The 2004 Northeast Universities Development Consortium Conference was held in Montreal 1-3 October. It has become the place to go for development economics research and provides a good overview of what is going on, especially among the junior researchers in the field. It is nice to see that micro-economics of development is still attracting a lot of attention and that interesting research is being done in the field. There is nothing like a conference like this to motivate me to go back and do research. I presented my paper, "Risk and Household Structure: Another Look at the Determinants of Fertility," in the session on Informal Social Protection. The paper's hypothesis is that children can act as imperfect substitutes for insurance and is based on the theory in my paper, "Children as Insurance," published in Journal of Population Economics, Vol. 14(1): 119-136, 2001. I use the ENCOVI 2000 survey from Guatemala combined with secondary data on the risk of hurricanes and earthquakes to estimate the effects of the risks natural disasters on fertility and education. The results show that increased risk of disasters that requires command of manpower to handle increase fertility and lower the education of children, while disasters where a larger family is of little use have a negative effect on fertility. Furthermore, actual shocks has a negative effect on both fertility and children's education, although there is some "catch-up" later. The paper still needs some work on the theoretical model, but the empirical part is converging (hopefully it is a stable path ;-) ) towards its final form. If you would like a copy of forthcoming versions please let me know or check this space.

I also discussed Sonia Laszlo's paper, "Education, Labour Supply and Market Development in Rural Peru," in the Economics of Schooling session. Its aim is very worthwhile: To estimate the return to education in a situation where most people either are self-employed or work on the family farm. Hence, contrary to standard returns to education papers it is not possible to use individual wages to access the return. Laszlo uses a combination of reduced form estimations combined with a structural model to back out the returns parameters. The main problem with the paper is that it is not clear that identification has been achieved, especially since the theoretical model, which is necessary for the structural parameters to be believable, is not convincing.