Tuesday 5 February 2013

Size Matters



The beauty of economics is that you can do research on almost anything. Among my favorites are research studies on the effects of physical characteristics. Hammermesh and Biddle (American Economic Review, 1994), for example, found that pretty people tend to have higher wages. Berggren, Jordahl and Poutvaara (Journal of Public Economics), similarly, showed that pretty politicians get more votes. And Johnston (Economic Letters, 2010), concluded that blondes not only have more fun but also get paid more (at least when they collect money for charitable causes).

Size is important as well. Tall men and women earn more, attain higher education and get promoted faster. They are happier and less likely to have a variety of psychological problems such as depression. Tall men (but not women) are more successful in the marriage market too. In a recent paper of mine (with Michèle Belot, Economics and Human Biology, 2010), we show that height helps account for asymmetries in interethnic marriages, such as the fact that black men are substantially more likely to have white spouses than black women.

Given the broad range of economists’ interests in human physiognomy, it was therefore only a matter of time until the ultimate topic would be tackled: penis size. This is exactly what a recent University of Helsinki discussion paper by Tatu Westling does. His work is, in effect, an empirical study on the determinants of economic growth. Specifically, he relates growth to a number of standard determinants of growth, such as investment in physical and human capital, population growth and political institutions, alongside the average physical endowment of the countries’ males.

The findings are perplexing. Westing finds a hump-shaped relationship between penis size and GDP per worker in 1985. In other words, economic development and prosperity seem to be associated with intermediate size: countries on either end of the size distribution tend to be poorer than those in the middle. He also finds a negative relationship between penis size and long-term economic growth (over a 25 year period, between 1960 and 1985). The latter result, however, may be driven by outliers, in particular, due to the presence of less well-endowed countries that recorded impressive rates of growth, such as Japan, Singapore and South Korea. Omitting these might make the relationship between size and growth also hump-shaped. Nevertheless, the statistical power of penis size is impressive: this variable alone explains 15% of the variation in output per worker and 20% of the variation in growth.

Now, should we take these findings seriously? Westling himself sees his work more as a methodological note than a fundamental contribution: he admits that his results are quite likely to be an example of spurious correlation or omitted variable bias. It is also unfortunate that he did not replicate his analysis with different periods, to see how stable these relationships are. Nevertheless, Westling does suggest a plausible explanation for the observed links: penis size is related to testosterone levels, which also determine risk-taking behavior in men. The latter has important implications for productivity, from health and safety during work or leisure to investment decisions. The hump-shaped relationship between output per worker and penis size might therefore indicate that there is an optimal degree of risk taking: too much or too little translate to less economic development. It would also suggest that differences in attitudes to risk among countries are not necessarily driven by culture but should perhaps be rather attributed to genes or differences in diets.

In other words, further research may be required to put this matter to rest.

After the Storm

Last year, Hurricane Sandy hit the Caribbean and the East Coast of the US. It caused widespread damage and scores of fatalities. In the US alone, some 130 people died and the economic damage is estimated at around $50 billion.

What will be the medium to long term impact of Sandy? Here, the news is not all that gloomy. A small yet interesting literature has looked into this question (two representative examples were published in the journal Economic Inquiry by Skidmore and Toya, in 2002, and Crespo, Hlouskova and Obersteiner, in 2008). They posit that the long-term impact may be in fact positive, at least for developed countries. Accordingly, natural disasters can be seen as akin to Schumpeterian creative destruction: the process in which old physical assets and technologies are discarded and replaced with new, more productive, ones. Natural disasters are thought to act this way because old physical assets are more vulnerable to their destructive forces than new assets. A new building, constructed using modern technologies and building materials, is probably more robust than one built 80 years ago. And a family with two cars fleeing an approaching hurricane, is likely to drive away in the newer and more valuable car; they are also likely to take their new i-pads and laptops with them but leave the old desktop computer behind. As a result, physical assets are replaced before their time is up – and their replacements are better and more modern.

As an added effect, high frequency of natural disasters encourages substitution from investing in physical capital towards human capital: high wind can blow away your roof but not your skills and education. I was reminded of this when visiting Japan where traditional buildings often consist of little more than wood, bamboo and paper. Such houses are bitterly cold in winter but tend to withstand earthquakes rather well; if they do not, they can be replaced relatively easily and cheaply. Living in an earthquake-prone region has limited the choice of construction technologies in Japan; perhaps it is not a coincidence that Japanese students consistently report also some of the highest test scores among OECD countries.

What about more immediate effects? Natural disasters, especially large ones such as Sandy, are followed by government relief and reconstruction. These can serve as a quasi-stimulus driving the economic recovery of the local economy – and this effect can spill over even beyond the affected region. We consider this possibility in a recent paper (“Government spending and the multiplier: New evidence from the US based on natural disasters,” Brunel Economics and Finance Working Paper No. 1224, available at http://www.brunel.ac.uk/__data/assets/pdf_file/0005/247577/1224.pdf), written jointly with Sugata Ghosh and Weonho Yang, also of Brunel University. Using detailed data on natural disasters and the associated emergency spending in US between 1977 and 2009, we find that natural disasters tend to be followed by an upsurge of federal non-defense spending and higher GDP at the national (US-wide) level and by a similar growth in state fiscal spending (largely reflecting increased transfers from the federal government) and a boost in personal incomes. Hence, the spending in the wake of natural disasters indeed displays stimulus-like effects.

Last but not least, natural disasters can help make an intellectual contribution too. My research with Ghosh and Yang was motivated by the desire to understand economic effects of government spending shocks. Much of the recent discussion on the virtues of fiscal stimulus spending evolved around the size of the so-called fiscal multiplier: a statistic that measures by how much the total GDP increases for every dollar (or pound) of new government expenditure. Ideally, this should be greater than one, although some studies suggest that it is lower than this benchmark or may even be negative.

The problem with estimating the fiscal multiplier is that one needs to identify changes to government fiscal policy which were unannounced and unexpected. Increased spending or tax cut that come into effect today would have a very different effect if they were implemented unexpectedly or if they were announced a year ago. In the latter case, much of their effect could have unfolded already after the announcement. This is where our interest in natural disasters comes in. By their very nature, natural disasters are unexpected. Sandy hit with a few days’ advance warning and even then its true magnitude was only revealed ex post. Importantly, for various reasons, the government response also varies: two disasters of similar size can be met by very different government responses. The aftermath of Hurricane Katrina that struck New Orleans in 2005, for instance, was marked by controversy over insufficient precautions and slow subsequent response on behalf of both the state and especially the federal government. Therefore, we can use the emergency spending after natural disasters as an instrument to identify fiscal shocks and to estimate the fiscal multiplier in a way that is free from potential anticipation effects.

Reassuringly, we find the fiscal multiplier to be safely above one: our estimate of the peak-effect multiplier is around 1.4, both at the national and state level. In other words, fiscal stimuli are indeed effective at reinvigorating the economy.