## 4.   Key Time Series

## Source details

**Canonical URL:** [4.   Key Time Series](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09195.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09195.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09195.pdf.json)

---

### Introduction
- Purpose: formally examine determinants of commercial bank deposit demand in Lebanon, assessing domestic versus external drivers using VECMs and bank-level panel regressions.
- Data:
  - Quarterly data 1993–2008 for macro models.
  - Annual bank-level panel 1997–2007 covering over 50 banks.
- Methodology highlights:
  - Vector error correction models (VECMs) to account for cointegration in non-stationary time series.
  - Johansen cointegration tests and sequential modified LR tests for lag selection.
  - Fixed-effects and robust panel regressions for bank-level analysis; bank-specific variables lagged to avoid endogeneity.
- Key macro variables: Banque du Liban coincident indicator (CI), Consumer Price Index (CPI), interest rate differential (DEPATTRACT), advanced economy industrial production (LADVECONIP), GCC real GDP (LGCC_GDP), GCC non-oil GDP (LGCC_GDP_NOIL), GCC outward remittances (LGCC_REMIT), crude oil price (LOIL), Goldman Sachs Risk Aversion Index (RISKAVERS), and other global liquidity/volatility proxies (VIX, MSCIWORLD, G7DEFL, EMCREDQ, EMPSCREDG).
- Stationarity: unit root tests point to non-stationarity of all macro variables; majority I(1) (Appendix).

### Commercial bank deposits in Lebanon: context and key statistics
- Fiscal and financial ratios:
  - Government debt-to-GDP ratio: 160 percent.
  - Total commercial bank assets to GDP ratio: 324 percent.
  - Interest payments: nearly 12 percent of GDP (2005–08 average).
  - Overall fiscal deficits: in the order of 10 percent of GDP (2005–08 average).
- Remittances:
  - Remittances: 21 percent of GDP in 2007.
  - Origin shares — GCC 54 percent, Australia 7 percent, United States 6 percent, Canada 3 percent.
- Observed correlations (illustrative):
  - Correlation between deposit growth and GCC real GDP growth: 0.51.
  - Correlation between deposit growth and interest differential: -0.27.

### Macroeconomic estimates (VECM results and interpretation)
- Core finding: domestic and international factors both help explain deposit demand; external variables often have relatively larger impact per impulse responses and variance decompositions.
- Domestic variables with significance in some specifications: coincident indicator (LCI_SA), consumer prices (LCPI_SA), interest differential (DEPATTRACT).
- External variables with significance in various specifications: advanced economy industrial production (LADVECONIP_SA), Goldman Sachs Risk Aversion Index (RISKAVERS), GCC GDP measures, crude oil prices (LOIL), GCC outward remittances (LGCC_REMIT_SA).
- Error correction dynamics (selected specifications):
  - Specification I: one cointegrating vector; ECM coefficient/CointEq1 = -0.198 (t-statistic -5.138) — "close to 20 percent of the deviation from the long-run equilibrium is corrected within one quarter."
  - Specifications II–V: ECM coefficients between -0.118 and -0.079 (t-stats -7.231 to -5.060) — error correction between 8 and 12 percent per quarter.
  - Specification VI (including RISKAVERS): ECM coefficient = -0.060 (t-statistic -3.653) — 6 percent per quarter.
- Selected cointegrating-vector coefficients for Deposits (LDEPOSITS_SA), t-statistics in parentheses:
  - Coincident indicator (LCI_SA):
    - Spec I = 0.419 (1.692)
    - Spec II = -0.373 (-0.942)
    - Spec III = -0.457 (-0.971)
    - Spec IV = 0.158 (0.281)
    - Spec V = 0.749 (1.867)
    - Spec VI = 0.885 (2.872)
  - Interest differential (DEPATTRACT):
    - I = 0.035 (2.882)
    - II = 0.015 (1.325)
    - III = 0.013 (1.047)
    - IV = 0.023 (1.539)
    - V = 0.012 (0.800)
    - VI = 0.015 (1.233)
  - Consumer prices (LCPI_SA):
    - I = 0.762 (2.450)
    - II = 2.236 (7.845)
    - III = 2.151 (6.021)
    - IV = 2.352 (5.387)
    - V = 2.257 (5.719)
    - VI = 2.703 (8.794)
  - Advanced econ ind. prod. (LADVECONIP_SA): I = 3.340 (2.836)
  - Real GDP in GCC (LGCC_GDP_SA): II = 1.130 (3.393)
  - Real non-oil GDP in GCC (LGCC_GDP_NOIL_SA): III = 1.202 (1.681)
  - Crude oil price (LOIL): IV = 0.329 (3.040)
  - GCC outward remittances (LGCC_REMIT_SA): V = 0.159 (0.656)
  - GS Risk Aversion Index (RISKAVERS): VI = -0.028 (-2.068)
  - Trend and constants reported per specification; adjusted R2 values:
    - I = 0.59; II = 0.63; III = 0.60; IV = 0.48; V = 0.47; VI = 0.49.
- Impulse Response Functions (responses of DEPOSITS_SA to one-standard-deviation shock after six periods, in percent):
  - VECM I (after six periods): LCI_SA = 1.00; DEPATTRACT = 1.38; LCPI_SA = 0.76; LADVECONIP_SA = 3.69.
  - VECM II: LCI_SA = 0.11; DEPATTRACT = 0.26; LCPI_SA = 1.35; LGCC_GDP_SA = 1.54.
  - VECM III: LCI_SA = 0.19; DEPATTRACT = 0.30; LCPI_SA = 1.20; LGCC_GDP_NOIL_SA = 1.52.
  - VECM IV: LCI_SA = 0.78; DEPATTRACT = 0.56; LCPI_SA = 1.27; LOIL = 2.18.
  - VECM V: LCI_SA = 1.34; DEPATTRACT = -0.33; LCPI_SA = 1.10; LGCC_REMIT_SA = 1.31.
  - VECM VI: LCI_SA = 1.77; DEPATTRACT = -0.76; LCPI_SA = 2.39; RISKAVERS = -0.69.
- Variance Decomposition (percent of variance in LDEPOSITS_SA after six periods due to each variable, selected entries):
  - VECM I:
    - Deposits (own shocks) = 11.23
    - LCI_SA = 11.85
    - DEPATTRACT = 7.63
    - LCPI_SA = 2.46
    - LADVECONIP_SA = 66.83
  - VECM II:
    - Deposits = 20.71
    - LCI_SA = 5.91
    - DEPATTRACT = 3.49
    - LCPI_SA = 3.07
    - LGCC_GDP_SA = 46.82
  - VECM III:
    - Deposits = 27.51
    - LCI_SA = 9.53
    - DEPATTRACT = 4.27
    - LCPI_SA = 19.09
    - LGCC_GDP_NOIL_SA = 39.60
  - VECM IV:
    - Deposits = 42.66
    - LCI_SA = 10.82
    - DEPATTRACT = 3.03
    - LCPI_SA = 9.34
    - LOIL = 34.15
  - VECM V:
    - Deposits = 42.75
    - LCI_SA = 25.35
    - DEPATTRACT = 5.89
    - LCPI_SA = 9.14
    - LGCC_REMIT_SA = 16.89
  - VECM VI:
    - Deposits = 13.85
    - LCI_SA = 32.82
    - DEPATTRACT = 13.08
    - LCPI_SA = 7.52
    - RISKAVERS = 2

### Bank-specific factors (panel regressions) — key findings and coefficients
- Data and estimation:
  - Annual 1997–2007, 357 observations, 53 banks (some consolidated, some unconsolidated; outliers handled).
  - Main estimators: Random Effects (RE), Fixed Effects (FE), Robust.
  - Bank-specific variables lagged by one period (-1).
- Bank-specific explanatory variables tested: z-score (perceived riskiness), profitability (ROE, net interest margin), total assets (USD billion), loans-to-assets ratio, liquidity ratio (liquid assets to deposits), government securities to total assets, general loss reserves/gross loans.
- Consistent, statistically significant positives across specifications:
  - Loans/Assets (-1): positive and highly significant in all columns. Examples:
    - Column (1) RE: 0.524 (p-value 0.000)***
    - Column (2) FE: 0.682 (0.000)***
    - Column (3) FE with macro controls: 0.891 (0.000)***
    - Column (7) Robust: 0.14 (0.029)**.
  - Liquidity (-1): positive and highly significant in all columns. Examples:
    - Column (1): 0.332 (0.000)***; Column (7): 0.096 (0.003)***.
  - Z-Score (-1): positive and significant in several specifications. Examples:
    - Column (1): 0.003 (0.011)**; Column (2): 0.004 (0.039)**; Column (7): 0.001 (0.144) (less significant in robust).
  - Net Interest Rate Margin (-1): positive; significant in some FE/robust specifications. Examples:
    - Column (2): 0.022 (0.005)***; Column (7): 0.008 (0.011)**.
- Insignificant or mixed results:
  - Profitability (ROE) (-1): generally insignificant across columns (e.g., Column (1) 0.017 (0.498)).
  - Total assets (-1): insignificant in most specifications; coefficients small with mixed sign and significance.
  - Government exposure (government securities/total assets) and asset-quality proxy (general loss reserves/gross loans): tested but statistically insignificant throughout.
- Macroeconomic controls (annual growth rates) in panel:
  - Coincident indicator: positive and significant in several FE specifications (e.g., Column (3): 0.702 (0.056)*) and in robust (Column (7)).
  - Inflation: positive and significant in multiple columns (e.g., Column (1): 0.01 (0.004)***).
  - Interest rate differential: mixed significance; p-values vary (some specifications show significance at 10 percent).
  - Advanced economy industrial production: positive but significance varies; robust estimation (column 7) shows significance: 1.553 (p-value 0.018)**.
  - GDP in GCC: positive and significant in one specification (column 4: 1.158, p-value 0.022)**.
  - Remittances from GCC and crude oil prices: generally not statistically significant in panel regressions.
- Column (7) robust estimation: seven variables statistically significant — loans-to-assets, liquidity ratio, interest margin (bank-specific); coincident indicator, inflation, interest rate differential (domestic macro); advanced economy industrial production (external macro).
- Model fit indicators: example R-squared values:
  - Column (1) 0.234; Column (2) 0.268; Column (3) 0.272.

### Conclusions and policy implications
- Empirical summary:
  - Both domestic and external variables significantly explain deposit demand in Lebanon; external variables often carry relatively higher weight (per impulse responses and variance decomposition).
  - Bank-specific factors (z-score, liquidity buffers, loan exposure, interest margins) significantly influence deposit growth at the bank level.
- Policy implications emphasized in the source:
  - Fiscal policy: pursue substantial adjustment to reduce large fiscal deficits and the debt-to-GDP ratio, thereby reducing dependence on deposit inflows.
  - Monetary policy: maintain a sufficient interest differential to the U.S. dollar to continue supporting deposit growth at comfortable rates.
  - Financial sector policy: focus on banking system soundness — encourage individual banks' stability, comfortable liquidity buffers, and allow healthy interest rate margins to support deposit attraction.
- Vulnerability highlighted:
  - High sensitivity to external factors creates a risk that adverse external conditions could slow deposit inflows and tighten government financing; Lebanon has limited control over these external variables and thus requires vigilance.

### Appendix: Macroeconomic dataset — unit root tests (selected results)
- Unit root tests (ADF and Phillips-Perron) indicate non-stationarity of variables; majority appear I(1). Selected entries (ADF statistic / Phillips-Perron statistic):
  - Deposits (LDEPOSITS_SA): 2.67 / -1.96 * ; first diff: 6.18 / -3.08 *
  - Coincident indicator (LCI_SA): 3.22 / -11.99 * ; first diff: 2.75 / -11.80 *
  - Consumer Prices (LCPI_SA): 2.41 / -3.63 * ; first diff: 3.60 / -3.55 *
  - Adv. Econ. Ind. Prod. (LADVECONIP_SA): -0.64 / -0.74 ; first diff: 1.71 / -1.46
  - G7 GDP deflator (LG7DEFL_SA): 1.73 / -0.70 ; first diff: 14.87 / -1.24
  - Real GDP in GCC (LGCC_GDP_SA): 4.02 / -1.96 * ; first diff: 5.68 / -3.33 *
  - Real non-oil GDP in GCC (LGCC_GDP_NOIL_SA): 2.36 / -0.19 ; first diff: 10.29 / -0.87
  - Goldman Sachs Risk Aversion Index (RISKAVERS): 1.35 / -4.07 * ; first diff: 1.85 / -3.97 *
- Note: despite some tests indicating I(2) for selected series (advanced econ ind. prod., G7 GDP deflator, GCC non-oil GDP), economic rationale suggests these are I(1) (quarterly percent changes stationary).

*Source: _wp09195 - 4.   Key Time Series.*

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

### References................................................................................................................................21

### Text Figures
- Fiscal Deficit and Change in Deposits.................................................................................4
- Deposit Growth, Interest Differential, and GCC Real GDP Growth...................................5

### Tables
- 1.   VECMs for Deposit Demand...............................................................................................8
- 2.   Impulse Response Functions................................................................................................9
- 3.   Variance   Decomposition....................................................................................................10
- 4.   Panel Estimations for Deposit Growth at the Bank Level .................................................13

### Figures
- 1.   VECM Actual and Fitted, Models I–III .............................................................................15
- 2.   VECM Actual and Fitted, Models IV–VI..........................................................................16
- 3.   Impulse Response Functions..............................................................................................17

*Source: _wp09195 - References..............................................................................................................*

### 4.   Key Time Series....................................................................................................

### _wp09195 - 4.   Key Time Series....................................................................................................

### Introduction
- Purpose: formally examine determinants of commercial bank deposit demand in Lebanon, assessing domestic versus external drivers using VECMs and bank-level panel regressions.
- Data: quarterly data 1993–2008 for macro models; annual bank-level panel 1997–2007 covering over 50 banks.
- Methodology highlights:
  - Vector error correction models (VECMs) to account for cointegration in non-stationary time series.
  - Johansen cointegration tests and sequential modified LR tests for lag selection.
  - Fixed-effects and robust panel regressions for bank-level analysis; bank-specific variables lagged to avoid endogeneity.
- Key variables (macro): Banque du Liban coincident indicator (CI), Consumer Price Index (CPI), interest rate differential between Lebanese pound deposit rate and U.S. dollar LIBOR (DEPATTRACT), advanced economy industrial production (LADVECONIP), GCC real GDP (LGCC_GDP), GCC non-oil GDP (LGCC_GDP_NOIL), GCC outward remittances (LGCC_REMIT), crude oil price (LOIL), Goldman Sachs Risk Aversion Index (RISKAVERS), and other global liquidity/volatility proxies (VIX, MSCIWORLD, G7DEFL, EMCREDQ, EMPSCREDG).
- Stationarity: unit root tests point to non-stationarity of all macro variables; majority I(1) (Appendix).

### Commercial bank deposits in Lebanon: context and key statistics
- Government debt-to-GDP ratio: 160 percent.
- Total commercial bank assets to GDP ratio: 324 percent.
- Interest payments: nearly 12 percent of GDP (2005–08 average).
- Overall fiscal deficits: in the order of 10 percent of GDP (2005–08 average).
- Remittances: 21 percent of GDP in 2007; origin shares — GCC 54 percent, Australia 7 percent, United States 6 percent, Canada 3 percent.
- Observed correlations (illustrative):
  - Correlation between deposit growth and GCC real GDP growth: 0.51.
  - Correlation between deposit growth and interest differential: -0.27.

### Macroeconomic estimates (VECM results and interpretation)
- Core finding: domestic and international factors both help explain deposit demand; external variables often have relatively larger impact per impulse responses and variance decompositions.
- Domestic variables with significance in some specifications: coincident indicator (LCI_SA), consumer prices (LCPI_SA), interest differential (DEPATTRACT).
- External variables with significance in various specifications: advanced economy industrial production (LADVECONIP_SA), Goldman Sachs Risk Aversion Index (RISKAVERS), GCC GDP measures, crude oil prices (LOIL), GCC outward remittances (LGCC_REMIT_SA).
- VECM specification diagnostics (selected):
  - Specification I (domestic + advanced econ ind. prod.): one cointegrating vector; error correction (ECM coefficient/CointEq1) = -0.198 (t-statistic -5.138) — "close to 20 percent of the deviation from the long-run equilibrium is corrected within one quarter."
  - Specifications II–V (adding GCC/Gulf proxies): ECM coefficients between -0.118 and -0.079 (t-stats -7.231 to -5.060) — error correction between 8 and 12 percent per quarter.
  - Specification VI (including RISKAVERS): ECM coefficient = -0.060 (t-statistic -3.653) — 6 percent per quarter.
- Selected cointegrating-vector coefficients solved for Deposits (LDEPOSITS_SA), t-statistics in italics (from Table 1):
  - Coincident indicator (LCI_SA): Specification I = 0.419 (1.692); II = -0.373 (-0.942); III = -0.457 (-0.971); IV = 0.158 (0.281); V = 0.749 (1.867); VI = 0.885 (2.872).
  - Interest differential (DEPATTRACT): I = 0.035 (2.882); II = 0.015 (1.325); III = 0.013 (1.047); IV = 0.023 (1.539); V = 0.012 (0.800); VI = 0.015 (1.233).
  - Consumer prices (LCPI_SA): I = 0.762 (2.450); II = 2.236 (7.845); III = 2.151 (6.021); IV = 2.352 (5.387); V = 2.257 (5.719); VI = 2.703 (8.794).
  - Advanced econ ind. prod. (LADVECONIP_SA): I = 3.340 (2.836).
  - Real GDP in GCC (LGCC_GDP_SA): II = 1.130 (3.393).
  - Real non-oil GDP in GCC (LGCC_GDP_NOIL_SA): III = 1.202 (1.681).
  - Crude oil price (LOIL): IV = 0.329 (3.040).
  - GCC outward remittances (LGCC_REMIT_SA): V = 0.159 (0.656).
  - GS Risk Aversion Index (RISKAVERS): VI = -0.028 (-2.068).
  - Trend and constants reported per specification; adjusted R2 values: I = 0.59; II = 0.63; III = 0.60; IV = 0.48; V = 0.47; VI = 0.49.
- Impulse Response Functions (responses of DEPOSITS_SA to one-standard-deviation shock after six periods, in percent; Table 2):
  - VECM I (after six periods): LCI_SA = 1.00; DEPATTRACT = 1.38; LCPI_SA = 0.76; LADVECONIP_SA = 3.69.
  - VECM II: LCI_SA = 0.11; DEPATTRACT = 0.26; LCPI_SA = 1.35; LGCC_GDP_SA = 1.54.
  - VECM III: LCI_SA = 0.19; DEPATTRACT = 0.30; LCPI_SA = 1.20; LGCC_GDP_NOIL_SA = 1.52.
  - VECM IV: LCI_SA = 0.78; DEPATTRACT = 0.56; LCPI_SA = 1.27; LOIL = 2.18.
  - VECM V: LCI_SA = 1.34; DEPATTRACT = -0.33; LCPI_SA = 1.10; LGCC_REMIT_SA = 1.31.
  - VECM VI: LCI_SA = 1.77; DEPATTRACT = -0.76; LCPI_SA = 2.39; RISKAVERS = -0.69.
- Variance Decomposition (percent of variance in LDEPOSITS_SA after six periods due to each variable, Table 3 — selected entries):
  - VECM I: Deposits (own shocks) = 11.23; LCI_SA = 11.85; DEPATTRACT = 7.63; LCPI_SA = 2.46; LADVECONIP_SA = 66.83.
  - VECM II: Deposits = 20.71; LCI_SA = 5.91; DEPATTRACT = 3.49; LCPI_SA = 3.07; LGCC_GDP_SA = 46.82.
  - VECM III: Deposits = 27.51; LCI_SA = 9.53; DEPATTRACT = 4.27; LCPI_SA = 19.09; LGCC_GDP_NOIL_SA = 39.60.
  - VECM IV: Deposits = 42.66; LCI_SA = 10.82; DEPATTRACT = 3.03; LCPI_SA = 9.34; LOIL = 34.15.
  - VECM V: Deposits = 42.75; LCI_SA = 25.35; DEPATTRACT = 5.89; LCPI_SA = 9.14; LGCC_REMIT_SA = 16.89.
  - VECM VI: Deposits = 13.85; LCI_SA = 32.82; DEPATTRACT = 13.08; LCPI_SA = 7.52; RISKAVERS = 2.

### Bank-specific factors (panel regressions) — key findings and coefficients
- Data: annual 1997–2007, 357 observations, 53 banks (some consolidated, some unconsolidated; outliers handled); main estimators: Random Effects (RE), Fixed Effects (FE), Robust.
- Bank-specific explanatory variables included (lagged): z-score (perceived riskiness), profitability (ROE, net interest margin), total assets (USD billion), loans-to-assets ratio, liquidity ratio (liquid assets to deposits), government securities to total assets (tested but insignificant), general loss reserves/gross loans (asset-quality proxy; insignificant).
- Consistent, statistically significant positives across specifications:
  - Loans/Assets (-1): positive and highly significant in all columns. Examples:
    - Column (1) RE: 0.524 (p-value 0.000)***
    - Column (2) FE: 0.682 (0.000)***
    - Column (3) FE with macro controls: 0.891 (0.000)***
    - Column (7) Robust: 0.14 (0.029)**.
  - Liquidity (-1): positive and highly significant in all columns. Examples:
    - Column (1): 0.332 (0.000)***; Column (7): 0.096 (0.003)***.
  - Z-Score (-1): positive and significant in several specifications. Examples:
    - Column (1): 0.003 (0.011)**; Column (2): 0.004 (0.039)**; Column (7): 0.001 (0.144) (less significant in robust).
  - Net Interest Rate Margin (-1): positive; significant in some FE/robust specifications. Examples:
    - Column (2): 0.022 (0.005)***; Column (7): 0.008 (0.011)**.
- Insignificant or mixed results:
  - Profitability (ROE) (-1): generally insignificant across columns (e.g., Column (1) 0.017 (0.498)).
  - Total assets (-1): insignificant in most specifications; small negative or positive coefficients with mixed significance.
  - Government exposure (government securities/total assets) and asset-quality proxy (general loss reserves/gross loans): tested but statistically insignificant throughout.
- Macroeconomic controls (annual growth rates):
  - Coincident indicator: positive and significant in several FE specifications (e.g., Column (3): 0.702 (0.056)*; Column (7) robust shows significance).
  - Inflation: positive and significant in multiple columns (e.g., Column (1): 0.01 (0.004)***).
  - Interest rate differential: mixed significance; p-values vary (some specifications show significance at 10 percent).
  - Advanced economy industrial production: positive but significance varies; in robust estimation (column 7) it is significant (1.553, p-value 0.018)**.
  - GDP in GCC: positive and significant in one specification (column 4: 1.158, p-value 0.022)**.
  - Remittances from GCC and crude oil prices: generally not statistically significant in panel regressions.
- Column (7) robust estimation: seven variables statistically significant — loans-to-assets, liquidity ratio, interest margin (bank-specific); coincident indicator, inflation, interest rate differential (domestic macro); advanced economy industrial production (external macro).
- Model fit indicators: R-squared ranges shown (e.g., Column (1) 0.234; Column (2) 0.268; Column (3) 0.272).

### Conclusions and policy implications
- Empirical summary:
  - Both domestic and external variables significantly explain deposit demand in Lebanon; external variables often carry relatively higher weight (per impulse responses and variance decomposition).
  - Bank-specific factors (z-score, liquidity buffers, loan exposure, interest margins) significantly influence deposit growth at the bank level.
- Policy implications emphasized in the source:
  - Fiscal policy: pursue substantial adjustment to reduce large fiscal deficits and the debt-to-GDP ratio, thereby reducing dependence on deposit inflows.
  - Monetary policy: maintain a sufficient interest differential to the U.S. dollar to continue supporting deposit growth at comfortable rates.
  - Financial sector policy: focus on banking system soundness — encourage individual banks' stability, comfortable liquidity buffers, and allow healthy interest rate margins to support deposit attraction.
- Vulnerability highlighted:
  - High sensitivity to external factors creates a risk that adverse external conditions could slow deposit inflows and tighten government financing; Lebanon has limited control over these external variables and thus requires vigilance.

### Appendix: Macroeconomic dataset — unit root tests (selected results)
- Unit root tests (ADF and Phillips-Perron) indicate non-stationarity of variables; majority appear I(1). Selected entries from Table A1 (ADF statistic / Phillips-Perron statistic):
  - Deposits (LDEPOSITS_SA): 2.67 / -1.96 * ; first diff: 6.18 / -3.08 *  (asterisk denotes rejection of unit root at 5 percent).
  - Coincident indicator (LCI_SA): 3.22 / -11.99 * ; first diff: 2.75 / -11.80 *.
  - Consumer Prices (LCPI_SA): 2.41 / -3.63 * ; first diff: 3.60 / -3.55 *.
  - Adv. Econ. Ind. Prod. (LADVECONIP_SA): -0.64 / -0.74 ; first diff: 1.71 / -1.46.
  - G7 GDP deflator (LG7DEFL_SA): 1.73 / -0.70 ; first diff: 14.87 / -1.24.
  - Real GDP in GCC (LGCC_GDP_SA): 4.02 / -1.96 * ; first diff: 5.68 / -3.33 *.
  - Real non-oil GDP in GCC (LGCC_GDP_NOIL_SA): 2.36 / -0.19 ; first diff: 10.29 / -0.87.
  - Goldman Sachs Risk Aversion Index (RISKAVERS): 1.35 / -4.07 * ; first diff: 1.85 / -3.97 *.
- Note in source: despite some test results indicating I(2) for selected series (advanced econ ind. prod., G7 GDP deflator, GCC non-oil GDP), economic rationale suggests these are I(1) (quarterly percent changes stationary).

*Source: _wp09195 - 4.   Key Time Series.*

### REFERENCES

### _wp09195 - REFERENCES

### Banking spreads, size, and performance
- Beck, Thorsten, and Heiko Hesse, 2009, “Why Are Interest Spreads So High in Uganda?,” Journal of Development Economics, Vol. 88 (March), pp. 192–204.
- Brock, Philip Lawton, and Liliana Rojas-Suárez, 2000, “Understanding the Behavior of Bank Spreads in Latin America,” Journal of Development Economics, Vo. 63 (October), pp. 113–134.
- Demirgüç-Kunt, Asli, and Harry Huizinga, 1999, “Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence,” World Bank Economic Review, Vol. 13 (May), pp. 379–408.
- De Nicoló, Gianni., 2000, “Size, Charter Value and Risk in Banking: An International Perspective,” International Finance Discussion Papers No. 689 (United States: Board of Governors of the Federal Reserve System, International Finance Division).
- Boyd, John H., and David E. Runkle, 1993, “Size and Performance of Banking Firms: Testing the Predictions of Theory,” Journal of Monetary Economics, Vol. 31 (February), pp. 47–67.
- Hesse, Heiko, and Martin Cihák, 2007, “Cooperative Banks and Financial Stability,” IMF Working Paper 07/2, (Washington: International Monetary Fund).

### Country- and region-specific studies (Lebanon, Uganda, Latin America)
- Beck, Thorsten, and Heiko Hesse, 2009, “Why Are Interest Spreads So High in Uganda?,” Journal of Development Economics, Vol. 88 (March), pp. 192–204.
- Brock, Philip Lawton, and Liliana Rojas-Suárez, 2000, “Understanding the Behavior of Bank Spreads in Latin America,” Journal of Development Economics, Vo. 63 (October), pp. 113–134.
- Poddar, Tushar, and others, 2006, “Interest Rate Determination in Lebanon,” IMF Working Paper 06/94 (Washington: International Monetary Fund).
- Schimmelpfennig, Axel, and Edward Gardner, 2008, “Lebanon—Weathering the Perfect Storms,” IMF Working Paper 08/17 (Washington: International Monetary Fund).
- Eken, Sena, and Thomas Helbling, 1999, “Back to the Future: Postwar Reconstruction and Stabilization in Lebanon,” IMF Occasional Paper No. 176 (Washington: International Monetary Fund).

### Sovereign debt and policy analyses
- Finger, Harald, and Mauro Mecagni, 2007, “Sovereign Debt Restructuring and Debt Sustainability: An Analysis of Recent Cross-Country Experience,” IMF Occasional Paper No. 255 (Washington: International Monetary Fund).

### Econometric methodology and cointegration
- Johansen, Soren, 1991, “Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models,” Econometrica, Vol. 59 (November), pp. 1551–80.
- Johansen, Soren, and Katarina Juselius,1990, “Maximum Likelihood Estimation and Inference on Cointegration, with Applications to the Demand for Money,” Oxford Bulletin of Economics and Statistics, Vol. 52 (May), pp. 169–210.

*Source: _wp09195 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09195.pdf_