5. Conclusion
We have presented a general model of household technology adoption and tailored it to the case of improved household lighting for unelectrified areas. To estimate the model and test hypotheses about patterns of tech nology adoption, we have used data from a field study of solar mircogrids in rural Uttar Pradesh, India. The results from the empirical analysis highlight the importance of income effects (affordability of quality lighting) and entrepreneurial spirit (willingness to experiment with new technologies), while pre-existing lighting expenditures on the conventional lighting alternative, kerosene fuel, appear to be less important as predictors. Similarly, trust in other people in the community and companies does not predict technology adoption. These results suggest productive directions for new research on the economics of household technology adoption. While the affordability result is broadly consistent with conventional accounts, the lack of a relationship between prior fuel expenditures is surprising and suggests that households do not view kerosene and solar power as ready substitutes. The importance of entrepreneurial spirit, but not risk acceptance, suggests that attitudes toward new technology cannot be reduced to risk aversion but instead constitute an additional dimension of technology adoption. Further developing and testing these hypotheses in other domains, such as communication technology or water-purifying equipment, would contribute to progress toward a more complete theory.
