Working Party on Agricultural Policies and Markets

Transcription

1 Unclassified TAD/CA/APM/WP(2014)14/FINAL TAD/CA/APM/WP(2014)14/FINAL Unclassified Organisation de Coopération et de Développement Économiques Organisation for Economic Co-operation and Development 02-Dec-2014 English - Or. English TRADE AND AGRICULTURE DIRECTORATE COMMITTEE FOR AGRICULTURE Working Party on Agricultural Policies and Markets TRANSITORY FOOD INSECURITY IN INDONESIA Contact person: Shingo Kimura ( shingo.kimura@oecd.org) English - Or. English JT Complete document available on OLIS in its original format This document and any map included herein are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

2 NOTE BY THE SECRETARIAT The PWB mandates to develop and use a framework for the analysis of transitory food insecurity under output area on policy coherence and global food security. This work on transitory food insecurity is the result of a consultation process with the Government of Indonesia and stakeholders, analytical work by the Secretariat in collaboration with consultants. Scenarios that could put food security at risk have been identified, their impacts have been quantified and policy options have been analysed. This work stream includes three reports that are all submitted to this meeting of the working party: 1) a Framework for the Analysis of Transitory Food Insecurity [TAD/CA/APM/WP(2013)35/FINAL]; and 2) a policy report on Transitory Food Insecurity in Indonesia [TAD/CA/APM/WP(2014)14/FINAL]; 3) a Technical Background Report on Transitory Food Insecurity in Indonesia [TAD/CA/APM/WP(2014)15/REV1]. This document is the policy report and examines the identified scenarios that could put food security at risk in Indonesia and provides a quantitative assessment of policy options. This document was declassified at the 64 th Session of the Working Party of Agricultural Policy and Markets on November

3 TABLE OF CONTENTS EXECUTIVE SUMMARY AND POLICY RECOMMENDATIONS... 5 TRANSITORY FOOD INSECURITY IN INDONESIA... 8 Introduction The state of food security in Indonesia... 9 Food Availability... 9 Access to food: poverty and prevalence of undernourishment Nutritional outcomes An assessment of food insecurity risks in Indonesia The quantification of each scenario Scenario I: International rice price spike Scenario II: Macroeconomic crisis Scenario III: An increase in the international price of fuel Scenario IV: Crop failure due to a plant diseases infestation Scenario V: Earthquake and tsunami in Sumatra A map of Indonesian food insecurity scenarios Policy options and their impact across scenarios Set of Policies in this analysis The impact of selected policies in the reference scenario The impact of selected policies across the scenarios Policy discussion and conclusion REFERENCES Tables Table 1. Composition of food expenditure, calorie intake and protein intake in Table 2. Composition of calorie intake by region Table 3. Food insecurity scenarios in Indonesia Table 4. Scenario I: International rice price spike (100%) Table 5. Scenario II: Macroeconomic crisis Table 6. Scenario III: Increase in fuel prices Table 7. Scenario IV: Crop failure Table 8. The most important earthquakes in Indonesia, Table 9. Scenario V: Earthquake in Sumatra Table 10. Frequency of scenarios and prevalence of undernourishment Table 11. Selected policy transfers in Indonesia Table 12. The reference scenario without a shock: estimated impacts of rice price support policies Table 12. Impacts of different programmes on the prevalence of undernourishment (percentage points) 37 Table 13. Impacts of policies on the average and standard deviation on the prevalence of undernourishment

4 Figures Figure 1. Impacts of rice price support on undernourishment Figure 2. Production and trade of soybeans in Indonesia Figure 3. Production Index of five main food commodities Figure 4. Poverty and food poverty ratios, and proportion of the transitory poor Figure 5. Prevalence of undernourishment in Indonesia, chronic and transitory Figure 6. Prevalence of undernourishment and poverty rate in Figure 7. Per capita expenditure and share of expenditure by region in Figure 8. Estimated average demand elasticities by expenditure class (quintiles) Figure 9. Impacts of rice price support on undernourishment Figure 10. International and domestic rice price in Indonesia and the Philippines (USD/kg) Figure 11. The recipients of Raskin (SUSENAS database) Figure 12. The impacts of transfers through food, cash and fertilizers Figure 13. The median expenditure response to social programmes (percentage points of total expenditure) Figure 14. The impacts of rice price support combined with other policies in different scenarios Figure 15. The impact of a selection of policies across scenarios Boxes Box 1. The risk assessment process Box 2. Distribution of price and income elasticity of food and non-food demand in Indonesia Box 3. Food security legal and institutional framework in Indonesia Box 4. Implications of rice trade restrictive measures and price support

5 EXECUTIVE SUMMARY AND POLICY RECOMMENDATIONS Indonesia has made major improvements in its food security situation, with the prevalence of undernourishment currently estimated at 9% of the population, half of what it was only a decade ago. Further reducing the prevalence of chronic undernourishment and building resilience to a variety of food insecurity risks remains a high priority for Indonesia. The New Food Law No. 18/2012 highlights a general principle of food security as kemandirian pangan (self-reliance) and establishes domestic production of staples as the priority action to be undertaken. The current strategic plan for established production targets for 39 products. For five food commodities (rice, corn, soybean, sugar and beef) the targeted levels aim to achieve self-sufficiency. Price support policy for rice is largely implemented by Bulog (a public logistic agency) through a combination of direct interventions in the domestic market, including delivery of rice at subsidised prices to poor households (Raskin), and trade restrictions. As a consequence of these policies, the OECD estimates that the domestic rice price was 60% higher than the reference international price in compared to 8% in Analysis of household expenditure data (SUSENAS) indicates that current rice policies increase the rate of undernourishment in Indonesia between 2 and 22%, depending on the degree of price transmission from international markets. A consultation process was implemented among Indonesian stakeholders and policy makers to identify the perceived threats to food security and possible policy responses to them. Five main scenarios were selected as main threats to food security in Indonesia: a rice price hike in the world market; a macroeconomic crisis; an increase in the world price of fuel; a crop failure due to a pest infestation affecting rice; and an earthquake on the island of Sumatra. This list is not exhaustive, but it allows a significant number of risk situations to be represented in the subsequent assessment of risk and identification of policy options. The study finds that domestic economic and natural disaster scenarios are more important than global price hikes both in terms of their likelihood and their potential impact on food insecurity. This fact should guide policy design. While developing early warning systems and disaster management strategies are beyond the scope of this study, they deserve strong policy attention in Indonesia. The five risk scenarios were used to examine the performance of existing agriculture and social policies: the rice price support measures, the rice consumer subsidies programme (Raskin), the social programme of unconditional cash transfers (BLT), and the fertilizer subsidy programme. This is the policy setting that currently exists in Indonesia and has not been changed by two recent framework laws: the 2012 Food Law and the 2013 Farmer s Empowerment Law. Other potential policy options have also been investigated: a stylized catastrophic crop insurance programme, a food aid programme distributing vouchers for staples, and a more targeted version of the cash transfers (BLT). The results of the analysis indicate that the current rice price support measures do not contribute to improve any dimension of food security, including stability, but instead they worsen the situation. Export restrictions can help avoid a surge of undernourishment only in the case of a price spike that is estimated to occur every 30 years. Import restrictions worsen the food security situation in all other scenarios, in 5

6 particular in the crop failure scenario, increasing the prevalence of undernourishment in Indonesia by 12 percentage points. Raskin delivers rice at a reduced price for the poor, but it has very high administrative costs and is not well targeted to the undernourished population. Indeed, almost half of Indonesian households benefit from this programme. The impacts of Raskin in terms of reducing the prevalence of undernourishment are not large, but they are robust across a diversity of scenarios. Its effectiveness is reduced most significantly in scenarios which imply a higher price for rice, such as under a price hike or a crop failure. The possibility of converting the programme into food vouchers that include other staples beyond rice should be explored. The analysis of this alternative policy option in this study looks promising: the food vouchers programme is more effective than Raskin in reducing undernourishment in all the scenarios. The reason is that households are more willing to increase expenditure on other staples than rice. This opens opportunities to increase the diversification in the diet, another policy objective of the government. The unconditional cash transfer programme BLT is slightly better targeted than Raskin. Its results across scenarios are more robust and the potential to improve its targeting is also larger. Because the programme provides income support, its impact on poverty is in general larger than on undernourishment: consumers can decide to spend extra income on non-food or luxury food goods. There is scope to improve the targeting of this unconditional cash transfer; if the transfers were concentrated in the lowest income quintile of the population its impacts on reducing undernourishment would double. Fertilizers subsidy programs are not effective in reducing food insecurity across any of the scenarios. Due to its low transfer efficiency, its poor targeting to the poor and undernourished and its small effects on lowering food price, the impact of fertilizer subsidies on reducing under nutrition is estimated to be small. Crop insurance is expensive and difficult to develop among very small producers, while its contribution to food security is only positive for crop failure scenarios. The objective of protecting vulnerable farmers from the effects of crop failures could be better served through social programmes. Cash transfers could be better tailored to the reality of farm households that are particularly affected by climate and production variability. The objective of developing risk management alternatives as part of a strategy to facilitate investment and growth in agriculture is fully complementary with food security objectives. The potential for developing market-based risk management tools such as futures and insurance deserves further technical assessment. However there are important prerequisites to develop such instruments: a well-functioning international market with a large number of private participants is needed for a futures market to exist; a good production risk information system, such as a dense network of weather stations, is required to develop actuarially sound insurance. The following policy recommendations are proposed to improve the portfolio of policies that serve to manage risks of food insecurity in Indonesia: 1. Dismantle the rice subsidy programme Raskin and substitute it with a Food Voucher programme that at the same cost could be better targeted to the most vulnerable part of the population. Food vouchers would be used to buy staples beyond rice, including other grains. The exact list of products covered by the voucher should be decided in consultation with regional groups and could be differentiated regionally to respond well to local food preferences. 2. Improve the targeting of unconditional cash transfers, including via triggers based on income and possibly special triggers based on weather or production losses for farmers. The convergence of social and food aid programmes should continue and the new food voucher programme should be 6

7 jointly managed with other social programmes to improve effectiveness and enable better monitoring of results. 3. Reform Bulog, by reducing its commercial activities and re-focussing on the neutral management of emergency food reserves. The floor purchasing price of rice should be phased out over time. Further analysis should be undertaken to define a good governance structure for the emergency reserve system and the links with the sub-national reserves and ASEAN+3 emergency rice reserve system (APTERR). 4. Reform of the administrative requirements for agro-food imports, including import permits for rice. Facilitating imports and the active participation of the Indonesian and foreign traders and investors can contribute to rural growth, incomes, and food supplies. 5. Promote a coordination agreement within ASEAN to restrain the use of export restrictions and eliminate the administrative requirement of export permits. Export restrictions are very damaging for global and regional food security and, when prices increase, they can create policy traps that can exacerbate price spikes. The ASEAN region includes large exporters and importers of rice and more open and reliable regional trade could conceivably do more to reduce the variability of rice prices and ensure availability in all countries. 6. Phase out fertilizer subsidies and use these budgetary outlays for strategic public investments, including in people (education, training, extension services) and in physical infrastructure. Priorities should be identified in consultation with regional groups. This report uses a risk management portfolio approach for policy assessment and highlights trade-offs between different policy options. In order to implement and design the details of the recommendations in this report, several policy areas deserve further analysis: practical ways to improve the targeting and effectiveness of cash transfers and food aid programmes; the governance structure and rules for a neutral emergency reserve system; modalities and benefits of reinforced regional trade agreements particularly on rice; and prioritisation of appropriate public investments in a resilient food and agriculture system in Indonesia. The proposed improvements in policy targeting can significantly contribute to the stability of food security in Indonesia and can complement other long-term policy objectives related to development and poverty reduction. 7

8 TRANSITORY FOOD INSECURITY IN INDONESIA Introduction 1. Indonesia has made impressive progress in improving its food security situation in the last two decades according to the FAO Food security indicators. The prevalence of undernourishment has been reduced from 20% of the population in to 13% in and 9.1% in Other indicators such as the depth of undernourishment or the depth of food inadequacy follow a similar trend. According to the Global Hunger Index (IFPRI 2013) Indonesia ranked 23 rd among 78 developing countries, behind China, Malaysia, Thailand and Vietnam, but ahead of the Philippines, Cambodia and Lao PDR. Indonesia has improved its GHI from 19.7 in 1990 to 10.1 in This evolution is the result of decades of high growth rates and reductions in poverty. 2. The 2012 Food Law puts self-sufficiency or self-reliance in the centre of the Indonesian approach to food security. The concept needs to be interpreted and translated into efficient policy options for Indonesia. In response to this challenge, the main focus of this report is on the analysis of agricultural policies and other policy options from the angle of food security objectives and, in particular, the management of risks for transitory food security of Indonesia. The objective is to investigate policy options that would improve Indonesians nutrition situation and make it less vulnerable and more resilient. The topic of this report is relevant for several countries in the region because they also follow self-sufficiency approaches to food security. 3. This report applies the framework for the analysis of transitory food insecurity [TAD/CA/APM/WP(2013)35/FINAL]. The main concepts of stability and transitory dimensions of food insecurity are used as defined in that framework. Following the framework, this study includes two steps: preparatory work and a comprehensive risk assessment including a policy assessment and dialogue. The presentation of the results is also aligned with the portfolio approaches suggested in the framework. Further details on the application of the framework to Indonesia and on the consultation process with stakeholders are available in the Technical Background Report on Transitory Food Insecurity in Indonesia [TAD/CA/APM/WP(2014)15/REV1]. 4. The report is organised in three sections: from the statistical analysis of the current situation of food security in Indonesia, to the scenarios of threats and risks, and, then, to the possible policy responses. Section 1 analyses the state of food security in Indonesia. This information constitutes the benchmark for the analysis of the stability dimension of food security. Section 2 defines five scenarios of food insecurity risk and compares their impacts on food insecurity to the benchmark. The main focus is on the prevalence of undernourishment, but other indicators are presented and discussed such as the depth of undernourishment, the protein intake, and poverty. Section 3 develops a preliminary policy assessment based on three existing policy measures: rice price support, rice consumption subsidies (Raskin) and unconditional cash transfers. A set of preliminary policy conclusions concludes this report. 8

9 1. The state of food security in Indonesia 5. This section analyses the food security situation of Indonesia in a benchmark year of Two main sources of data are used: 1 the production balance sheets and the food security indicators of FAO (2013) and the National Socio Economic Survey of Indonesia (SUSENAS). SUSENAS is a Nation-wide household survey started in , representing 33 provinces and seven groups of islands (Sumatra, Jawa, Nusa Tenggara, Kalimantan, Sulawesi, Maluku, and Papua) in Indonesia. 2 In 2007, the sample size of the panel survey was expanded to cover around households, including about individuals. It contains a core questionnaire which consists of household characteristics (location of residence, the sex, age, marital status, and education and employment of all household members, and receipt of social security or subsidized rice programme) and a consumption module questionnaire which gathers household s consumption information. Every three years a wider SUSENAS survey is undertaken with other information such as health, education, and housing. The panel part of the surveys is repeated on the same households, allowing the construction of a panel data in certain periods such as the period that is frequently used as reference is this paper. Food Availability 6. Indonesia is a net agro-food exporting country in value terms. Food exports are concentrated in perennial crops such as palm oil, coffee, coconuts and cocoa beans, while non-food agricultural exports include also natural rubber. Food imports include wheat, soybeans, sugar and animal products. Rice is a staple and its production has grown to more than 60 million tonnes in the last decade, keeping pace with domestic consumption. Imports have been historically below 4% of total consumption, except in certain years such as in the second half of the 1990s with imports above 10%. Rice imports and exports are currently controlled by a system of import licenses. 1. Another important source of information for local analysis of transitory food insecurity in Indonesia is the Food Security and Vulnerability Atlas of the World Food Program (WFP 2009, 2012). It provides detailed information on the vulnerability to food security risk in different districts. The Cost of Diet Methodology (Baldi et al., 2013) allows to estimate the local monetary needs for a good diet. 2. This report defines nine regions for seven groups of islands, separating Jakarta and Bali from Jawa Island. 9

10 Tonnes Tonnes TAD/CA/APM/WP(2014)14/FINAL Figure 1. Impacts of rice price support on undernourishment Production Imports Stock variation Exports Domestic supply Source: FAO Stat. 7. Soybeans are an important source of protein for Indonesians and imports have increased in the last two decades. Geo-climatic conditions are not particularly suitable for this crop. The production of soybeans fell by half, from a peak of more than 1.5 million tonnes during the 1990s, and has been maintained at around 0.7 million tonnes, less than 50% of domestic consumption. Figure 2. Production and trade of soybeans in Indonesia Production Imports Exports Domestic supply Source: FAO Stat. 8. According to the Agriculture Strategic Plan the government attentively follows the production and self-sufficiency performance of five main food commodities: rice, soybeans, sugar, maize 10

11 and beef. Maize is mainly used for feed in Indonesia and its production has significantly increased in the last decades following the pace of demand. The production of sugar and beef has stagnated in the last two decades and imports have been growing to meet the demand. Figure 3. Production Index of five main food commodities (1960 = 100) Source: FAO Stat. Access to food: poverty and prevalence of undernourishment 3 9. Two indicators of access are widely used in the existing literature: the rate of poverty (or food poverty) and the rate of undernourishment. Food poverty is an indicator on whether an individual has the purchasing power to buy sufficient amounts of food to maintain a minimum level of calories. Individuals are considered as food poor when their household expenditure per capita per day is below the food poverty line set by BPS 4. The food poverty line is set as the minimum expenditure required to purchase a basket of food items to satisfy kcal per capita per day. The poverty line includes other minimum non-food expenditures. In this sense, the food poverty ratio is a more restrictive definition of poverty: all the food poor are also poor. Although this indicator is widely used and practical to illustrate the long-term trend of food insecurity, it is not well-suited to capture transitory food insecurity since normally external shocks, such as natural disasters, are accompanied with an increase in non-food expenditure, so that the food poverty ratio is seemingly improved on such an occasion. 10. Figure 4 presents the poverty ratio and its composition in based on the official poverty line published by the BPS. The incidence of poverty is further categorised as chronic and transitory using the panel information. The definition of transitory poverty used in our analysis is different from the conventional definition. The conventional definition of chronic poverty includes households or individuals that are identified as poor in five consecutive years. Our definition is based on a shorter panel of just threeyears. We define as chronic poor the households that are poor in a certain year and are poor or near-poor in the other two years. The near-poor households are conventionally defined as those whose total expenditure is less than 1.2 times the poverty line. The households in poverty which are not categorised as chronically poor are defined as transitory poor. Therefore our definition of chronic poor is less restrictive than the conventional one, while our definition of transitory poor is more restrictive. 3. All data and figures in this section on poverty and undernourishment and in the next section on nutritional outcomes are based on data from SUSENAS database and calculations by the OECD Secretariat. 4. BPS sets the regional poverty line and food poverty line for urban and rural households separately. When the household is located in an urban area, its total expenditure is compared with the urban (food) poverty line to judge whether it is (food) poor. 11

12 Headcount ratio (percentage) TAD/CA/APM/WP(2014)14/FINAL Figure 4. Poverty and food poverty ratios, and proportion of the transitory poor Transitory poor Chronically poor Food poor Source: SUSENAS Database. 11. The poverty ratio, i.e. the percentage of population below the poverty line, in Indonesia is around 15%, decreasing every year between 2008 and 2010 (Figure 4). This trend is confirmed in the official rate of poverty published by the Indonesian statistical agency. The data also show a much higher poverty incidence for farm households (21.5%) than non-farm households (5.3%) in Of the poor, 67% has been in and out of poverty during this time period. Despite our restrictive definition, transitory poverty accounts for more than 60% of the poverty incidence in Indonesia. The reduction of the poverty ratio from 2008 to 2010 comes mainly from the reduction of population under transitory poverty. The incidence of food poverty can also be calculated from the SUSENAS dataset and, by definition, is lower than poverty, in the range of 4-8% of the population. The surge in the number of food poor in 2009 is hard to interpret and may partly be the result of a statistical effect of a re-classification of urban and rural households in that year. Food poverty is concentrated intensively in less developed regions such as Papua, Nusa Tenggara and Maluku. The incidence of food poverty among farm households (5.2%) was higher than non-farm households (2.2%) in The second indicator of access to food is the rate of undernourishment, which is determined by the household calorie consumption. An individual is defined as undernourished when daily calorie intake is less than a certain threshold. 6 The distribution of calorie intake per day per capita in the 2010 SUSENAS survey is presented in Figure 5. The ratio of the undernourished applying the re-calibrated threshold is 5. A household is classified as farm household when the occupation of the household head is an employer in agriculture (including self-employment). 6. The FAO sets the official threshold for Indonesia at kcal. Directly applying the FAO threshold to the survey distribution leads to a prevalence of undernourishment that is systematically higher than the official FAO estimation. To avoid this mismatch, this report recalibrates the threshold in order to reproduce a ratio of undernourishment that is consistent with the official number published by the FAO and known by policy makers and stakeholders: 13% in Please see more detailed discussion on this point in Box 2 of the Technical Background Report on Transitory Food Insecurity in Indonesia [TAD/CA/APM/WP(2014)15/REV1]. 12

13 13%, which replicates the FAO s estimation for There is a large share of the total population that is just above the threshold and, therefore, vulnerable to become undernourished after external shocks. Applying to undernourishment the same definition of chronic and transitory as that applied to poverty. According to this definition, no more than 25% of the undernourished were chronic in Indonesia in the period (Figure 5). 7 Due to the high concentration of population near the threshold, a large share of the undernourished population moves in and out of the threshold in consecutive years. This is a factor of vulnerability and instability of food security for the Indonesian population. 13. The mean of calorie intake for the 2010 sample is kcal and the standard deviation is 596 kcal. The situation of undernourishment is different across islands and regions. Using the national threshold, a higher proportion of the population in Maluku and Papua (24% and 18.7% respectively) are undernourished compared with those who live in other regions (Figure 6). Although farm households have a higher incidence of poverty and food poverty, the incidence of undernourishment is smaller among farm households (12.1%) than among non-farm households (13.3%). Similarly, the rate of poverty is higher in rural areas (17% as compared to 10% for urban), but the prevalence of undernourishment is lower (12% as compared to 15% for urban). This could be explained by lower incomes in rural areas, but easier access to food through self-production or community networks that are less frequent in urban areas. Figure 5. Prevalence of undernourishment in Indonesia, chronic and transitory Source: SUSENAS Database. 14. There is not a strong geographical correlation between the poverty and undernourishment indicators (Figure 6). For instance, Jakarta, a big city, has a very low poverty rate of below 5% as measured by expenditure and income, but it also has high rates of undernourishment beyond 15%. On the contrary, Papua has much larger poverty indices than prevalence of undernourishment. 7. Figure 6 shows an increase in the rate of undernourishment from 2008 to The FAO record a reduction in the same period. This is the result of the different threshold and distribution between our calculations and FAOs. Additionally the FAO publishes its results in the form of three year moving averages, which captures better trends than variability. 13

14 Figure 6. Prevalence of undernourishment and poverty rate in 2010 per cent Source: SUSENAS Database. 15. There is no systematic relationship between poverty and undernourishment. Only 5% of households are at the same time poor and undernourished. The additional 8% of households suffering from undernourishment are not poor. Nine per cent of all households are poor but have sufficient intake of energy from food. 16. The consumption module of SUSENAS shows on average around a 60% share of expenditure dedicated to food, with higher per capita expenditure and lower share of food in expenditures among nonfarm households (Figure 7). In Jakarta, the per capita expenditure is the highest, while the share of food in total expenditure is the lowest (45%). At the other extreme, farm households in Papua have very low per capita expenditure and a very high share for food (more than 70%). Figure 7. Per capita expenditure and share of expenditure by region in 2010 Source: SUSENAS Database. 14

15 Nutritional outcomes 17. Even if a household has potential access to food due to its income level, consumption decisions lead to different nutritional choices. 8 The largest share of food expenditure in Indonesia (Table 1) is rice for farm households (16%) and readymade food (16%) for non-farm households, followed by rice. On average, rice has by far the largest share in both calorie and protein intake in Indonesia with 44% and 50% of household calorie intake for non-farm and farm households, respectively. 9 On the other hand, rice accounts for a smaller share of food expenditure. Seafood and soybean/nuts are important sources of protein intake accounting for around 15% and 8% in protein intake, respectively. Table 1. Composition of food expenditure, calorie intake and protein intake in 2010 Share in food expenditure Share in calorie intake Share in protein intake Nonfarm Farm Nonfarm Farm Nonfarm Farm Grains/Tubers 0.9% 2.7% 2.8% 6.8% 1.7% 4.2% Rice 10.6% 15.9% 44.1% 50.1% 36.9% 44.4% Seafood 4.9% 6.1% 2.5% 2.6% 13.9% 15.1% Meat 1.7% 1.7% 2.1% 1.3% 4.3% 2.7% Eggs/Milk 3.0% 2.4% 3.1% 1.8% 6.0% 3.9% Vegetables 4.7% 6.5% 2.0% 2.5% 4.5% 6.8% Soybeans/Nuts 1.8% 1.9% 2.8% 2.5% 8.8% 7.9% Fruits 2.5% 2.8% 2.1% 2.4% 0.9% 1.0% Oil/Fats 2.3% 3.1% 12.2% 12.8% 0.6% 1.1% Ready-made food 16.4% 11.9% 22.4% 14.1% 18.9% 9.9% Seasoning 1.3% 1.6% 0.8% 0.8% 1.2% 1.2% Other food 1.4% 1.5% 3.1% 2.3% 2.3% 1.8% Non-food 48.4% 41.8% Source: SUSENAS Database. 18. The share of calorie intake from different types of food is different in different regions (Table 2). For the non-poor households in Jakarta, 28% of the energy intake comes from readymade food. At the other extreme, the poor in Maluku or Papua consume a lot of tubers, which represent, respectively, 20% or 50% of daily calorie intake. 8. Child malnutrition is one of the more extensible used indicators of nutritional outcomes. According to FAO the percentage of children under five years of age who are underweight has been reduced in Indonesia from 30% in 1992 to 25% in 2000 and 19% in This information is not available in SUSENAS database. 9. The large shares of rice in calorie and protein intake are confirmed with similar shares of supply calculated by FAO. 15

16 Table 2. Composition of calorie intake by region Sumatra Jawa Jakarta Bali Tenggara Kalimantan Sulawesi Maluku Papua Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Gains/Tubers 2% 2% 3% 7% 2% 1% 2% 4% 7% 13% 3% 2% 5% 10% 13% 20% 15% 50% Rice 48% 61% 44% 55% 36% 52% 49% 64% 55% 60% 46% 61% 49% 57% 38% 37% 36% 22% Seafood 3% 3% 2% 1% 2% 1% 2% 1% 2% 1% 4% 3% 4% 3% 5% 4% 4% 2% Meat 2% 1% 2% 0% 4% 1% 4% 1% 2% 0% 3% 1% 1% 0% 1% 1% 4% 2% Egg/Milk 3% 2% 3% 1% 5% 3% 3% 1% 2% 0% 4% 2% 3% 1% 2% 1% 4% 1% Vegetables 2% 2% 2% 2% 1% 1% 2% 2% 3% 3% 2% 2% 2% 2% 2% 3% 3% 3% Soybeans/Nuts 2% 1% 4% 3% 3% 3% 3% 2% 3% 2% 2% 1% 2% 1% 1% 0% 3% 3% Fruits 2% 2% 2% 1% 2% 1% 3% 2% 2% 2% 2% 1% 4% 3% 5% 4% 3% 2% Oil/Fat 15% 13% 12% 11% 12% 11% 10% 8% 9% 8% 12% 11% 12% 11% 18% 18% 14% 10% Ready-made/Drinks 17% 11% 23% 14% 28% 21% 20% 13% 15% 10% 19% 13% 16% 10% 13% 10% 11% 4% Seasoning 1% 1% 1% 1% 1% 1% 1% 0% 1% 0% 1% 1% 1% 1% 1% 1% 1% 0% Other foods 3% 2% 3% 2% 4% 4% 2% 1% 2% 1% 4% 3% 2% 1% 2% 2% 3% 1% Source: SUSENAS Database. 2. An assessment of food insecurity risks in Indonesia 19. Emergency situations due to specific shocks may require policy responses: a price spike of a staple food or sudden food inflation, a nation-wide drought that generates a crop failure, an economic slowdown reducing the income of the poor, a local natural disaster (e.g. an earthquake) that destroys assets and livelihoods, etc. These contingent situations, with their potential damage and likelihood define the risks to the stability dimension of food security. A good policy design can contribute to an efficient management of potential risk scenarios or threats to food security and it can mitigate possible transitory surges in the prevalence of food insecurity The assessment of risks of food insecurity in Indonesia was undertaken following the predefined process: risk identification, risk analysis and risk evaluation in Framework for the Analysis of Transitory Food Insecurity (see Figure 4 in TAD/CA/APM/WP(2013)35/FINAL). Box 1 summarizes the risk assessment process and selection of scenarios. 10. Although there are several indicators of food insecurity, some of which have been discussed and presented in Section 1, the analysis in Sections 2 and 3 is focused on the prevalence of undernourishment. The demand simulation model allows similar analysis and quantifications for other indicators, such as poverty, food poverty and those indicators based on other nutrients (e.g. protein intake). 16

17 Box 1. The risk assessment process Risk assessment is one of the crucial steps in the Framework for the Analysis of Transitory Food Insecurity. The objective is not developing an exhaustive list of risks and scenarios, but to have a short list that is recognized as relevant and plausible by stakeholders and experts. This is why respecting a process to identify the scenarios is as important as the scenarios that are finally retained. More details on the application of the process in Indonesia can be found in the technical background paper [TAD/CA/APM/WP(2014)15/REV1]. Risk identification. In September 2013 an electronic survey was carried out among policy makers and experts to gather information on their perceptions of the risks of food insecurity. In October 2013, two roundtables among stake holders and experts were organized in Indonesia to identify the perceived risks and threats to the food security. The first roundtable took place in the Ministry of Agriculture in Jakarta and gathered policy makers and officials from different units, ministries and agencies. The second roundtable took place in Bogor with the participation of several experts from academia and international organizations. As a result of the survey and the discussions, a set of 11 risk scenarios was identified for further analysis and evaluation. Risk analysis. The eleven scenarios have been quantified in two modelling approaches. The existing empirical literature and data was used as main sources for the scenario definitions in the corresponding background paper. A general equilibrium model was used to ensure the quantitative consistency of each scenario. The SUSENAS database was used to estimate food demand systems that allow replicating the food consumption response to external shocks for all the households in the survey. Risk evaluation. The estimated demand systems have been used to micro-simulate the shocks in each risk scenario and their implications for the nutrition status of all the Indonesians represented in the survey. The risk assessment result of each scenario was summarised in a one page note, which was used to facilitate the discussion at the consultation seminar in Bogor on 26 February More than 60 participants from government including high level officials, international organisations (WB, WFP, UN-CAPSA), research centers (ICASEPS) and universities (Bogor Agricultural University) attended. Of the initial eleven scenarios, six have been retained in this report. The main reason for reducing the number of scenarios was the result of the discussions during the consultation seminar. The initial number of scenarios was found to be too large to be able to have a comprehensive discussion with policy makers. It became obvious that in addition to the benchmark status quo scenario, at least one example should be retained for each of the three typologies of scenarios: price shocks from world markets, macroeconomic shocks, and domestic natural disasters. Only the price shock of the most important staple, rice, was retained. Soybean price hike scenarios were analysed, but its impact in terms of the prevalence of undernourishment was relatively minor, even if it had slightly larger impacts on protein consumption. The scenarios on financial crisis and domestic macroeconomic downturn have been combined into a single scenario with an economic crisis that occurred in1997 as the reference base. An increase in international fuel prices has also been retained. Two natural disaster scenarios are also analysed: a systemic crop failure and a more local event like an earthquake in Sumatra. During the process there was discussion about considering scenarios that combine several shocks at the same time. One of the initial scenarios included a general increase in all main commodity prices, rather than just rice. However the profile of impacts on food insecurity in the combined scenario was fully dominated by the price of rice. It was also argued that the macroeconomic crisis could be combined with a commodity price hike like in However this later event was less dramatic for Indonesia as compared with the 1997 financial crisis, which was separated from a commodity shock. For all these reasons only six scenarios were retained. Of course the robust policy approach in section 3 calls for a consideration of the impacts of policy on all scenarios, even if they are not considered as occurring simultaneously. The quantification of each scenario 21. The quantification focuses on estimating and analysing the prevalence of undernourishment in different scenarios. This proceeds in two steps: a quantification of the economic shock, and a quantification of the impact in terms of households food consumption. 11 Table 3 shows the quantification of each food insecurity scenario identified in the consultation process. 11. A general equilibrium model (Indonesia-E3) was used to ensure the consistency of each scenario (Warr and Yusuf, 2014). The quantification exercise consisted in estimating the consequences of each scenario in 17

18 Table 3. Food insecurity scenarios in Indonesia 0: Reference without shock Probability: Once in 2 years I: Rice Price Spike in International Markets international price of rice by 100% Probability: Once in 30 years II: Macroeconomic crisis 11% in factor supplies (except rural land & capital) Probability: Once in 25 years III: An increase in international price of fuel international price of fuel by 114% Probability: Once in 20 years 12% reduction of factor productivity of paddy land Probability: Once in 15 years IV: Crop failure due to insect or plant disease infestation V: Earthquake and tsunami in Sumatra 10% in all factor supply (capital, land, labor) in Sumatra Probability: Once in 20 years 22. The SUSENAS database is used to estimate a food demand system model that allows replicating the food consumption response to external shocks of all households in the survey (Box 5). 12 This in turn enables the assessment of implications of external shocks for the nutritional status, such as the increase in the prevalence of undernourishment among urban and rural households, depth of undernourishment, protein intake and regional impacts. 23. This demand system is used to estimate the food demand response of each household and the impact on the distribution of calories intake and undernourishment across households. Other indicators such as poverty and protein intake are also analysed. The reference level is defined as the current food security situation of Indonesia that has been described in Section 1. The level of undernourishment is 13%, according to the FAO number for The rate of undernourishment is larger in urban than in rural areas. terms of income for each rural and urban centile household and in terms of relative prices. Section 2 of the technical background report [TAD/CA/APM/WP(2014)15/REV1]. 12. The equivalence between the scenarios in Table 3 in Latin numbers and those in the Box 3 of the technical background report [TAD/CA/APM/WP(2014)15/REV1] in Arabic numbers are: I=1; II=5; III=6; IV=7; and V=9. 18

19 Box 2. Distribution of price and income elasticity of food and non-food demand in Indonesia The econometric estimation of the demand system allows to compute elasticities for each of the households. Section 4 of the technical background report [TAD/CA/APM/WP(2014)15/REV1].documents the method and estimated results of a household food demand system in Indonesia. Figure 8 shows the average own price elasticities and expenditure elasticities of some selected goods by quintile class of household expenditure. The own price elasticity of rice is rather small, below 0.5 in the five quintiles. Rice price elasticity exhibits an increasing relationship with income: a 10% increase in rice prices causes a 4% decline of demand among poor households, but only 0.3% decline of demand in the richest class. On the contrary, the own price elasticity of meat is very high across all income classes. Other product elasticities tend to be in the middle between these extreme values. In general, the higher the expenditure class, the lower the income elasticity of food demand. Non-food commodities and meats are, as expected, luxury or superior goods for the lowest expenditure class: their demand increases more than proportionally with total expenditure because their expenditure elasticities are higher than one. Expenditure elasticities of meat are higher than 0.8 for all quintiles except for the 20% households with higher income. On the other hand, the demand for rice is much less elastic to income change than other food and non-food commodities. Indeed, rice becomes an inferior good for the richest 60% of households that have a negative average elasticity. Soybeans, one important source of proteins, are in the middle between rice and the rest of the commodities, with positive but lower than one income elasticities. Figure 8. Estimated average demand elasticities by expenditure class (quintiles) Own price elasticity Expenditure elasticity 24. The levels of undernourishment are directly affected by policies, such as the Raskin domestic food aid programme and existing trade measures that impede the market response of imports or exports of rice. The six scenarios considered in this report assume that existing policies do not change and remain in place as they are. This means that Raskin continues to be distributed to the same households and that exports and imports of rice continue to be restricted in each of the scenarios. However, the price support measures have an impact on price transmission between international and domestic markets which are contingent to the situation in each specific scenario. This is particularly relevant for scenarios that imply large price shocks in the domestic and international rice markets: rice crop failure and international price spike. In order to isolate the impact of the external shock from the impact of policy, two alternative policy settings are investigated in these two scenarios: the current rice price support system with restrictive trade measures and the elimination of trade restrictions. Figure 9 shows that the current price support system has impacts on increasing undernourishment by 2-22 percentage points in a reference scenario with no shock Box 3 summarizes the impacts of rice trade restrictive measures and price support on the prevalence of undernourishment under alternative assumptions 19

20 Figure 9. Impacts of rice price support on undernourishment 33% 11% Calorie intake per capita per day Without Price support Price support (full price transmission) Price support (partial price transmission) Scenario I: International rice price spike 25. Since the 1960s, three instances of sharp increases in international rice prices were observed: , and For example, between September 2007 and May 2008 the monthly rice price on the world market shot up by around 200%. In the Philippines, the peak monthly retail price of rice in June 2008 was 62% higher than the annual average retail price in 2007 (Figure 10). The main channel that affects food insecurity in this scenario is rice prices: there will be an increase in household expenditure on rice and other staples, and lagged increase in domestic production. Most households in Indonesia are net buyers of rice and will be negatively affected by higher prices. 20

21 Feb-08 May-08 Aug-08 Nov-08 Feb-09 May-09 Aug-09 Nov-09 Feb-10 May-10 Aug-10 Nov-10 Feb-11 May-11 Aug-11 Nov-11 Feb-12 May-12 Aug-12 USD per kg TAD/CA/APM/WP(2014)14/FINAL Figure 10. International and domestic rice price in Indonesia and the Philippines (USD/kg) Indonesia, Retail Philippines, Retail INTERNATIONAL PRICE, Thailand: 25% broken, Export* *International price: USD 20 per tonne of transportation cost and 10% marketing cost is added to Thai export price. Source: FAO Global Information and Early Warning System. 26. However, almost no price transmission to the domestic rice market was observed in Indonesia. This was because of various policy interventions that Indonesia put in place, including a partial ban on rice imports since 2004, prohibition of rice exports, and management of domestic rice stocks. This type of scenario is described as 1-a with a spike of 100% in the international price of rice that is hardly transmitted to domestic prices (column 1-a in Table 4). On the contrary if rice exports are permitted like in scenario I-b the impact is nation-wide due to an increase in the domestic price of rice of 44% 14 and a decline in real income across all households, in particular those that are net buyers of rice. Some farm households that commercially sell rice will benefit from the situation, but they are a minority. 27. On the other hand, the negative consequences of restricting exports and other trade measures are spread along the time: the domestic price of rice has been structurally higher than the international price in almost every year since 2000, chronically affecting the prevalence of undernourishment in the reference scenario. This is a one in thirty-year scenario and that it assumes that only one country applies export restrictions during the price spike. Export restrictions by all countries at the same time would be expected to directly exacerbate the price spike and its negative consequences for the poor (FAO et al., 2011). There are additional long-term effects of constraining exports, which in turn discourages investment and producer s response to market signals. 14. These price transmission parameters were derived from simulations using the Indonesia E-3 general equilibrium model (see [TAD/CA/APM/WP(2014)15/REV1]). The almost zero price transmission in scenario 1-a is due to policy, that is import and export bans. The almost 50% price transmission in scenario 1-b reflects both policy aspects (import restrictions) and other physical or economic factors influencing price transmission. 21

22 Table 4. Scenario I: International rice price spike (100%) Availability Production Imports and stocks I-a: Trade restrictions Rice production virtually unchanged because the domestic producer price of rice is almost unchanged. Access Food price Increase in real consumer price of rice by 0.1%. Household income and assets Household real expenditures in urban and rural areas decline by 0.05% and 0.02% on average, respectively. I-b: Exports permitted Rural households increase paddy production by 18.1%, but 22% of domestic rice is exported. Increase in real consumer price of rice and soybean by 44.1% and 3.3%, respectively. Rural and urban households experience decline in real income, mainly due to higher consumer price of rice. The magnitude depends on the share of expenditure on rice. Poor households suffer largest decline in real income. B. Impacts of Scenario I: International rice price spike Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Before shock All Indonesia Urban households Rural househods a) After rice price hike with trade restrictions All Indonesia Urban households Rural househods b) After rice price hike, export permitted All Indonesia Urban households Rural househods Results presented in the bottom of Table 4 show a strong impact on undernourishment that increases by 10 percentage points to 23% of the population in case export of rice is permitted. The median calorie intake is significantly reduced by 160 kcal per day per capita, while the depth of the food deficit 15 more than doubles. Scenario II: Macroeconomic crisis 29. Recessions occur from time to time in Indonesia. The most severe was the recession that accompanied the Asian financial crisis. In this case real GDP per person declined by 11%. This scenario looks at a broadly based macroeconomic slowdown caused by an economy-wide collapse in 15. The extent of the food deficit indicates how many calories per capita would be needed to lift the undernourished from their status, everything else being constant. The numbers do not match with the FAO numbers because we use real household data while FAO calculations are based on assumed distributions of calorie intake. 22

23 production. For convenience, it is represented as a contraction in all factor supplies of 11%, with the exception of capital and land used in agriculture. 30. The loss of real incomes leads to reduced capacity to purchase goods and services, even necessities. Food consumption was affected less than most of other goods, but it has more severe impacts on low income households with higher shares of food expenditure. Urban households are more affected than rural households which produce food (Table 5). Real GDP per capita declines by 9.7%. Availability Table 5. Scenario II: Macroeconomic crisis Production imports and stocks Production of rice and other staples remains almost constant. Imports of food grains declines by 10%. Access Food price Real price of rice increases by 9.8 per cent. Real price of staples in general increases by 8%. Household income and assets Real incomes of urban and rural household decline by 15.3% and 11.2%, respectively. B. Impacts of Scenario II: Macroeconomic crisis Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Before shock All Indonesia Urban households Rural househods After the Economic shock All Indonesia Urban households Rural househods The macroeconomic crisis has very strong consequences for food security, with the rate of undernourishment nearly doubling. The impact is only slightly larger in urban areas where the fall in economic activity is stronger. However the reduction in the median calorie intake is smaller in urban areas where the middle classes of income seem to resist better to the macroeconomic shock. Scenario III: An increase in the international price of fuel 32. Since the 1960s three instances of especially sharp increases in international petroleum prices have occurred: , and For example, between September 2007 and May 2009, the international price of petroleum on the world market shot up by around 114%. The Indonesian government subsidises petroleum products. The analysis in this scenario assumes that rates of subsidy remain constant as petroleum prices rise, meaning that international price increases are transmitted to domestic petroleum product prices. 33. Increases in petroleum prices affect food security first by raising transport costs within Indonesia for all goods, especially food, which is bulky and costly to transport, increasing domestic retail prices. Second, transport costs for people are raised, on both private and public transport, and this affects the living costs of all Indonesians. Third, the cost of producing food is particularly sensitive to petroleum prices because fertiliser, pesticides, herbicides and fuel for farm machinery are all dependent on petroleum and as the cost of petroleum rises, the cost of producing food also rises (Warr, 2011). These linkages negatively affect food security, but the impacts are transmitted with some lag. 23

24 34. This scenario takes the 114% increase in the price of fuel from 2007 to 2009 as the basis for the shock. The analysis then estimates the effect of this price increase on food security, assuming that trade policy with respect to food remains at its present settings (Table 6). Availability Production imports and stocks Table 6. Scenario III: Increase in fuel prices Real costs of food production increase by 4.3 %. Total grain consumption declines slightly by 0.02%. Access Food price Real consumer prices of rice and soybeans increase by 36.1% and 3.7%, respectively. Household income and assets Real incomes of urban and rural households decline by 0.2% and 3.0%, respectively. B. Impacts of Scenario III: An increase in the international price of fuel Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Before shock All Indonesia Urban households Rural househods After the Fuel price hike All Indonesia Urban households Rural househods The increase in undernourishment is not as dramatic as in the previous scenarios. The prevalence of undernourishment increases by 8 percentage points to 21% of the population. The impact is a bit larger in rural areas due to the negative impacts on transportation and agricultural input prices. Scenario IV: Crop failure due to a plant diseases infestation 36. Rice is grown as a monoculture and is therefore especially susceptible to disease and pest infestations. Pests such as the brown plant hopper (BPH), known locally as wereng coklat, have been a major problem in the past. Major outbreaks occurred during the planting season and there was an especially severe outbreak in More recently, outbreaks occurred in 1998 and A major outbreak is possible in 2014 according to the Department of Plant Protection at the Bogor Agricultural University. The loss of 6 million tonnes of paddy output is predicted, equivalent to about 4 million tonnes of milled rice, roughly 12% of the annual crop. 37. Insect and plant disease infection may spread across a wide area of Indonesia. Rural households in infected regions are the most severely affected. Rural households in other regions gain from a higher rice price if they are net sellers of rice but lose if they are net buyers. There may be severe impacts on urban households, particularly for low income households. Research has linked BPH infestations to excessive use of pesticides. The BPH is not itself susceptible to insecticides, but its natural biological enemies are. The BPH population has gradually increased. 38. A reduction in rice yield would force a combination of increased imports and higher domestic price of rice. The combination that would result will depend on government policies. Increased imports may not be enough to prevent an increase in real domestic rice prices. Quantification in sub-scenario IVa assumes constant import quotas and leads to high increases in undernourishment, while IVb allows imports to occur at a constant tariff leading to a much smaller impact on undernourishment. 24

25 Availability Production imports and stocks Table 7. Scenario IV: Crop failure IV-a: Trade restrictions Access Food price Increase in real consumer price of rice by 48% Household income and assets IV-b: Imports permitted 12 % loss of production of rice 12% loss of production of rice Imports increase to supplement domestic demand Rural and urban households experience 1.2% and 2% decrease in real income on average, respectively. B. Impacts of Scenario IV: Crop Failure Increase in consumer price of rice by 16.1 % Household income remains almost constant on average. Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Before shock All Indonesia Urban households Rural househods a) After crop failure All Indonesia Urban households Rural househods b) Import allowed after crop failure All Indonesia Urban households Rural househods A crop failure will have a very large impact on undernourishment, up to a prevalence rate of 25% if imports do not rapidly find their way into Indonesia (scenario IV-a in Table 7). Rural households will be the most affected because of the higher impact of the crop failure in the rural economy. This high impact on undernourishment would not occur if trade restrictions did not impede imports (scenario IVb in Table 7). In this case, the prevalence of undernourishment would be mitigated to an increase of 3 percentage points. Scenario V: Earthquake and tsunami in Sumatra 40. Indonesia is located on the seam of four major tectonic plates in one of the most seismically active regions in the world. As a consequence, earthquakes occur frequently in the region and the country experiences some of the strongest earthquakes in the world (Table 8). Since Indonesia is composed of islands and the majority of the population live in coastal areas or nearby, the country is highly vulnerable to tsunamis. Sumatra Island and the surrounding small islands comprise one of the regions where important earthquakes have occurred most frequently over the past decade. The region accounts for 23% of rice production. 41. According to Gignoux and Menéndez (2013), earthquakes of intensity higher than 7 are likely to have effects on individuals, with the most important economic damages involving asset and household losses. They estimated that an earthquake with an intensity level higher than 7 reduces per capita consumption expenditure by, on average, 10% for rural households and 6% for urban households in the short term. 25

26 Date Table 8. The most important earthquakes in Indonesia, Place Magnitude (Richter scale) Number of deaths Number of people affected Damage (000 USD) Bali Flores Region Sumatra Andaman Islands Java Southern Sumatra Southern Sumatra Source: The OFDA/CRED International Disaster Database 42. The channel through which an earthquake will affect food security is the loss of income, private assets and public infrastructure. There is a temporary loss of physical access to food and water in the affected regions and higher food prices in local markets. Impacts are catastrophic, but more location specific. Households in remote areas which lose transportation infrastructure will be more severely damaged. Higher food prices have a more modest impact on low income households in non-affected regions, operating through higher food prices. Table 9. Scenario V: Earthquake in Sumatra Availability Production imports and stocks Rice production in Sumatra declines by 9.8% and rice production in Indonesia declines by 1.7%. Access Food price Increase in consumer price of rice, soybean and meat by 3.5%, 3.3% and 3.3%, respectively. Household income and assets Nominal income of urban and rural household decline by 3.0% and 1.9%, respectively. B. Impacts of Scenario V: Earthquake in Sumatra Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Before shock All Indonesia Urban households Rural househods Sumatra Urban households Rural househods After the Earthquake All Indonesia Urban households Rural househods Sumatra The effect is an increase in undernourishment of 1 point in the whole of Indonesia, but 4 points within Sumatra (bottom of Table 9). This is a relatively frequent event that has a small impact on the nation-wide prevalence of undernourishment, but with a much larger effect at the local level. The local prevalence of undernourishment can be a more relevant indicator of the importance of these types of events for policy makers and Indonesian society. 26

27 A map of Indonesian food insecurity scenarios 44. The reference scenario reflects the current levels of undernourishment with no particular extreme shock in place. Similar circumstances are likely to occur at least every other year and are then estimated to represent the most likely situation (Table 10). According to estimates of experts and the discussion in the consultation seminar, the most frequent among the risk scenarios are likely to be crop failures (once every 15 years), followed by earthquakes, and increase in fuel prices (both occurring once every 20 years). Economic crisis and price spikes are the most unlikely of the six scenarios. Table 10. Frequency of scenarios and prevalence of undernourishment Scenario A. Frequency (No. of years) B. Prevalence of undernourishment C. Expected of undernourishme nt (1/A)*(B-B 0) 0: Reference without shock Ib: Rice Price Spike in International Markets II: Macroeconomic Crisis III: Increase in international fuel price IVa: Crop failure due to plant infestation V: Earthquake and tsunami in Sumatra D. The worst local prevalence of undernourishment Maluku 24 Nusa Tenggara 26 (+14) Java 25 (+11) Sulawesi 20 (+8) Java 29 (+15) Sumatra 14 (+4) 45. The frequency of local events such as the earthquake in Sumatra requires some explanation. The likelihood is small in a specific location, but the probability of small events in different locations can be larger. The frequency of 20 has been retained following the advice of experts. This study focuses on food insecurity risk which impacts at the national level, although there may be regional specific disasters which require regional level response. 46. The analysis of the prevalence of undernourishment in different scenarios needs to account for the estimated frequency of each scenario. The simplest way of doing this is by multiplying the expected prevalence of undernourishment in each scenario by the likelihood of this scenario. This expected damage indicator is calculated in column C of Table Both the frequency and the prevalence of undernourishment are highest in the crop failure scenario. This makes this scenario most relevant for food security as represented by the expected damage indicator. All other scenarios have smaller prevalence of undernourishment in the range 21%-23%. The lowest impact on undernourishment is observed in the local scenario of an earthquake in Sumatra, with the prevalence increasing to 14% in the whole country, but with highest impacts on the island itself. 3. Policy options and their impact across scenarios 48. Following the assessment of risk scenarios, a policy analysis and dialogue was undertaken with Indonesian stake holders. 16 The starting point for the policy analysis is the set of relevant regulations and institutions, and policy programmes already existing in Indonesia. A preliminary policy analysis is 16. The policy dialogue began with a consultation seminar in February Section 3 of the technical background report [TAD/CA/APM/WP(2014)15/REV1] summarizes the discussion on the policy options and responses with the participation of high level policy makers.. 27

28 undertaken in this section, beginning with policy measures already in place, and moving then into a broader policy discussion. 49. Recent framework laws related to agriculture have emphasised food security and food selfsufficiency (Box 2). The 2012 Food Law strengthens the principles of food sovereignty (kedaulatan pangan) and food self-reliance (kemandirian pangan) as dominating approaches to food security. The term kemandirian pangan is sometimes translated as independence or self-reliance, and at other times as selfsufficiency, all of which causes confusion. The same terms are also used in the Farmers Empowerment Law of 2013, but with no reference to food security. The Food Law specifically mentions food production threats to food availability. The list of threats in the Law includes climate change, invasive species, natural disasters, social disasters, environmental pollution, degradation of land and water, shifts on land use and economic disincentives. This provides a direct link between the Law and food security stability, in particular across a set of scenarios many of which are analysed in this report. Food crisis is mentioned in the Law without a definition of its meaning. 50. Since 2000, several import restrictions including tariffs, quotas, licensing requirements, product registration and entry port, have been introduced in Indonesia for different products, including sugar, rice, meat, cereals and horticulture (OECD 2012 and 2013b). Even if some of them have been rolled back after negative domestic impacts, they are an indication of a possible anti-trade interpretation of the self-reliance concept in the Food Law (Box2). Set of Policies in this analysis 51. The major agricultural policies in place in Indonesia are analysed and discussed in depth in the OECD Review of Agricultural Policies in Indonesia (2012). The main policies include price support measures, food consumption subsidies (Raskin), and fertiliser subsidies. All are somehow related with the Indonesian self-sufficiency or self-reliance approach to food security. These policies are quantified in the current policy analysis, together with other alternatives such as unconditional cash transfers, food vouchers and agricultural catastrophic insurance. 52. Market price support constitutes the largest share of agricultural support in Indonesia. It affects several commodities, in particular poultry, rice and sugar, which are the commodities with highest transfers as percentage of gross farm receipts in (OECD, 2013). The focus of policy analysis in this report is rice because of its major importance as a staple in Indonesia, being the first source of both energy and proteins (Figure 10). Other agricultural commodities are not subjected to such support policies because there are few barriers to trade (soybeans), or because it is hardly tradable (tubers). Market price support is sustained with restrictive trade measures. The costs of these measures have the form of efficiency losses and costs for consumers, which are to a great extent invisible because they are not reflected in budgetary costs. Alternative trade policy approaches would focus on building freer and more trustable access to regional and global trade. But trade restrictive measures are subject to strong political economy pressures that make them difficult to reform. 28

29 Box 3. Food security legal and institutional framework in Indonesia On 18 October 2012, Indonesia s House of Representatives passed a new Food Law No 18/2012. The law replaced the previous one voted in The Food Law contains provisions for staple foods establishing that exports will be implemented only after fulfilling National Food Reserve and domestic food consumption needs, and imports will be implemented only if domestic food production plus National food reserves are not sufficient according to the government. Other aspects of the Food Law related to the priority to domestic production and consumption of staples is a reflection of what was already the practice. Since 2004 the government limits imports and exports of rice according to an import and export license system. The current strategic plan already established production targets for 39 products. For five food commodities (rice, corn, soybean, sugar and beef) the targeted levels intend to achieve self-sufficiency to meet forecasted consumption. But no commodity is specifically mentioned in the food law and no specific self-sufficiency objective is stated, only the general principle of kemandirian pangan (self-reliance). In terms of food affordability the government is called to stabilize supply and prices of staples, and to provide and distribute staple food to the poor. The law also fixes the objective of improving nutrition and diversifying consumption and deals with food safety and labelling. The Law also mandates the creation of a new food-security government institution reporting to the president and with the task to execute government s orders to production, procurement, storing and/or distribution of staple food. The current Food Security Agency under the Ministry of Agriculture is the predecessor of this future super agency. It currently follows up food availability and balances, in particular for the five self-sufficiency commodities; it builds early warning systems; 17 develops systems for rice distribution and food reserves; it analyses prices, looking at volatility of producer prices and to increases in consumer prices; it promotes food diversification and Food safety. On 9 July 2013 the Indonesian House of representatives passed a new Farmer s Empowerment Law 19/2013. This law is based on the same principles of food sovereignty (kedaulatan pangan) and food self-reliance or independence (kemandirian pangan), but it has no reference to food security. This law has a chapter on farmers empowerment focused on education and training, technical assistance, marketing facilities, small farmers land tenure and access to science and technology. The chapter on farmers protection allows the use of import tariffs that adjust to favor prices for farmers. In terms of risk management the Farmers Empowerment Law has several provisions, including the establishment of an early Warning System. It allows the central and local government to provide compensation after a crop failure, and obligates the government to assign state or local own enterprises to implement agricultural insurance. The Indonesian rice reserves are required by the Food Law. There is an objective of reaching 10 million tons of reserves in 2015, including public and private. There are four levels of public reserves: national food reserve, province government reserve, district government reserve and community food reserve. The Food security Agency develops guidelines for the management of district, province and village reserves. Most of the public reserves are managed by Bulog and used for two objectives: price stabilization and disasters. There are also several sources of private reserves including farm households, industry and traders. According to AMIS, rice reserves in Indonesia at the end of the marketing year have been steadily increasing in the last years, more than tripling from 2 million tons in 2005/06 to 6.3 million tons in 2013/14. According to joint estimations from the Food Security Agency, BPS and the survey conducted by Sucoffindo (2010), in March 2011 about 20-25% of the reserves were from Bulog, another 20-25% from farm households and the rest from the industry and traders. The Coordinating Ministry for economy leads periodical meetings with other relevant ministers and decides on reserve interventions by Bulog. The East Asia Emergency Rice Reserve (currently ASEAN Plus Three Emergency Rice Reserves APTERR) is a system of humanitarian coordination based on national stocks of member countries. Each country earmarks a given amount of rice and cash reserves to be used in case of disasters. It is not thought to substitute trade. The amount earmarked in Indonesia is very modest. Most earmarked reserves are from China, Japan and South Korea. This reserve has been used in very rare occasions and its use is limited only to natural disasters. 53. Rice price support is the result of a combination of a large set of trade and domestic measures, most of them managed by the public logistic agency Bulog: trade restrictions, interventions in the rice market and management of a reserve system. These three elements are deeply interlinked. Imports and exports of rice are subjected to strict quantitative controls. A price band (guaranteed floor price called purchasing price for producers and a ceiling price for consumers) is used to decide the public procurement or release of rice from the national reserve managed by Bulog. The Food Security Agency follows the prices and interventions on food reserves are called if the producers price gets out of the +/- 15% band of the government purchasing price for two weeks in a row. The Food security agency can send a 17. For example, work with the WFP on the Food security and Vulnerability Atlas 2009 and

30 recommendation to buy or sell reserves of rice. The recommendation is then approved by the Minister coordinator for the economy in coordination with other relevant ministers, and then executed by BULOG. The only commodity on which there is a proper reserves system is rice. 54. The result of this combination of rice price support measures is a domestic price that is systematically higher than the world price since This harms access to food by the poor because most of the poor in Indonesia are net buyers of rice. Additionally, domestic price is supposedly stabilised. During a price hike, such as in 2008, exports are not allowed. This impedes the transmission of high international prices to the domestic market, but it also exacerbates the size of the world market spike. If prices are lower than the minimum purchasing price, imports are controlled or banned. This isolates Indonesia from the world market and impedes the development of economic opportunities and commercial relationships that can be crucial for the quick delivery of imports in the case of an urgent need, for instance after a disaster such as those described in scenarios IV (crop failure due to a plant infestation) and V (earthquake). 55. Fertiliser subsidies are a major agricultural policy in Indonesia. The subsidies are provided to fertiliser companies which are state owned enterprises (SOE) required to sell their products below a Highest Retail Price to small farmers producing on less than 2 hectares. This indirect support only partially reaches farmers due to potential rents and inefficiencies of SOEs. Cheaper fertilisers should contribute to more production and lower consumer prices, but the savings are unlikely to be transmitted to consumers due to the existence of rice minimum purchasing prices. These subsidies are also supposed to increase the economic returns of farm households, but are not targeted to the poor There are other agricultural programmes that can contribute to food security. This is the case of ad hoc payments after natural disaster and subsidised insurance. However, the size of these programmes is currently small and are thus not analysed in this preliminary exercise. 57. Raskin is a domestic food aid programme that delivers rice at subsidised prices, prioritising poor or near-poor households. Raskin is the largest assistance programme to the poor in Indonesia. It is fully managed by Bulog and it provides relatively small amounts of no more than 10 kg/household per month, which represents less that 10% of the average household consumption. Raskin rice accounts for about 8% of total annual rice sales in Indonesia. Raskin is thought of as a consumer subsidy that could compensate for the negative impacts of price support on access to food by the poor. Bulog would normally get the Raskin rice from domestic public procurement and from imports and, in principle, National reserves cannot be used for Raskin. Roughly 50% of the entire Indonesian population buys Raskin rice at least once a year. The SUSENAS database contains detailed information about the households that benefit from Raskin. The same agency Bulog manages both the market price support measures and Raskin, which creates a strong inter-linkage between these measures. An alternative way of providing domestic food aid through food vouchers is also analysed. It has several advantages: it allows households to freely spend the subsidy on a diversified set of food items; and it breaks the inter-linkage between rice support and domestic food aid. 58. The most important Indonesian social assistance programme after Raskin is the Bantuan Langsung Tunai (BLT), which is one of the largest targeted cash transfer programmes in the developing world. It was first introduced as a compensation for the poor for reductions in the fuel subsidies in 2005 and was used for the same reason in 2008 and It provides transfers which change from one year to next and is normally identified as an Unconditional Cash Transfers (UCT). According to the World Bank (2012), BLT provides temporary protection for poor households with a very lean administrative apparatus, but it is not appropriate for the long-term poverty reduction goal. For this, conditional cash transfers, health 18. OECD (2014b) provides more description and quantitative impacts of fertilizer subsidy on agricultural production and income in Indonesia. 30

31 insurance and school scholarships are considered better instruments. In this sense, the BLT programme is considered from the perspective of alleviating transitory food insecurity. 59. The transfers generated by existing Indonesian policies in recent years are presented in Table 11. The amount of price support to rice is very variable depending on the international price of rice: it was negative during the price spike of , benefiting consumers rather than producers; and it became positive again from 2010 onwards, taxing consumers with higher prices. Unconditional cash transfers under BLT program was conceived as an emergency income support and, therefore, it also has a very variable budget that adjusts to circumstances. Raskin and fertilizer subsidies have had a steadily growing budget, with a very similar average expenditure in the last eight years. Table11. Selected policy transfers in Indonesia Billion Rupiah Avergae Rice Market Price Support* -5,927 24,180 34,329-68,679-27,484 71,126 90, ,055 32,259 RASKIN** 6,357 5,320 6,584 12,096 12,987 13,925 15,267 20,926 11,683 Fertilizers Subsidies** 2,527 3,166 6,261 15,182 18,533 18,412 16,345 13,959 11,798 Unconditional Cash Transfers BLT*** 4,487 18, ,966 3,844 n.a. n.a. n.a. 8,183 * Estimated transfers to producers according to the PSE methodology. Source: OECD PSE Database 2013 ** Budgetary transfers to consumers. Source: OECD PSE Database 2013 *** Budgetary transfers to households. Source: World Bank (2012b) 60. Agricultural insurance is another instrument oriented to help farms and farm-households to manage their risk. Insurance programmes that cover significant parts of the agricultural sectors are typically highly subsidized and can have a large budgetary cost. There is not a large government agricultural insurance programme in Indonesia. There are only two pilot insurance projects in the rice and cattle sectors since A stylized rice insurance programme for systemic catastrophic risks is simulated for rice. 61. The risk of food insecurity implies a financial risk for the government. Food production shocks, natural disasters and import price hikes can have a negative impact on the balance of payments and foreign currency reserves, and they can worsen the ability of the country to import food and implement agricultural and social policy measures that respond to risky scenarios. Market based mechanisms can help to manage this Government risk and they have been used by governments in other countries. Weather index insurance focuses on production risk associated to weather conditions and has already been used by the government of Malawi in Call option contracts in the futures markets to lock a maximum price of maize have been used iby the government of Mexico. Physical Over the Counter (OTC) instruments could also be used (FAO et Al 2011). All these market tools for government risk have potential to complement other policies and ensure their financial viability across all scenarios. Specific scenarios of difficulties in the balance of payments causing food insecurity were not selected in the risk assessment consultation process. In order to facilitate a more stable access to imports and in the context of the recent deterioration in Indonesian external balance (OECD 2014a) it may advisable to explore this type of maket tools for the management of Government risk. The impact of selected policies in the reference scenario 62. The impact of rice price support is estimated to depend on the assumptions about price transmission and degree of adjustment in food demand (Box 3). The impact on the prevalence of undernourishment is estimated to be at least 2 percentage points and could be up to 21 percentage points or 31

32 more. 19 In the analysis we are going to use the most conservative number of 2 percentage points of impact on undernourishment. Box 4. Implications of rice trade restrictive measures and price support The government of Indonesia applies a restrictive trade policy to rice OECD (2012). Imports are limited by quotas and exports are controlled by a strict license system and, de facto, banned. The state agency Bulog manages the market through a monopoly on trade, a system of public procurement and reserves and a price band (guaranteed floor price for producers and a ceiling price for consumers). As a consequence of this set of measures the domestic price of rice in Indonesia has been significantly higher than the world price in recent years. In 2010 the price of rice in the world market was 30% lower. The impact of the shocks in different scenarios and of rice price support policy depends on the assumptions about the degree of price transmission between world prices and domestic prices 20 and on the degree of adjustment in the diet through the demand system. Table 12 shows the results of the simulations using the estimated demand system. Price support increases the prevalence of undernourishment in the reference scenario which has the highest probability. The increase in undernourishment can be very large. For instance, with full price transmission, prevalence of undernourishment could increase by 22 percentage points up to 33%, due to the higher domestic prices derived from the price support and related trade measures. The increase in undernourishment is likely to be in the interval (2 22 percentage points). The table also presents the impact on the depth of food deficit in Indonesia. These higher prices make access to food by the poor more difficult. Table 12. The reference scenario without a shock: estimated impacts of rice price support policies Rate of undernourishment Median calorie intake Depth of Food deficit (percentage) (kcal per day per capita) (kcal) Estimated impacts of price support measures Without price support Partial price trasmission Price support with full diet adjustment Price support with no diet adjustment Full price trasmission Price support with full diet adjustment Price support with no diet adjustment As a response to higher rice prices, households change their food expenditure and diet. According to the median elasticities, there is a reduction of 5% in the consumption of rice and an increase in consumption of other food items such as dairy and vegetables. Rice price support also reduces expenditure in non-food items. The nutritional consequences of these changes for the intake of different nutrients deserve further research. Price support measures have very negative impacts on food security in the base case situation of no particular shock. These negative impacts prevail across several scenarios, in particular all climate or disaster scenarios that reduce domestic yields and production, and all macroeconomic shocks that have negative broad effects on economic growth. Restrictions on imports increase undernourishment in any domestic food shortage scenario and strongly damage food security. 63. The degree of targeting of Raskin and BLT is often put into question (World Bank, 2012a and 2012b). The degree of targeting depends on the decisions made at the community level on the selection of eligible households (OECD 2014). It is often the case that once a community receives support from the government that this support is distributed equally among the households in the community, following an egalitarian social norm. The SUSENAS database includes information on the degree of targeting. Among 19. The highest level in the interval is 21. It was estimated by introducing market price support that increases the prevalence of undernourishment. It is likely that for the elimination of price support the reductions of undernourishment would be proportionally smaller. 20. The partial price transmission scenario is simulated according to the results of the general equilibrium model for Indonesia known as INDONESIA E-3 (Warr and Yusuff, 2013). 32

33 the poor, 70% of the households receive Raskin (Figure 11). But interpreting this figure as good targeting is misleading because more than 40% of the non-poor also receive Raskin and there are recipients of the programme even in the highest expenditure decile. Targeting to the undernourished is more difficult because the energy intake is less observable than income and consumption. This is why the proportion of undernourished that received Raskin is 50%, only marginally higher than the proportion of nonundernourished. The same result applies to the quantity of Raskin rice that is only significantly larger for the poor households, but not for the undernourished. Figure 11. The recipients of Raskin (SUSENAS database) 64. In practical terms, the data from SUSENAS shows that the relative percentage of population benefiting from BLT across expenditure deciles follows a similar profile, even if Raskin benefits a larger number of households in all income classes. The demand system for all households in the SUSENAS sample is used to simulate the impacts of these social programmes on poverty and undernourishment, assuming the same budgetary outlays of BLT as Raskin (Figure 12) 21. An alternative domestic food aid programme that distributes vouchers to the recipients of Raskin is also simulated. The value of the vouchers per household is equal to the value of the rice subsidy, but the recipient has the right to spend it in other staple food items such as grains, tubers, oils and fats. The transactions costs of this voucher programme would normally be smaller than those of Raskin, further increasing its impact on reducing undernourishment If Raskin were eliminated the percentage of poor increases only by a few decimals of a point, while the prevalence of undernourishment is increased by more than 1.2 percentage point (Figure 12). Raskin is directly subsidising rice consumption and having a particular impact on calorie intake so that it is better targeted to undernourishment than to poverty. The opposite is true for the unconditional cash transfers BLT that reduce poverty in 2.1 percentage points, but undernourishment by less than 1. The possibility of better targeting BLT is also simulated by providing the transfer only to those in the lowest two expenditure deciles. The improved results of more targeted BLT indicate a scope to improve the targeting of the programme, setting aside its feasibility in practice. Finally Food vouchers that can be spent on all staple are more effective than Raskin in reducing undernourishment because voucher can be spent in 21. Per household cash transfer was calculated so that total budget including 5.4% programme administrative cost equals 2010 budget outlay of Raskin programme. 22. In the simulations in this paper the transfer efficiency of Raskin and food vouchers is assumed to be the same. The estimated gains on reducing undernourishment are just due to the possibility of using the subsidy for food items with higher demand elastic than rice. 33