## _wp14190

## Source details

**Canonical URL:** [_wp14190](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14190.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14190.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14190.pdf.json)

---

### I. Background and paper structure
- Purpose: Investigate monetary transmission mechanisms in Sri Lanka and discuss possible policy ramifications.
- Paper approach and structure:
  - Section II: Literature review.
  - Section III: Relates changes in the policy rate to changes in money market and lending and deposit rates.
  - Section IV: Data and empirical strategy.
  - Section V: Role of each channel using a Vector Autoregressive (VAR) model; includes Granger causality tests and impulse response functions (IRF).
  - Section VI: Alternative simulation comparing IRFs for two models (endogenous vs. exogenous transmission mechanism); difference measures quantitative strength of each channel.
  - Section VII: Eight-variable VAR to assess bank lending channel scope and role of Treasury bill rates in inflation formation.
  - Section VIII: Concludes with policy implications.
- Methodological augmentations:
  - Augmented by Toda and Yamamoto (1995) technique to accommodate uncertainty about the correct order of integration.
  - Builds on previous work summarized in Annex II.

### II. Key findings (main results)
- Dominant channel:
  - No other channel is as strong as the interest rate channel in Sri Lanka.
- Exchange rate and asset price channels:
  - Ineffective and do not have Granger effects on either output or prices, owing to limited movements in the exchange rate, limited foreign ownership, and less active equity market financing for borrowers.
- Bank lending channel:
  - Impacts output with a lag of about 5 quarters.
  - On inflation, the bank lending channel is effective within about 5–10 quarters.
- Quantified impulse-response effects to repo-rate innovations:
  - Output declines by about 0.6 percent in the second quarter after an innovation to the repo rate.
  - Output declines by about 0.5 percent during the entire period of nearly three years after repo-rate innovations.
  - The repo-rate effect on prices is about 0.2 percent via the money market rate with a lagged response embedded in the nominal interest rate (price puzzle).
- Additional quantified findings:
  - The baseline policy-rate shock produced about a 0.5 percent decrease in output; the bank lending channel adds another 0.2 percentage point to this baseline, making the total impact 0.3 percent during the three-year period.
  - A one percent shock to the oil price will increase the price level by about 0.9 percent within 10 quarters.
  - Variance decomposition: almost 6 percent of the change in GDP is explained by the variance in the policy rate within a three-year period.
  - For inflation, money supply explains about 4 percent of fluctuations; the rest is explained by inflation inertia.
  - Variance decomposition (expanded VAR with 12-month Treasury bill rate): 20 percent of variance in GDP and inflation explained by changes in money supply (12 percent) and 12-month Treasury bill rate (8 percent).

### III. Methodology and empirical strategy
- Identification and techniques:
  - Uses VAR models including specific variables for each channel (e.g., stock of bank credit or bank lending rate).
  - Significance and timing tested via Granger causality and impulse response functions (IRF).
  - Alternative simulation follows Ramey (1993) and Disyatat and Vongsinsirkul (2003) to compare IRFs under endogenous vs. exogenous transmission mechanisms.
  - Robustness to integration order via Toda–Yamamoto (1995): estimate VAR with p + k lags (k = maximum order of integration) and test first p lags.
- Data and testing:
  - Sample: quarterly seasonally adjusted data from 2000Q1 to 2013Q3; number of observations: 54 quarterly observations.
  - Data source: IMF’s International Financial Statistics (IFS).
  - Variables: natural logarithms except interest rates (in level form); seasonally adjusted by multiplicative and additive MA or AR terms; seasonal dummies where observed.
  - Unit-root and DGP concerns: distinction between I(1) and trend stationary processes; ADF, PP, NP (Ng-Perron), KPSS, ERS used; structural-break tests (Perron/Chow) applied where appropriate.
  - Data-inspection: select DGP by autocorrelation and lag length, respecify if misspecified, conduct unit root tests on selected DGP; conclusions reported in Annex I, Table 3.
- Baseline VAR specification:
  - Five-variable VAR (2000Q1–2013Q3): output, prices, money supply, interest rates, exchange rates; variables in levels or first-differences per stationarity guidance.
  - Reduced-form VAR: Yt = A0 + A1 Yt−1 + · · · + Ak Yt−k + ut, with ut serially uncorrelated disturbances and k lags.
- Identification of monetary policy shocks:
  - Choleski decomposition with policy variables ordered after nonpolicy variables.
  - Policy-variable ordering: money supply first, policy rate second, exchange rate third.
  - Alternative orderings and reserve money experimented with; sensitivities discussed in Section V.

### IV. Channel-specific empirical evidence and statistics
- Interest rate channel (money view):
  - Identified as the strongest channel; policy rate has significant predictive value for output.
  - Output responses: about 0.6 percent decline in second quarter post-repo shock; about 0.5 percent decline over nearly three years.
  - Policy variables jointly significant for output and prices.
- Bank lending channel (credit view):
  - Operational on output with impact around 5 quarters; on prices around 5–10 quarters; effects weak and significantly lagged.
  - Granger findings:
    - Weak Granger causality between private credit and output; stronger Granger relationship between the prime lending rate and output.
    - Private credit contributes about 0.2 percent to the interest rate channel starting in the second quarter (in the model with exchange rates).
    - A policy tightening that reduces private credit associated with a 0.7 percent reduction in output starting from the second quarter when reduction in private credit is associated with real exchange rate appreciation.
    - The prime rate reduces output by about 0.1 percent more after about five quarters.
  - Variance decomposition: about 8 percent of GDP shocks explained by the prime lending rate; about 6 percent by the policy rate; about 7 percent by private credit.
  - No Granger causality from bank credit or lending rate on core and headline inflation in the bank-lending-channel model.
  - Limiting factors: slow pass-through; banks purchasing securities rather than increasing lending; limited competition; off-balance-sheet lending.
- Exchange rate channel:
  - Weak under limited floating and constrained capital mobility.
  - No Granger effect of the exchange rate on output or prices in basic specifications.
  - Including exchange rate augments bank lending channel influence on output by about 0.2 percent.
  - Variance decomposition: about 8 percent of output fluctuations explained by exchange rate changes (in models including exchange rate).
  - Exchange rate strongest impulse from 12-month Treasury bill rate model; cumulative effects fade after about 4 quarters.
- Asset price channel:
  - Proxied by the Colombo stock exchange index.
  - No meaningful impact on output and prices in the estimated models.
  - Inclusion does not change that policy rate reduces output by about 0.5 percent over the shock period.
- Pass-through and contemporaneous coefficients (examples):
  - Policy rate -> money market rate contemporaneous pass-through in Sri Lanka: 0.35 percentage point increase in money market rate for a one percentage point increase in policy rate.
  - Money market rate -> lending rates: 0.68 percentage point increase for a one percentage point increase in money market rate.
  - Money market rate -> deposit rates: 0.19 percentage point increase for a one percentage point increase in money market rate.
  - Comparator (Malaysia) contemporaneous pass-through:
    - Policy rate -> money market rate: 0.94 percentage point.
    - Money market rate -> lending rate: 0.96 percentage point.
    - Money market rate -> deposit rate: 0.98 percentage point.
- Long-term bond yields:
  - Impact on long-term bond yields is weak; some perverse contemporaneous relationship between policy rates and bond yields that dissipates when using the money market repo rate; long-run meaningful impact observed for 3, 5, and 10-year maturities.

### V. VAR diagnostics, Granger causality and selected test statistics
- Sample: 2000Q1 to 2013Q3 for core exercises; also run for 1995Q1 to 2013Q3 with no major differences.
- Data summary (quarter-on-quarter difference in logarithm; seasonally adjusted):
  - Repo rate: Mean 9.444, Median 8.581, Maximum 19.337, Minimum 2.776
  - Three-month T-bill rate: Mean 11.321, Median 10.12, Maximum 21.36, Minimum 6.98, Standard Deviation 3.892
  - Twelve-month T-bill rate: Mean 11.882, Median 11.17, Maximum 20.21, Minimum 7.14, Standard Deviation 3.891
  - Prime rate: Mean 13.69, Median 12.82, Maximum 22.24, Minimum 8.94, Standard Deviation 3.762
  - ∆ CPI: Mean 0.024, Median 0.021, Maximum 0.065, Minimum -0.021, Standard Deviation 0.017
  - ∆ Real GDP: Mean 0.013, Median 0.010, Maximum 0.087, Minimum -0.105, Standard Deviation 0.031
  - ∆ Reserve money: Mean 0.029, Median 0.029, Maximum 0.095, Minimum -0.079, Standard Deviation 0.031
  - ∆ Broad money: Mean 0.037, Median 0.038, Maximum 0.070, Minimum 0.001, Standard Deviation 0.013
  - ∆ Overall credit: Mean 0.038, Median 0.044, Maximum 0.124, Minimum -0.135, Standard Deviation 0.04
  - ∆ Private credit: Mean 0.037, Median 0.040, Maximum 0.107, Minimum -0.025, Standard Deviation 0.027
- Selected Granger causality excerpts (F-stat and p-values preserved as in source):
  - Baseline (Table 7 and 8 excerpts):
    - Money supply 1/: F = 1.21, p = 0.31
    - Interest rate (repo): F = 3.18*, p = 0.05
    - Prime lending rate 2: F-stat and p-values indicate significance (source table formatting indicates prime lending rate significance for output)
    - Private credit: F = 2.82*, p = 0.07
    - Policy variables jointly: F = 3.7***, p = 0.00
  - Expanded VAR (Table 10 excerpts; impact on output):
    - Money supply 1/: 1.21 0.31
    - Interest rate (repo): 3.18* 0.05
    - Exchange rate: 0.15 0.86
    - Private credit: 2.82* 0.07
    - Policy variables jointly: 1.62 0.12
  - Expanded VAR with 12-month T-bill (Table 11 excerpts; selected):
    - Impact on output:
      - Money supply 1/: 3.08** 0.04
      - Interest rate (repo): 1.47 0.24
      - Twelve-month Treasury bill 1: 20.41 0.75 5.85*** 0.00
      - Policy variables jointly: 3.22*** 0.01
    - Impact on prices:
      - Interest rate (repo): 2.32* 0.09
      - Twelve-month Treasury bill 12: 4.08** 0.01
      - Policy variables jointly: 2.82*** 0.01
  - Notes:
    - 1/ Defined as broad money (M2b). Retested with reserve money with no significant difference in findings.
    - 2/ Only the prime lending rate has a Granger effect on output; weighted average lending rate was not significant with private credit included.
    - Critical values and significance codes reported in source tables (*** at 1%; ** at 5%; * at 10%).

### VI. VAR experiments: exogeneity tests and channel blocking
- Exercise: Evaluate each channel with and without being endogenized in the VAR; compare IRFs.
- Findings:
  - Interest rate channel is the single strongest channel.
  - Bank lending channel meaningful for output at 5 quarters and for prices at 5–10 quarters when endogenized; its contribution to policy-rate affecting inflation is strongest among channels when endogenized.
  - Exchange rate and asset price channels show limited marginal contributions when treated endogenously versus exogenously.
- Interpretation:
  - When banks redirect policy impulses into securities markets (owing to excess liquidity), policy impact on output is dampened and inflation effects via market rates may strengthen.
  - Longer convergence between policy and market rates implies policy-rate or money-supply effects on inflation are observable only in later periods when rates converge.

### VII. Policy-relevant implications and recommendations
- Monetary policy emphasis:
  - Given dominance of the interest rate channel, policy rate adjustments remain the primary conduit for influencing output and prices.
- Strengthening the bank lending channel and pass-through:
  - Develop alternative sources of financing, such as the capital market.
  - Reduce the role of state banks in the financial system to avoid pitfalls associated with state banking, including lack of trust in transparency.
  - Create avenues to connect bank financing with the real economy.
- Strengthening the exchange rate channel:
  - Limit foreign exchange interventions to smoothing excessive fluctuations and allow flexibility to the Sri Lankan rupee.
  - Gradually increase the degree of capital mobility and allow more transparent and active foreign participation in domestic secondary securities and deposit markets, consistent with macroeconomic and financial stability.
  - Further develop a competitive export base to increase the value added of the exchange rate channel.
- Making the asset-price channel operational:
  - Adopt institutional reforms to increase transparency and entry into equity and property markets for both residents and nonresidents.
  - Encourage participation in asset markets so nonbanking assets grow as a proportion of total wealth, reducing dependence on bank financing and enhancing the asset price channel.
- Operational note:
  - If fully operational, bank lending, exchange rate, and asset price channels can significantly contribute to the effectiveness of monetary policy transmission in Sri Lanka.

*Source: _wp14190 (extracted content from the supplied IMF staff paper).*

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

### _wp14190 - References

### I. Background and paper structure
- Purpose: Investigate monetary transmission mechanisms in Sri Lanka and discuss possible policy ramifications.
- Paper approach and structure:
  - Section II: Literature review.
  - Section III: Examines first steps in the transmission mechanism—relating changes in the policy rate to changes in money market and lending and deposit rates.
  - Section IV: Discusses the data and empirical strategy.
  - Section V: Assesses the role of each channel using a Vector Autoregressive (VAR) model; includes Granger causality tests and impulse response functions (IRF).
  - Section VI: Alternative simulation comparing IRFs for two models of monetary policy impact on output and prices; one model permits each channel to respond endogenously to a monetary policy shock, the other treats the transmission mechanism as an exogenous variable. The difference between the two IRFs measures the quantitative strength of each transmission channel.
  - Section VII: Uses an eight-variable VAR model to assess the scope of the bank lending channel in neutralizing policy signals on output and the role of Treasury bill rates in inflation formation.
  - Section VIII: Concludes with policy implications.
- Methodological augmentations:
  - Augmented by Toda and Yamamoto (1995) technique to accommodate uncertainty about the correct order of integration.
  - Builds on previous work summarized in Annex II.

### II. Key findings (main results)
- Dominant channel:
  - No other channel is as strong as the interest rate channel in Sri Lanka.
- Exchange rate and asset price channels:
  - Ineffective and do not have Granger effects on either output or prices, owing to limited movements in the exchange rate, limited foreign ownership, and less active equity market financing for borrowers.
- Bank lending channel:
  - Has an impact on output but with a lag of about 5 quarters.
  - On inflation, the bank lending channel is effective within about 5–10 quarters.

### III. Methodology and empirical strategy
- Identification and techniques:
  - Uses VAR models to include specific variables for each channel (for example, stock of bank credit or bank lending rate to assess bank lending channel).
  - Significance and timing tested via Granger causality and impulse response functions (IRF).
  - Alternative simulation approach follows Ramey (1993) and Disyatat and Vongsinsirkul (2003) to compare IRFs under endogenous vs. exogenous transmission mechanisms.
- Data and testing:
  - Annex I: Data Generation Process and Unit Root Test Results (listed in contents).
  - Comparison with previous work provided in Annex II.

### IV. Figures and tables (inventory of analytic materials listed)
- Selected figures (titles preserved):
  - 1. Sri Lanka: Ratio of Bank Loans to GDP
  - 2. Sri Lanka: Selected Financial Indicators
  - 3. Sri Lanka: Banks’ Liquid Asset Holdings
  - 4. Sri Lanka: Selected Financial Indicators
  - 5. Sri Lanka: Banks’ Liabilities
  - 6. Sri Lanka: Repo and Prime Lending Rates
  - 7. Sri Lanka: Pass-through from Repo Rate to Money Market Rate
  - 8. Sri Lanka: Pass-through from Money Market to Prime Lending Rates
  - 9. Sri Lanka: Pass-through from Repo Rate to Lending-Deposit Spread
  - 10. Sri Lanka: Pass-through from Money Market Rate to Lending-Deposit Spread
- Selected tables (titles preserved):
  - 1a. Sri Lanka: Correlations Between Changes in Repo Rate and Changes in Money Market and Securities Returns
  - 1b. Sri Lanka: Correlations Between Changes in Repo Rate and Changes in Money Market and Securities Returns
  - 2. Sri Lanka: Correlations Between Changes in the Money Market Rates and Changes in the Lending and Deposit Rates
  - 3. Sri Lanka: Impact of Policy and Money Market Rates on Long-term Domestic Rates
  - 4. Selected Asian Countries: Correlations Between Changes in the Repo Rate and Changes in Money Market and Securities Returns
  - 5. Selected Asian Countries: Correlations Between Changes in the Money Market Rates and Changes in the Lending and Deposit Rates
  - 6. Sri Lanka: Baseline Model
  - 7. Sri Lanka: Credit Channel
  - 8. Sri Lanka: Credit Channel
  - 9. Sri Lanka: Asset Price Channel
  - 10. Sri Lanka: Expanded VAR Model
  - 11. Sri Lanka: Expanded VAR Model with 12-Month Treasury Bill Rate

### V. Policy-relevant implications (derived from results)
- Monetary policy emphasis:
  - Given the dominance of the interest rate channel, policy rate adjustments remain the primary conduit for influencing output and prices.
- Bank lending considerations:
  - The bank lending channel affects output with a lag of about 5 quarters and inflation within about 5–10 quarters, implying monetary policy effects on credit take several quarters to materialize in the real economy and prices.
- Exchange rate and asset market interventions:
  - Limited effectiveness of exchange rate and asset price channels suggests these tools are less reliable transmitters of monetary policy in the Sri Lankan context due to structural market characteristics.

*Source: _wp14190 - References (content extracted from the provided PDF content unit).*

### Section VI). Later, Alam and Waheed (2006) also used recursive VARs both at the aggregate

### _wp14190 - Section VI)

### Current challenges in monetary policy transmission
- Intermediate target narrowed to a monetary aggregate (with reserve money serving as the operating target); main policy instruments are (a) policy interest rates (interest rates on overnight repurchase and reverse repurchase agreements) and open market operations (OMO) and (b) the statutory reserve requirement (SRR) on commercial bank deposit liabilities (Central Bank of Sri Lanka, 2013).
- Foreign exchange operations and liquidity management associated with the issuance of international bonds have indirectly become part of monetary policy tools.
- Banks are the dominant financial intermediaries in Sri Lanka; the bank lending channel should be the main vehicle for transmission but is constrained by rigidities, imperfect links between policy rates and domestic and international capital markets, imperfect signals in the term structure, and lack of competition.

Key observed frictions and statistics
- Ratio of bank loans to GDP: the average has remained close to 25 percent over time; post-conflict average ratio is slightly higher reaching 27 percent.
- Banks’ asset composition has shifted toward holding government securities (“excess liquidity”), reducing incentives to lend despite policy rate increases.
- Share of deposits has not much increased in the last 15 years while the share of demand deposits in total deposits has significantly declined recently.
- Weak correlation and slow reaction of market and bank rates to policy changes; examples of contemporaneous pass-through:
  - An increase in the policy rate by one percentage point increases the money market rate by only 0.35 percentage point.
  - An increase in the money market rate by one percentage point increases lending rates by 0.68 percentage point and deposit rates by 0.19 percentage point.
- Comparator (Malaysia) contemporaneous pass-through for comparison:
  - Policy rate -> money market rate: 0.94 percentage point.
  - Money market rate -> lending rate: 0.96 percentage point.
  - Money market rate -> deposit rate: 0.98 percentage point.
- Longer-term effects in Sri Lanka: spread narrows over time so longer-term effects on lending rates can be more significant than contemporaneous effects.
- Impact on long-term bond yields is weak: some perverse contemporaneous relationship between policy rates and bond yields that dissipates when using the money market repo rate; only long-run meaningful impact observed for 3, 5, and 10-year maturities.

Observed dynamics and recent shifts
- Significant recent shifts in pass-through from policy rates to money market and to lending and deposit rates (Figures 7 and 8).
- Money market rates magnify the spread between lending and deposit rates recently; deposit rates change negligibly as banks use abundant excess liquidity to offset policy changes.
- Examples from impulse-response / correlation tables:
  - Tables show contemporaneous, short-term, and long-term correlations between changes in repo rate (or call market rate) and changes in various money market, lending, deposit, and Treasury bill rates (specific coefficients are in Tables 1a, 1b, 2, and 3 as presented).

### Data, sample, and inspection strategy
- Sample frequency and period: quarterly seasonally adjusted data from 2000Q1 to 2013Q3.
- Number of observations: 54 quarterly observations noted as relevant to unit root test power considerations.
- Data source: International Monetary Fund’s (IMF) International Financial Statistics (IFS).
- Data transformation and seasonality:
  - Variables expressed in natural logarithms except interest rates (in level form).
  - Seasonally adjusted by multiplicative and additive MA or AR terms; seasonal dummy variables added where seasonality observed.
- Stationarity and DGP considerations:
  - Distinction between I(1) unit-root processes and trend stationary processes (TSP) emphasized.
  - Traditional unit root tests (ADF, PP) have low power in short samples; Christiano and Ljungqvist (1988) caution against unnecessary differencing.
  - Structural breaks (e.g., 2008–09 global financial crisis) can bias unit-root test conclusions; P-break tests and Chow tests used where appropriate.
- Data-inspection procedure:
  - Select candidate DGP equations based on autocorrelation behavior and lag length, include seasonal dummies where needed.
  - Respecify if misspecification detected; conduct unit root tests on the “true” DGP equations selected using AIC, SBC, lowest SSR, and Q statistics.
  - Conclusions on differencing for baseline VAR reported in Annex I, Table 3.

### Methodology and identification
- Baseline approach: Sims (1972) Granger-causality framework with policy variables ordered after nonpolicy variables (policy variables determined on knowledge of contemporaneous shocks to output and prices; output and prices respond with a lag).
  - Advantages stated:
    - Characterizes stylized facts about policy–output–price relationships in Sri Lanka.
    - Requires minimal a priori assumptions; includes three measures of monetary policy tools (money supply, interest rate, exchange rate).
- Baseline VAR specification:
  - Five-variable VAR from 2000Q1 to 2013Q3: output, prices, money supply, interest rates, exchange rates; variables are in levels or first-differences per stationarity test guidance.
  - Reduced-form VAR notation provided: Yt = A0 + A1 Yt−1 + · · · + Ak Yt−k + ut, with ut serially uncorrelated disturbances and k lags.
- Robustness to integration order:
  - Toda–Yamamoto (1995) modification used to accommodate uncertain order of integration: estimate VAR with p + k lags (k = maximum order of integration) and test first p lags.
  - Test statistic has a χ2 asymptotic distribution but power is reduced by inclusion of k extra lags.
- Identification of monetary policy shocks:
  - Identify shocks via Choleski decomposition assuming lower-triangular θ; policy variables ordered after nonpolicy variables.
  - Policy-variable ordering: money supply first, policy rate second, exchange rate third to reflect degrees of endogeneity.
  - Alternative orderings and replacement of broad money with reserve money were experimented with; sensitivities discussed in Section V.

### Baseline VAR results: interest rate channel (money view)
- The money view posits that reducing money supply (and increasing policy rate) reduces investment and output via cost of capital and affects durable consumption.
- Empirical findings:
  - The policy rate has significant predictive value for output in Sri Lanka.
  - Money supply weakly Granger-causes (helps predict) prices (Table 6).
  - Quantified impulse-response effects to repo-rate innovations:
    - Output declines by about 0.6 percent in the second quarter after an innovation to the repo rate.
    - Output declines by about 0.5 percent during the entire period of nearly three years after repo-rate innovations.
    - The repo-rate effect on prices is about 0.2 percent via the money market rate with a lagged response embedded in the nominal interest rate (price puzzle).
  - Policy variables jointly have significant predictive value for both output and prices.
- Additional notes:
  - Using market repo rates in lieu of announced repo rates for 2004Q3 to 2013Q3 increases the impact of direct money supply on output relative to the policy rate itself (footnote observation).
  - Including the U.S. refinancing rate in the model reduces the impact of the policy rate on output (not fully quantified in the supplied extract).

*Source: _wp14190 - Section VI).*

### 0.2 percentage point, making the total impact 0.3 percent during the  three-year period.

### _wp14190 - 0.2 percentage point, making the total impact 0.3 percent during the  three-year period.

### Key empirical findings: overall transmission and shocks
- A one percent shock to the oil price will increase the price level by about 0.9 percent within 10 quarters.
- The variance decomposition shows that almost 6 percent of the change in GDP is explained by the variance in the policy rate within a three-year period.
- For inflation, the money supply explains about 4 percent of fluctuations and the rest is explained by inflation inertia.
- The baseline policy-rate shock produced about a 0.5 percent decrease in output; the bank lending channel adds another 0.2 percentage point to this baseline, making the total impact 0.3 percent during the three-year period.

### Interest rate channel (money view)
- Identified as the strongest monetary policy channel in Sri Lanka.
- Directly affects decision making of economic agents and is the dominant contributor to transmission in impulse-response comparisons (interest rate channel versus all channels).
- Policy-rate effects on inflation and output show notable timing: significant effects manifest over multiple quarters (plots indicate responses over at least 14 quarters).

### Bank lending channel (component of the credit view)
- Described as banks’ response (via supply of loanable funds) to changes in deposit base or policy rate.
- Operational on output with impact around 5 quarters and on prices around 5–10 quarters, but effects are weak and significantly lagged.
- Observations and quantitative results:
  - Weak Granger causality between private credit and output; stronger Granger relationship between the prime lending rate and output (Tables 7 and 8).
  - Private credit contributes to the interest rate channel by about 0.2 percent starting in the second quarter (only in the model with exchange rates).
  - A policy tightening that reduces private credit is associated with a 0.7 percent reduction in output starting from the second quarter when reduction of private credit is also associated with real exchange rate appreciation.
  - The prime rate reduces output by about 0.1 percent more after about five quarters.
  - Variance decomposition: about 8 percent of GDP shocks are explained by the prime lending rate; about 6 percent by the policy rate; about 7 percent by private credit.
  - No Granger causality from either bank credit or the lending rate on core and headline inflation in the bank-lending-channel model.
- Factors limiting bank lending channel effectiveness:
  - Slow pass-through in Sri Lanka.
  - Banks attracting external funds may purchase securities rather than increase lending.
  - Degree of competition among banks affects pass-through; in a noncompetitive environment banks may not lower loan rates when costs fall.
  - Off-balance-sheet lending and other bank activities have weakened the channel.

### Credit channel—Granger causality and statistics (Tables 7 and 8 excerpts)
- Impact on output (selected F-stat and p-values):
  - Money supply 1/: F = 1.21, p = 0.31
  - Interest rate (repo): F = 3.18*, p = 0.05
  - Prime lending rate 2: F = /, p = 3.58**, p = 0.04 (note: table formatting in source shows prime lending rate F-stat and p-values indicating significance)
  - Private credit: F = 2.82*, p = 0.07
  - Policy variables jointly: F = 3.7***, p = 0.00
- Impact on prices (selected F-stat and p-values):
  - Money supply 1: F = 1.42, p = 0.25
  - Interest rate (repo): F = 2.15, p = 0.13
  - Prime lending rate: F = 0.77, p = 0.47
  - Private credit: F = 0.47, p = 0.63
  - Policy variables jointly: F = 2.64**, p = 0.06
- Notes from tables:
  - 1/ Defined as broad money (M2b). Tests with reserve money did not change main findings.
  - 2/ Only the prime lending rate has a Granger effect on output; weighted average lending rate was not significant with private credit included.

### Exchange rate channel
- Under limited floating and constrained capital mobility, exchange rate channel is weak on its own.
- No Granger effect of the exchange rate on either output or prices in basic specifications.
- Presence of the exchange rate in the model with the bank lending channel augments its influence on output by about 0.2 percent.
- Variance decomposition: about 8 percent of output fluctuations are explained by changes in the exchange rate (in models including exchange rate).
- The exchange rate responds to changes in short-term rates on government bonds; cumulative effects fade after about 4 quarters and impact on output fades even sooner.
- For the exchange rate channel, strongest impulse on the REER is received from the model with the 12-month Treasury bill rate (tested with 3, 6, and 12-month T-bill rates and money market rates).

### Asset price channel
- Proxied by the Colombo stock exchange index.
- Has no meaningful impact on output and prices in the estimated models.
- Inclusion of the asset price channel did not change earlier results where the policy rate reduced output by about 0.5 percent during the entire shock period.

### VAR models, excess liquidity, and Treasury bill effects
- Expanded VARs include real GDP, CPI, money supply (or reserve money), policy interest rate, outstanding deposits (or securities in banking system), 12-month Treasury bill rate, exchange rate, and private credit.
- Impact of policy rates on output can be dampened when bank lending channel is muted and policy impulses land in securities markets.
- Granger causality findings:
  - Policy rate shows a significant Granger effect on output when deposits are available in the model.
  - Without deposits in the model, the policy rate lacks a Granger effect; money supply still weakly Granger-causes output.
  - The 12-month Treasury bill rate Granger-causes prices: a one percent reduction in the 12-month Treasury bill rate will increase prices by about 0.2 percent within two quarters.
- Variance decomposition:
  - 20 percent of variance in GDP and inflation is explained by changes in money supply (12 percent) and 12-month Treasury bill rate (8 percent).
- Interpretation:
  - Longer convergence between policy and market rates and the stronger role of Treasury bill and money market rates imply that policy-rate or money-supply effects on inflation are observable only in later periods when rates converge.
  - Purchases of securities by banks increase liquidity and thereby affect inflation dynamics through market rates.

### Exogeneity tests and channel blocking (Section VI)
- Each monetary policy channel was evaluated with and without being endogenized in the VAR (baseline period 2000Q1 to 2013Q3; exercise also run for 1995Q1 to 2013Q3 with no major differences).
- Blocking individual channels and comparing IRFs indicates:
  - Interest rate channel is the single strongest channel.
  - Bank lending channel shows meaningful contributions to output at 5 quarters and to prices at 5–10 quarters when endogenous; its contribution to policy rate affecting inflation is strongest among channels when endogenized.
  - Exchange rate and asset price channels show limited marginal contributions when treated endogenously versus exogenously.

### Conclusions and policy implications (Section VIII)
- The interest rate channel is the most important monetary policy transmission channel in Sri Lanka.
- The roles of the bank lending, exchange rate, and asset price channels should be strengthened going forward.
- If fully operational, each of these channels can significantly contribute to the effectiveness of the monetary policy transmission mechanism in Sri Lanka.
- Quantified example: the bank lending channel adds another 0.2 percentage point to the baseline 0.5 percent decrease in output in response to innovations in the policy rate.

*Source: IMF staff calculations, from the supplied IMF staff paper content.*

### 0.2 percent reduction of output to the increase in the policy rate.

### _wp14190 - 0.2 percent reduction of output to the increase in the policy rate.

### Key empirical findings on monetary transmission
- Estimated accumulated response: "0.2 percent reduction of output to the increase in the policy rate."
- Evidence on channels:
  - Investment channel (policy rates → decision-making of economic agents) identified as the strongest channel in Sri Lanka.
  - Interest rate and credit channels are important; bank lending channel shows weaknesses.
  - Money growth weakly Granger-causes inflation in baseline tests.
  - In an expanded model with securities and Treasury bill rates, "20 percent of inflation is explained by changes in money growth and Treasury bill rates."
- Banks’ behavior may neutralize policy actions via counteractive transactions in the government securities market, potentially redirecting policy effects from output to inflation.
- The paper found no difference between results using CPI and core inflation.

### Granger causality and VAR results (selected statistics preserved)
- Table 10. Sri Lanka: Expanded VAR Model (impact on output; F-stat and p-values):
  - Money supply 1/ : 1.21 0.31
  - Interest rate (repo): 3.18* 0.05
  - Exchange rate: 0.15 0.86
  - Securities: 0.20 0.82 0.21 0.81
  - Private credit: 2.82* 0.07 1.89 0.16
  - Deposits: 0.93 0.40 0.02 0.98
  - Policy variables jointly: 1.62 0.12
- Table 10. Impact on prices (F-stat and p-values):
  - Money supply 1/: 1.42 0.25
  - Interest rate (repo): 2.15 0.13
  - Exchange rate: 0.18 0.84
  - Securities: 1.63 0.21
  - Private credit: 0.47 0.63
  - Deposits: 0.13 0.88
  - Policy variables jointly: 1.57 0.13
- Notes for Table 10:
  - p values show the significance of lagged monetary policy variables for output and prices
  - F statistic at (2, 36)
  - Critical values: 5.28 at 1% level; 3.26 at 5% level; 2.46 at 10% level
  - *** significant at 1%; ** significant at 5%; * significant at 10%
  - The optimal lag is selected based on LR=2
  - F statistic for joint significance is at (14, 37)
  - Critical values are 2.62 at 1% level, 1.97 at 5% level and 1.69 at 10% level
  - 1/ Defined as broad money (M2b). Retested with reserve money with no significant difference in findings.
  - Source: IMF staff calculations.
- Table 11. Sri Lanka: Expanded VAR Model with 12-Month Treasury Bill Rate (selected):
  - Impact on output (F-stat and p-values):
    - Money supply 1/: 3.08** 0.04
    - Interest rate (repo): 1.47 0.24
    - Exchange rate: 0.78 0.51
    - Securities: 0.36 0.78 2.04 0.12
    - Private credit: 2.12 0.11 1.51 0.23
    - Twelve-month Treasury bill 1: 20.41 0.75 5.85*** 0.00
    - Policy variables jointly: 3.22*** 0.01
  - Impact on prices (F-stat and p-values):
    - Money supply 1/: 0.97 0.41
    - Interest rate (repo): 2.32* 0.09
    - Exchange rate: 0.51 0.68
    - Private credit: 1.32 0.28
    - Twelve-month Treasury bill 12: 4.08** 0.01
    - Policy variables jointly: 2.82*** 0.01
  - Notes for Table 11:
    - p values show the significance of lagged monetary policy variables for output and prices
    - F statistic at (3, 28)
    - Critical values: 4.57 at 1% level; 2.95 at 5% level; 2.29 at 10% level
    - *** significant at 1%; ** significant at 5%; * significant at 10%
    - The optimal lag is selected based on LR=3
    - F statistic for joint significance is at (21, 29)
    - Critical values are 2.58 at 1% level, 1.95 at 5% level and 1.69 at 10% level
    - 1/ Defined as broad money (M2b). Retested with reserve money with no significant difference in findings.
    - Source: IMF staff calculations.

### Policy recommendations to strengthen transmission
- To address weaknesses in the bank lending channel and increase short-run pass-through between policy rates and market/lending rates:
  - Develop alternative sources of financing, such as the capital market.
  - Reduce the role of state banks in the financial system to avoid pitfalls associated with state banking, including lack of trust in transparency.
  - Create avenues to connect bank financing with the real economy.
- To strengthen the exchange rate channel:
  - Limit interventions in the foreign exchange market to smoothing excessive fluctuations and allow flexibility to the Sri Lankan rupee.
  - Gradually increase the degree of capital mobility and allow more transparent and active foreign participation in domestic secondary securities and deposit markets, consistent with Sri Lanka’s macroeconomic and financial stability.
  - Further develop a competitive export base to increase the value added of the exchange rate channel.
- To make the asset channel operational:
  - Adopt institutional reforms to increase transparency and entry into equity and property markets for both residents and nonresidents.
  - Encourage participation in asset markets so nonbanking assets grow as a proportion of total wealth, reducing dependence on bank financing and enhancing the asset price channel.

### Data, methods, and diagnostics (selected)
- Data sample for descriptive statistics: "The sample covers 2000Q1 to 2013Q3."
- Descriptive statistics (quarter-on-quarter difference in logarithm; seasonally adjusted):
  - Repo rate: Mean 9.444, Median 8.581, Maximum 19.337, Minimum 2.776
  - Three-month T-bill rate: Mean 11.321, Median 10.12, Maximum 21.36, Minimum 6.98, Standard Deviation 3.892
  - Twelve-month T-bill rate: Mean 11.882, Median 11.17, Maximum 20.21, Minimum 7.14, Standard Deviation 3.891
  - Prime rate: Mean 13.69, Median 12.82, Maximum 22.24, Minimum 8.94, Standard Deviation 3.762
  - ∆ CPI: Mean 0.024, Median 0.021, Maximum 0.065, Minimum -0.021, Standard Deviation 0.017
  - ∆ Real GDP: Mean 0.013, Median 0.010, Maximum 0.087, Minimum -0.105, Standard Deviation 0.031
  - ∆ Reserve money: Mean 0.029, Median 0.029, Maximum 0.095, Minimum -0.079, Standard Deviation 0.031
  - ∆ Broad money: Mean 0.037, Median 0.038, Maximum 0.070, Minimum 0.001, Standard Deviation 0.013
  - ∆ Overall credit: Mean 0.038, Median 0.044, Maximum 0.124, Minimum -0.135, Standard Deviation 0.04
  - ∆ Private credit: Mean 0.037, Median 0.040, Maximum 0.107, Minimum -0.025, Standard Deviation 0.027
- Unit root and model diagnostics:
  - Models describing DGP and model selection reported with AIC, Schwarz, SSR, and Ljung-Box Q(p) statistics for residuals.
  - Table 2 summary of stationarity tests indicates which series should be differenced; CPI is reported as stationary in the tests and thus "should not be differenced in the model," though the paper also runs VARs with differenced CPI.
  - Tests used: ADF, PP, NP (Ng-Perron), KPSS, ERS (DF-GLS with structural break Perron).

### Comparison with previous literature
- Findings align broadly with previous work that interest rate and credit channels are dominant; investment channel emphasized as strongest.
- Previous studies:
  - Fernandez and others (2004): found shocks to output and inflation explain most variability; downplayed bank lending channel importance.
  - Perera and Wickramanayake (2013): found prices responsive to monetary shocks but presence of a "price puzzle".
  - Ratnasiri (2009): money growth affects inflation in the long run; output gap not significant for inflation.
- This paper:
  - Finds money growth weakly Granger-causes inflation.
  - Finds in expanded models a larger role for Treasury bill and money market rates in explaining inflation.
  - Highlights timing and pass-through issues: slow adjustment of money and other rates to policy rates implies long-run inflation impact may operate through those rates rather than directly from policy variables.

*Source: IMF staff calculations from _wp14190 - 0.2 percent reduction of output to the increase in the policy rate.*

### References

### _wp14190 - References

### Major themes and topics covered
- Monetary transmission mechanisms and channels
  - Credit channel and financial accelerator: Bernanke (1993a), Bernanke and Gertler (1995), Bernanke, Gertler, and Gilchrist (1994), Gertler and Gilchrist (1993, 1994).
  - Bank lending channel and monetary policy and bank lending: Kashyap and Stein (1994); Perera, Ralston, and Wickramanayake (2014).
  - Money, interest rates, and aggregate demand: Bernanke and Blinder (1988); Christiano and Ljungqvist (1988); Boschen and Mills (1992).
  - Inflation indicators and inflation expectation formation: Cecchetti (1995); Pfajfar and Santoro (2007).

- Empirical methods and time-series inference
  - VAR and SVAR approaches to monetary policy identification: Christiano, Eichenbaum, and Evans (1999); Cushman and Zha (1995); Peersman and Smets (2001).
  - Unit root, cointegration, and inference in time series: Dejong et al. (1992); MacKinnon (1996); Perron (1989); Perron and Ghysels (1994); Nelson and Kang (1981); Toda and Yamamoto (1995).
  - Seasonal and filtering issues: Franses (1991); Perron and Ghysels (1994).

- Monetary policy in low-income and emerging market contexts
  - Effectiveness and transmission in low-income countries: Mishra and Montiel (2012); Mishra, Montiel, and Spilimbergo (2010).
  - Case studies and country-focused analyses: Pakistan — Agha et al. (2005); Alam and Waheed (2006); India — Mallick (2009); Mohan (BIS Papers No 35); Sri Lanka — Fernandez et al. (2004); Perera and Wickramanayake (2013); Ratnasiri (2009); Vinayagathasan (2013); Thailand — Bank of Thailand (BIS Papers No. 35); Disyatat and Vongsinsirikul (2003); China — Yi (BIS Papers No 35); Fiji and PNG — Ahmad (2008); Pacific Island Countries — Yang et al. (2011); South Africa — Smal and de Jager (2001); CIS — Starr (2005); Euro area — European Central Bank (2010); Eurozone heterogeneity and financial fragility — Ciccarelli, Maddaloni, and Peydró (2013).

- Asset prices, expectations, and learning
  - Asset prices and learning in monetary policy: Airaudo, Nisticò, and Zanna (2012); Gilchrist and Saito (2006); Pfajfar and Santoro (2007).

- Liquidity, open-economy, and exchange rate aspects
  - Liquidity and open-economy models: Grilli and Roubini (1995).
  - Exchange rate targeting and implicit exchange rate targeting: Mallick (2009).

- Debt sustainability and stochastic simulation
  - Assessing debt sustainability in emerging markets using stochastic simulation methods: Hostland and Karam (2005).

### Representative methods and empirical contributions (selected entries preserved verbatim)
- Bernanke, Ben, 1993a, “Credit in the Macroeconomy,” Federal Reserve Bank of New York Quarterly Review (Spring 1992-93), pp. 50-70.
- Bernanke, Ben, and Mark Gertler, 1995, "Inside the Black Box: The Credit Channel of Monetary Policy Transmission," NBER Working Papers 5146 (Cambridge, Massachusetts: National Bureau of Economic Research).
- Christiano, L.J., Martin Eichenbaum, and Charles Evans, 1999, “ Monetary Policy Shocks: What Have We Learned and to What End?” The Handbook of Macroeconomics, ed. by John Taylor and Michael Woodford, Vol. I (Amsterdam: North Holland Publishing Co.), pp. 65–148.
- Mishra, Prachi, and Peter Montiel, 2012, “How Effective Is Monetary Transmission in Low-Income Countries? A Survey of the Empirical Evidence,” IMF Working Paper 12/143 (Washington: International Monetary Fund).
- Toda, Hiro Y., and Taku Yamamoto, 1995, “Statistical Inference in Vector Autoregressions with Possibly Integrated Processes,” Journal of Econometrics, Vol. 66 (1-2), pp. 225-50.
- European Central Bank, 2010, “Monetary Policy Transmission in the Euro Area, A Decade After the Introduction of the Euro,” ECB Monthly Bulletin (May), pp.85-98.

### Cross-cutting empirical evidence and geographic coverage
- Geographic coverage includes advanced economies and a broad set of emerging and low-income countries: United States, Euro area, China, India, Pakistan, Thailand, Sri Lanka, Fiji, Papua New Guinea, Pacific Island Countries, South Africa, CIS countries.
- Empirical approaches include VAR analysis, structural VAR identification, time-series unit root and cointegration testing, stochastic simulation, computable general equilibrium analysis, and micro evidence on small firms and bank lending.

*References list from _wp14190 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14190.pdf_
