yet the saying leaves a question unanswered, which is how long the sunshine will last. If the weather is fair for a generation, then the repairs seem wasteful, the funds are raided and the memory of the last storm fades. It is not clear how much roof is enough, or whether a society that has not seen a storm in living memory can be persuaded to believe in one. The lesson may only be learnt in the rain, which is when it cannot be applied.