The Design and Implementation of Effective Safety Nets December 2 - - - PowerPoint PPT Presentation

The Design and Implementation

• f Effective Safety Nets

December 2 - 13, 2002 Washington, DC The World Bank

Public Intervention in Responses to the Shock of the 1998 Flood in Bangladesh

Carlo del Ninno

Protecting the Vulnerable: The Design and I mplementation of Ef f ective Saf ety Nets December 2 - December 13, 2002 - Washingt on, D.C.

Bangladesh in Summer of 1998

The 1998 Floods in Bangladesh

• From early July to the end of September,

1998, floods covered much of Bangladesh

• At their peak on 7 September, 1998, 51

percent of the country was under water

• The floods caused 2.04 million MTs losses,

equal to 17.9 percent of target national rice production for the first half of 1998/99

• Yet, no reported deaths from starvation
• ccurred after the flood, and food market

prices were stable

Outline of the Presentation

• Availability of Food

– Volume of public food grain distribution following the flood – Market availability and private sector imports

• Impact on household and coping strategies
• Government transfers

– Effectiveness of targeting – Determinants of program participation

• Policy impacts on calorie consumption

– Econometric estimates of calorie consumption – The contribution of rice prices and government transfers to post-flood calorie consumption

• Conclusions

Food Aid for Flood Relief and Rehabilitation

• After a GOB appeal for international help on

26 August, donors committed 1.081 million MTs for flood relief to Bangladesh.

• By 30 June 1999, total food aid received

was 1.22 million MTs, about 620 thousand MTs for flood relief.

• The remainder of the flood relief food aid

later was deferred to fiscal year 1999-2000.

Public Food Grain Distribution

• Initial efforts to increase public food grain

distribution were constrained by a shortage of rice and wheat stocks.

• Actual distribution from July through

December 1998 was only 26 thousand MTs (4.3 percent) greater than originally planned for these months.

• The arrival of government commercial

imports and food aid later permitted a 415 thousand MT expansion in public distribution.

Foodgrain Distribution by Channels, Budget and Actual 1998/99

Budget 1998/99

EP/OP Rice 8% VGF Rice 1% FFE Rice 9% EP/OP Wheat 6% FFW Wheat 22% FFW/VGD Rice* 18% TR/VGD Wheat** 11% VGF Wheat 1% OMS Rice 12% FFE Wheat 12% Rice = 813 Th MTs Wheat = 905 Th MTs Total Foodgrain = 1.78 Ml MTs

Actual 1998/99

FFE Rice 3% FFW/VGD Rice* 7% EP/OP Wheat 5% FFW Wheat 33% VGF Wheat 14% FFE Wheat 11% OMS Rice 0% TR/VGD Wheat** 13% EP/OP Rice 6% VGF Rice 8%

Rice = 529 Th MTs Wheat = 160 Th MTs Total Foodgrain = 2.32 Ml MTs

Trade Liberalization and Private Food Grain Imports

• In the early 1990s, Bangladesh liberalized

food grain trade, allowing private sector imports of wheat in 1992/93 and rice in early 1994.

• In October 1994, India also liberalized its

private rice trade, permitting private sector exports of non-basmati rice, (though an export quota remained).

Rice Price Stabilization

• At the same time, as part of its price

stabilization strategy, the government encouraged private sector imports through the elimination of a 2.5 percent import tax

• n rice and other measures.
• In response to market incentives, hundreds
• f private sector traders imported an

estimated 2.42 million tons of rice from July 1998 to April 1999 (according to official data).

Rice Price Stabilization

• Private sector rice imports thus set a ceiling
• n domestic rice prices, equal to the import

parity price of rice, ( equal to the cost of rice in Indian markets plus trade and transport margins), preventing a major price increase.

• Adequate levels of government stocks may

have also helped stabilize markets through influencing private traders’ expectations of the ability of the government to intervene in local rice markets

Rice Prices and Quantity of Private Rice Imports in Bangladesh, 1993-2000

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 Jul-93 Jan-94 Jul-94 Jan-95 Jul-95 Jan-96 Jul-96 Jan-97 Jul-97 Jan-98 Jul-98 Jan-99 Jul-99 Jan-00 Price (Taka/kg)

50 100 150 200 250 300 350 400

Imports ('000 MT)

Private Sector Imports Import Parity(ex: Bongaon) Dhaka Wholesale Price Import Parity(ex: Delhi) Note : Price data for April 2000 is up to the fourth week only; private sector imports are as of 26th April, 2000. From November 1998, the carrying cost has increased by 1.1 Tk/kg to 4.1 Tk/kg. From January 2000, 5% tax imposed on rice import. Source : Dorosh (1999), calculated using data from FPMU and MIS, DG Food, CMIE (1999, 1998, 2000) and Baulch, Das et. al, (1998).

1998 Market Prices in the Absence of Imports From India

• In the absence of trade with India, the next lowest

cost source of imports would likely have been Thailand.

• From December 1997 to November 1999 import

parity prices from Thailand averaged 16.1 Tk/kg, 21 percent higher than the Dhaka price.

• If private trade with India had not been feasible,

(and assuming no change in GOB interventions), rice consumption would have fallen by an estimated 4.2 to 6.3 percent and rice imports (from Thailand) would have been 0.7 - 1.0 million tons less

Increasing Access to Food by Flood Exposed Households

• Even though the contribution of public

food grain distribution was small in terms of increasing availability of food grain, it nonetheless played an important role in increasing access to food for poor, flood-exposed households.

Impact on Households and Coping Strategies

• Employment and Income

– Agricultural production – Wealth (asset losses)

• Health and Nutritional

– environment and illness – status of children and women

• Food consumption

– Caloric consumption

• Household coping mechanisms
• Government transfers
• Effect of rice price increase

FMRSP-IFPRI Household Survey

• FMRSP-IFPRI conducted a survey of 757

rural households in seven flood affected thanas at three points in time: Nov 98, Apr 99 & Nov 99

• 71.1 percent of households surveyed were

exposed to the flood, 57.6 percent were severely exposed to the flood (defined in terms of depth of water in the house, depth

• f water in the homestead, and number of

days away from home due to the flood).

Food consumption

• The price of rice, vegetables and many other

foods that were in short supply increased, and as a consequence caloric consumption

• f flood-exposed households was 272

cal/person/ day less than non flood exposed households

• After the flood, the mean level of total

household expenditure decreased from Taka 4,001 in the first round to Taka 3,663 in the second round and remained relatively stable at Taka 3,508 in the third round

Food consumption

• The decrease of total expenditure was due to

the decrease of non-food expenditure from Taka 1,293 in the first round to Taka 842 in the second round and Taka 855 in the third round.

• As a consequence of the change in the

expenditure pattern, the resulting consumption of calories per capita per day increased across the three rounds from 2,249 to 2,518 and 2,526 respectively

Per capita daily calorie consumption

1,000 2,000 3,000 4,000 Nov-98 May-99 Nov-99 Per Capita Daily Calories

Poor Middle Non-Poor

Government Transfer Programs

• 47.2 percent of all households received

government transfers.

• 24.4 percent of households received

Gratuitous Relief (GR); 21.8 percent of households received Vulnerable Group Feeding (VGF).

• For the August through November period,

government transfers were generally equal to only a small share of expenditures.

10 20 30 40 50 60 70 80 90 100 Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Household Expenditure Quintile Average Value of Transfers (Taka/Month/Household) Gratuitous Relief (GR) Vulnerable Group Feeding (VGF) Other Government Transfers NGO Transfers

(5.3) (2.9) (2.6) (1.8) (1.0)

*The Numbers In Perentheses Denote The Percentage Share of All Transfers On Total Household Expenditutres

Value of Transfer Received and Share of Total Household Expenditures

Targeting the Poor

• 67.1 percent of households in the lowest per

capita expenditure quintile received government transfers.

• 38.8 percent of households in the lowest per

capita expenditure quintile received VGF transfers; 17.2 percent and 11.2 percent of households in the top two expenditure quintiles, also received VGF.

• GR transfers were less well targeted to the

poor.

Households Receiving Transfer By Expenditure Quintile (Percent)

5 10 15 20 25 30 35 40 Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 Household Expenditure Quintile Percentage of Households Gratuitous Relief (GR) Vulnerable Group Feeding (VGF) Other Government Transfers NGO Transfers

Targeting by Flood Exposure

• GR was better targeted towards flood-exposed
• households. Only 11.4 percent of GR households

(as compared to 24.7 percent of VGF households) were not directly exposed to the flood.

• 36.7 percent of very severely flood-exposed

households received GR transfers, while only 24.5

• f this category of households received VGF

transfers.

• Coverage was small in both programs. 69.3

percent of flood-exposed households did not receive GR; 76.6 percent did not receive VGF.

Households Receiving Transfer By Severity of Flood Exposure (Percent)

5 10 15 20 25 30 35 40 Very Severely Exposed Severely Exposed Moderately Exposed Not Exposed Severity of Flood Exposure Percentage of Households Gratuitous Relief (GR) Vulnerable Group Feeding (VGF) Other Government Transfers NGO Transfers

Leakages and Program Coverage

• The average value of VGF grain transfers was

202 Tk/household/month in October and November 1998 (equal to 5.0 percent of household expenditures).

• For VGF participants in the lowest quintile,

VGF transfers equaled 10.5 percent of total expenditures.

• 75 percent of the value of VGF transfers went

to the poorest 60 percent of households.

Determinants of Program Participation

• Probit regressions on VGF and GR participation

indicate:

• For GR, flood exposure (particularly severe and

very severe exposure) is a highly significant explanatory variable; ownership of a house with a tin roof (an evidence of household wealth) is negatively associated with participation.

• For VGF, flood exposure variables are not
• significant. Landlessness and housing

characteristics are significant explanatory variables.

Determinants of Participation in GR and VGF Programs, Probit Regression Results

GR Participants VGF Participants Dependent Variable

• Coef. Std. Err

z P>|z|

• Coef. Std. Err

z P>|z| Flood Exposed-Moderate 0.528 0.215 2.457 0.014

• 0.002

0.208

• 0.009

0.993 Flood Exposed-Severe 0.901 0.177 5.089 0.000

• 0.125

0.176

• 0.708

0.479 Flood Exposed-Very Severe 0.947 0.203 4.673 0.000

• 0.163

0.210

• 0.776

0.438 Observations 737 737 Chi-squared (33) 145.06 61.91 Prob > Chi-squared 0.0017 Pseudo R-squared 0.177 0.099

Coping strategies, flood exposure and poverty in November 1998

Flood Exposed Not All Bot 40% Mid 40% Top 20% All exposed households Monthly households expenditure (Tk) 2,414.3 3,973.9 7,720.6 4,063.6 3,843.5 4,000.5 Share of food expenditures (percent) 72.4 69.5 62.2 67.4 68.4 67.7 Households in debt (percent) 68.2 58.9 62.6 63.5 53.5 60.6 Share of monthly expenditure (percent) 186.1 138.7 131.3 144.4 140.2 143.6 Household purchasing food on credit (percent) 56.7 54.1 50.5 54.4 29.5 47.3 Share of monthly expenditure (percent) 37.6 27.2 17.3 25.8 20.0 25.0 Households receiving government transfers (percent) 60.7 54.1 32.7 52.6 33.6 47.2 Share of monthly transfer on expenditure (percent) 3.4 2.4 0.8 2.0 2.1 2.0 Households selling assets (percent) 25.2 21.3 15.9 21.9 20.3 21.4 Share of monthly expenditure (percent) 45.5 51.3 75.3 51.9 44.2 49.9 Number of households 226 207 107 540 217 757

Household coping mechanisms

• Private borrowing was the dominant coping

mechanism for flood-exposed households

• Public and NGO interventions, though well

targeted, were too small to have a major impact

• n household food security
• Poor households bore the weight of substantially

increased debt more than a year after the floods

• Given the sharp increase in debt for the poor,

there was a need for additional efforts in providing rural employment, transfers and credit to poor households following the flood

Policy impact on calorie consumption

• The trade liberalization policy enabled private

sector imports to stabilize rice prices; government targeted food transfers increased the income of households

• Here we estimate the contribution of those

policies to maintaining household calorie consumption using price and income elasticities

• f demand from econometric analysis of

household data

• Finally, we use the estimated parameters to

simulate the effects of alternative rice prices and transfers on calorie consumption

Simulated Change in Calorie Demand due to Price Increase

Calories/person/day Change in Rice Price Change in Income Price Income (percent) (percent) Effect Effect Total Import Parity (ex: Bangkok) Inelastic parameters 21.00%

• 3.40%
• 44
• 20
• 64

Elastic parameters 21.00%

• 3.40%
• 109
• 25
• 133

No Private Imports Inelastic parameters 30.00%

• 3.40%
• 60
• 20
• 80

Elastic parameters 30.00%

• 3.40%
• 148
• 25
• 171

Source: Authors' calculations Notes: Inelastic parameters: calorie-rice price elasticity = -0.142; calorie income elasticity = 0.363. Elastic parameters: calorie-rice price elasticity = -0.362; calorie income elasticity = 0.445.

Simulation Results

• Given an 18.7 percent decrease in the average

rice price from 16.04 to 13.04 Tk/kg and a 19.3 percent increase in total expenditures from 422 to 503 Tk/capita/month, the calorie demand regression coefficients imply a 160 to 272 calorie increase in daily per capita consumption

• If private sector rice trade from India had been

prohibited and the rice price were 19.4 percent higher (equal to the import parity price of rice imports from Thailand) per capita consumption

• f the rural poor would have been 44 to 109

calories/day less

• In the absence of any private sector imports, if

government imports were constrained by administrative delays or shortages of funds and prices increased by 30 percent, per capita calorie consumption could have fallen by 60 to 148 calories/day

• Government transfer programs, by comparison,

added an estimated 20 to 25 calories/person/day to calorie consumption by flood-exposed rural poor households.

Simulation Results (2)

Conclusions

• The Bangladesh flood illustrates the crucial role

that private markets and appropriate government investments and policies can play in maintaining food availability, limiting price increases and supplementing household access to food, thereby helping to avoid a major food crisis

• Food aid did not have a major impact on foodgrain

availability; rather food aid’s main contribution to food security was to enable the GOB to increase access to food of poor households.

• Nonetheless, the poor did suffer both in the short

term through reduced consumption and increased illness, as well as in the medium term through increases in household debt and lingering nutritional consequences

Conclusions

• Government programs were, in general, well-

targeted to flood-exposed households in the first two months (through GR) and to poor households thereafter (through VGF)

• Without the private cross-border rice trade, rice

prices would likely have been at least 19 percent higher and total calorie consumption of the poor could have fallen by 44 to 109 calories/ person/day below their already reduced levels.

• Reducing the negative impact of natural disasters

will require effective policies and increased resources, both at the time of disasters, and perhaps more important, over the long term to reduce poverty and malnutrition, as well

Fine

References

del Ninno, C., P. Dorosh and N. Islam, 2002 . “Reducing Vulnerability to Natural Disasters: Lessons from the 1998 Floods in Bangladesh”. IDS Bulletin 33(4): 98-107 del Ninno, C., P. Dorosh, L. Smith. 2002. “Public Policy, Markets and Household Coping Strategies in Bangladesh: Avoiding a Food Security Crisis Following the 1998 Flood”. Paper presented at the conference on “Crisis and Disasters: Measurement and Mitigation of they Human Costs”. World Development (Forthcoming) del Ninno, C. and P. Dorosh, 2002 . “Maintaining Food Security in the Wake of a Natural Disaster: Policy and Household Response to the 1998 Floods in Bangladesh”. Journal of Bangladesh Studies No 4(1):12-24 del Ninno, C. and P. Dorosh, 2002 . “In-Kind Transfers and Household Food Consumption: Implications for Targeted Food Programs in Bangladesh”. FCND- IFPRI Discussion Paper 134. Washington D.C. (Co-author with P. Dorosh) del Ninno, C., P. Dorosh, L. Smith and Dilip Roy. 2001. “Household Food Security, Welfare Losses and Recovery in the Aftermath of the 1998 Floods in Bangladesh”. IFPRI Research Report 122. Washington D.C. (Co-author with P. Dorosh,). del Ninno, C. and P. Dorosh, 2002. “Averting a Food Crisis: Private Imports and Public Targeted Distribution in Bangladesh after the 1998 Flood”. Agricultural

• Economics. No 25:337-346.