## _wp1157

## Source details

**Canonical URL:** [_wp1157](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1157.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1157.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1157.pdf.json)

---

### Introduction and objectives
- Inflation in Chad reached 8–10 percent over 2008 and 2009, while it averaged 3½ percent in neighboring members of the Central African Economic and Monetary Community (CEMAC).
- Reducing the inflation gap is important to comply with the CEMAC inflation convergence criterion of 3 percent or less and to protect the purchasing power of the poor.
- This paper provides the first empirical analysis of inflation determinants focusing on Chad using quarterly data from 1983:Q1 to 2009:Q3.
- The analysis is based on monetary and external theories of inflation and accounts for supply-side constraints (captured by agricultural shocks), institutional factors (episodes of political instability), the 1994 devaluation of the CFAF, and the start of oil production in 2003.

### Empirical strategy and methodology
- Estimation methods combined:
  - Vector error correction model (VECM).
  - Single-equation model, augmented with error correction terms from the VECM.
  - Structural vector autoregression (SVAR).
- Comparative evidence and prior studies cited: Diouf (2007), Loening, Durevall, and Birru (2009), Moriyama (2008), Klein and Kyei (2009), Kandil and Morsy (2009), Coleman (2010).

### Theoretical framework
- Three candidate theories considered (following Ubide, 1997): (i) external sector theory; (ii) monetary theory; (iii) markup theory.
  - External sector and monetary theories are pertinent to Chad; markup theory is less relevant given labor market characteristics.
- IS-LM small open economy formulation:
  - General price level is weighted average of tradable and nontradable prices with 0 < δ < 1 representing tradable share.
  - Tradable prices follow purchasing power parity and depend on foreign prices and the exchange rate.
  - Nontradable prices determined by money market equilibrium where real money demand depends on real income, inflation expectations, and rate of return (BEAC discount rate minus annual time deposit rate).
- Reduced-form expectations: depreciation, higher foreign prices, higher money supply, and higher real income expected to increase inflation; higher rate of return expected to reduce inflation via lower money holdings.
- Empirical model augmented with structural factors specific to Chad: agricultural production shocks (national rainfall), political instability (rebel assaults or major social tensions), CFAF devaluation, and start of oil production.

### Data and empirical model specification
- Sample: quarterly data from 1983:Q1 to 2009:Q3.
- Quarterly real GDP and quarterly government expenditures interpolated from annual data using proportional Denton method.
- Quarterly rainfall: averages of monthly DREM data for 2000–08; FAO annual rainfall for 1983–1999; seasonal pattern from 2000–08 applied to derive 1983–1999 quarterly rainfall.
- Political instability variable constructed from multiple sources measuring rebel assaults.
- Time series properties: Augmented Dickey-Fuller and Phillips-Perron tests show all variables nonstationary in level but stationary in first difference.
- Single-equation inflation specification (Equation 8) models ∆(log CPI) as function of lags of ∆(NEER), ∆(foreign prices), ∆(real money), ∆(real GDP), interest rates, rainfall (with lags), event dummies, quarterly dummies, and error correction terms. General-to-specific approach led to a parsimonious specification that excludes real income as main determinant but includes public expenditure.

### Long-run equilibrium estimates (VECM)
- Money market VECM (Table 1) highlights:
  - Log(Real GDP) coefficients 3.39 and 5.13 (t-statistics (1.83) and (2.23)).
  - Discount minus deposit rate coefficient -0.49 (t-statistic (-3.93)).
  - Error correction: -0.10 (t-statistic (-2.41)) and -0.05 (t-statistic (-1.52)).
  - Observations: 114.
- External market VECM (Table 2) highlights:
  - Log(NEER) coefficient 1.18 (t-statistic (1.60)).
  - Log(Foreign price) coefficient 1.46 (t-statistic (3.02)).
  - Error correction: -0.10 (t-statistic (-2.92)).
  - Observations: 114.
- Johansen tests do not reject one cointegrating vector in both money and external market systems.

### Single-equation inflation model — main determinants (Table 3)
- ∆(foreign price) t-1:
  - Coefficients 0.183, 0.198, 0.209 (robust t-statistics (1.69)*, (1.82)*, (1.93)*).
  - Interpretation in source: a 10 percent increase in foreign prices increases domestic inflation by 2 percent one quarter later.
- ∆(NEER) t-1:
  - Coefficients -0.252, -0.262, -0.269 (robust t-statistics (4.05)***, (4.17)***, (4.30)***).
  - Interpretation in source: a 10 percent appreciation (depreciation) of the NEER decreases (increases) inflation by around 2.6 percent one quarter later.
- ∆(Public exp.) t-1:
  - Coefficients 0.093, 0.093, 0.095 (robust t-statistics (2.10)**, (2.03)**, (2.09)**).
  - Interpretation in source: a 10 percent increase in public spending during one quarter increases inflation by 1 percent the following quarter.
- Rainfall t-3:
  - Coefficients -0.036, -0.033, -0.035 (robust t-statistics (2.82)***, (2.58)**, (2.70)***).
  - Interpretation in source: a 10 percent increase in rainfall reduces inflation by 0.35 percent three quarters later.
- Error correction (external market) EC_ext:
  - Coefficients -0.281, -0.266, -0.291 (robust t-statistics (2.40)**, (2.25)**, (2.44)**).
  - Interpretation in source: a 10 percent increase in external market disequilibrium reduces inflation by 3 percent with a one-quarter lag.
- Seasonality and devaluation:
  - Quarter dummies: Quarter 1 coefficients ~0.19–0.20; Quarter 2 coefficients ~0.26–0.28; Quarter 3 coefficients ~0.178–0.188; Quarter 4 omitted (reference). Many quarter coefficients significant at 1%.
  - Devaluation dummy: 0.018 (t-statistic (2.07)**) in column (1); 0.013 (t-statistic (1.44)) in specification where dummy captures one year after devaluation.
- Goodness of fit and diagnostics:
  - R-squared: 0.49, 0.47, 0.48 across specifications.
  - Observations: 114.
  - Diagnostic tests (misspecification, normality, heteroskedasticity, ARCH) do not reveal major problems (Table A.6).
- Political instability and oil-era dummies are not significant in all regressions.

### Structural VAR findings on shocks and persistence
- Variables in SVAR system (first differences and expectations): rainfall (Rain), foreign prices, NEER, public spending, domestic prices.
- Identification/ordering:
  - Rainfall treated as exogenous; foreign prices affect variables except rainfall; NEER shocks affect public spending and domestic inflation; public spending shocks affect only domestic inflation disturbances.
- Impulse responses and persistence:
  - Rainfall shock:
    - Reduces inflation; effect persists over six quarters.
    - Maximum impact at three quarters after the shock.
  - Foreign prices shock:
    - Increases inflation; effect persists over six quarters.
    - Maximum impact in the first quarter after the shock.
  - Nominal appreciation shock:
    - Reduces inflation over four quarters.
    - Highest effect at two quarters.
  - Public spending shock:
    - Increases inflation between the first and third quarter.
    - Maximum impact around the second and third quarters.
- SVAR corroborates importance of rainfall, foreign prices, exchange rate, and public spending and highlights persistence (three to six quarters) of these effects.

### Recent macroeconomic context and dynamics
- Oil production since 2003 materially transformed Chad’s macroeconomic outlook by increasing government revenue and enabling a surge in public expenditure.
- Public expenditure as a share of non-oil GDP has doubled since 2003, reaching 46 percent in 2009.
- The non-oil primary balance reached 28 percent of non-oil GDP in 2009, noted as being well above its estimated long-term sustainable level (which is in the low single digits).
- The current account balance improved by more than 50 percent of GDP between 2002 and 2009 due to higher oil export revenue.
- Despite oil revenues, the monetization of the economy remains low and credit to the private sector limited; fragile security conditions impede private sector growth.
- The real exchange rate appreciated over 2008 and 2009; inflation in Chad exceeded that of trading partners, contributing to real appreciation and reduced competitiveness.
- After an initial surge at the start of the oil era, real growth weakened since 2005 primarily due to a decreasing trend in oil production.
  - Real GDP contracted by 1.6 percent in 2009 because of contraction in the oil sector and collapse of agricultural production owing to poor rainfall.
- Credit to the private sector and broad money remain limited.

### Inflation dynamics and historical context
- Historical inflation in Chad:
  - Averaged 4.5 percent from 1983 to 2009, with sizeable year-to-year fluctuations.
  - Moderated since 2000 (text truncated in source).
- Major events and observed inflation responses:
  - Devaluation of the CFAF by 50 percent in January 1994: inflation surged from minus 8 percent in 1993 to plus 42 percent in 1994.
  - Floods in 1999 and 2007 and drought in 2001 were associated with large increases in inflation.
- CPI coverage and composition:
  - CPI covers only N’Djamena and includes 332 products.
  - CPI weights based on a survey of 1,024 households in N’Djamena in 2003–04 (ECOSIT2).
  - Food and nonalcoholic beverages represent 46 percent of the index.
- Recent drivers: the global food crisis of 2008 and a significant rise in food prices chiefly explain inflation developments in recent years.
- Coleman (2010) highlighted that inflation is persistent in Chad, Cote d’Ivoire, and Niger.

### Conclusions and policy recommendations
- Main drivers of inflation in Chad: rainfall (proxy for agricultural production), foreign prices, exchange rate movements, and public spending.
  - Rainfall increases reduce inflation over six quarters with peak effect in third quarter.
  - Foreign prices raise inflation over six quarters with peak impact in first quarter.
  - Nominal appreciation reduces inflation over four quarters with peak effect in second quarter.
  - Public spending raises inflation between first and third quarters with peak around second and third quarters.
- Policy recommendations to mitigate inflationary pressures:
  - Reduce sensitivity to rainfall shocks:
    - Improve efficiency in agriculture, storage, and transportation.
    - Improve food storage facilities to prevent large food price increases after droughts or floods.
    - Enhance agricultural productivity via modern techniques (irrigation, improved seeds) and better access to rural credit.
  - Reduce market segmentation and trade frictions:
    - Liberalize internal and external trade through improved transportation infrastructure such as rural roads.
  - Strengthen fiscal discipline and public financial management:
    - Better control of public expenditures through improved investment planning and procurement practices.
    - Integrate the inflationary effect of public expenditures when designing public spending programs.
- Modeling implication: single-equation model captures main drivers but does not fully capture persistence; SVAR is used to analyze persistence and interrelations among shocks.

### Appendix highlights (selected figures, tables, and diagnostics)
- CPI weights (Table A.1):
  - Food and non alcoholic beverages 46.18
    - Bread and Cereals 17.24
    - Meat 7.85
    - Fish 3.64
    - Milk, Cheese, and Eggs 1.49
    - Oils and Fats 3.74
    - Vegetables and Tubers 3.67
  - Alcoholic beverages and tobacco 3.08
  - Clothing and shoes 10.88
  - Housing, water, gas, electricity, and other fuels 13.31
  - Furniture, houseware, and house maintenance 4.23
  - Health 2.73
  - Transportation 6.36
  - Communications 2.35
  - Leisure and culture 3.55
  - Education 0.96
  - Other goods and services 1.74
  - Restaurants et hotels 4.63
  - Total 100
- Unit root and cointegration tests:
  - In level, Log(CPI) ADF -0.18 (P-value 0.94); in first difference, Log(CPI) ADF -8.99 (P-value 0.00).
  - Johansen cointegration tests for money and external markets: *The null hypothesis that there is one cointegrating equation is not rejected.
- Diagnostic statistics for single-equation inflation models (Specification 1 example):
  - Breusch-Pagan / Cook-Weisberg test for heteroskedasticity: 0.86 (p-value 0.35).
  - Ramsey RESET: 0.56 (p-value 0.64).
  - ARCH (1 lag): 0.027 (p-value 0.87).
  - Notes: p-values reported in source; diagnostics indicate no major misspecification or heteroskedasticity issues.

*Source: _wp1157 (IMF working paper content unit _wp1157 - REFERENCES).*

### References .............................................................................................................

### _wp1157 - References

### Introduction and objectives
- Inflation in Chad reached 8–10 percent over 2008 and 2009, while it averaged 3½ percent in neighboring members of the Central African Economic and Monetary Community (CEMAC).
- Reducing the inflation gap is important to comply with the CEMAC inflation convergence criterion of 3 percent or less and to protect the purchasing power of the poor.
- This paper provides the first empirical analysis of inflation determinants focusing on Chad using quarterly data from 1983:Q1 to 2009:Q3.
- The analysis is based on monetary and external theories of inflation and accounts for supply-side constraints (captured by agricultural shocks), institutional factors (episodes of political instability), the 1994 devaluation of the CFAF, and the start of oil production in 2003.

### Empirical strategy and methodology
- The empirical strategy combines three estimation methods:
  - Vector error correction model (VECM).
  - Single-equation model, augmented with error correction terms from the VECM.
  - Structural vector autoregression (SVAR).
- Comparative evidence and prior studies cited in the literature review include:
  - Diouf (2007) for Mali.
  - Loening, Durevall, and Birru (2009) for Ethiopia.
  - Moriyama (2008) for Sudan.
  - Klein and Kyei (2009) for Angola.
  - Kandil and Morsy (2009) for GCC countries.
  - Coleman (2010) on inflation persistence in 13 CFA member states.

### Key empirical findings
- Single-equation model (with VECM error correction terms) identifies the main drivers of inflation in Chad:
  - Supply-side constraints (agricultural shocks) affect domestic prices with a three-quarter lag via agricultural production channels.
  - Public spending has a significant impact on inflation with a one-quarter lag, operating through demand pressures on nontradables.
  - Changes in trading partner prices and exchange rate movements are passed through to domestic prices one quarter later.
- SVAR results corroborate the importance of:
  - Rainfall, foreign prices, the exchange rate, and public spending as main determinants of inflation.
  - Persistence of effects: the impact of a rainfall shock is highest three quarters after the shock and persists until the sixth quarter; similar persistence (three to six quarters) is found for foreign prices, the exchange rate, and public spending.

### Recent macroeconomic context and dynamics
- Oil production since 2003 materially transformed Chad’s macroeconomic outlook by increasing government revenue and enabling a surge in public expenditure.
- Public expenditure as a share of non-oil GDP has doubled since 2003, reaching 46 percent in 2009.
- The non-oil primary balance reached 28 percent of non-oil GDP in 2009, noted as being well above its estimated long-term sustainable level (which is in the low single digits).
- The current account balance improved by more than 50 percent of GDP between 2002 and 2009 due to higher oil export revenue.
- Despite oil revenues, the monetization of the economy remains low and credit to the private sector limited; fragile security conditions impede private sector growth.
- The real exchange rate appreciated over 2008 and 2009; inflation in Chad exceeded that of trading partners, contributing to real appreciation and reduced competitiveness.
- After an initial surge at the start of the oil era, real growth weakened since 2005 primarily due to a decreasing trend in oil production.
  - Real GDP contracted by 1.6 percent in 2009 because of contraction in the oil sector and collapse of agricultural production owing to poor rainfall.
- Credit to the private sector and broad money remain limited; income velocity (M2) and exchange rate indices are noted in the descriptive figures.

### Inflation profile and persistence
- Historical inflation in Chad:
  - Averaged 4.5 percent from 1983 to 2009, with sizeable year-to-year fluctuations.
  - Moderated since 2000 (text truncated thereafter).
- Coleman (2010) highlighted that inflation is persistent in Chad, Cote d’Ivoire, and Niger.

### Interpretation and policy-relevant implications (inferred from findings)
- Supply-side interventions to mitigate agricultural shocks (timely response to rainfall-related production shocks) could reduce food-price inflation with effects materializing around three quarters after shocks.
- Fiscal policy and public spending management are important for stabilizing inflation, given the one-quarter lagged inflationary impact of public spending through nontradable demand.
- Exchange rate and external price stability matter for domestic inflation; import price pass-through occurs with about a one-quarter lag.
- The persistence of effects for three to six quarters implies that policy responses may need to be sustained for multiple quarters to fully offset shocks.

*Source: _wp1157 - References (source PDF filename provided in content).*

### 3.8 percent, but is still subject to significant shocks. Figure 2 illustrates the inflation

### _wp1157 - 3.8 percent, but is still subject to significant shocks. Figure 2 illustrates the inflation

### Inflation dynamics and historical context
- Major events and observed inflation responses:
  - Devaluation of the CFAF by 50 percent in January 1994: inflation surged from minus 8 percent in 1993 to plus 42 percent in 1994.
  - Floods in 1999 and 2007 and drought in 2001 were associated with large increases in inflation.
- CPI coverage and composition:
  - CPI covers only N’Djamena and includes 332 products.
  - CPI weights based on a survey of 1,024 households in N’Djamena in 2003–04 (ECOSIT2).
  - Food and nonalcoholic beverages represent 46 percent of the index.
- Recent drivers:
  - The global food crisis of 2008 and a significant rise in food prices chiefly explain inflation developments in recent years.

### Theoretical framework used to explain inflation in Chad
- Three candidate theories (following Ubide, 1997): (i) external sector theory; (ii) monetary theory; (iii) markup theory.
  - External sector and monetary theories are pertinent to Chad; the markup theory is less relevant given labor market characteristics.
  - Labor market context: only 9 percent of workers are paid employees; 63 percent are self-employed in agriculture.
- IS-LM small open economy formulation:
  - General price level is a weighted average of tradable and nontradable prices with 0 < δ < 1 representing tradable share.
  - Tradable prices follow purchasing power parity and depend on foreign prices and the exchange rate.
  - Nontradable prices determined by money market equilibrium where real money demand depends on real income, inflation expectations, and rate of return (BEAC discount rate minus annual time deposit rate).
- Reduced-form expectations:
  - Depreciation expected to fuel inflation via higher import prices.
  - Higher foreign prices, money supply, and real income expected to increase inflation.
  - Higher rate of return expected to reduce inflation via lower money holdings.
- Empirical model augmented with structural factors specific to Chad:
  - Agricultural production shocks (national rainfall), political instability (rebel assaults or major social tensions), CFAF devaluation, and start of oil production.

### Data and empirical model specification
- Sample and data construction:
  - Quarterly data from 1983:Q1 to 2009:Q3.
  - Quarterly real GDP and quarterly government expenditures interpolated from annual data using proportional Denton method.
  - Quarterly rainfall: averages of monthly DREM data for 2000–08; FAO annual rainfall for 1983–1999; seasonal pattern from 2000–08 applied to derive 1983–1999 quarterly rainfall.
  - Political instability variable constructed from multiple sources measuring rebel assaults.
- Time series properties:
  - Augmented Dickey-Fuller and Phillips-Perron tests: all variables nonstationary in level but stationary in first difference.
- Modeling approach:
  - Long-run equilibria estimated via VECMs for money and external markets; Johansen tests do not reject one cointegrating vector in both markets.
  - Single-equation inflation model constructed by combining error correction terms from long-run VECMs and first differences of explanatory variables.
  - Equation 8 specification: ∆(log CPI) as function of lags of ∆(NEER), ∆(foreign prices), ∆(real money), ∆(real GDP), interest rates, rainfall (with lags), dummies for events, quarterly dummies, and error correction terms.
  - General-to-specific approach leads to parsimonious specification that ultimately excludes real income as a main determinant but includes public expenditure to capture fiscal expansion during the oil era.

### Long-run equilibrium estimates (money and external markets)
- Money market VECM results (Table 1):
  - Log(Real GDP): coefficients 3.39* and 5.13** (t-statistics (1.83) and (2.23)).
  - Discount minus deposit rate: coefficient -0.49*** (t-statistic (-3.93)).
  - Error correction: -0.10*** and -0.05 (t-statistics (-2.41) and (-1.52)).
  - Observations: 114.
- External market VECM results (Table 2):
  - Log(NEER): coefficient 1.18 (t-statistic (1.60)).
  - Log(Foreign price): coefficient 1.46*** (t-statistic (3.02)).
  - Error correction: -0.10*** (t-statistic (-2.92)).
  - Observations: 114.

### Single-equation inflation model results (Table 3) — main determinants
- Variables with consistent and significant effects:
  - ∆(foreign price) t-1: coefficients 0.183, 0.198, 0.209 (robust t-statistics (1.69)*, (1.82)*, (1.93)*).
    - Interpretation: a 10 percent increase in foreign prices increases domestic inflation by 2 percent one quarter later (text statement).
  - ∆(NEER) t-1: coefficients -0.252, -0.262, -0.269 (robust t-statistics (4.05)***, (4.17)***, (4.30)***).
    - Interpretation: a 10 percent appreciation (depreciation) of the NEER decreases (increases) inflation by around 2.6 percent one quarter later.
  - ∆(Public exp.) t-1: coefficients 0.093, 0.093, 0.095 (robust t-statistics (2.10)**, (2.03)**, (2.09)**).
    - Interpretation: a 10 percent increase in public spending during one quarter increases inflation by 1 percent the following quarter.
  - Rainfall t-3: coefficients -0.036, -0.033, -0.035 (robust t-statistics (2.82)***, (2.58)**, (2.70)***).
    - Interpretation: a 10 percent increase in rainfall reduces inflation by 0.35 percent three quarters later.
- Error correction (external market) EC_ext: coefficients -0.281, -0.266, -0.291 (robust t-statistics (2.40)**, (2.25)**, (2.44)**).
  - Interpretation: a 10 percent increase in external market disequilibrium reduces inflation by 3 percent with a one-quarter lag (text statement).
- Seasonality and devaluation:
  - Quarter dummies: Quarter 1 coefficients ~0.19–0.20; Quarter 2 coefficients ~0.26–0.28; Quarter 3 coefficients ~0.178–0.188; Quarter 4 omitted (reference). Many quarter coefficients significant at 1%.
  - Devaluation dummy: 0.018 (t-statistic (2.07)**) in column (1); 0.013 (t-statistic (1.44)) in specification where dummy captures one year after devaluation.
- Goodness of fit and diagnostics:
  - R-squared: 0.49, 0.47, 0.48 across specifications.
  - Observations: 114.
  - Tests for misspecification, normality, heteroskedasticity, and autoregressive heteroskedasticity do not reveal major problems (Table A.6).
  - Political instability and oil-era dummies are not significant in all regressions.

### Structural Vector Autoregression (SVAR) findings on shocks and persistence
- SVAR confirms main determinants: rainfall, public spending, nominal exchange rate, and foreign prices; effects persist over several quarters.
- Identification and recursive ordering:
  - Variables in system (first differences and expectations): rainfall (Rain), foreign prices, NEER, public spending, domestic prices.
  - Rainfall treated as exogenous; foreign prices affect variables except rainfall; NEER shocks affect public spending and domestic inflation; public spending shocks affect only domestic inflation disturbances.
- Impulse response and persistence:
  - Rainfall shock:
    - Reduces inflation; effect persists over six quarters.
    - Maximum impact at three quarters after the shock.
  - Foreign prices shock:
    - Increases inflation; effect persists over six quarters.
    - Maximum impact in the first quarter after the shock.
  - Nominal appreciation shock:
    - Reduces inflation over four quarters.
    - Highest effect at two quarters (one quarter later than single-equation prediction).
  - Public spending shock:
    - Increases inflation between the first and third quarter.
    - Maximum impact around the second and third quarters (one or two quarters later than single-equation prediction).

### Conclusions and policy recommendations
- Main drivers of inflation in Chad:
  - Rainfall (proxy for agricultural production), foreign prices, exchange rate movements, and public spending.
  - Rainfall increases reduce inflation over six quarters with peak effect in third quarter.
  - Foreign prices raise inflation over six quarters with peak impact in first quarter.
  - Nominal appreciation reduces inflation over four quarters with peak effect in second quarter.
  - Public spending raises inflation between first and third quarters with peak around second and third quarters.
- Policy recommendations to mitigate inflationary pressures:
  - Reduce sensitivity to rainfall shocks:
    - Improve efficiency in agriculture, storage, and transportation.
    - Improve food storage facilities to prevent large food price increases after droughts or floods.
    - Enhance agricultural productivity via modern techniques (irrigation, improved seeds) and better access to rural credit.
  - Reduce market segmentation and trade frictions:
    - Liberalize internal and external trade through improved transportation infrastructure such as rural roads.
  - Strengthen fiscal discipline and public financial management:
    - Better control of public expenditures through improved investment planning and procurement practices.
    - Integrate the inflationary effect of public expenditures when designing public spending programs.
- Modeling implications:
  - Single-equation model captures main drivers but does not fully capture inflation persistence; SVAR used to analyze persistence and interrelations among shocks.

*Source: IMF working paper content unit _wp1157 - 3.8 percent, but is still subject to significant shocks. Figure 2 illustrates the inflation*

### REFERENCES

### _wp1157 - REFERENCES

### References
- Barnichon, Regis, and Shanaka J. Peiris, 2008, “Sources of Inflation in Sub-Saharan Africa,” Journal of African Economies, 17(5): 729–46.
- Blavy, Rodolphe, 2004, “Inflation and Monetary Pass-Through in Guinea,” IMF Working Paper 04/223 (Washington: International Monetary Fund).
- Bloem, Adriaan M., Robert J. Dippelsman, and Nils O. Maehle, 2001, Quarterly National Accounts Manual: Concepts, Data Sources, and Compilation (Washington: International Monetary Fund), available via the Internet: http://www.imf.org/external/pubs/ft/qna/2000/Textbook/index.htm
- Chibber, A., and N. Shafik, 1990, “Exchange Reform, Parallel Markets, and Inflation in Africa: A Case of Ghana,” World Bank Policy Working Papers 47 (Washington: World Bank).
- Coleman, Simeon, 2010, “Inflation Persistence in the Franc Zone: Evidence from Disaggregated Prices,” Journal of Macroeconomics, 32(1): 426–442.
- Diouf, Mame Astou, 2007, “Modeling Inflation for Mali,” IMF Working Paper 07/295 (Washington: International Monetary Fund).
- Gottschalk, Jan, Kadima Kalonji, and Ken Miyajima, 2008, “Analyzing Determinants of Inflation When There Are Data Limitations: The Case of Sierra Leone,” IMF Working Paper 08/271 (Washington: International Monetary Fund).
- Juselius, K., 1992, “Domestic and Foreign Effects on Prices in an Open Economy: The Case of Denmark,” Journal of Policy Modeling, 14(4): 401–428.
- Kandil, Magda, and Hanan Morsy, 2009, “Determinants of Inflation in GCC,” IMF Working Paper 09/82 (Washington: International Monetary Fund).
- Klein, Nir, and Alexander Kyei, 2009, “Understanding Inflation Inertia in Angola,” IMF Working Paper 09/98 (Washington: International Monetary Fund).
- Loening, Josef L., Durevall Dick, and Birru Yohannes A., 2009, “Inflation Dynamics and Food Prices in an Agricultural Economy: The Case of Ethiopia,” Policy Research Working Paper 4969 (Washington: The World Bank).
- Miguel, Edward, Shanker Satyanath, and Ernest Sergenti, 2004, “Economic Shocks and Civil Conflict: An Instrumental Variables Approach,” Journal of Political Economy, 112(4): 725–53.
- Montenegro, Claudio E., and Maximilian L. Hirn, 2009, “A New Disaggregated Set of Labor Market Indicators Using Standardized Household Surverys from Around the World,” World Development Report 2009 (Reshaping Economic Geography) Background Paper (Washington: World Bank).
- Moriyama, Kenji, 2008, “Investigating Inflation Dynamics in Sudan,” IMF Working Paper 08/189 (Washington: International Monetary Fund).
- Nassar, Koffie, 2005, “Money Demand and Inflation in Madagascar,” IMF Working Paper 05/236 (Washington: International Monetary Fund).
- Ubide, Angel, 1997, “Determinants of Inflation in Mozambique,” IMF Working Paper 97/145 (Washington: International Monetary Fund).

### Appendix — Chad: CPI composition and data definitions
- Table A.1. Chad: Weights of the Components/Subcomponents of the CPI
  - Food and non alcoholic beverages 46.18
    - Bread and Cereals 17.24
    - Meat 7.85
    - Fish 3.64
    - Milk, Cheese, and Eggs 1.49
    - Oils and Fats 3.74
    - Vegetables and Tubers 3.67
  - Alcoholic beverages and tobacco 3.08
  - Clothing and shoes 10.88
  - Housing, water, gas, electricity, and other fuels 13.31
  - Furniture, houseware, and house maintenance 4.23
  - Health 2.73
  - Transportation 6.36
  - Communications 2.35
  - Leisure and culture 3.55
  - Education 0.96
  - Other goods and services 1.74
  - Restaurants et hotels 4.63
  - Total 100

- Table A.2. Chad: Definitions and Sources of Variables
  - Log (CPI): Logarithm of consumer price index — INSEED
  - Log (NEER): Logarithm of nominal effective exchange rate — IMF Effective Exchange Rate Facility
  - Log(Foreign price): Logarithm of foreign effective price level — IMF Effective Exchange Rate Facility
  - Discount rate: BEAC discount rate — International Financial Statistics (IFS)
  - Deposit rate: Interest rate on time deposits — IFS
  - Log(Real Money): Logarithm of real broad money (money and quasi money) — IFS
  - Log(Government expenditures): Logarithm of government expenditures — World Economic Outlook (WEO)
  - Log(Real GDP): Logarithm of Real GDP — WEO
  - Log(ToT): Logarithm of terms of trade — WEO
  - Log(Rainfall): Logarithm of Average national rainfall in mm. — Direction des Ressources en Eau et de la Météorologie, Division de la Climatologie (Octobre 2009) and Miguel, Satyanath, and Sergenti (2004).
  - Political instability dummy: Dummy equal 1 during period of political instability such as rebels’ assaults — Polity IV and http://www.jeuneafrique.com/Chronologie-pays_84_Tchad
  - Devaluation dummy: Dummy equal 1 after the devaluation in 1994
  - Oil era dummy: Dummy equal 1 after the start of oil production in 2003
  - Quarter: Quarterly dummies

### Appendix — Chad: Unit root and cointegration test results
- Table A.3. Chad: Unit Root Test (ADF and Phillips-Perron)
  - In level
    - Log(CPI): ADF -0.18, P-value 0.94; Phillips-Perron -0.31, P-value 0.92
    - Log (NEER): ADF -1.75, P-value 0.40; Phillips-Perron -1.81, P-value 0.37
    - Log(Foreign price): ADF -2.71, P-value 0.07; Phillips-Perron -2.86, P-value 0.05
    - Discount rate: ADF -1.60, P-value 0.48; Phillips-Perron -0.96, P-value 0.77
    - Deposit rate: ADF -1.13, P-value 0.70; Phillips-Perron -1.91, P-value 0.33
    - Log(Real Money): ADF -0.79, P-value 0.82; Phillips-Perron -0.44, P-value 0.90
    - Log(Gov. expenditures): ADF -1.02, P-value 0.74; Phillips-Perron -1.03, P-value 0.74
    - Log(Real GDP): ADF 1.38, P-value 0.99; Phillips-Perron 0.368, P-value 0.98
    - Log(ToT): ADF -1.58, P-value 0.49; Phillips-Perron -1.57, P-value 0.50
  - In first difference
    - Log(CPI): ADF -8.99, P-value 0.00; Phillips-Perron -8.88, P-value 0.00
    - Log (NEER): ADF -10.91, P-value 0.00; Phillips-Perron -10.91, P-value 0.00
    - Log(Foreign price): ADF -8.50, P-value 0.00; Phillips-Perron -8.33, P-value 0.00
    - Discount rate: ADF -9.76, P-value 0.00; Phillips-Perron -9.74, P-value 0.00
    - Deposit rate: ADF -9.50, P-value 0.00; Phillips-Perron -9.49, P-value 0.00
    - Log(Real Money): ADF -11.08, P-value 0.00; Phillips-Perron 11.40, P-value 0.00
    - Log(Gov. expenditures): ADF -3.58, P-value 0.01; Phillips-Perron -4.15, P-value 0.00
    - Log(Real GDP): ADF -2.34, P-value 0.15; Phillips-Perron -3.41, P-value 0.01
    - Log(ToT): ADF -3.0, P-value 0.03; Phillips-Perron -3.99, P-value 0.00

- Table A.4. Chad: Johansen Cointegration Test, Money Market Equilibrium
  - Trace test (statistics and 5% critical values)
    - Rank 0: parms 84, LL 559.9; Trace statistic 64.1; critical value (5%) 63.0
    - Rank 1: parms 92, LL 574.3, Eigenvalue 0.2; Trace statistic 35.3*; critical value (5%) 42.4
    - Rank 2: parms 98, LL 583.3, Eigenvalue 0.2; Trace statistic 17.3; critical value (5%) 25.3
    - Rank 3: parms 102, LL 589.2, Eigenvalue 0.1; Trace statistic 5.4; critical value (5%) 12.3
    - Rank 4: parms 104, LL 591.9, Eigenvalue 0.1
  - Maximum eigenvalue test (statistics and 5% critical values)
    - Rank 0: parms 84, LL 559.9; Max statistic 28.9; critical value (5%) 31.5
    - Rank 1: parms 92, LL 574.3, Eigenvalue 0.2; Max statistic 18.0; critical value (5%) 25.5
    - Rank 2: parms 98, LL 583.3, Eigenvalue 0.2; Max statistic 11.9; critical value (5%) 19.0
    - Rank 3: parms 102, LL 589.2, Eigenvalue 0.1; Max statistic 5.4; critical value (5%) 12.5
    - Rank 4: parms 104, LL 591.9, Eigenvalue 0.1
  - *The null hypothesis that there is one cointegrating equation is not rejected.

- Table A.5. Chad: Johansen Cointegration Test, External Market Equilibrium
  - Trace test (statistics and 5% critical values)
    - Rank 0: parms 105, LL 1058.0; Trace statistic 71.1; critical value 68.5
    - Rank 1: parms 114, LL 1071.8, Eigenvalue 0.2; Trace statistic 43.4*; critical value 47.2
    - Rank 2: parms 121, LL 1080.6, Eigenvalue 0.2; Trace statistic 26.0; critical value 29.7
    - Rank 3: parms 126, LL 1088.6, Eigenvalue 0.1; Trace statistic 9.8; critical value 15.4
    - Rank 4: parms 129, LL 1092.4, Eigenvalue 0.1; Trace statistic 2.3; critical value 3.8
    - Rank 5: parms 130, LL 1093.5, Eigenvalue 0.0
  - Maximum eigenvalue test (statistics and 5% critical values)
    - Rank 0: parms 105, LL 1058.0; Max statistic 27.7; critical value 33.5
    - Rank 1: parms 114, LL 1071.8, Eigenvalue 0.2; Max statistic 17.4; critical value 27.1
    - Rank 2: parms 121, LL 1080.6, Eigenvalue 0.2; Max statistic 16.1; critical value 21.0
    - Rank 3: parms 126, LL 1088.6, Eigenvalue 0.1; Max statistic 7.6; critical value 14.1
    - Rank 4: parms 129, LL 1092.4, Eigenvalue 0.1; Max statistic 2.3; critical value 3.8
    - Rank 5: parms 130, LL 1093.5, Eigenvalue 0.0
  - *The null hypothesis that there is one cointegrating equation is not rejected.

### Appendix — Model diagnostics and figures (Chad)
- Figures included (names as in source):
  - Figure A.1. Chad: Evolution of the Main Variables (Log (NEER); Log(Government expenditures); Discount and deposit rate; Log(Foreign price); Log (Real broad money); Log(Real GDP); Deposit rate; Discount rate)
  - Figure A.2. Chad: Four-Quarter Lag Moving Average of the National Rainfall and CPI (CPI; Rainfall)
  - Figure A.3. Chad: Actual and Predicted Quarterly Inflation (Specification 1)
  - Figure A.4. Chad: Residuals (Specification 1) — Predicted inflation; Actual inflation
  - Figure A.5. Chad: Actual and Predicted Quarterly Inflation (Specification 2)
  - Figure A.6. Chad: Residuals (Specification 2) — Predicted inflation; Actual inflation
  - Figure A.7. Chad: Actual and Predicted Quarterly Inflation (Specification 3)
  - Figure A.8. Chad: Residuals (Specification 3) — Actual inflation; Predicted inflation

- Table A.6. Chad: Diagnostic Statistics for the Single-Equation Inflation Models
  - Specification 1
    - Breusch-Pagan / Cook-Weisberg test for heteroskedasticity: 0.86 (p-value 0.35)
    - Ramsey RESET: 0.56 (p-value 0.64)
    - Skewness: 6.08 (p-value 0.81)
    - Kurtosis: 1.46 (p-value 0.23)
    - ARCH (1 lag): 0.027 (p-value 0.87)
    - ARCH (2 lags): 0.173 (p-value 0.92)
    - ARCH (3 lags): 0.223 (p-value 0.97)
  - Specification 2
    - Breusch-Pagan / Cook-Weisberg test for heteroskedasticity: 0.28 (p-value 0.60)
    - Ramsey RESET: 1.29 (p-value 0.28)
    - Skewness: 4.78 (p-value 0.91)
    - Kurtosis: 1.07 (p-value 0.30)
    - ARCH (1 lag): 0.035 (p-value 0.85)
    - ARCH (2 lags): 0.129 (p-value 0.94)
    - ARCH (3 lags): 0.118 (p-value 0.99)
  - Specification 3
    - Breusch-Pagan / Cook-Weisberg test for heteroskedasticity: 0.54 (p-value 0.46)
    - Ramsey RESET: 1.05 (p-value 0.37)
    - Skewness: 4.87 (p-value 0.90)
    - Kurtosis: 1.27 (p-value 0.26)
    - ARCH (1 lag): 0.018 (p-value 0.89)
    - ARCH (2 lags): 0.065 (p-value 0.97)
    - ARCH (3 lags): 0.057 (p-value 0.99)
  - Notes:
    - p-values in parentheses.
    - Null hypothesis for the Breusch-Pagan/Cook-Weisberg test for heteroskedasticity is constant variance.
    - Null hypothesis for the Ramsey RESET test is the absence of omitted variables.
    - Null hypothesis for skewness and kurtosis tests is that the distribution of residuals has skewness and kurtosis corresponding to those of a normal distribution.
    - Null hypothesis for autoregressive conditional heteroskedasticity (ARCH) tests is the absence of autocorrelation.

*Content derived from _wp1157 - REFERENCES (PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1157.pdf_