Lecture 8: Risk and Insurance

1 / 25

Risk and the Poor

In 2009, Robert Zoellick then president of the World Bank warned the world that the economic crisis could be a disaster for the worlds poor. The global economic crisis [sparked by the collapse of Lehman Brothers in September 2008] threatens to become a human crisis in many developing countries unless they can take targeted measures to protect vulnerable people in their communities. While much of the world is focused on bank rescues and stimulus packages, we should not forget that poor people in developing countries are far more exposed if their economies falter. Surprisingly, the crisis did not end up being so important. It is not that jobs were not lost, and prices did not drop, but compared to the risk that the worlds poor usually face it was not a huge event.

2 / 25

Risk and the Poor

50% of the urban poor run a non agricultural business. 30-98% have of the rural poor have a farm depending on the country. Those who dont have a business are often casual labourers, paid a daily wage. Daily wage workers lose their jobs easily and moreover the wage volatility is much higher in poor countries that in rich countries. See Wage Volatility and GDP . The price of food is volatile (see the poor more.
Food Price Volatility

) and this aects

3 / 25

Shocks

A very small business may not be protable, while a somewhat larger one is. One shock may plunge a household in a poverty trap. Can we see this in the S shaped world that we studied in the rst lecture?

4 / 25

Stress

Risk can cause poverty. Stress can exacerbate it as well:

Under stress body produces cortisol, which partly aects decision making ability. Cortisol in pregnancy can aect the development of a fetus babies of stressed out mothers do less well in life.

5 / 25

Coping with Risk

Saving and borrowing.

However, as we have seen, the ability to borrow is limited. Suppose you cannot borrow, and your income follows an i.i.d process. Nearly optimal rule of thumb you consume everything when you have very little cash on hand, and otherwise save a fraction above some threshold. Even if you save a lot, there will be years when you consume very little.

6 / 25

1230
0 @

ANGUS DEATON

is N(100,10), r=0.05, d=0.10, p=3 ~~y

0 aL

consumption-40

C',

=-o 0
Lo

',,

,M!,

10
FIGURE

20

40

60

80

100

120

140

160

180

200

time
2.-Simulations
source: Deaton (1991) and assets, with white noise income. of income, consumption,

was no faster, and althoughthe same resultswere obtained,there are a number 7 / 25

Coping with Risk

Try to work more. But when everybody tries to work more at the same time, (say, because there is a drought), what will happen to wages? The poor are more likely to want to work harder when they face a shock, because the drop in consumption is more costly to them. Jayachandran 2006 Selling Labor Low shows that this leads to an increase in risk Not only does risk cause poverty, but poverty causes risk!

8 / 25

Table 3 Banking & Wage Volatility Dependent variable: Log agricultural wage, 1956-1987 Measure of Banking Bank deposits per capita (1) Log crop yield Banking Log crop yield * Banking -.091 ** (.036) 7678 -.075 * (.044) 7614 .162 ** (.083) Bank credit per capita (2) .158 * (.083) Bank branches per capita (3) .138 * (.082) -.049 ** (.021) -.033 * (.019) 8080

District and year FE? Y Y Y Instruments = RainShock, RainShock * Banking, RainShock * % of workforce in agriculture source: Jayachandran (2006)
Notes: Standard errors, which are in parentheses below coefficient, allow for clustering within regionyears. *** indicates p<.01; ** indicates p<.05, * indicates p<.10. Variables interacted with log crop yield have been transformed to be mean 0, standard deviation 1. All regressions include as an endogenous control variable the % of workforce in agriculture in 1961 interacted with log crop yield. The district fixed effect absorbs the level effect of time-invariant measures of banking. Sources for banking measures: columns 1-2, Census 1981; column 3, Reserve Bank of India.

9 / 25

Avoiding Risk

Knowing that risk is ever-present and would be very costly, poor try to avoid it. This would lead them to avoid more protable but more risky investment (new seeds, etc.) prot rates would be 30% higher if there was no weather risk. Diversify portfolio of activities (each person in a family has several jobs) no investment in specic skills. Migrate for short periods of time, keeping the family home no career. These strategies contribute to maintain people in poverty.

10 / 25

Example: Weather Insurance

Pilot program in Ghana for a simple weather insurance programe. At full actuarially fair price, demand for the product is almost zero. At almost fully subsidized price, every farmer demands it. Large eects on adoption of fertilizer and food security afterwards, when insurance was combined with credit. We can see that risk is costly for the poor Insurance would greatly help them. Why are they not insured? By the private sector? Through friends, relatives, etc.?

11 / 25

Why is Formal Insurance Missing?

Forbes magazine presents insurance for the poor as an unpenetrated natural market. Many micro nance institutions would like to see in micro insurance the next micro credit revolution. And yet the attempts have been disappointing rst, very low demand, and second, low use even among those who are insured. Insurance is inherently dicult Moral hazard, adverse selection, and outright fraud.

12 / 25

Example: Health Insurance

Moral hazard too much curative care expenditures already. What if they had insurance? Too little preventive care already. Adverse selection the sick may sign up if the insurer cannot accurately observe a persons health. Outright fraud Doctors may charge for services they do not perform. The solution may be to oer simple products: To avoid fraud and moral hazard oer only catastrophic health insurance hospitalisation insurance, usually in a networked hospital which has been vetted. To avoid adverse selection make insurance compulsory for a pre-selected group of people (e.g. clients of a micro credit organization).

13 / 25

Another Example: Crop Insurance

Moral hazard An insured farmer has an incentive to take less care of his crop. Adverse selection A farmer may choose insurance for a eld that he knows is prone to disaster or has low quality soil. Simple products: Based on a rainfall index at a nearby weather station. People can buy as much insurance as they like. No fraud, little administrative cost, no adverse selection or moral hazard.

14 / 25

An oer in search of a demand

Poor have little interest in these products and the take up is very low. Lack of understanding. Mistrust: How to explain the rules when they are at the advantage of the insurer? The product is not what they actually need or want. Dicult to think ahead about a negative event few young people have burial insurance in South Africa, even though burial is a major cost for their surviving relatives. The market may be missing for a reason.

15 / 25

Market is Missing

The market is limited in what it can oer. Moreover it is limited precisely in ways that makes what it can oer relatively unattractive to the poor. This is very dierent from what we will learn about micro credit where the market is also very limited in what it can oer, but what it can oer suits at least some people.

16 / 25

Why is Informal Insurance Missing?

Is there scope for informal insurance? Two farmers Ahmad and Bijou grow crops, their income can be either high or low both with probability 1/2. When Ahmad has high and Bijou has low, or vice versa, they agree to share. With probability 1/2 they now have average income, with probability 1/4 low, and with probability 1/4 high. Does this increase their welfare? Chris Udry spent a year in Nigeria and recorded the term of the loans that villagers give each other. He found that borrowers pay back less when they have an adverse shock and pay more when the lender has an adverse shock.

17 / 25

TABLE I

shocksreceived Realizedtermsvs. borrowver and lender Samplemeans Adverseshock receivedby (A) Borrower -no shock -adverse shock Impactof shock -on mean: _(t) (B) Lender -no shock -adverse shock Impactof shock -on mean:
-(t)

Monthlyinterest Simpleinterest ratesb ratesa 0-5% -4 0% lower (1-58) -7 5% 2-6% higher

(4 56)

Repayment period in days 67 72 longer (1 03) 89 80 shorter

(1.89)

20-4% -0-6% lower (2 20) -5 0% 11-8% higher

(3 06)

Notes. The impactof the shocks is judged by a two-sidedt-test for equalmeans is in parentheses. The absolutevalueof the t-statistic (P noshock - P shock). The definitionof 'adverseshock'is: 1. A respondent (borrower or lender)is judgedto have receivedan adverseshock if he reported an unexpected adverseeventon any of the fieldshe farmsduringthe term of the loan. Commonevents were flooding,wind damage,or infestationby insects. an adverseshock or lender)is judgedto havereceived 2. The otherparty(borrower seriousevent that occurredin the other if the respondent reportedan unexpected, eventswerefarming eventsas in (I), duringthe termof the loan.Common household and medicalproblems,rain damageto houses,and other 'householdemergencies'. a This is a standard monthlycompoundinterestrate. b This is the simpleratio of the amountrepaidminusthe amountborrowedto the amountborrowed.
source: Udry (1994) 18 / 25

Relative Strengths of Informal Insurance

Moral hazard What if income depends on eort on farm? Adverse selection What if a farmer prefers to insure a plot of land that isnt very productive? Fraud? Informal insurance has the ability to leverage on the ground information and the value of social ties for mitigating some of these problems.

19 / 25

Relative Weaknesses of Informal Insurance

Limited commitment when you have high income realization there is a temptation to lie about it. What prevents you from yielding to this temptation? Is insurance potentially fragile? What is the eect of migration? Aggregate shocks farmers can only insure each other against shocks that dont aect them all at the same time. Limited size In practice most informal insurance arrangements seem limited to bilateral relationships, not a pool. This is not very well designed to deal with large shock such as a major illness.

20 / 25

Role of Policy

This opens an obvious role for public policy step in to complement the market to insure social protection. But this could take several forms
Food aid. Cash transfers (conditional or not). Free primary health care. Oer publicly provided insurance. Subsidize insurance oered by the market. And many more!

21 / 25

Week 11 Debate

There is a conict between commercialization of micro-credit and its social role.

22 / 25

Recommended Readings

Chapter 6, Poor Economics Jayachandran Selling Labor Low: Wage Responses to Productivity Shocks in Developing Countries JPE 2006 Udry Risk and Insurance in a Rural Credit Market: An Empirical Investigation in Northern Nigeria REStud 1994

23 / 25

Wage Volatility and GDP

Figure 1: Agricultural Wage Volatility Versus Gross Domestic Product

source: Jayachandran (2006) 24 / 25

Food Price Volatility

25 / 25