,000 children born between 1979 and 1987 in the Canadian province of Manitoba. These children are followed until 2006, and their … records are linked to provincial registries with outcomes data. We compare children with health conditions to their own …

between two possible explanations. The first is that low-SES children are less able to respond to a given health shock. The … second is that low SES children experience more shocks. We show, using panel data on Canadian children that: 1) the gradient … we estimate in the cross section is very similar to that estimated previously using U.S. children; 2) both high and low …

, second we use a variety of measures of infant health, and finally we track children through their schooling years and into …