Monday, December 14, 2009

Education feedbacks

Last week I found a 2003 document from Ludger WöBmann about an analysis of European education.

These estimates show that each additional year of education goes hand in hand with, for example, a 10.9% increase in wages in Ireland, or of slightly more than an 8% wage increase averaged across the European countries. Ireland, UK and Portugal give great incentives to study. Obviously something not accounted here is that it also depends on your initial level of studies; it’s not the same to increase one more year of studies when you have no studies at all than when you already have a MSc.

The following graph shows the unemployment rate by country and by studies achieved. It is clear that those countries who incentives the most to study are those that have a higher difference between ‘Below upper secondary education’ and ‘Tertiary education’, Germany, UK, Finland and France.

By both methods UK, Finland and Germany are the ones who give more incentives to study. Moreover, they are three countries that already have high education levels.

Figure 3 plots European countries’ average educational performance against a measure of the size of the effect that family background has on students’ performance in each country. The size of the family-background effect can be viewed as a measure of the equality of educational opportunities for children from different backgrounds. As there does not seem to be a clear relationship between average performance and the size of the family-background effect across European countries, the cross-country pattern suggests that there is no apparent trade-off between achieving efficiency in educational production and equality of educational opportunity. This is good news for governments aiming to jointly achieve both economic efficiency and social cohesion.

Original document :

Friday, December 4, 2009

Presidency Analysis

A more enjoyable way than hearing US presidency speeches is to analyze their speeches by words and sentences. This is what I tried with the last presidency debate in fall 2008.
A beautiful way of showing the results of the debate is the following (with Wordle):

These were the most used words by each candidate (excluding those shorter than four characters). We can see that McCain’s two most used words were ‘Senator’ and ‘Obama’ (111 times together). Senator refers clearly to Senator Obama. However, its much more difficult to find words referring to McCain from Obama (63 times). This is something for what McCain was much criticized from Republican’s side moreover if we consider, see below, that McCain said less words in total.

The rest of the words said by them were, in the case of McCain, ‘America’, ‘need’, ‘now’ and ‘going’ and in the case of Obama, ‘think’, ‘people’, ‘going’ and ‘now’. ‘Now’ was used by both due to the emergency situation, economically. ‘Going’ is commonly used to construct future sentences so naturally they were whether exposing their respective plans and programs or their thought about the future. The first difference was ‘America’ against ‘people’, those are political spectrum words pretty linked to Republicans and Democrats respectively. The second difference is ‘need’ from McCain versus ‘think’ from Obama. I would say that the second word is much more evocative, freer and opened. In the other hand, ‘think’ is more reflexive and ‘need’ is basically an action verb. Also here those are words quite linked to Democrats (said to be more reflexives) than Republicans (said to be more pragmatics).

The rest of analysis is pretty fun. Assuming that they were assigned with the same number of minutes to speech (a common rule in all debates) we can conclude that Obama was able to say 7% more words than McCain (7194 vs 6710, including all words), a faster talking. Moreover, Obama used a larger variety of words, 3% more variety (1408 vs 1364). Again words from Obama were a little bit larger, 1% larger, than those from McCain (4.47 vs 4.42 characters by word). Finally, Obama used less sentences than McCain what means that Obama sentences were larger than McCain, 23% larger to be precise.

In conclusion, Obama showed a better grammar and lexical proficiency, but we don’t know if that’s why he won the elections, I guess not.

Wednesday, December 2, 2009

Simpson's Paradox

How can be possible that unemployment rate become higher now than in 1980’s but if we broke down population by education grade each component’s unemployment rate have surged.

Currently unemployment rate is 10.8 per cent in US and in 1980’s arrived to 10.3. However, by population education’s unemployment rates were and are the following:
(Invented data)

The solution to this apparently paradox is weights. Today there are more graduates than in 1980’s so 7% weights more than the 5% did for their respective Total rates.

You can know more at WSJ > The Numbers Guy

Tuesday, December 1, 2009

Too big to fail

Due the financial crisis two questions has become repeatedly asked. Are some financial institutions too big to let them bankrupt? Do governments have to bailout financial institutions?

On one hand it is true that the spillover effects of a big bankruptcy could be disastrous for the national financial system and at the end all the economy. Even thought, a NYTimes article argued that bigger financial institutions have become a historical process and there is no turning back. Moreover, he argues, it wouldn’t be any easy to impose limits to banks.

But the truth is that first, if governments bailout financiers, those big enough will have an incentive to become over leveraged. And second and more important, something wrong exists in financial systems that when they are freely auto controlled they become unstable. This is well explained on a piece from Sharyn OHalloran, on March 11th 2009, where she exposes the coincidences between the tragedy of Commons wrote in 1968 by Garrett Hardin and the Financial system. In its origins the tragedy of Commons is a 19th century phenomenon. English cows, sheep, pigs and other livestock grazed on the commons (shared pastures which anyone could use). But the commons were always depleted, with barely enough grass to everyone’s detriment. Everyone would be better off if fewer animals graze on the commons, but each individual user has incentives to increase the size of his herd past the socially efficient point. Beyond this social efficient point, gains and only for the individual and costs are share by all. So individuals have an incentive to exceed the socially efficient point. The same happens to financial systems. When people become over confident on the economic situation they want to get into debt. Banks in order to respond extend new and greater credits. Each bank is better off extending this new credits and avoiding the rest to get it, but the whole system as the commons in 19th century is worse, too risky. This failure is also implicit in prisoner’s dilemma, telling the other is guilty makes you free but telling both other’s guilt makes their situation worst. What about if a bank, knowing this, waits until the storm pass through? As a consultancy exposed during the crisis: or you play the game or your clients won’t accept your less returns and at the end you will be forced to close down.

The solution proposed by 19th century herdsman was to assign private rights to grazing land but no barrier would be enough to separate financial world from real economy. Solutions are not easy, but necessary. An article from The Economist (23rd Nov.) pointed two papers having a proposition: ‘The role of macroprudential policy’ and ‘The Warwick Commission on International Financial Reform’.

So, we said that to fix these bailout problems, first, financial system must be regulated and second something must be done with incentives behind ‘too big banks’. One would say that big banks are necessary but an empirical study (very few there are) from T.Beck, R. Levine in 2005 conclude that while there is a link between bank concentration systems and and less risk of financial systemic crisis, the also positive relation between competitiveness an banking stability suggests that “bank concentration is an insufficient measure of bank competitiveness” and therefore insufficient to measure risk probabilities.

Finally, we know that systemic financial crises are produced by a under regulated systems but we don’t know how to fix it properly (at least consensus is far away) neither if we have to bailout big banks. What is clear is once done, once a big bank is bailout we should demand for considerations towards the society.