The paradoxical effect of omitted variables

Omitted variables are terrible. If you are beset by them then (& unless you are lucky that they are orthogonal to whats included) you are condemned to regression hell: your coefficients are biased and inconsistent, you cannot derive policy relevant conclusions and your girlfriend won't love you anymore.
So the conclusion is to get better data. So say you do and you now have a previously omitted variable in your data. You should include it, right?
Wrong actually, if the paper below is correct which it looks like being. The problem is that the standard results in this area are based on there being only one omitted variable. If you have two omitted variables the bias on whats included depends in a messy way on all the correlations between the X's.
Say the model is:
Y=b1*X1 + b2*X2 + b3*X3 [ignoring the constant & disturbance term]

So you don't observe X2 and X3 initially so your estimate of "b1" is biased. It may seem counter-intuitive but adding X2 does not necessarily get you a better estimate of "b1". Actually, its quite intuitive: say omitting X2 was biasing b1 upwards and omitting X3 was having the reverse effect. So its quite possible you could have a small [or even zero] bias and adding in one of them makes things worse. I don't think you don't actually need these opposing biases for the result to hold because there is also the X2,X3 correlation.
Its rather analogous to the Second Best Theorem in Welfare Economics due to Lipsey & Lancaster.
The practical problem is that there may always be an "X3", that is typically you cannot be sure that you have all the relevant variables. Its all rather disturbing.

The Phantom Menace: Omitted Variable Bias in Econometric Research , Kevin Clarke
http://www.rochester.edu/college/psc/clarke/CMPSOmit.pdf

Peace, love, house prices

Many people estimate hedonic models of house prices. Explanatory variables usually include characteristics of the house like size, the number of bedrooms or proximity to amenities - a recent Australian study looked at the effect of proximity to brothels. A new IFS paper looks at the effect of killings in Northern Ireland on house prices.

Estimating the peace dividend: the impact of violence on house prices in Northern Ireland
Tim Besley, Hannes Mueller
This paper exploits data on the pattern of violence across regions and over time to estimate the impact of the peace process in Northern Ireland on house prices. We begin with a linear model that estimates the average treatment effect of a conflict-related killing on house prices .showing a negative correlation between house prices and killings. We then develop an approach based on an economic model where the parameters of the statistical process are estimated for a Markov switching model where conflict and peace are treated as a latent state. From this, we are able to construct a measure of the discounted number of killings which is updated in the light of actual killings. This model naturally suggests a heterogeneous effect of killings across space and time which we use to generate estimates of the peace dividend. The economic model suggests a somewhat different pattern of estimates to the linear model. We also show that there is some evidence of spillover effects of violence in adjacent regions.
http://www.ifs.org.uk/wps/wp0918.pdf

Richard Feynman video

I can't recommmend this highly enough. It's an hour long profile of theoretical physicist, bongo drummer, safecracker, codebreaker, reluctant Nobel Laureate, adventurer and raconteur, Richard Feynman. As hinted at in this description, there is something here for everyone
http://video.google.com/videoplay?docid=3164300309410618119#

Social Status and Health

This is a tricky area due to various concerns about endogeneity etc, but there are some papers which provide experimental evidence for the relationship between social status and susceptibility to illness. In each of the following participants were exposed to the cold virus and tracked to examine which individuals developed symptoms. The first, Sociability and Susceptibility to the Common Cold (Cohen et al 2003), found that sociability was negatively associated with the probability of developing a cold following infection. The second, Objective and Subjective Socioeconomic Status and Susceptibility to the Common Cold (Cohen et al 2008), found that subjective SES was negatively associated with the development of symptoms. Particularly interesting was the fact that this was independent of objective SES. There was some evidence that the relationship was mediated by sleeping patterns. You still have to wonder exactly what “sociability” and socioeconomic status are measuring, but interesting all the same.

Daily default option link

Several US Senators have concerns about the privacy of personal data on Facebook. In a letter to Facebook's CEO they assert that the FTC will likely investigate these issues in due course. However:

"In the meantime, we believe Facebook can take swift and productive steps to alleviate the concerns of its users," they wrote. "Providing opt-in mechanisms for information sharing instead of expecting users to go through long and complicated opt-out processes is a critical step."

Assorted Links

1. The Economist on British schools

2. Are "Dublin graduates" more likely to find jobs?

3. Poor students 'just as likely to gain good degree'

4. Volunteering: something that people feel morally obligated to do when asked, but which they would just as soon let someone else do?

5. Curb Your Enthusiasm: The Economist on the dangers of getting too excited

6. Speaking of curbing enthusiasm, below is a clip from the TV show where Larry David teaches Christian Slater the unwritten rules of hors d'oeuvres consumption. Peter Martin explains how this video illustrates the work of Elinor Ostrom for which she won this year's Economics Nobel Prize (on rules for managing common-pool resources).

VOX Article on Social Determinants of Well-Being

International evidence on the social context of wellbeing  Christopher Barrington-Leigh   Anthony Harris   John Helliwell   Haifang Huang

What accounts for life satisfaction differences across countries? This column presents new findings from the Gallup World Poll of more than 140,000 respondents worldwide. It suggests the happiest nations are those with strong social support from family and friends, freedom in making life choices, and low levels of corruption.

Assorted Links

1. Labor Supply of New York City Cabdrivers: One Day At A Time

2. Do Doctoral Students' Financial Support Patterns Affect Their Times-To-Degree and Completion Probabilities?

3. Why labour market experiments?

4. Searching for the Opportunity Cost of an Individual's Time

5. Student achievement and birthday effects

6. MumPanel opens for business

7. Behavioural Economics and Tort Law

8. Are students illegally boosting their brainpower by using "smart drugs" like Ritalin and Aderall?

9. Did Thierry Henry's handball cost the Irish economy $150 million?

Esther Duflo Bates Clark Medal

The awarding of the Bates Clark medal to Esther Duflo of MIT has been covered on many websites at this stage. Details of her work are available at her IDEAS page here and personal webpage here Much of her work combines experimental assignment and insights from behavioural economics to design and evaluate interventions in developing country contexts.

Two Teaching Assistant Posts - University of Limerick

(Via Head of School Anthony Leddin) 

Two vacancies for teaching assistants in the Department of Economics which commence in September 2010.  

The University of Limerick (UL) with over 11,000 students and 1,300 staff is a young, energetic and enterprising University with a proud record of innovation in education and excellence in research and scholarship. UL is situated on a superb riverside campus of over 130 hectares with the River Shannon as a unifying focal point. Outstanding recreational, cultural and sporting facilities further enhance this exceptional learning and working environment

Title of Post: Teaching Assistant in Economics (2 posts)

Details can be found at:-  Web:  http://www.ul.ie

I would be grateful if you could circulate this information and if you or your colleagues can recommend any potential applicants for these posts we will be delighted to hear from them.

Edgeworth Mathematical Psychics

Following Denis Conniffe's masterful lecture on Francis Edgeworth at the Irish Economics Association this weekend, I am currently engaged in a lot of head-scratching reading back through Edgeworth's 1881 work Mathematical Psychics. The talk was particularly fascinating for me in highlighting the strong connection of Edgeworth to contemporary experimental psychologists such as Fechner. It is a time in the history of economic thought with a lot of parallels to the present day, and it is remarkable how much material one can find in the 1880's and 1890's that is very similar in spirit to a lot of modern behavioural economics.

Book club to follow. A full copy of this book is available here

The socio-economic gradient in students expectations

One of the most striking, not to say depressing, features of education everywhere is the presense of a strong socio-economic gradient: poorer kids get less education and fare worse in general than richer kids by pretty much any criterion in any country. Understanding why is a mammoth task.
Given that gradient, one would expect it to be reflected in students expectations: low SES students should expect to do worse than high SES one's. Its simply the rational thing to do.
Good comparative data on such expectations is hard to come by [I think] but the PISA 2000 data [and possibly other waves] has a measure of how students expect to end up and one can compare it with where their parents actually are. The index is due to Ganzeboom et al (1992) and though it may be not how an economist would do things it seems to make sense.
So below I graph students expected SES against their father's for Ireland and Sweden. As predicted, it slopes up. Its interesting that there are flat bits at the bottom. Rightly or wrongly, students there don't think SES matters. Perhaps its a case of how, in Bob Dylan's memorable words, "When you got nothing, you got nothing to lose".
Sweden, of course, is a much more egalitarian country than Ireland. Other evidence shows that the dependence of one's education on one's parents is particularly high in Ireland compared to other OECD countries [shameless plug for my own work below]. So I would have expected the gradient to be steeper in Ireland than Sweden. I bet you would too.
Wrong! Sweden's is noticebly steeper: in numerical terms the correlation is 0.25 in Sweden and only 0.19 in Ireland. Why I don't know.

Chevalier, A., K.Denny, and D.McMahon (2009) “Intergenerational Mobility and Education Equality”. In Education and Inequality across Europe. Eds P. Dolton, R. Asplund, and E. Barth. London: Edward Elgar.

Ganzeboom, H.B.G., De Graaf, P. and Treiman, D.J. (with De Leeuw, J.) (1992), “A standard international socio-economic index of occupational status”, Social Science Research, 21(1), 1-56.

World Bank data

This week has provided free access to over 2000 data series

http://data.worldbank.org/

Seminar on Sanitary Reform in 19th century Irish Hospitals

Spring-Summer Seminar Series 2010

‘Pure air, pure water and good light’

Sanitary reform in the Dublin hospitals, 1858-1898

Prof. Gerard Fealy

Director, UCD Irish Centre for Nursing and Midwifery History

All welcome

Thursday 29^th April, 2010 – 5.30pm

Room B116, UCD Health Sciences Centre

School accountability: good or bad?

How do we get good schools? How do we get the best out of our schools? Questions like this have interested policy makers, researchers, and indeed parents for years in many countries. Accountability is one of the keywords that has emerged. We need to hold schools accountable. We hold this truth to be self evident. The idea is central in the No Child Left Behind act in the US.
But how? Well for a start one needs to measure school's peformance and compare them against benchmarks. At the very least, this means publishing measures of schools' academic outcomes, "league table" to you and me. This is done in many countries 'though not in Ireland where it is against the law.
But its not that simple. While it might be legitimate to measure the value-added by a school, simple measures of output will not measure that since the inputs (like the students) will differ. If we reward schools with high output then they have an incentive to engage in cream-skimming: exclude special-needs students, non-nationals and anyone else who might lower one's scores. Even with the information vacuum in Ireland, this occurs through various, not to say rather devious, means. The paper below addresses this important question head-on.

School accountability: (how) can we reward schools and avoid cream-skimming?

Erwin OOGHE, Erik SCHOKKAERT
Introducing school accountability may create incentives for efficiency. However, if the performance measure used does not correct for pupil characteristics, it will lead to an inequitable treatment of schools and create perverse incentives for cream-skimming. We apply the theory of fair allocation to show how to integrate empirical information about the educational production function in a coherent theoretical framework. The requirements of rewarding performance and correcting for pupil characteristics are incompatible if we want the funding scheme to be applicable for all educational production functions. However, we characterize an attractive subsidy scheme under specific restrictions on the educational production function. This subsidy scheme uses only information which can be controlled easily by the regulator. We show with Flemish data how the proposed funding scheme can be implemented.
http://www.econ.kuleuven.be/eng/ew/discussionpapers/Dps09/Dps0922.pdf