Karlan, D$,$ R Osei, I Osei$-$Akoto and C Udry $(2014),$Agricultural decisions after relaxing credit and risk constraints$,$ The Quarterly Journal of Economics $129(2)$: $597652.$

To compare cash grants with free insurance, it would be desirable if the expected cost of the two interventions would be the same. How would you assess this? Is this the case for the Year $1$ free insurance and cash grant?

What is base risk in this context? What are the sources of base risk? What is the implication of a high base risk?

On page $15$ the authors present the theory result: $$As long as the demand for insurance is positive, access to the insurance market implies a neo$-$classical separation results in which investment in the risky and hedging inputs is independent of farmer preferences and wealth. Provide the intuition behind this result.

Table IV shows heterogenous treatment effects with respect to wealth. How would you expect these vary, and do the results confirm your expectations?

From p $29$: $$The income effect interpretation, however, is not consistent with the finding that capital grants in one$$s social network increase insurance demand: unobserved transfers should reduce insurance demand.$$ Why? I think they mean transfers to others in network, rather than unobserved.Modeling the Demand for Insurance and Investment

In years two and three, we provided access to insurance at randomized prices for a random set of

farmers. In year two, this insurance pricing experiment was crossed with the capital grants experiment

The empirical results above lead us to focus the model on an environment in which farmers are not

confronted with binding credit constraints but do face incomplete insurance. The second part of the

model in the appendix shows how farmers in such an environment respond to treatments of $($a$)$ access

to insurance at varying prices, $($b$)$ grants of capital and $($c$)$ their interaction.

First, we consider the demand for insurance and the average response of investment to access to

insurance and to capital grants. If the index insurance product has no basis risk, farmers with access to

actuarially fair insurance will fully insure and then invest in the risky and hedging inputs at the profit$-$

maximizing level. At prices higher than actuarially fair, farmers will purchase less than full insurance,

invest less in the risky input and invest more in the hedging input. As long as the demand for insurance is

positive, access to the insurance market implies a neoclassical separation result in which investment in

the risky and hedging inputs is independent of farmer preferences and wealth. This implies that

investment is the same for those who receive a capital grant in addition to access to index insurance at a

given price as for those who only receive access to insurance. At higher prices of insurance, demand falls

to zero, at which point the separation result fails and the capital grant has the additional impact of

increasing $($decreasing$)$ investment in the risky $($hedging$)$ input. We discuss more in sections $6\mathrm{.}1$ and $6\mathrm{.}2.$

Second, we examine the heterogeneous treatment effect of access to insurance at varying prices. The

effect on investment of access to insurance for a given farmer varies with the insurance price: At higherTable IV: Reallocation of Investments

IV

$($Column $1)$ includes sum of chemicals, land preparatory costs $($e$.$g$.,$ equipment rental, but not labor$),$ hired labor, and family labor $($valued at

gender$/$community$/$year specific wages$).$ Columns $4,5$ and $6($regarding non$-$farm income$)$ are derived from a survey question which asks if each

person in the household "spent the most time during the past $12$ months" $($and if so$,$ how many months$)$ on fishing, construction, working as a hired

agricultural laborer, working in non$-$farm labor, self$-$employed, employed in a household enterprise, or salaried regular employment. Harvest value

includes own$-$consumed production, valued at community$-$specific market value. All specifications include controls for full set of sample frame and year

interactions.