Empirical Evidence On Insurance (Karlan 4 Groups, Roz & Udry State Contingence, Lane ECL)

Question 1

Karlan et al: Credit and Insurance in Ghana
Research design

Answer 1

Focus group with farmers to identify constraints to further investment.

Then sees whether what they say is true in experiment.

Question 2

What was the main reasons farmers claimed for not investing? (2)

Answer 2

Lack of capital (money)

Unpredictable rainfall

Question 3

Why should we improve financial markets (3)

Answer 3

1. marginal return on investments may be high. (Fosu & Dittoh)
2. Agriculture in Ghana is almost exclusively rain-fed, thus weather risk is significant, so rainfell index insurance is promising.
3. Study found rainfall shocks translate directly to consumption fluctuations.

So thus, improving financial markets can address these issues and create large gains.

Question 4

Fosu & Dittoh: Marginal returns to investment high

Answer 4

Investment of $60 per acre on fertiliser generates $215 of additional output per acre.

Question 5

Why use rainfall index rather than crop insurance

Answer 5

Eliminates moral hazard and adverse selection because payouts are dependent on only observable rainfall.

(Hard to measure alternatively how much crop would need to be compensated for in crop insurance, e.g finding value of crops prices change, and hard to quantify individually, would require regular monitoring of crop levels! )

Question 6

What is the poverty trap in Karlan:

Answer 6

Risks are large relative to their income, and lack access to credit, barriers to saving and limited risk sharing (group insurance)

So do not invest are not insured risks and remain poor!

Question 7

So after the focus group identifying proclaimed reasons for underinvestment (lack of capital and unpredictable rainfall)

What were the 4 groups farmers were split into?

Answer 7

Control: business as usual
Free cash grants
Free rainfall insurance
Cash grants + insurance

Question 8

Karlan’s Theory

(Back to fluctuations in production, income, consumption)

Answer 8

If markets are well-functioning, fluctuations in production or income should not influence consumption or investment.

Where insurance markets are imperfect, the cash grants or insurance may influence investment.

Question 9

Example of a risky vs hedging asset in terms of farming

Answer 9

Risky - investments in fertiliser or cultivating a larger plot (since returns can be wiped out by flood droughts etc)

Hedging - investment in irrigation (ensure right amount of water on farm)

Question 10

Findings: what is the most effective grant?

Answer 10

Investment/output increased most from insurance grant, small effects of the cash grants. (So capital wasn’t the biggest issue)

Therefore UNINSURED RISK is a big constraint on farmers’ investment (hence why insurance had the biggest impact)

Question 11

Rosenzweig and Udry caveat

Answer 11

Impact of an intervention is state contingent

The results in Karlan of increased output from insurance was down to weather being good! Had it not been, results may be completely different!

Question 12

So what is the main implication we can gather from Rosenzweig & Udry? (impact is state contingent)

Answer 12

Consider the year to year variation in the study
Or if single year: is it a “normal” year?

E.g imagine looking at the economy during COVID, some results may be unrepresentative of normal.

Question 13

What do results suggest in terms of risk attitude?

Answer 13

High risk aversion (since insurance is the biggest driver of investment and output)

Question 14

Other risk-management tools used in developing countries (3)

Answer 14

Flood tolerant crops (can withstand excess water)

Access to weather forecasting (predict weather & inform farmers’ decision making)

Emergency credit lines

Question 15

Unpredictable weather seems to be a key issue

Households adopt costly coping strategies, and lower-risk activities that limit their long-run earning potential.

So how can emergency credit lines overcome this? (Lane)

Answer 15

Emergency loans are agreed in advance (removing credit availability uncertainty!), where they are guaranteed access to credit if a flood occurs!

So allows them to invest an maximise earning potential

Question 16

Lane’s research design (Bangladesh)

Answer 16

Informed over 300,000 clients in advance of crop decisions they have an emergency credit line.

Question 17

Why are credit lines better than insurance (2)

Answer 17

Overcomes credit market failure i.e adverse selection preventing lending

Overcoming insurance market failure- demand for insurance is low due to upfront costs for uncertain benefits. This avoids it (no upfront fee, only pay after)

Question 18

Why pre-approved emergency loans? (3)

Answer 18

Overcomes limitations with large scale infrastructure programs, since substantial investment & coordinated effort needed (now can pool money towards fixing the emergency)

Overcomes failures in markets for climate resistant agricultural technologies. (E.g may not need flood tolerant crops if able to invest in irrigation etc)

Farmers avoid having to make costly changes to crop/input choices (since change often lowers yield if extreme weather doesn’t occur; we only change when emergency has occurred!)

Question 19

Lane’s hypothesis: ex-ante and ex-post outcomes

Answer 19

Ex-ante outcomes: guaranteed access to credit may increase investment

Ex-post outcomes: production levels should be more sensitive to flood shocks, but not consumption.

Question 20

Lane findings:

Answer 20

Emergency loan is effective in lowering the impact of flood shocks.

Question 21

The Emergency Loan is attractive to both sides of the market. Explain

Answer 21

Supply side - guaranteed credit for borrowers improve their repayment rates, thus improving profits for MFI.

Demand side - households value loan after a shock 1.8x more than receiving loan pre-period (no upfront costs with uncertain benefits!)

Question 22

Karlan, Rosenzweig and Udry (caveat to Karlan results), Lane