Sunday, July 28, 2013

International wars

I have heard many times that the world that we live in is a much more dangerous place than before, that the number of international conflicts is much higher than last century.

The truth is that even if the WWII is ignored the number of conflicts since the end of the Cold War has been dramatically decreased.

The Systemicpeace website provides a great amount of data about the number of international conflicts. According to this data the turning point in the number of international conflicts was the end of the Soviet Union in 1991. 


Before the end of the Cold War, more and more countries were involved in armed conflicts. There were at least three factors involved for this increase, I reckon. The first factor is the end of colonialism and the wars of independence from metropolis. The second reason is the indirect confrontation between the two super powers (USA and USSR) in third parties´ territories like Angola, Namibia, Vietnam, Central America, Chile, Afghanistan,… And finally, the third reason is pure logic, the more countries there are in the world the more likely to see a conflict between two of them.

Why the huge drop after the end of the USSR? A part from the obvious political reason, economically, one country would only be involved in a conflict if what it takes from it is more than the investment needed. As the West has seen, invasions are very expensive and apparently the investment to stabilise countries after the war is even more expensive. But also, there’s a cost of opportunity, today land is not a big stake of GDP, instead information, innovation and creativity is and those do not need land. Moreover, international trade is greater than ever and a war would mean a huge loss for the two involved.

It may seem then that the need for economic growth was the reason behind the big drop in conflicts. However, both concepts are related to a third one. GDP is linked to freedom a higher freedom higher GDP and conflicts are negatively correlated to freedom:


Monday, July 15, 2013

Google correlate

 For those who don’t know Google Correlate it’s a wonderful tool. Given a temporal or spatial data, Google Correlate determines which queries at Google best mimic the data. It allows for automated query selection across millions of candidate queries for any temporal or spatial pattern of interest. Similar to Trends and Insights for Search, Google Correlate is an online system and can surface its results in real time.

For example let’s say you want to find what queries are more correlated to mortgage applications in America. Then, you just need to upload the data and Google Correlate will find what query is most correlated to it. According to this, the search volume of “refinancing calculator” has a very high correlation with the actual value of mortgage applications. 


Monday, July 1, 2013

Auto-motivation and professional success

A very interesting paper (or here) by  Carmit Segal shows that one of the best predictors of professional success is intrinsic motivation  (auto-motivation) in absence of incentive.

Dr. Segal compared standardized tests taken by university students to their professional careers several years later. To make the correlation meaningful, those tests had to be, firstly, unincentivized (to measure auto-motivation) and, secondly, not linked to knowledge (to rule out any potential IQ effect).

First, in order to rule out incentives, tests needed to be unrewarded or unbinding on university marks:

“Classical economic theory predicts that if an action, like taking a standardized test, requires effort, then without performance-based incentives individuals will invest the lowest effort possible in performing the action. Therefore, one might expect that scores on low-stakes tests (tests administered without performance-based incentives) would be uninformative about an individuals cognitive ability, and thus that low-stakes tests scores would be uncorrelated with test-takersfuture economic outcomes.”

Secondly, in order to rule out IQ effects, tests had to be easy.

“Ideally, to investigate whether test-takers differ in their intrinsic motivation to take a test, we would like to find a test such that all test-takers have the knowledge necessary to correctly answer all questions if they so desire. The coding speed test seems a likely candidate to fulfill this requirement. It seems likely that everyone who knows how to read has the knowledge to correctly answer questions on the test. Therefore, due to its simplicity, speed becomes a good dummy of intrinsic motivation”


Results of the comparison were clear: “nearly all researchers find strong positive correlations between high scores on unincentivized tests and future labor market success.” The paper suggests that those unincentivized scores capture intrinsic motivation. Moreover, it suggests that intrinsic motivation is associated with favorable personality traits, such as conscientiousness, which lead test-takers to put effort into low-stakes tests and to future labor market success.


A good explanation might be that adult labour markets are often uncertain about rewards and therefore auto-motivation must play a key role.