Big Data and the Pursuit of Better Disaster Relief

As Hurricane Irma ripped through the Caribbean and the United States, leveling homes, crippling energy and water systems, and putting ports and airports out of commission, government officials looked ahead to the immense relief challenges that lay in store. First, they had to take stock: What sectors of the economy were most devastated? And where should relief aid go to satisfy people’s immediate survival needs?

Such are the quandaries policymakers face after virtually every natural catastrophe, and they can prove overwhelming. Aggregate statistics on agricultural or industrial production and GDP can take months to compile. They are of little use in the aftermath of a disaster when getting supplies to people immediately may be a question of life and death.

But policymakers have a new tool: big data. Data from online retailers can provide objective and useful information on which products are available and which are not and where disruptions in supply are occurring.

This may help solve a perennial problem when dealing with disaster relief: how to best allocate the scarce resources available for aid? Economists are tempted to search for the information that can be derived from market prices. If the prices of bottled water, diapers, and flashlights soar in the days following a hurricane, for example, a policymaker might rationally assume that the producers of those goods had been hard hit and that government intervention was needed to restore supply.

Unfortunately, it’s rarely that simple. As I discuss in a paper published with Alberto Cavallo and Roberto Rigobon, retailers’ may not be able to raise prices, either because they are prohibited from doing so (for example because of anti-price gouging laws) or because they themselves want to project an image of fairness and altruism. These may prevent them from charging higher prices even when disruptions in production or transport mean they are being charged more by their suppliers. Fearing accusations of price gouging, they may simply decide not to restock, making essential goods unavailable. This may not be so much the case with perishable goods, like eggs and meat, given consumer’s understanding of how a disaster can harm food supplies. But it is certainly true when it comes to easy-to-stock, non-perishable goods like pasta, diapers, and powdered milk.

Whatever the case, our study, examining the catastrophic 2010 earthquake in Chile and the even more powerful earthquake the following year in Japan, shows that prices for specific goods in the aftermath of natural disasters are a poor indication of supply disruptions. Indeed, we found that while the number of goods available for sale fell 32% in Chile and 17% in Japan from the moment of the disaster to their lowest point, prices for goods did not increase for months after the earthquakes, even for goods that were running desperately short.

This is where the big data available from the advent of online shopping provides a potentially excellent tool. When we looked at the post-disaster situation in Japan and Chile, for example, we were able to access data from the Billion Price Project (BPP), an initiative at MIT that uses software to monitor the public web pages where retailers list product and price information on a daily basis. Examining such pages, allows one to see both how prices change and what products disappear from day to day as they become unavailable.

On-line transactions, of course, are only still a small share of retail sales in most countries. But since research suggests that on-line sales are representative of offline ones, we can feel confident that examining retailer’s public web pages will tell us what products are running scarce and what sectors of the economy are likely affected in the event of a natural disaster so that effective relief can be delivered.

Governments in affected countries face a long road ahead as they begin to dig out from the tumbled mass of crumpled roofs, fallen trees, and toppled telephone polls and try to resurrect their economies. But first and foremost they must attend to the survival needs of their people. To the extent that they can rely on big data, now or after future natural disasters, they will have a useful tool.


Print Friendly, PDF & Email

The Author

Eduardo Cavallo

Eduardo Cavallo

Eduardo Cavallo is currently a Lead Economist at the Research Department of the Inter-American Development Bank (IDB) in Washington DC. Prior to re-joining the IDB, Eduardo was a Vice-President and Senior Latin American Economist for Goldman Sachs in New York. Eduardo had already worked at the IDB as a Research Economist between 2006 and 2010. Before that he served as a research fellow at the Center for International Development (CID), a visiting scholar at the Federal Reserve Bank of Atlanta, and a member of the faculty at the Kennedy School of Government's Summer Program. In Argentina he co-founded Fundación Grupo Innova. Eduardo’s research interests are in the fields of international finance and macroeconomics with a focus on Latin America. He has published in several academic journals, and is the co-editor of the IDB book Dealing with an International Credit Crunch: Policy Responses to Sudden Stops in Latin America. He holds a Ph.D. in Public Policy and an MPP from Harvard University, and a B.A. in Economics from Universidad de San Andres (UdeSA) in Buenos Aires, Argentina.

Leave a Reply

Your email address will not be published. Required fields are marked *

Ideas matter © 2016