Facebook Google Menu Linkedin lock Pinterest Search Twitter



Nov 4, 2012

Natural disasters: does government matter?

The loss of life is tragic and the economic impact enormous, but places like New York are much better able to deal with natural disasters than most of the world's population.

User login status :


Predicted annual deaths from disasters by GDP (hypothetical nation with population set at 100 million and year at 1990). Source: Kahn, 2005

So far there are a confirmed 182 deaths from hurricane Sandy. All deaths are tragic but given New York City’s population is around 8 million, it’s a welcome relief that so far only 48 New Yorkers died as a direct result of Sandy.

That’s not likely to be the result of good luck though. It’s the result of good government, good institutions and a high standard of living.

Harvard economist Ed Glaeser argues that Sandy highlights the importance and value of good government:

Death tolls from comparable natural disasters are substantially lower in countries with more capable governments. In 2010, massive earthquakes hit both Chile and Haiti. In competent Chile, deaths numbered in the hundreds, while in chronically mismanaged Haiti, tens of thousands died.

His numbers are conservative. An estimated 300,000 people died in Haiti compared to 550 in Chile in the following month, even though the Chilean earthquake was an order of magnitude more powerful – 7.0 vs 8.8 on the (logarithmic) Richter Scale.

Professor Glaeser cites an important paper published in 2005 by Matthew Kahn from UCLA, The death toll from natural disasters: the role of income, geography and institutions. Professor Kahn begins his paper by comparing the second and third most populous countries in the world.

Between 1980 and 2002, India experienced fourteen major earthquakes that killed a total of 32,117 people while the United States experienced eighteen major earthquakes that killed only 143 people.

His key message is that “economic development provides implicit insurance” against disasters. Nations that are democratic and have “higher-quality institutions suffer less death from natural disaster.”

Professor Kahn analysed a data set of annual deaths from disasters in 73 countries between 1980 and 2002. It records more than 4,000 disaster events that resulted in the deaths of over 800,000 people.

He sought to discover how the relative number of deaths from earthquakes, floods, windstorms, extreme temperature events, landslides and industrial disasters (like Bhopal) depend on a country’s geography, income and institutions.

He found geography makes a difference. For example, African nations and those closer to the equator suffer fewer disaster events and fewer disaster-related deaths than Asian nations. Deaths are inversely related to elevation (fewer windstorms) but increase with distance from the equator (more earthquakes).

Wealth makes a bigger difference. While the income of a nation has little to do with the number of natural disasters it experiences, richer nations suffer many fewer deaths (see exhibit):

If a nation with a population of 100 million experienced a GDP per capita increase from $2,000 to $14,000, that nation would suffer 764 fewer natural-disaster deaths a year.

There’s a host of possible reasons for this. Richer countries “can provide implicit disaster insurance through effective regulation and planning and by providing quality infrastructure.”

For example, they have higher quality emergency services and better immediate and long term health care. They have better early warning systems and more durable buildings.

Good government and high social capital also make a difference (and in most cases are the foundation for higher incomes). He says less democratic nations and those with higher income inequality suffer more deaths. Corruption might lead to more deaths through “lack of enforcement of building codes, infrastructure quality, and zoning.”

Democratic nations on the other hand are more accountable and more responsive to their citizen’s welfare and desires. Moreover, “nations with stronger institutions suffer lower national death counts from disasters.”

As Professor Kahn says, his findings have implications for how the world deals with the expected increase in the frequency of natural disasters arising from climate change.

And for all its failings (e.g. New Orleans), it’s likely the US is much better equipped to deal with the long-term economic and health impact of Sandy than the Caribbean nations.

P.S. Contrary to what the Twitterverse thinks, there’ve been more deaths in the US and Canada so far from Sandy than in the Caribbean (112 vs 70 at the time of writing).

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

The Urbanist is edited by Dr Alan Davies, a principal of Melbourne-based economic and planning consultancy, Pollard Davies Consulting.

Get a free trial to post comments
More from Alan Davies


We recommend

From around the web

Powered by Taboola


Leave a comment

8 thoughts on “Natural disasters: does government matter?

  1. Neil Hunt wrote directly to Crikey Insider (the subscriber newsletter) to say:

    Re. “Hurricane Sandy: why good government, not luck, saves lives” (yesterday). I just wanted to correct Alan Davies on his use of 7.0 vs 8.8 magnitude. This is measured on the Moment Magnitude Scale, not the Richter scale. Seismologists have not used the Richter scale since the early 1980s. While the two measure at a similar magnitude for a particular earthquake, how they’re derived is completely different.

  2. I don’t know that Indian people will appreciate the “undemocratic” slur, or for that matter that India’s massive public service machine (inherited from Britain) equates with a puacity of public institution.
    Another factor that seems to have been missed is population density in concert with proximity do disaster prone areas. India’s people are equally more likely to live in large numbers on river deltas (flood) or mountain sides (quake/landslide). While Manhattan probably compares in terms of population density, it’s location is (or was, until the ice-caps started to dissappear) fairly ideal – and that has nothing to do with the current system of government.

  3. Ari Sharp #5:

    I think that’s spot on. Acemoglu and Robinson have a lot to say about corruption and poor institutions in their new book. Here’s my summary of their argument from a few months ago:

    Acemoglu and Robinson’s contention is that the pattern of world inequality is determined primarily by differences in political systems. Poor nations are commanded by “extractive” elites – Kings, Priests, Dictators, merchants – who appropriate wealth. Whether in Nigeria or North Korea, these elites aren’t about to give up their extractive power.

    Rich countries however got wealthy by having inclusive political and economic institutions that facilitate and sustain economic growth and encourage innovation. They have secure property rights, freedom to contract and exchange, and pluralistic political institutions.

  4. Perhaps a little beyond the scope of a piece on natural disasters, but I strongly suspect man-made disasters also rise in poorer places or those without high governance standards.

    Here in Indonesia, bridges collapse at an alarming rate, dilapidated boats on hops between remote islands, basic houses disappear amid mudslides and buses crash along poor-quality roads.

    A better building code and less corruption among officials charged with enforcing it would surely go along way to keeping people safe.

  5. IkaInk #3:

    I expect its poorly expressed. Here’re the passages I relied on for my (passing) reference to distance from the equator:

    Larger nations, those that are more elevated, and those that are farther from the equator are more likely to experience shocks (280).

    This ZINB regression model’s results indicate that a nation’s elevation has a negative effect on death counts, whereas distance from the equator raises death counts. (280)

    As shown there, elevation has a negative and statistically significant effect in reducing deaths from windstorms, whereas distance from the equator raises deaths from earthquakes (280).

    Unlike other measures of economic performance, natural-disaster deaths are lower for nations closer to the equator. (283)

    Maybe in one breath he’s talking about disasters collectively and in another about a specific type of event e.g. earthquake?

    Kahn is a scholar of high international standing. I’d be amazed if he got the method wrong.

    Update: Now I see the problem! I’m using the final version as published in The review of economics and statistics, May 2005. It’s gated so I’ve linked to a working paper version. This is a great way many US academics have of making their work available without contravening copyright. In this case it seems the author’s made some small changes for the published version.

  6. That paper you’re discussing either has some typos in it; or I’m misreading it, and its written in a very confusing manner.

    “The zero inflated negative binomial regression estimates reported in columns (2) through (5) in Table Seven also provide evidence on the importance of geography in explaining deaths from natural disasters… Nations further away from the equator suffer less death from natural disaster. (p18.)

    Death counts differ sharply by continent. African nations experience fewer
    natural disasters and all else equal, suffer less death from natural disasters. Unlike other measures of economic performance, natural disaster deaths are lower for nations closer to the equator.

    Either I’ve somehow missed some context which is providing different results for different data or something, or his conclusion has statements directly contradictory to this thesis.

  7. Mark, you can see for yourself. The link to the paper’s in the text. Although, rather than “more earthquakes”, probably should be more exposure to earthquakes.

  8. “Deaths…increase with distance from the equator (more earthquakes)” WTF? If there’s any relationship between latitude and earthquakes at all, it must only be in a very limited sense, otherwise the poles would be the most seismically active areas on the planet. Certainly this doesn’t accord with any tectonic theory I’ve ever heard of, let alone the orthodox one.