Interpreting DiD results

[Daniel Kosoko]

I am trying to interpret DiD results for an environmental policy analysis. It is evaluate if a PES programme had any effect on the rate of deforestation (defor variable is Absolute change from forest cover to non-forest in ha) between enrolled (treatment =1) and non-enrolled parcels (treatment =0) of land. Am I right in my interpretation that the enrolled parcels of land had seen higher of deforestation 117.7 ha before the programme and that after the programme, they saw a reduction in 55 ha? and that -49 represents the counterfactual meaning deforestation would have reduced in the enrolled parcels by 49.6 if the program did not exist?

[Clyde Schechter]

Am I right in my interpretation that the enrolled parcels of land had seen higher of deforestation 117.7 ha before the programme

Yes, in the sense that the rate of deforestation before intervention among the treated was 117.7 ha higher than among the untreated.

and that after the programme, they saw a reduction in 55 ha?

No, that's wrong. It's the untreated that experienced a 55 ha reduction following the intervention.

and that -49 represents the counterfactual meaning deforestation would have reduced in the enrolled parcels by 49.6 if the program did not exist?

Wrong again. The -49.7 is the DID estimate of the treatment effect. It a decrease in deforestation of 49.7 ha that is attributable to the intervention.

The "counterfactual" decrease is the decrease that would have been expected in the treated group in the absence of the treatment, and that's just the same as the decrease in the untreated group, which is, as noted, 55 ha.