I have a new working paper up coauthored with former masters student Jack Gregory. The paper analyses data from two tribal area villages in Maharashtra State. Jack organized and supervised the survey when he worked for the NGO WOTR in India a few years ago. The surveyed was intended to provide data on greenhouse gas emissions and so had some shortcomings as an information source for understanding energy demand and the choice of fuel. Still, we thought it was worthwhile presenting this information to a wider audience.
We found that there were really big differences between the villages. In the village of Purushwadi income had a big effect on energy use, but there was little relationship in the nearby village of Kohane. Unfortunately, we don't have any explanation for this difference. In Purushwadi the relationship between energy use and income was also different than that found in national surveys.
The national surveys show fairly constant use of biomass across income groups and increased use of modern fuels in the upper half of the income distribution in rural India (Khandker et al., 2012):
The numbers refer to income deciles. Unfortunately, we don't have data on electricity use, though we do have data on electricity connections. Here is the data for per capita energy use that we do have by income quintile in each village (P = Purushwadi, K = Kohane):
It seems to me that research in this area is mostly about coming up with common patterns but understanding more about the differences between villages might be also of interest.
Besides this, our modelling shows modest support for the energy ladder or rather "energy stacking" hypothesis. Energy stacking implies that rural households continue to use traditional fuels but add more and more of the modern ones as their income rises. Also we find that using higher quality energy sources reduces energy use, ceteris paribus. We also find that household size, stove ownership, and season influence rural energy choices. However, the effects of improved stoves are small and not consistent across the villages. This fits with recent evidence for modest or even perverse impacts of improved stoves.
Here are some pictures to give you an idea of what is involved in measuring energy use in rural India:
(a) Measuring rice with a 5 kg basket scale; (b) Measuring a headload of branches with a 25 kg hanging scale; (c) Measuring kerosene with a 200 ml graduated cylinder