## sipea2025028

## Source details

**Canonical URL:** [sipea2025028](https://www.imf.org/-/media/files/publications/selected-issues-papers/2025/english/sipea2025028.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2025/english/sipea2025028.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2025/english/sipea2025028.pdf.json)

---

### A. Context and key statistics on Madagascar’s rice sector
- National self-sufficiency targets:
  - Produce 6 million tons of rice in 2024.
  - Reach an annual rice production of 11 million tons by 2030.
  - Presidential priority for full rice self-sufficiency by 2027 (SNDR III).
- Consumption and production:
  - Average Malagasy individual consumes 153.5 kilograms of rice annually.
  - Annual rice production increased by 76 percent since 2002, reaching 4.6 million tons in 2022.
  - Population grows at an annual average rate of 3 percent; rice production has increased by an average of 2 percent per year since the 1960s.
  - Rice paddy accounts for 24 percent of total agricultural gross production value, totaling $1,006,030 in 2022.
- Trade and vulnerability:
  - India and Pakistan accounted for 98 percent of total imported rice to Madagascar in 2022.
  - Imported rice prices have remained lower than local white rice varieties since 2011.
- Policy- and fiscal-related figures:
  - VAT exemption (previously at 18 percent) on both imported and locally produced rice in place since July 2008.
  - Estimated fiscal cost of VAT exemptions on local and imported rice: around 1.3 percent of GDP (1.1 percent from local rice and 0.2 percent from imported rice).
  - Exemption of customs duties on imported rice since 2005 estimated to cost another 0.2 percent of GDP annually.
- Farming practices and constraints:
  - System of Rice Intensification (SRI) introduced in Madagascar in the 1980s; adoption remains limited due to perceptions of higher labor/time requirements, lack of tools and organic fertilizer, inadequate access to fast-growing seeds, and lack of technical knowhow.
  - Low mechanization and limited fertilizer use prevail among smaller-scale rice plantations.

### B. Empirical strategy and events exploited
- Identified policy events used as quasi-natural experiments:
  - 2005 one-off imposition of tariffs on imported rice: average effectively applied rate increased from 0 to 17.5 percent in 2005 (weighted by partner import shares).
  - 2008 VAT exemption on imported and locally produced rice (VAT previously at 18 percent), implemented since July 2008.
- Data sources:
  - Time series of local rice production and consumption from FAOSTAT, USDA Foreign Agricultural Service, Rice Observatory, authorities' data, and IMF staff calculations.
- Estimation approach:
  - Local projections (LP) models used to estimate impulse responses of local rice production and local rice consumption to the respective price shocks.

### C. Estimated elasticities and dynamic responses
- Response to the 2005 tariff increase (local production):
  - Table 2: Percent change in local production from 2005
    - 2005: 0
    - 2006: 2.8%
    - 2007: 5.9%
    - 2008: 15.4%
    - 2009: 33.9%
    - 2010: 39.7%
  - Price elasticity of supply (PES):
    - Estimated around 2 at a 4-to-5-year horizon: "a 1 percent increase in rice price would lead to an increase in local rice production of around 2 percent."
  - Short-term dynamics:
    - Local production relatively inelastic in the short-term following the tariff shock; gains materialize over several years.
- Response to the 2008 VAT exemption (consumption):
  - Price elasticity of demand (PED):
    - Estimated close to zero; local rice consumption is inelastic to rice price given rice is the main staple.
  - Rationale for using VAT exemption:
    - VAT exemption affects prices of both imported and domestically produced rice, providing a better indicator of overall consumer price changes.

### D. Methodology details (Box 1)
- Estimation approach: dynamic multipliers estimated using Local Projections (LP).
- LP specification (for horizon h = 0,1,2,..., H−1):
  - y_{t+h} = ∅_h x_t + z_t δ + u_{t+h}
  - Coefficient of interest: dynamic multiplier coefficient ∅_h.
- Reporting and sample:
  - Responses: ln_production
  - Impulses: tariff_shock
  - Number of controls = 0
  - Number of responses = 1
  - Number of impulses = 2
  - Sample: 1962 thru 2018
  - Number of obs = 57
- Interpretation guidance:
  - Dynamic multipliers ∅_h are presented for each horizon following the shock; these are the IRF coefficients for the exogenous variables.
  - Tables present coefficient, Std. err., z, P>|z|, and [95% conf. interval] for each reported horizon and variable.

### E. Policy implications and recommended approaches
- Tradeoffs from removing exemptions or raising tariffs:
  - Removing customs duty exemptions or raising tariffs may incentivize local production over the medium term (via higher producer prices and PES ≈ 2 at 4–5 years).
  - Immediate impact of higher rice prices would lower household welfare because demand is inelastic (PED ≈ 0).
- Suggested targeted policy options:
  - Consider removing VAT exemptions on high-quality, specialty rice consumed by affluent households while maintaining exemptions for low-quality broken rice consumed by poorer households to:
    - Raise competitiveness of locally produced non-broken rice.
    - Shield low-income households from higher prices.
  - Caveats:
    - Enforcing tax-regime distinctions between high- and low-quality rice could be challenging and would require stronger controls to limit tax evasion and misclassification.
- Broader VAT reform on agriculture-related products:
  - Reconsider VAT exemptions on imported rice to free fiscal space.
  - Use gained fiscal room for targeted VAT exemptions on productivity-enhancing agricultural inputs and mechanization.
  - Target VAT exemptions on farming inputs to farmers with revenues below a given threshold to better incentivize productivity-boosting investments.
- Complementary supply-side measures:
  - Promote sustainable agricultural practices (e.g., SRI, permaculture) and climate-smart agricultural practices.
  - Invest in infrastructure, irrigation, access to quality seeds, fertilizers, machinery, and agricultural extension services.
  - Strengthen coordination among government agencies and reduce bureaucratic delays to improve implementation of the SNDRIII strategy.

### F. Constraints, risks, and research priorities
- Main constraints and risks:
  - Poor transport infrastructure severely limits market access of rice producers, disincentivizing investments to raise productivity.
  - Natural disaster risks (cyclones, drought, locust infestations) would still pose significant risks even with productivity improvements.
  - Environmental risks from hybrid rice adoption: elevated seed costs, greater need for chemical fertilizers, and potential erosion of resilience from monoculture.
- Research, piloting, and safeguards before scaling interventions:
  - Further studies on the efficacy and unintended effects of hybrid rice adoption are needed.
  - Ongoing/planned field comparisons:
    - At the start of 2025, the World Bank, in collaboration with the FAO, the Rice+ Project will install field schools in the Alaotra Mangoro Region to compare the performance of the hybrid rice with that of high-yielding local varieties, at various levels of organic and chemical fertilizations. Farmers in the region will then be invited to these field schools to choose the rice variety and the cultivation method best suited to their farms.
  - Policy design recommendation:
    - The rice self-sufficiency goal should be balanced against macro-economic, fiscal, and environmental sustainability and should avoid siloed policy-making.

*Source: IMF staff paper "RICE PRODUCTION IN MADAGASCAR—CHALLENGES TO SELF-SUFFICIENCY" (selected issues chapter).*

### 1. Estimating PES and PED using Local Projections (LP) _______________________________ 11

### 1. Estimating PES and PED using Local Projections (LP)

### A. Context and key statistics on Madagascar’s rice sector
- National self-sufficiency targets:
  - Produce 6 million tons of rice in 2024.
  - Reach an annual rice production of 11 million tons by 2030.
  - Presidential priority for full rice self-sufficiency by 2027 (SNDR III).
- Consumption and production:
  - Average Malagasy individual consumes 153.5 kilograms of rice annually.
  - Annual rice production increased by 76 percent since 2002, reaching 4.6 million tons in 2022.
  - Population grows at an annual average rate of 3 percent; rice production has increased by an average of 2 percent per year since the 1960s.
  - Rice paddy accounts for 24 percent of total agricultural gross production value, totaling $1,006,030 in 2022.
- Trade and vulnerability:
  - India and Pakistan accounted for 98 percent of total imported rice to Madagascar in 2022.
  - Imported rice prices have remained lower than local white rice varieties since 2011.
- Policy- and fiscal-related figures:
  - VAT exemption (previously at 18 percent) on both imported and locally produced rice in place since July 2008.
  - Estimated fiscal cost of VAT exemptions on local and imported rice: around 1.3 percent of GDP (1.1 percent from local rice and 0.2 percent from imported rice).
  - Exemption of customs duties on imported rice since 2005 estimated to cost another 0.2 percent of GDP annually.
- Farming practices and constraints:
  - System of Rice Intensification (SRI) introduced in Madagascar in the 1980s; adoption remains limited due to perceptions of higher labor/time requirements, lack of tools and organic fertilizer, inadequate access to fast-growing seeds, and lack of technical knowhow.
  - Low mechanization and limited fertilizer use prevail among smaller-scale rice plantations.

### B. Empirical strategy and events exploited
- Identified policy events used as quasi-natural experiments:
  - 2005 one-off imposition of tariffs on imported rice: average effectively applied rate increased from 0 to 17.5 percent in 2005 (weighted by partner import shares).
  - 2008 VAT exemption on imported and locally produced rice (VAT previously at 18 percent), implemented since July 2008.
- Data sources:
  - Time series of local rice production and consumption from FAOSTAT, USDA Foreign Agricultural Service, Rice Observatory, authorities' data, and IMF staff calculations.
- Estimation approach:
  - Local projections (LP) models used to estimate impulse responses of local rice production and local rice consumption to the respective price shocks.

### C. Estimated elasticities and dynamic responses
- Response to the 2005 tariff increase (local production):
  - Table 2: Percent change in local production from 2005
    - 2005: 0
    - 2006: 2.8%
    - 2007: 5.9%
    - 2008: 15.4%
    - 2009: 33.9%
    - 2010: 39.7%
  - Price elasticity of supply (PES):
    - Estimated around 2 at a 4-to-5-year horizon: "a 1 percent increase in rice price would lead to an increase in local rice production of around 2 percent."
  - Short-term dynamics:
    - Local production relatively inelastic in the short-term following the tariff shock; gains materialize over several years.
- Response to the 2008 VAT exemption (consumption):
  - Price elasticity of demand (PED):
    - Estimated close to zero; local rice consumption is inelastic to rice price given rice is the main staple.
  - Rationale for using VAT exemption:
    - VAT exemption affects prices of both imported and domestically produced rice, providing a better indicator of overall consumer price changes.

### D. Policy implications and recommended approaches
- Tradeoffs from removing exemptions or raising tariffs:
  - Removing customs duty exemptions or raising tariffs may incentivize local production over the medium term (via higher producer prices and PES ≈ 2 at 4–5 years).
  - Immediate impact of higher rice prices would lower household welfare because demand is inelastic (PED ≈ 0).
- Suggested targeted policy options:
  - Consider removing VAT exemptions on high-quality, specialty rice consumed by affluent households while maintaining exemptions for low-quality broken rice consumed by poorer households to:
    - Raise competitiveness of locally produced non-broken rice.
    - Shield low-income households from higher prices.
  - Caveats:
    - Enforcing tax-regime distinctions between high- and low-quality rice could be challenging and would require stronger controls to limit tax evasion and misclassification.
- Broader VAT reform on agriculture-related products:
  - Reconsider VAT exemptions on imported rice to free fiscal space.
  - Use gained fiscal room for targeted VAT exemptions on productivity-enhancing agricultural inputs and mechanization.
  - Target VAT exemptions on farming inputs to farmers with revenues below a given threshold to better incentivize productivity-boosting investments.
- Complementary supply-side measures:
  - Promote sustainable agricultural practices (e.g., SRI, permaculture) and climate-smart agricultural practices.
  - Invest in infrastructure, irrigation, access to quality seeds, fertilizers, machinery, and agricultural extension services.
  - Strengthen coordination among government agencies and reduce bureaucratic delays to improve implementation of the SNDRIII strategy.

*Source: IMF staff paper "RICE PRODUCTION IN MADAGASCAR—CHALLENGES TO SELF-SUFFICIENCY" (selected issues chapter).*

### Box 1. Estimating PES and PED using Local Projections (LP)

### Box 1. Estimating PES and PED using Local Projections (LP)

### Methodology
- Estimation approach: dynamic multipliers estimated using Local Projections (LP).
- Response variable: local rice production or consumption.
- Exogenous variable (shock): tariff or VAT shocks.
- LP specification (for horizon h = 0,1,2,..., H−1):
  - y_{t+h} = ∅_h x_t + z_t δ + u_{t+h}
  - Coefficient of interest: dynamic multiplier coefficient ∅_h.
- Additional controls z_t include further lags of the endogenous variables.
- The estimation tables report dynamic multipliers at each horizon following the event.
- Note: IRF coefficients for exogenous variables are referred to as dynamic multipliers.

### Estimation details and data reported
- Responses: ln_production
- Impulses: tariff_shock
- Number of controls = 0
- Number of responses = 1
- Number of impulses = 2
- Sample: 1962 thru 2018
- Number of obs = 57
- Labeling in tables: "Local-projection impulse–responses"
- The box explicitly identifies the coefficient ∅_h as the dynamic multiplier and notes that additional controls z_t are included (e.g., further lags of the endogenous variables).

### Reporting and interpretation guidance
- Dynamic multipliers ∅_h are presented for each horizon following the shock; these are the IRF coefficients for the exogenous variables.
- The local-projection framework treats the shock x_t as exogenous and tracks the response y_{t+h} across horizons h = 0,1,2,...,H−1.
- Tables present coefficient, Std. err., z, P>|z|, and [95% conf. interval] for each reported horizon and variable (labels preserved from source).
- The source notes that the estimation tables below give the dynamic multipliers at each horizon following the event (tables contain detailed numeric coefficient estimates for multiple horizons and for ln_production responses to tariff_shock impulses).

*Source: IMF staff estimates using USDA IPAD data, WITS*

### 23.  While rice production in Madagascar has lagged its population needs over the past

### sipea2025028 - 23.  While rice production in Madagascar has lagged its population needs over the past

### Findings on local rice supply and prices
- Local rice supply displays positive price elasticity, rising significantly when rice price increases, albeit with several years lag.
- Rice demand in Madagascar is inelastic; policies that raise rice prices may erode consumer welfare.
- Rice yields in Madagascar are far below their potential and under the right input conditions, could be substantially improved and allow the country to meet its self-sufficiency goals.
- The country aims to reach rice self-sufficiency by 2027.

### Constraints and risks to self-sufficiency
- Poor market access and storage:
  - Poor transport infrastructure severely limits market access of rice producers, disincentivizing investments to raise productivity and preventing producers from meeting rice demand.
  - Improvements to the transport and storage of rice buffer stocks (procured locally or abroad) would help Madagascar tackle these challenges.
- Natural disaster risks:
  - Idiosyncratic natural disaster shocks including cyclones, drought, and locust infestations would still pose significant risks even with productivity improvements.
- Environmental and socio-economic risks of proposed measures:
  - Planting of hybrid rice seeds involves elevated cost of procuring hybrid rice seeds and a greater need for chemical fertilizers.
  - The emergence of a rice monoculture could erode the resilience of Madagascar’s rice production.
  - The goal of rice self-sufficiency should be balanced against macro-economic, fiscal, and environmental sustainability.

### Policy implications and recommendations
- Caution on price-raising demand-side policies:
  - Policies resulting in higher rice prices, including import tariffs or VAT increases, must carefully weigh their impact on local production and local food prices because of rice demand inelasticity.
- Prioritize supply-side productivity-enhancing policies:
  - Supply-side policies aimed at increasing the productivity of local rice farmers could raise local rice production without a spike in rice prices.
  - Where possible, adopt CSA practices that concurrently improve rice yields, climate resilience and limit the negative environmental impact of rice cultivation.
- Infrastructure and buffer management:
  - Invest in transport infrastructure and storage to improve market access and reduce vulnerability to shocks.
  - Improve transport and storage capacity for rice buffer stocks procured locally or abroad.

### Research, piloting, and safeguards before scaling interventions
- Hybrid rice adoption:
  - Further studies on the efficacy and unintended effects of hybrid rice adoption are needed.
  - More research should be conducted before the use of hybrid rice seeds is scaled up, given their elevated cost, greater need for chemical fertilizers, and the lack of studies comparing hybrid seeds to local rice varieties under comparable input conditions.
- Ongoing and planned field comparisons:
  - At the start of 2025, the World Bank, in collaboration with the FAO, the Rice+ Project will install field schools in the Alaotra Mangoro Region to compare the performance of the hybrid rice with that of high-yielding local varieties, at various levels of organic and chemical fertilizations. Farmers in the region will then be invited to these field schools to choose the rice variety and the cultivation method best suited to their farms.
- Avoid siloed policy-making:
  - The rice self-sufficiency goal should not be viewed in silos but instead be balanced against other considerations, including macro-economic, fiscal, and environmental sustainability.

*sipea2025028 - 23.  While rice production in Madagascar has lagged its population needs over the past*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2025/english/sipea2025028.pdf_