Measuring economic growth from outer space

I love the satellite picture of the world at night - the one where Europe and America are seas of light, North Korea is invisible and the only lights visible in much of Africa are in South Africa and the oil flares in the Gulf of Guinea. Obviously, there is a strong correlation with economic development, or at least people having stuff to do at night.
J. Vernon Henderson, Adam Storeygard and Vernon N. Weil at Brown University take this insight a step further: why not use lighting levels to measure changes in development over time? This turns out to be a particularly useful method for countries where statistics are erroneous or missing, for example because of civil war. Governments can manipulate figures, but lighting never lies. Here are the two main findings (also summed up by the Wall Street Journal and Marginal Revolution):

1. Lighting is indeed a proxy for economic development and it goes down as well as up: there are some great pictures of changes in lighting levels in Eastern Europe in the 1990s. Whereas Poland experienced economic growth of 56% and an increase in lighting of 80%, neighbouring Ukraine saw its economic activity decline by 35% and lighting fall by 47%.

2. Increases in agricultural productivity (from high rainfall years)raise economic activity and hence lighting in nearby cities. Are these farmers rushing to sell their goods at the market, buy TVs with the earnings or simply celebrating their good fortune at the bar?

I'd love to see some further work on this data set. One question I have is whether the data might be distorted by certain high-light activities: mining, for example, or oil refining. This might matter for, say, DR Congo, where light levels seemed to increase in the 1990s, in the middle of a devastating civil war. I'd also like to see the impact of power cuts, such as California experienced a few years ago or South Africa in 2008. Most power cuts don't last long enough to show up in GDP figures, but their short-run effect might be severe: think of what happened in Europe this January when Gazprom turned the heating off. Next time that happens, we might be able to measure its effect from outer space.