Res econ 262- 4/4/17
Low levels of crime are followed by high levels of crime.
How does one identify causality? –
Lead exposure and brain damage- in the early 90s, we predicted lead exposure caused brain
damage but then later found out that exposure is associated with general plasticity issues and
pre-frontal cortex deterioration.
Where was the lead coming from? The culprit was a gas additive – tetraethyl lead invented in
1920s. Lead was coming from emissions. Lead settled in soil also doesn’t decay quickly.
Childhood lead exposure can seriously reduce IQ.
Correlation vs. causation: correlation is not causation. Just because two curves are moving with
each other does not mean they are correlated.
First best method: lab experiment. Randomly sample from population of children, then
randomly assign to control and treatment, then observe behavior over time and compare.
Second best method: exploit “exogenous variation in nature” – random experiment and you
cant change anything about it.
Some states have more lead exposure than others (variation)
Clean air act- 1970s – remove lead from the atmosphere. The reductions were state specific.
Each state got a different level of emissions reductions (exogenous variation.) It seems that the
EPA is investing money differently into each state. This is exogenous! Identification- when you slice and dice everything you don’t want and look at only everything
333you want to look at.
Reyes (2007) - predict crime from lead exposure.
Pr(crime) = b(childhood blood lead) +a(other factors) + e
^lead level ^control ^efficiency