Monday, August 31, 2009

Competition is key to getting good politicians

Wouldn't we all love to have competent politicians? Vincenzo Galasso and Tommaso Nannicini show you how to get them. First the theory: suppose you care about having people in power who are experts and hard-workers and favor your policy choices. Parties allocate candidates to districts, knowing that some districts will be more contested than others. For district that are safe bets, they will allocate relatively incompetent candidates. In toss-ups, the best candidates will compete.

Does this really happen? Galasso and Nannicini also look at data. They used parliament elections in Italy, where they know a fair bit about each elected candidate, including education, past appointments, pre-election income and absenteeism. And these variable vary in expected ways with the safeness of the districts, as measure by previous election margins. Also, they find that left-wing and right-wing candidates look very similar to each other in close elections, whereas they differ much in safe elections, following predictable patterns: right-wing candidates have private sector experience, are late starters in politics and are more educated, left-wing ones are have more local political experience, are more female and are older.

In conclusion, as a voter you should vote against sure winners in order to attract better candidates. And these candidates will more likely be centrists.

Friday, August 28, 2009

The long-term impact of hurricanes

Hurricanes create havoc and large costs, as nobody doubts. Some of those costs are easy to measure: life are lost, property is damaged, insurance companies pay out claims. But there can be a more subtle, long-term cost that goes beyond the physical replacement of destruction. Hurricane Katrina still has a profound impact on New Orleans, four years after landfall and will have for many years to come, as the city is still much less populated than before and its labor market is profoundly changed. Four years is still short, though, to analyze the long term.

Makena Coffman and Ilan Noy study the case of the Hawaiian island of Kauai that was affected by hurricane Iniki in 1992, while neighboring Maui was not. Comparing the two islands, it appears clearly that while externally there are few signs of the hurricane seventeen years ago, Kauai is still suffering, mostly because its labor market still has not recovered, despite massive transfers right after the storm. Population took a permanent hit, at least partly as a consequence of a sudden drop in the housing stock and a spike in unemployment, and never recovered compared to the neighboring island.

New Orleans shows similar signs, if not stronger, given that many are saying this city is poorly located and its partial destruction should be taken as an opportunity to relocate. It appears the same happened on Kauai, although one has to wonder, because the other islands of Hawaii face very similar risks.

Thursday, August 27, 2009

Unemployment: Culture matters more than policy

Why do some areas have higher unemployment? There are some obvious explanations, but could cultural aspects matter? For example, the social stigma attached to being unemployed could vary from a region to the other. This is very difficult to measure and typically ends up in the error term or the regional fixed effect of a regression.

Beatrix Brügger, Rafael lalive and Josef Zweimüller use data from Switzerland. There are two mains cultural groups in this country, the French and Italians speaking, called them the Latins, who have more of a Latin nonchalance and express it by not working more than necessary, and the German speaking, with a strong work ethic. (As a side-note, it is remarkable that Calvinism started in French-speaking Geneva). Of particular interest here is that the boundary between Latin and German speaking regions is well-defined, but does not overlap with political or labor market boundaries. A perfect setting to study difference in border regions using a micro dataset.

And the differences are important. For example, the duration of unemployment is 7 weeks longer in the Latin regions. This has nothing to do with discrimination, as the German speaking workers are much more likely to find a job on their ones, instead of through an employment office. And looking at those who moved from one community to the other, it appears these values are rather shared with those you live with rather than those you grew up with. Finally, the impact of these values is stronger than what policy changes would typically yield.

Wednesday, August 26, 2009

Strategic tax auditing

There is a well published on tax competition between authorities, where the object of competition are tax rates and occasionally public services. But there is another neglected aspect: competition on the thoroughness of tax auditing. This is relevant in countries where local authorities are also in charge of collecting the taxes of the central government.

Alexander Libman and Lars Feld consider the case of modern Russia and find that indeed the willingness to audit depends on the strength of the Moscow. This has two channels: the first is about where audit revenues are directed and whether audit happen in the first place. If there were no audit variations, the share of revenue going to the local government should not vary across regions. Yet, Libman and Feld find that there are systematic variations that can be explained by the power structure, in particular during the loose administration of Boris Yeltsin.

Tuesday, August 25, 2009

Saving for the nursing home and aggregate capital

Looking at when health expenses are made, in particular out-of-pocket ones, a large share of them are made towards the end of one's life, in particular for nursing home stays. Thus when studying savings in an economy, and in particular savings for retirement, it is critical to understand well savings for the nursing home. Of great importance is the institutional context: who pays for the nursing home. In the United States, there is no significant private insurance for nursing home costs, however, Medicaid picks up the tab if some means tests are met. This leads to perverse effects wherein retirees divest their savings anticipating the need for a nursing home.

Karen Kopecky and Tatyana Koreshkova study this in more detail with a life cycle model with uncertainty. They notice a splitting of the population: the rich, whose savings behavior is driven by the large and persistent nursing home costs, and the poor, who only save towards the smaller medical costs (and avoid getting too rich to avoid losing Medicaid benefits).

Introducing generalized public health care obviously reduces the incentives to save, their is less uncertainty. The reduction in saving is drastic for the richer households if one keeps the interest constant. But with lower savings, the latter must increase, which induce more savings. In general equilibrium, the decline in aggregate savings is modest, but welfare is enhanced, due to the dramatic reduction in uncertainty for households who do not face the risk of bankruptcy due to health shocks. Doing the opposite (removing all Medicaid benefits) would dramatically increase aggregate savings (+134%), almost all on account of the poor households. But again, higher savings do not necessarily means households are better off.

Monday, August 24, 2009

Risk taking and the menstrual cycle

We tend take preferences as given and constant, but there is mounting evidence that preferences change over the life cycle and over external circumstances, as I reported before. There may now even evidence that preferences that preferences follow a predictable cycle, at least for women.

Matthew Pearson and Burkhard Schipper asked an unusual question to the female participants in an otherwise standard auction experiment: what stage of your menstrual cycle are you in? It turns out the answer can explain bidding behavior: while otherwise indistinguishable from men, women in their menstrual and premenstrual phase bid significantly higher, thus exhibiting more risk taking. Pearson and Schipper argue this can be explained by biology and evolution. This is when women are most likely to be fertile and take risks to increase the probability to procreate.

The experiment is not entirely clean, it would be preferable to have the same women participate repeatedly during a menstrual cycle, but even this indirect evidence is intriguing. And it should tell women to take their cycle into consideration when taking decisions.

Friday, August 21, 2009

Intellectual protection should be decreasing, not increasing

Have you noticed how the grip on intellectual property law keeps expanding: copyright periods lengthen, the scope of patentable "innovations" widens, and the enforcement of intellectual property become the topic of international trade negotiations. But should we expect this?

Michele Boldrin and David Levine look at this using the age-old trade-off in intellectual property protection: long protection provides the innovator with monopoly rents and thus incentives to create more innovations, whereas should protection allows society to benefit earlier and more widely of these innovations. As you move the protection duration (or scope or enforcement), the question really is how many new innovations one gains or loses at the margin. The distribution of innovation thus matters a lot as the marginal idea (in terms of quality) will be pursued. Ultimately, you want to measure the elasticity of revenue with respect to the marginal idea.

Boldrin and Levine do this with various empirical strategies that all come to a similar conclusion: the elasticity mentioned above increases a lot with the quality of the marginal idea. Also, they find that the growth rate of ideas is lower that that of population, there are decreasing returns to scale. What this means is that protection for intellectual property should be decreasing with the scale of the market. And as globalization has dramatically increased that scale, protection should be decreasing rather than being reinforced.

Thursday, August 20, 2009

Why are so many elderly widows living alone?

It used to be the case that most elderly people, once widowed, would live with one of their children. Very few do so now, and one can have several conjecture for the reasons thereof: societal changes, government sponsored care for the elderly, better retirement pensions, emancipation of the children or the elderly, etc.

Carlos Bethencourt and José-Víctor Ríos-Rull find that the increases in incomes of the elderly can explain the most in this change. Indeed, their income has increased twice as fast as that of their children in the 1970's and 1980's. They come to this conclusion after estimating a rich model that takes into account household economies of scale, the game that parent and child play when investing in the home (they may have different preferences) and differences in income. Many specifications come to basically to the same conclusion: it is the elderly who want to avoid living with their children, and they started doing so as soon as they had sufficient income. Thus the story is not about ungrateful children who push their parents away.

Wednesday, August 19, 2009

Mortgage innovation was good for home ownership

In the United States, there have been two periods during which home ownership rates increased markedly: 1940's-1950's (47% to 65%) and since the mid 1990's (65% to now 70%). Why would there be so drastic increases interrupted by a long period of stable homeownership rates. Clearly this is difficult to justify with demographic changes, which are much more gradual than that. Another justification could be changes in the tax treatment of homeownership, but the timing is not right.

Matthew Chambers, Carlos Garriga and Donald Schlagenhauf say this is all due to innovations in mortgages. In particular, the creation of new mortgages that are more accessible to young households (low down-payment) during the 1990's help Americans to buy their first houses earlier. Also, pre-WWII, mortgages were for 10 years, interest only and thus implied a massive bill at the end. In the 1940's, the government encouraged longer maturities and amortization, which made buying a house less risky.

The real contribution of this article is to determine how much of the increase in homeownership can be attributed to new mortgage options. Chambers, Garriga and Schlagenhauf address this with a frighteningly rich model and they show that they can explain half of the ownership increase in the earlier period, and three quarters in the more recent one. So these new mortgage tools did something good, as long as we believe that home ownership should be encouraged, which is not obvious.

Tuesday, August 18, 2009

I made it in the SSRN top ten!

I just learned that one of my papers made it to the exclusive list of the SSRN top ten. Well, that was a few weeks ago, as my listed contact details are wrong. But I learned it from my co-author, who is actually not listed either. He stumbled on the listing while looking for similar material. And to add to our confusion, neither of us submitted the paper...

In any case, should I be proud of this achievement and rush to my dean to announce the news? Well, I counted on the SSRN web site 83 different topical journals, each having a top ten list each month. Assuming no overlap, that makes 83 times 10 times 12, or 9960 top ten papers a year on SSRN, just for economics. I could not figure out how many new papers in economics are uploaded on SSRN, or what the current stock is, but I guess my paper may have been downloaded more than half of the papers. Not that much of a distinction. However, given the fact that I did not know about the paper's presence on SSRN, I cannot possibly have manipulated the download count. My numbers are absolutely clean.

Oh, and the download count? Barely double digit.

Monday, August 17, 2009

Limit foreigners in sports leagues

I have always found it difficult to root for the local sports teams knowing that very few of their players are actually locals. In professional sports, teams carry the name of a location only because this is where they happen to play half their games, or possibly because management is local, but certainly not because of the players. In fact, I have always been in favor of some rule that would force teams to have at least some local content, so that local residents can better identify with these teams: "the players are from us."

It turns out that there can be economic reasons to do so beyond local loyalty. Markus Lang, Alexander Rathke and Marco Runkel show that imposing such constraints provides for more parity in a sports league, provides more profits to leagues and local players are better paid. But such constraints need to be imposed, as teams always want to deviate from such constraints. Similar constraints have been imposed in European football, apparently with success.

Friday, August 14, 2009

Open source and private firms can coexist

Open source is a mystery to many, given that contributors give away their innovations and competitors can just scoop them up. One would thus think that an industry would either be proprietary or open source, but not both at the same time.

Gastón Llanes and Ramiro de Elejalde show that it is possible. The critical features are that the open and proprietary goods not be perfect substitutes and that open source firms need to sell for a price a complementary good to the open one. That does not seem to be very constraining, as they need this anyway to survive, even without competition from proprietary goods. A perfect example for this is the database management industry, where the free MySQL is doing very well despite Oracle and Microsoft SQL.

One consequence of this is that some industry associations that like to pretend they represent the whole industry should stop chasing those that support an open source model. The music industry seems to be a perfect example here.

Thursday, August 13, 2009

One more argument against banning child labor

Everybody agrees that child labor is a problem, including the working child's parents. Some years ago, various groups, including the International Labour Organization, called for a ban on child labor. Yet it appears resistance to a ban is mounting, none the least due to pressure from economists who have argued that child labor is not a choice. Indeed, poverty often forces parents to send their kids to work, consciously neglecting the future returns of schooling for immediate survival. Others have argued that without a proper infrastructure in place, kids would not go to school anyway. Finally product boycotts against products thought to be the results of child labor are now thought to be counterproductive as they impoverish even further the targeted populations.

Matthias Doepke and Fabrizio Zilibotti add further arguments against bans and boycotts by considering their political economy aspect. A boycott or a ban from outside directly impacts only the export sector. Children are thus pushed to the non-traded sector, typically local agriculture where children engage in physically less demanding work than adults. Thus specialization occurs and children do not compete with adults. There is no local support for a child labor ban.

If there is no outside intervention, then children stay in the export sector and are in competition with the local unskilled adults. Those will now want to support a local ban of child labor so as to get rid of this competition for jobs. Thus, paradoxically, an attitude of laissez-faire in the rest of the world would lead to a child labor ban. Intervening would prevent the ban from happening.

Wednesday, August 12, 2009

Using oaths to elicit true preferences

Economists always find data elicited from surveys very suspect. Unless money or actual decisions are on the line, people may say anything. For non-monetary matters, this is particularly important problem, as there is no market observation that can provide more reliable data.

Nicolas Jacquemet, Robert-Vincent Joule, Stéphane Luchini and Jason Shogren look whether a simple trick can provide a more truthful revelation of preferences: a survey under oath. They test this with a second-price auction and find it works better than anything else. Very interesting and intriguing. One wonders how the cultural background of respondents may matter here. For example, religious people may take oaths more seriously, and they may also have preferences that differ in a systematic way.

Tuesday, August 11, 2009

How well does the flat tax work?

The debate about introducing a flat tax regularly flares up, but rarely is it based on a serious case study. The problem is that this debate is overly politicized and camps are formed on ideological lines, not an actual evaluation of the flat tax. Yet, there is data out there, as a flat tax has been introduced in Russia in 2001 and subsequently in six other Eastern European countries.

The flat tax in Russia was quite revolutionary. It is set at 13%, and replaced a system with marginal rates of 13%, 21% and 31%. So just looking at this, tax revenue has to decrease, yet it increased massively. This is not due to a Laffer curve, but rather to a drastic reduction in tax evasion, argue Yurly Gorodnichenko, Jorge Martines-Vazquez and Klara Sabirianova. Using micro data, they find that by measuring in a micro dataset tax evasion by the gap between reported income and expenses. They attribute most of the gain in tax revenue to changes in tax compliance, not to increases in economic activity. Russia's GDP grew by 5% in 2001, nothing unusual for the period.

What does this mean for a flat tax in OECD countries? First, the efficiency gains of such a reform do not seem to be that important. Second, tax compliance is already relatively high, for example in the US, which makes the gains less important in this respect. In the case of the US, a flat tax has, however, still the merit of simplifying an incredibly complex tax system that is the bread and butter of too many accountants.

Monday, August 10, 2009

The industrial organization of extortion

Extortion is illegal, but still tolerated in some parts of the world. Understanding it well is key to eradicate it. Benjamin Olken and Patrick Barron have recently published a fascinating article that describes extortion at weigh stations and roadside check points in Aceh, Indonesia. They managed to record about 6000 transactions at these location by having observers ride along on trucks on the two routes that lead into Aced from the Province of Sumatra. Of particular interest is that the "market" there differs by province, because the density of check points was higher in Aceh (there was a civil war) than in Northern Sumatra, and it has significantly dropped in Aceh after a truce led to major troop withdrawals.

The experiment is fascinating, and so are the data analysis. Indeed, it looks like extortions (to avoid various fines, mostly on truck weight) follow extremely well what economic theory would tell you. To take a few examples: "fees", or should we call them tolls, are implicitly negotiated and are higher for heavier, newer trucks with more valuable cargo at stops with more officers with visible arms. These tolls are also higher closer to the destination, a nice example of the hold-up problem. Also, with the reduction in the number of check points on the roads, the fees at each point have increased significantly. Thus market structure matters a lot. In particular, decentralization can lead to higher bribes, thus rooting out corruption at the top is not the solution.

Friday, August 7, 2009

Exploiting sunk costs

Sunk costs are cost that do not have an impact on decisions once spent. I have have recently come across some websites that use this to their advantage, so-called penny auctions. Examples are Bid Cactus and Bid Rodeo. There you bid in simple auctions that increment in very small steps, one cent for smaller items, ten cents for larger items like consumer electronics. Items typically sell at a fraction of their retail value.

But, -- there is a but -- bidding is not free. On Bid Cactus, each bid costs $0.75 and is sunk. Thus it appears to be a good idea when you bid on a $50 gift card at $1.00, but you may lose your bid the next second and need to bid again at $1.02. Auctions have a predetermined end time, but it gets extended by a few seconds after each bid. I watched a few and it is fascinating how some "players" keep getting outbid and see that their costly bids bring them nothing. I saw a $400.00 camera go for $90.00, which means there have been 900 bids, or $675.00 in revenue for Bid Cactus. Bid Rodeo has more useful information. A $670.00 laptop was sold for $19.57, thus generating 1957 bids at $0.72 or $1409.04. The winning bidder paid $295.41 for his bids, thus still got a good deal, but plenty of people paid for nothing. In terms of return, this seems similar to a state lottery...

There is also a huge potential for fraud. Clearly, the auction houses here have every interest to have people bid often, and thus could have robotic bidders participating. If by any change an item is won by a robot, they still have the income from the bid fees, and can put the item for sale again. Absent this kind of illicit behavior, I do not think these penny auctions can be legally called a scam, but they sure like a suspiciously profitable way to sell stuff.

Thursday, August 6, 2009

Public pensions vs. public investment

While being civil servant does not pay much in current wages (usually), it typically pays off handsomely in other benefits: stabily of job and pay, health care benefits and especially retirement benefits. This can be viewed as a long-term contract, where the state gives less now for the the peace of mind and future incomes. Few private firms offer this kind of benefit, one likely reason is that they do not have the power to tax in the future to meet such obligations. One can thus ask whether it is a good thing for governments to offer such generous pensions. Could they do better with that money, say, invest in public infrastructure or education?

This is precisely what Gerhard Glomm, Juergen Jung, Changmin Lee and Chung Tran address in the context of emerging economies. That context tilts obviously the answer, given that these economies precisely lack in public infrastructure and education. In addition, public service wages are quite high in those countries, making generous benefits even less appropriate. Thus, there is no doubt on the result, and the interest is in the size of the effects. The authors calibrate their model to Brazil and find indeed that public investment or simple reductions in taxes would be very beneficial (on average, of course some have reasons to complain). Indeed, cutting public pensions makes civil servants save more, and this increases GDP by 4% through added capital accumulation. The rest is just gravy: reducing distorting labor income taxes increases GDP by a total of 15% or increasing public infrastructure bring it to 10%. It is, however, highly unlikely that those numbers would apply to developed economies.

Wednesday, August 5, 2009

Explaining high unemployment and low mobility in Europe

There is an endless stream of papers trying to understand why, on average, unemployment rates are higher in Europe than in North America. I have reported here about several of the recent ones, and there seems no shortage of new explanations. In fact, if one were to build a model with all those explanations, one would probably be left to explain why after all the unemployment rate is not even higher in Europe...

So what is the latest explanation? Peter Rupert and Etienne Wasmer pick up the ball where several left it: high unemployment is due to low mobility: Europeans are much more attached to their region and are less willing to move for a new job. This begs the question as to why. This calls for a model that explains both unemployment and mobility, based on some friction that differentiates North America from Europe (and does not involve taste shocks, the catch-all for the unexplained). Rupert and Wasmer argue that differences in unemployment insurance benefits and taxes are not sufficient to explain the differential, one needs also to factor in commuting costs. While commuting time is a little shorter in, say, France, fuel costs are much higher, which explains the shorter commute and the lower mobility.

Calibrating this labor search model, Rupert and Wasmer find that indeed they can explain both the unemployment rate and mobility differentials. But I have a feeling this is not the end of the story. If the cost of commuting is so high, why not move closer to the job? European housing markets are much less liquid than in the US. Why? It seems the economic force discussed here should make them more liquid.

Tuesday, August 4, 2009

The impact of wage subsidies

Several European countries, including Germany, have wage subsidies to encourage employer to hire unemployed workers. While there is no doubt that, at the margin, this can reduce unemployment, the real question is whether those hired with a subsidy also stay employed after the subsidy expires, and have longer employment spells than those hired without a subsidy.

Gesine Stephan looks at the German case and finds the subsidy works in the sense that after 3.5 years the cumulated wages are higher for the subsidized workers. But does it work well enough to justify the cost of the subsidy. On average, the subsidy amounts to €2500 to €3000. This results in cumulative wages increased by €2200 to €5000. Is this worthwhile? First, this generates more tax and social security income, second, less needs to be spent on unemployment insurance. In a carefully worded appendix, Stephan argues it looks like wages subsidies are fiscally beneficial at least to a first approximation. While this result is subject to all sorts of qualification, I am glad that, for once, a cost-benefit analysis of a policy is performed, instead of only showing that something is statistically significant.

Monday, August 3, 2009

Patriotism and tax compliance

Nobody likes paying taxes, especially when one disagrees what is done with them. But one can suspect that the more one identifies with the taxing state, the less one would want to evade taxes. While this hypothesis makes intuitively sense, the real question is whether this effect is economically relevant.

Kai A Konrad and Salmai Qari test this using data on the size of shadow economies and a measure of patriotism. You can rightfully argue that these two variables are very imperfectly measured, as they are based on polls. They also use an aggregate measure of the shadow economy, a notoriously difficult object to measure. When regressing patriotism on tax compliance, they include of course a series of controls, but miss the most important ones: the tax rate and penalties for non-compliance. There is plenty of evidence that tax compliance will depend on them. This can explain why the shadow economy is so small in the US: tax rates are low and penalties are unusually high. This makes you wonder whether the positive impact of national pride is really measuring what the author think it is. Arguably, the tax rate and penalty would be included in the country fixed effects, but this misses the actual marginal tax rate of the polled person, especially when taxes are more progressive. Also, one cannot exclude that patriotism is correlated with tax rates or tax penalties.

Economics has taught us that people respond to economic incentives, too bad that was forgotten in this study.