## _wp07286

## Source details

**Canonical URL:** [_wp07286](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07286.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07286.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07286.pdf.json)

---

### I. Introduction
- Algeria experienced brisk economic growth, low inflation, and a markedly strengthened current account over the five years prior to 2006.
- The government launched a five-year public investment program, the Growth Consolidation Plan (GCP), to improve infrastructure, housing, and public services.
- Concern: increased domestic use of hydrocarbon revenues, especially via capital spending, may appreciate the real exchange rate and undermine competitiveness of the nonhydrocarbon tradable sector (Dutch disease).
- Empirical finding summarized: increases in government capital spending have had a limited appreciating impact on the real exchange rate in Algeria so far, partly because public spending has high import content generating foreign exchange outflows that attenuate appreciation.
- Policy attention: fiscal sustainability, quality of capital spending, and implications of the investment program for future recurrent expenditure, as many of these goods are nontradables.

### II. Fiscal policy stance in Algeria (context and risks)
- Hydrocarbon dependence and revenue flows:
  - Hydrocarbon sector accounted for almost 45 percent of GDP in 2005 and 98 percent of exports of goods.
  - Sonatrach repatriates and surrenders foreign currency receipts to the Bank of Algeria; on average, two-third of export receipts accrue to the treasury.
- Historical pattern of expenditure:
  - Capital expenditure rose in the 1970s and early 1980s with high hydrocarbon prices; fell after price collapse in the late 1980s while current expenditure remained high.
  - Fiscal tightening occurred during 1994–98 under a Fund-supported program.
  - Capital expenditure increased from about 8 percent of GDP in 2001 to 11 percent of GDP in 2004.
- Growth Consolidation Program (GCP) 2005–09:
  - Public investment program for 2005–09 amounts to about $140 billion (about 140 percent of 2005 GDP).
  - Execution rates of capital spending were low during 2005; capital spending remained about 11 percent of GDP in 2005 though composition shifted toward capital.
- Identified risks:
  - Limited planning, implementation, and management capacity could channel resources to inefficient and unsustainable capital projects and current expenditure.
  - Fiscal expansion could generate inflationary pressures and real exchange rate appreciation, reducing competitiveness of the nonhydrocarbon traded goods sector.

### III. Analytical framework (model structure and implications)
- Model foundations:
  - Builds on Cashin and others (2004) by incorporating government spending into a BEER-style model linking the real exchange rate to real commodity prices.
- Key components:
  - Three goods: nontradable, exportable (primary commodity), and final tradable.
  - Two agents: household/firms and government (government preferences may differ from households).
  - Labor is sole factor, freely mobile across sectors; labor supply inelastic.
- Parameters and definitions (as used in the source):
  - REER defined via equation (1) and log-linearized in equation (2).
  - γ: elasticity of household consumption to the nontraded good.
  - g: elasticity of government consumption to the nontraded good.
  - uP: government’s share in total aggregate demand.
  - a_x^* / a_x and a_N^* / a_N capture productivity differentials (Balassa-Samuelson effects).
- Comparative static implications:
  - Equation (3): a rise in the relative price of the primary commodity leads to appreciation of the REER.
  - Equation (4): increasing government spending has an appreciating effect on the REER; magnitude depends on the share of nontraded goods in government spending.
  - Equation (5): an increase in the relative share of nontraded goods in public spending with unchanged total public spending leads unambiguously to REER appreciation.
- Interpretation:
  - If government marginal propensity to import is lower than private sector’s, fiscal expansion shifts demand toward domestic goods and leads to REER appreciation (Sachs and Wyplosz, 1984–type result).

### IV. Empirical model specification
- Explanatory variables selected for a commodity-dependent economy:
  - Real price of the primary commodity (real world oil price) — proxy for external terms of trade.
  - Differential productivity (PROD) — captures Balassa-Samuelson effects.
  - Government capital spending as a share of GDP (GK).
- Functional representation:
  - REER_t = f(PROD_t, Oil_t, GK_t)  (equation (6)).
- Econometric framework:
  - VAR / VECM based on Johansen (1988, 1995) methodology (equations (7)–(9)).
  - ω_t assumed iid normal with zero mean and covariance matrix Ω.
  - Number of lags chosen using AIC procedure: 4 lags.

### V. Data, estimation approach, and robustness adjustments
- Data coverage:
  - Annual series covering the period 1970–2006.
  - Data drawn mostly from the World Economic Outlook (WEO) database.
- Unit-root testing:
  - Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests do not reject the null of a unit root in levels; all series are integrated of order one (Table 1).
- Cointegration testing and small-sample correction:
  - Johansen cointegration tests applied with lag length q = 4 (AIC).
  - Small-sample bias addressed using Reinsel and Ahn (1988) multiplicative scaling factor T/(T-nj) to adjust critical values.
  - After adjustment, both Trace and Max-Eigen statistics reject the null of zero cointegrating equations; tests indicate one cointegrating vector (Tables 2a and 2b).
  - Additional small-sample checks: bootstrap/BDS tests used to check whether residuals are iid; results show residuals are iid (discussion in footnote).

### VI. Empirical results and numerical estimates
- Sample period: 1970–2006.
- Cointegration: one cointegrating vector linking LREER, LOIL, LPROD, LGK.
- Long-run parameter estimates (Table 3, normalized on LREER):
  - LOIL: 0.26***  (standard error (0.05))
  - LPROD: 1.36***  (standard error (0.04))
  - LGK: 0.13**  (standard error (0.07))
- Adjustment dynamics:
  - Speed of adjustment (error-correction term): -0.78**  (standard error (0.2))
  - Adjusted R^2 = 0.82
- Interpretive statements from the source text:
  - "The REER appreciates as oil prices increase. For example, a one percent increase in real oil prices will appreciate the real effective exchange rate by 0.3 percent."
  - "The parameter associated with the differential productivity suggests ... A one percent worsening in Algeria’ productivity differential has more than one-to-one depreciating effect on the exchange rate."
  - "An increase in government capital expenditure is associated with an appreciation of the exchange rate. However, the impact appears relatively small: for example, a 10 percent increase in capital spending to GDP ratio will appreciate the real effective exchange rate by 0.3 percent, suggesting that a non negligible proportion of government spending is directed toward tradable goods."
- Structural interpretation:
  - Capital expenditure accounts for one-third of total expenditure and often includes equipment and goods not domestically produced; hence increases in capital spending boost imports and attenuate REER appreciation.
  - Causality from capital spending to imports is supported by a Granger causality test.

### VII. Key findings (concise bullets)
- All series (LREER, LOIL, LPROD, LGK) are I(1) over 1970–2006.
- Johansen tests (with Reinsel and Ahn small-sample correction) indicate one cointegrating relationship among the variables.
- Long-run elasticities (from Table 3):
  - LOIL = 0.26*** (REER response to a 1 percent change in real oil price).
  - LPROD = 1.36*** (REER response to a 1 percent change in productivity differential).
  - LGK = 0.13** (REER response to a 1 percent change in government capital spending to GDP ratio).
- Error-correction dynamics: adjustment speed = -0.78** (convergence toward long-run equilibrium).
- Econometric and sectoral evidence: government capital spending has an appreciating effect on REER, but magnitude is limited because a non-negligible share of capital spending imports tradable goods and equipment.

### VIII. Policy implications highlighted in the source
- Core recommendations:
  - Ensure fiscal sustainability.
  - Improve the quality and planning/implementation capacity for capital projects to avoid inefficient and unsustainable programs.
  - Carefully consider implications of the public investment program on future recurrent expenditure, since these are largely nontradables and can contribute to REER appreciation.
- Additional policy guidance:
  - Implement structural policies encouraging private sector saving.
  - Invest part of the foreign exchange inflows in foreign assets.
  - Ensure good coordination between fiscal and monetary policies to help avoid a strong real exchange rate appreciation.
  - Damp potential pressure on the real exchange rate by making current expenditure, especially recurrent spending, make more use of production factors, goods, and services that are in excess supply in the economy.
  - Complement exchange rate and infrastructure reforms with productivity-enhancing reforms to secure high and sustained growth rates of the nonhydrocarbon sector.

### IX. Co-movements between public expenditure and imports (Appendix 3 summary)
- Strong correlation between public expenditure and imports, with notable co-movements during recent years (Figure 3).
- Granger Causality test results (first differences; 1969 test frame):
  - Ho: DLGK does not Granger Cause DLREAL_M — F-Statistic: 3.578** (rejects at 5 percent).
  - Ho: DLREAL_M does not Granger Cause DLGK — F-Statistic: 0.598 (does not reject).
  - Note: “Granger causality captures precedence and information content and does not always have the general meaning of the term.”

### X. Theoretical model (Appendix 1 summary)
- Framework: small open economy producing two goods — a nontraded good and an exportable (primary commodity); extends Cashin and others (2004) by incorporating government spending effects on the real exchange rate.
- Key model relationships:
  - Relative price relation (equation (9)): P_N = P_X (a_X / a_N) — relative price of the nontraded good with respect to the primary commodity determined by technological factors.
  - Private consumption aggregator: C = κ C_N^γ C_T^(1−γ) with unit cost P = P_N^γ P_T^(1−γ).
  - Government consumption aggregator price: P_G = P_N^g P_T^(1−g).
  - Law of one price for imported good: P_T = E P_T^*.
  - Real exchange rate expression derived in equation (13) combining (9)–(12).

### XI. Data (Appendix 2: variables and descriptive statistics)
- Variable definitions and sources:
  - REER: CPI based Real effective exchange rate — Information Notice System, IMF.
  - Oil: Real oil price: UK Brent price deflated by Manufacturing Unit value (MUV). — WEO, IMF, DEC Database, World Bank.
  - DPROD: Productivity differential: Real GDP per employee relative to trading partners. — WEO, IMF.
  - GK: Government capital as share of GDP — WEO, IMF.
  - REAL_M: Real Imports: Imports of Goods deflated by the import unit value index. — WEO, IMF.
- Descriptive statistics (logs where applicable):
  - LREER
    - Mean: 5.146
    - Median: 5.449
    - Maximum: 5.871
    - Minimum: 4.392
    - Std. Dev.: 0.511
  - LOIL
    - Mean: 0.165
    - Median: 0.133
    - Maximum: 1.141
    - Minimum: -0.928
    - Std. Dev.: 0.525
  - LPROD
    - Mean: 0.352
    - Median: 0.345
    - Maximum: 0.761
    - Minimum: -0.079
    - Std. Dev.: 0.292
  - LGK
    - Mean: 3.403
    - Median: 3.371
    - Maximum: 3.630
    - Minimum: 3.246
    - Std. Dev.: 0.104

*Source: _wp07286 - 3. Econometric Results; Appendixes 1–3.*

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

### _wp07286 - References..............................................................................................................

### Tables
- 1.    Unit-Root Tests .................................................................................................................11
- 2a.  Trace Statistics for Cointegrating Rank ............................................................................11
- 2b.  Max-Eigen Statistics for Cointegrating Rank ...................................................................11
- 3.    ECM Parameter Estimates ................................................................................................12

### Figures
- 1. Public Expenditure and Oil Prices, 1970–2005 .....................................................................4
- 2. Composition of Government Expenditure,1970–2005 ..........................................................5
- 3. Imports and Government Expenditures, 1995–2005 ...........................................................13

### Appendixes
- 1.  Theoretical Model ...............................................................................................................17
- 2.  Data .....................................................................................................................................20

*Source: _wp07286 - References..............................................................................................................*

### 3.  Econometric Results ................................................................................................

### 3.  Econometric Results

### I. Introduction
- Algeria experienced brisk economic growth, low inflation, and a markedly strengthened current account over the past five years prior to 2006.  
- The government launched a five-year public investment program, the Growth Consolidation Plan (GCP), to improve infrastructure, housing, and public services.  
- There is concern that increased domestic use of hydrocarbon revenues, especially via capital spending, may appreciate the real exchange rate and undermine competitiveness of the nonhydrocarbon tradable sector (Dutch disease).  
- The paper’s empirical work suggests that increases in government capital spending have had a limited appreciating impact on the real exchange rate in Algeria so far, partly because public spending has high import content generating foreign exchange outflows that attenuate appreciation.  
- Policy attention should focus on fiscal sustainability, quality of capital spending, and implications of the investment program for future recurrent expenditure, as many of these goods are nontradables.

### II. Fiscal policy stance in Algeria (context and risks)
- Hydrocarbon sector share and fiscal dependence:
  - Hydrocarbon sector accounted for almost 45 percent of GDP in 2005 and 98 percent of exports of goods.
  - Sonatrach repatriates and surrenders foreign currency receipts to the Bank of Algeria; on average, two-third of export receipts accrue to the treasury.
- Historical pattern:
  - Capital expenditure rose in the 1970s and early 1980s with high hydrocarbon prices; fell after price collapse in the late 1980s while current expenditure remained high.
  - Fiscal tightening occurred during 1994–98 under a Fund-supported program.
  - Capital expenditure increased from about 8 percent of GDP in 2001 to 11 percent of GDP in 2004.
- Growth Consolidation Program (GCP) 2005–09:
  - Public investment program for 2005–09 amounts to about $140 billion (about 140 percent of 2005 GDP).
  - Execution rates of capital spending were low during 2005; capital spending remained about 11 percent of GDP in 2005 though composition shifted toward capital.
- Risks identified:
  - Limited planning, implementation, and management capacity could channel resources to inefficient and unsustainable capital projects and current expenditure.
  - Fiscal expansion could generate inflationary pressures and real exchange rate appreciation, reducing competitiveness of the nonhydrocarbon traded goods sector.

### III. Analytical framework (model structure and implications)
- Model basis and extension:
  - Builds on Cashin and others (2004) by incorporating government spending into a BEER-style model linking the real exchange rate to real commodity prices.
- Structure:
  - Three goods: nontradable, exportable (primary commodity), and final tradable.
  - Two agents: household/firms and government (government preferences may differ from households).
  - Labor is sole factor, freely mobile across sectors; labor supply inelastic.
- Key parameters and definitions (as used in the source):
  - REER defined via equation (1) and log-linearized in equation (2).
  - γ: elasticity of household consumption to the nontraded good.
  - g: elasticity of government consumption to the nontraded good.
  - uP: government’s share in total aggregate demand.
  - a_x^* / a_x and a_N^* / a_N capture productivity differentials (Balassa-Samuelson effects).
- Comparative statics:
  - Equation (3): a rise in the relative price of the primary commodity leads to appreciation of the REER.
  - Equation (4): increasing government spending has an appreciating effect on the REER; the magnitude depends on the share of nontraded goods in government spending.
  - Equation (5): an increase in the relative share of nontraded goods in public spending with unchanged total public spending leads unambiguously to REER appreciation.
- Interpretation:
  - If government marginal propensity to import is lower than private sector’s, fiscal expansion shifts demand toward domestic goods and leads to REER appreciation (Sachs and Wyplosz, 1984–type result).

### IV. Empirical model specification
- Explanatory variables chosen (for commodity-dependent economy like Algeria):
  - Real price of the primary commodity (real world oil price) – proxy for external terms of trade.
  - Differential productivity (PROD) – captures Balassa-Samuelson effects.
  - Government capital spending as a share of GDP (GK).
- Functional representation:
  - REER_t = f(PROD_t, Oil_t, GK_t)  (equation (6))
- Econometric framework:
  - Vector autoregression (VAR) / Vector error correction model (VECM) based on Johansen (1988, 1995) methodology (equations (7)–(9)).
  - ω_t assumed iid normal with zero mean and covariance matrix Ω.
  - Number of lags chosen using AIC procedure (4 lags).

### V. Data, estimation approach, and robustness adjustments
- Data:
  - Annual series covering the period 1970–2006.
  - Data drawn mostly from the World Economic Outlook (WEO) database.
- Unit-root testing:
  - Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests do not reject the null of a unit root in levels; all series are integrated of order one (Table 1).
- Cointegration testing:
  - Johansen cointegration tests applied with lag length q = 4 (AIC).
  - Small-sample bias concern addressed using Reinsel and Ahn (1988) multiplicative scaling factor T/(T-nj) to adjust critical values.
  - After adjustment, both Trace and Max-Eigen statistics reject the null of zero cointegrating equations; tests indicate one cointegrating vector (Tables 2a and 2b).
  - Additional small-sample checks: bootstrap/BDS tests used to check whether residuals are iid; results show residuals are iid (discussion in footnote).

### VI. Empirical results and numerical estimates
- Sample period: 1970–2006.
- Cointegration: presence of one cointegrating vector linking LREER, LOIL, LPROD, LGK.
- Key long-run parameter estimates (Table 3, normalized on LREER):
  - LOIL: 0.26***  (standard error (0.05) reported in table)
  - LPROD: 1.36***  (standard error (0.04))
  - LGK: 0.13**  (standard error (0.07))
- Adjustment dynamics:
  - Speed of adjustment (error-correction term): -0.78**  (reported standard error (0.2))
  - Adjusted R^2 = 0.82
- Interpretive statements from the source text:
  - "The REER appreciates as oil prices increase. For example, a one percent increase in real oil prices will appreciate the real effective exchange rate by 0.3 percent."
  - "The parameter associated with the differential productivity suggests ... A one percent worsening in Algeria’ productivity differential has more than one-to-one depreciating effect on the exchange rate."
  - "An increase in government capital expenditure is associated with an appreciation of the exchange rate. However, the impact appears relatively small: for example, a 10 percent increase in capital spending to GDP ratio will appreciate the real effective exchange rate by 0.3 percent, suggesting that a non negligible proportion of government spending is directed toward tradable goods."
- Supporting structural interpretation:
  - Capital expenditure accounts for one-third of total expenditure and often includes equipment and goods not domestically produced; hence increases in capital spending boost imports and attenuate REER appreciation.
  - Causality from capital spending to imports is confirmed by a Granger causality test (referred to in the text).

### VII. Key findings (summary bullets)
- All series (LREER, LOIL, LPROD, LGK) are I(1) over 1970–2006 (unit-root tests).
- Johansen tests (with Reinsel and Ahn small-sample correction) indicate one cointegrating relationship among the variables.
- Long-run elasticities (from Table 3):
  - LOIL = 0.26*** (REER response to a 1 percent change in real oil price).
  - LPROD = 1.36*** (REER response to a 1 percent change in productivity differential).
  - LGK = 0.13** (REER response to a 1 percent change in government capital spending to GDP ratio).
- Error-correction dynamics: adjustment speed = -0.78** (convergence toward long-run equilibrium).
- Econometric and sectoral evidence suggests government capital spending has an appreciating effect on REER, but the magnitude is limited because a non-negligible share of capital spending imports tradable goods and equipment.

### VIII. Policy implications highlighted in the source
- While public investment can promote infrastructure and employment, the authorities should:
  - Ensure fiscal sustainability.
  - Improve the quality and planning/implementation capacity for capital projects to avoid inefficient and unsustainable programs.
  - Consider implications of the public investment program on future recurrent expenditure, since these are largely nontradables and can contribute to REER appreciation.
- Given the import content of capital spending, a significant portion of public investment can be absorbed by imports, mitigating but not eliminating appreciation pressures; policymakers should monitor composition and import content of public spending.

*Source: _wp07286 - 3.  Econometric Results*

### Appendix 3).

### _wp07286 - Appendix 3)

### Co-movements between public expenditure and imports
- The text documents a strong correlation between public expenditure and imports, with notable co-movements during recent years (Figure 3).
- The results of the Granger Causality test are described as suggestive rather than conclusive: “Granger causality captures precedence and information content and does not always have the general meaning of the term.”

### Key conclusions and policy recommendations (from main text)
- The decayed and overstretched economic infrastructure and inadequate institutional and regulatory framework constrain emergence of a vibrant non-hydrocarbon sector.
- Fixing and extending economic infrastructure is required to promote an alternative growth engine to the hydrocarbons sector; the current ambitious investment program is an important step.
- Empirical finding: an increase in government capital expenditure is associated with a relatively small appreciation of the exchange rate, suggesting an important proportion of capital spending is directed toward imported goods.
- Possible future dynamic: the envisaged increase in capital spending, if well designed and implemented, might in the long-run translate into rising operations and maintenance expenditure—mostly nontradable goods—thereby causing a higher real appreciation, with potential adverse implications for the nonhydrocarbon sector.
- Recommended policy actions:
  - Ensure fiscal sustainability and the quality of capital spending.
  - Carefully consider implications of the public investment program on recurrent expenditure.
  - Implement structural policies encouraging private sector saving.
  - Invest part of the foreign exchange inflows in foreign assets.
  - Ensure good coordination between fiscal and monetary policies to help avoid a strong real exchange rate appreciation.
  - Damp potential pressure on the real exchange rate by making current expenditure, especially recurrent spending, make more use of production factors, goods, and services that are in excess supply in the economy.
  - Complement exchange rate and infrastructure reforms with productivity-enhancing reforms to secure high and sustained growth rates of the nonhydrocarbon sector.

### Theoretical model (summary of Appendix 1)
- Framework: small open economy producing two goods — a nontraded good and an exportable (primary commodity). Model extends Cashin and others (2004) by incorporating government spending effects on the real exchange rate.
- Production functions:
  - Export (primary commodity): YaL = X X (presented as XXX YaL=, with X L as labor per unit and x a as labor productivity in the export sector).
  - Nontradable good: YaL = N N (presented as NNN YaL=, with N L and N a as employment and labor productivity in the nontradable sector).
- Profit maximization conditions produce:
  - X X w P a = 
  - N N w P a =
- Relative price relation (equation (9)):
  - P_N = P_X (a_X / a_N)   (presented as X N N a P P a =  in the source)
  - “Thus, the relative price of the nontraded good (P_N) with respect to the primary commodity (P_X) is solely determined by technological factors and is independent of demand conditions.”
- Aggregate demand and preferences:
  - Private consumption aggregator: C = κ C_N^γ C_T^(1−γ) (presented as 1 NT CCC γγ κ − = ), with minimum cost of one unit of private consumption P = P_N^γ P_T^(1−γ) (presented as 1 CNT PPP γγ− =).
  - Government consumption aggregator: G price P_G = P_N^g P_T^(1−g) (presented as 1 gg GNT PPP − =), with u as share of public spending in aggregate demand and aggregate domestic price level specified (equation (10)).
- Law of one price for imported good: P_T = E P_T^*.
- Foreign region: three sectors (nontraded, intermediate, final); foreign wage-price condition analogous to domestic (equation (11) in source).
- Final tradable good production in foreign region: Y_T = (Y_1^β Y_X^(1−β)) (presented as 1 ** * TX Y v Y Y ββ− = ), and cost of tradable good in foreign currency given by P_T^* = P_1^*^β P_X^*^(1−β) (presented accordingly).
- Real exchange rate expression (equation (13) from combining equations (9)–(12)):
  - (1) E P = P^* [expression in source: * * ** * 11 PugPu xN X N aa P E P P aaP γ−+ (presented verbatim in the source)].

### Data (Appendix 2)
- Variable definitions and sources:
  - REER: CPI based Real effective exchange rate — Information Notice System, IMF.
  - Oil: Real oil price: UK Brent price deflated by Manufacturing Unit value (MUV). — World Economic Outlook (WEO), IMF, DEC Database, World Bank.
  - DPROD: Productivity differential: Real GDP per employee relative to trading partners. — WEO, IMF.
  - GK: Government capital as share of GDP — WEO, IMF.
  - REAL_M: Real Imports: Imports of Goods deflated by the import unit value index. — WEO, IMF.
- Descriptive statistics (variables noted as logs where applicable):
  - LREER (log of the real effective exchange rate)
    - Mean: 5.146
    - Median: 5.449
    - Maximum: 5.871
    - Minimum: 4.392
    - Std. Dev.: 0.511
  - LOIL (log of the real price of oil)
    - Mean: 0.165
    - Median: 0.133
    - Maximum: 1.141
    - Minimum: -0.928
    - Std. Dev.: 0.525
  - LPROD (log of the productivity differential)
    - Mean: 0.352
    - Median: 0.345
    - Maximum: 0.761
    - Minimum: -0.079
    - Std. Dev.: 0.292
  - LGK (log of the ratio of capital expenditure to GDP)
    - Mean: 3.403
    - Median: 3.371
    - Maximum: 3.630
    - Minimum: 3.246
    - Std. Dev.: 0.104

### Econometric results (Appendix 3)
- Granger Causality Tests (1969) — variables are I(1); tests run on first differences.
- Capital Expenditure and Total Imports of Goods:
  - Ho: DLGK does not Granger Cause DLREAL_M — F-Statistic: 3.578**
  - Ho: DLREAL_M does not Granger Cause DLGK — F-Statistic: 0.598
- Notes from source:
  - DLGK and DLREAL_M stand for the first log difference of the ratio of capital expenditure to GDP and the first log difference of real imports, respectively.
  - The F-statistics displayed are the Wald statistics for the joint assumptions: 0 = l δ with l = 1, 2, ..., r, and 0 = s δ with s = 1, 2..., u.
  - ** denotes significance of the estimates at 5 percent critical level; * denotes significance at 10 percent critical level.

*Source: _wp07286 - Appendix 3).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07286.pdf_