## 1. Decomposition of the Money Multiplier and Velocity in the EAC

## Source details

**Canonical URL:** [1. Decomposition of the Money Multiplier and Velocity in the EAC](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1339.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1339.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1339.pdf.json)

---

### I. Purpose and key contributions
- Study objective: assess effectiveness and heterogeneity of the monetary transmission mechanism (MTM) across EAC Partner States (Burundi, Kenya, Rwanda, Tanzania, Uganda) to inform harmonization for an East African Monetary Union (EAMU).
- Three contributions:
  - Exhaustive and critical review of existing empirical literature of MTM in the EAC.
  - Application of Bayesian VAR (BVAR) and Factor Augmented VAR (FAVAR) methods previously unused for MTM in the EAC.
  - Methodology to quantify relative importance of various MTM channels in each EAC country.

### II. Conduct of monetary policy in the EAC — instruments, targets, goals
- Instruments used by EAC central banks:
  - Reserve Requirement
  - Discount Rate/Policy Rate
  - Open Market Operations
- Operating target:
  - Reserve Money
  - Short-term Interest Rate
- Intermediate targets:
  - Broad Money
  - Exchange Rate
  - Inflation Forecast
- Policy goals:
  - Inflation
  - Output
  - Financial Stability
- For the sample period, the dominant framework was reserve money targeting (Reserve Money Program, RMP). Broad money is the intermediate target; reserve money is the operating target.
- Practical RMP implementation has become more flexible to:
  - accommodate shifts in money multipliers and velocity;
  - incorporate unanticipated shocks to output and inflation.
- Notable framework changes (excluded from empirical work):
  - Central Bank of Uganda declared "inflation targeting Lite" in July 2011.
  - Central Bank of Kenya adopted a monetary policy framework giving more prominence to its policy interest rate in November 2011.

### III. Stylized MTM channels
- Main channels:
  - Money Channel
  - Interest Rate Channel
  - Asset Price Channel
  - Exchange Rate Channel
  - Credit Channel
  - Expectation Channel
- Endogenous feedback: output and inflation feed back into monetary policy, complicating empirical identification of exogenous policy shocks.

### IV. Stability and predictability of velocity and the money multiplier
- Aggregate stylized findings:
  - "All EAC countries show declines in velocity and increases in money multiplier, consistent with increase in monetization or financial deepening experienced by all countries."
  - "Therefore, neither velocity nor multiplier is stable, 'constant' parameters, but both are predictable."
- Correlation statistics reported:
  - "Correlation coefficient between velocity and lagged velocity exceeds 0.6 for each country, while correlation coefficient between money multiplier and lagged money multiplier exceeds 0.65 for all but two EAC countries (Burundi and Uganda have correlation coefficients of 0.4 and 0.3, respectively)."
- Country-specific remark:
  - "Rwanda has the most stable velocity in the EAC, followed by Kenya and Uganda, while Tanzania and Burundi show the largest decline in velocity (a decline of" [source continues beyond excerpt].

### V. Table 1 — Decomposition of the Velocity, January 2000–December 2010
- Decomposition of Velocity, January 2000- December 2010 (table lines as presented)
  - Burundi: -0.4 1.0 1.5 3175 -5.5 0.1 -0.8
  - Kenya: -0.1 0.9 1.0 47493.6 0.0 -0.7
  - Rwanda: 0.0 1.3 1.4 1883 -1.1 0.0 -0.9
  - Tanzania: -0.6 1.1 1.7 28710.4 0.0 -0.8
  - Uganda: -0.2 1.3 1.5 17766.1 -0.1 -0.9
- Notes in table:
  - "1 In percent of total variance of velocity."
  - "Variables defined in first differences of log levels. v, GDP and BM refer to income velocity of broad money, nominal GDP and broad money. Cov and Corr refer to covariance and correlation coefficient, respectively between variables listed in each column."
- Average Monthly Growth and Variance columns are in the table but presented in the source with the above numeric lines.

### V. Table 1 — Decomposition of the Money Multiplier, January 2000–December 2010
- Decomposition of Money Multiplier, January 2000- December 2010 (table lines as presented)
  - Burundi: 0.3 1.5 1.5 38151 -88.0
  - Kenya: 0.3 1.0 0.9 12107 -18.9
  - Rwanda: 0.6 1.4 1.0 2599 -24.3
  - Tanzania: 0.3 1.7 1.6 1696 -11.4
  - Uganda: 0.3 1.5 1.6 1699 -15.0
- Notes in table:
  - "1 In percent of total variance of money multiplier."
  - "Variables are in first differences of log levels. m, BM, and RM refer to broad money multiplier, broad money and reserve money, respectively. Cov and Corr refer to covariance and correlation coefficient, respectively between variables listed in each column."
- Correlation column entry shown as:
  - Corr. BM,RM: 0.6 (Burundi), 0.3 (Kenya), 0.2 (Rwanda), 0.1 (Tanzania), 0.2 (Uganda)

### VI. Implications for RMP and monetary harmonization
- Because velocity and the money multiplier are not constant but are predictable, RMP implementation requires:
  - accommodating shifts in multipliers and velocity (ex-post adjustments or flexible RMP);
  - frequent use of high-frequency data and more regular Monetary Policy Committee meetings to fine tune stance.
- Heterogeneity across Partner States implies challenges for a one-size-fits-all common monetary policy; a common instrument must have similar effects across countries to be effective in a monetary union.

### VII. Channels — brief definitions preserved from source
- Money Channel: reserve money → money multiplier → broad money → intermediate target.
- Interest Rate Channel: policy rate → short-term nominal rate → real interest rate → investment → aggregate demand.
- Exchange Rate Channel: policy rate cut → currency depreciation (via UIP augmented by risk premium) → net exports → output.
- Credit Channel: bank-lending and balance-sheet effects via external finance premium and loan supply.
- Asset Price Channel: monetary shocks → asset prices (equity, housing) via discount rates → Tobin’s q and wealth effects.
- Expectation Channel: expectations of future policy rates affecting medium/long-term rates and inflation expectations.

### VIII. Review of empirical literature — main summary points
- General finding:
  - More studies on Kenya than other EAC countries.
  - MTM is strong in Kenya (mainly for prices) but generally weak in the rest of the EAC for output or price.
- Selected country highlights from the literature (methodologies and main findings summarized in source):
  - Kenya: evidence of price effects from interest rate and monetary aggregate shocks; mixed output effects; bank-lending and credit channels documented.
  - Tanzania: weak interest rate transmission; mixed evidence on reserve money effects depending on specification and sample.
  - Uganda: evidence that excess liquidity weakens transmission; mixed results with some short-lived output effects but weak interest rate pass-through.
  - Rwanda and Burundi: limited studies focused on money demand stability; scant published monetary transmission analysis, especially for Burundi.

### IX. Empirical methodology (as implemented)
- Main approaches:
  - Recursive Structural VAR (Choleski identification with ordering: output, price level, reserve money, policy/short-term interest rate, credit to private sector, exchange rate).
  - Bayesian VAR (Minnesota priors).
  - Factor-Augmented VAR (principal components to extract information from many series).
- Data and benchmark model:
  - Benchmark period: January 2000 through December 2010.
  - Variables: real GDP, CPI, reserve money, short-term interest rate, credit to private sector, nominal effective exchange rate (NEER) — logs except interest rates in percent.
  - Four exogenous variables: global oil price index, global food price index, U.S. federal funds rate, U.S. industrial production.
  - FAVAR: first principal component of the four exogenous variables as an exogenous factor; first principal component of monetary indicators (credit, M1, M2, M3, NEER) as an endogenous factor (M3 not used for Burundi).

### X. Data issues and frequency choices
- Quarterly national accounts start dates: Kenya (2000 Q1), Rwanda (2006 Q1), Uganda (1999 Q4), Tanzania (2001 Q1); Burundi lacks quarterly GDP.
- Maximum of 40 quarterly observations per country limits degrees of freedom.
- Strategies for missing GDP series:
  - Rwanda pre-2006: apply post-2006 quarterly seasonality factors to annual real GDP.
  - Burundi: use Rwanda’s quarterly seasonality factors on Burundi’s annual GDP.
  - Monthly GDP estimated by interpolating quarterly GDP using a cubic spline and seasonally adjusting with X-12 ARIMA.

### XI. Estimation design and inference
- VAR lag length: maximum three lags chosen by standard criteria; three lags sufficient to render serially uncorrelated VAR errors.
- Shocks in impulse-response analysis: one standard deviation; vertical axes are percentages, horizontal axes are months; shaded areas denote the 90% confidence bands in main reported results.
- Confidence interval choices:
  - Appendix II reports the 95 percent confidence interval, showing that MTM is weak in the EAC.
  - The main reported results use a 90 percent confidence interval, showing a somewhat stronger MTM.
- Monthly data often displayed heteroscedasticity; authors avoided extensive intervention-dummy saturation to prevent distortion.

### XII. Country-by-country empirical highlights (main reported results at 90% confidence interval)
- Burundi:
  - Reserve money shock: output responds positively and significantly with a lag spanning almost a year.
  - T-bill rate shock: output response peaks in 6 months but not statistically significant.
  - FAVAR: more muted output responses; more pronounced price response to interest rate shock versus SVAR.
  - Interest rate channel tests: exogenizing the T-bill rate switches a negative output response to reserve money into a positive and statistically significant response; interest rate movements amplify reserve money impacts.
  - Exchange rate: appreciation associated with expansionary output response (contrary to Keynesian expectation).
- Kenya:
  - Positive policy rate shock: significant and persistent effect on CPI peaking 9 to 11 months after the shock.
  - Reserve money shock: positive impact on CPI peaking at 15 months.
  - Both shocks: no statistically significant effect on GDP.
  - Appendix II: shocks to reserve money and policy rate have significant impacts on NEER and credit; NEER and credit shocks significantly impact CPI → indirect evidence for exchange rate and credit channels.
  - Endogeneity tests: allowing credit or NEER endogenous magnifies impact of monetary shocks on CPI.
- Rwanda:
  - Reserve money shock: positive and statistically significant output effect; no price effect in some specifications.
  - Positive repo rate shock: positive price response (price puzzle) lasting a long time in recursive SVAR; FAVAR reduces extent of price puzzle and confirms positive output response.
  - Private sector credit shocks: significant effect on output for first 5 months and on price level for first 15 months.
  - Reserve money significantly influences private sector credit → important credit channel. Exogeneity tests confirm strong credit channel.
- Tanzania:
  - Baseline (2000–2010) recursive VAR: reserve money shock increases CPI in first 1.5 years but not statistically significant.
  - Longer sample (1993m1–2010m12): impact on CPI becomes highly significant.
  - FAVAR: reserve money shock has statistically significant positive output effects.
  - Overall confidence intervals often include zero → weak MTM.
  - Possible causes: shallow financial markets, oligopolistic banking structure, instability in velocity rather than money multiplier.
  - Policy note: planned removal of capital controls by 2015 should strengthen exchange rate and interest rate channels.
- Uganda:
  - Reserve money shock: output responds significantly in short term; prices largely unaffected.
  - Policy rate shock: ambiguous output effects; persistent deflationary impact on prices.
  - FAVAR: interest rate increases reduce inflation more than in SVAR/BVAR, suggesting use of larger information sets by central bank.
  - Sensitivity: longer samples increase magnitude and persistence of shocks, suggesting weakening transmission over time.

### XIII. Results from variance decomposition (36 month horizon, percent of variance)
- Table 2 — percent of variance in GDP or CPI at 36 month horizon accounted by shocks to reserve money and interest rate (baseline SVAR)
  - Burundi: GDP 8.3, 7.6; CPI 4.5, 0.3
  - Kenya: GDP 5.4, 3.8; CPI 19.3, 12.8
  - Rwanda: GDP 12.7, 3.1; CPI 3.9, 4.4
  - Tanzania: GDP 0.7, 0.7; CPI 1.4, 11.1
  - Uganda: GDP 11.8, 2.7; CPI 0.3, 7.2
- Major finding:
  - Changes in output are more due to shocks to reserve money than to the interest rate; largest in Uganda and Rwanda.
  - Inflation is more due to shocks to the interest rate than shocks to reserve money; more pronounced in Tanzania and Uganda.

### XIV. Discussion and conclusions — synthesized evidence (90 percent confidence interval)
- Key empirical findings:
  - Expansionary monetary policy (positive shock to reserve money) increases output significantly in Burundi, Rwanda, and Uganda but has no statistically significant effect on prices in any EAC country.
  - Reserve money shocks:
    - Short lags in Uganda (statistically significant output effects for the first six months only).
    - Long lags in Burundi and Rwanda (statistically significant output effects from 6 to 15 months).
  - Policy rate shocks (expansionary monetary policy as negative shock to policy rate) increase prices significantly in Kenya and Uganda and output in Burundi, Kenya, and Rwanda.
  - Policy-rate shocks have long lags for prices and output across countries (varying from 5 months in Burundi to some 36 months in Uganda); Rwanda has the shortest lag (3 to 5 months).
  - Channels differ across countries:
    - Exchange rate and credit channels important in Kenya;
    - Credit channel important in Rwanda;
    - Interest rate channel important in Burundi.
  - Policy rate matters more for price evolution in countries with deeper financial markets and more competitive banking systems (Kenya and Uganda).
  - Interest rates typically decline in response to a positive shock to reserve money; statistically significant at the 90 percent confidence interval in Burundi, Kenya, and Uganda.
  - In some cases (Tanzania and Rwanda), reserve money and policy rate shocks may exert offsetting impulses, producing statistically insignificant impacts on prices.
  - Simultaneous inconsistent choices of prices and quantities can weaken MTM and development of market-determined rates.

### XV. Policy recommendations and challenges to strengthen MTM
- Ensure monetary targets and interest rate policy are consistent; interest rates should be allowed to be endogenous when money is the operating target.
- Address high share of currency in circulation in reserve money, which reduces central bank influence on cost conditions.
- Reduce constraints from a large informal economy; improve public awareness of the cost of holding idle money balances.
- Deepen financial markets and improve access to finance.
- Improve integration and depth of interbank FX and money markets to enhance exchange rate and interbank interest rate transmission.
- Promote competition in the banking sector to improve interest rate pass-through.
- Reduce high commercial bank excess reserves that weaken central bank regulation of reserve markets.
- Reconsider capital controls that can weaken exchange rate and interest rate channels.
- Improve high quality and high frequency data to reduce noise and aid timely policy responses.

### XVI. Transition issues for a future East African Monetary Union
- Key questions for monetary union design:
  - How to measure monetary policy stance across five countries with different instruments?
  - What does tightening or loosening policy mean when an instrument tightens in one country but loosens in another?
  - To what extent do impacts on inflation and output depend on specific transmission channels, and does heterogeneity of channels across countries matter if outcomes align with monetary principles?
  - Would monetary policy be a blunt instrument if factors of production are not mobile across borders (i.e., without a fully functional common market)?
  - What should be the pace of deepening domestic financial markets given disparities among member countries?
  - Can countries with narrow financial markets piggy-back on those with deeper markets?
- Recommendation: More research using methodologies beyond VARs is needed to better understand MTM in the EAC.

*Source: IMF Working Paper — "1. Decomposition of the Money Multiplier and Velocity in the EAC" (section and tables as provided in the source content).*

### 1. Decomposition of the Money Multiplier and Velocity in the EAC .....................................12

### 1. Decomposition of the Money Multiplier and Velocity in the EAC

### I. Purpose and key contributions
- Study objective: assess effectiveness and heterogeneity of the monetary transmission mechanism (MTM) across EAC Partner States (Burundi, Kenya, Rwanda, Tanzania, Uganda) to inform harmonization for an East African Monetary Union (EAMU).
- Three contributions:
  - Exhaustive and critical review of existing empirical literature of MTM in the EAC.
  - Application of Bayesian VAR (BVAR) and Factor Augmented VAR (FAVAR) methods previously unused for MTM in the EAC.
  - Methodology to quantify relative importance of various MTM channels in each EAC country.

### II. Conduct of monetary policy in the EAC — instruments, targets, goals
- Instruments used by EAC central banks:
  - Reserve Requirement
  - Discount Rate/Policy Rate
  - Open Market Operations
- Operating target:
  - Reserve Money
  - Short-term Interest Rate
- Intermediate targets:
  - Broad Money
  - Exchange Rate
  - Inflation Forecast
- Policy goals:
  - Inflation
  - Output
  - Financial Stability
- For the sample period, the dominant framework was reserve money targeting (Reserve Money Program, RMP). Broad money is the intermediate target; reserve money is the operating target.
- Practical RMP implementation has become more flexible to:
  - accommodate shifts in money multipliers and velocity;
  - incorporate unanticipated shocks to output and inflation.
- Notable framework changes (excluded from empirical work): Central Bank of Uganda declared "inflation targeting Lite" in July 2011; Central Bank of Kenya adopted a monetary policy framework giving more prominence to its policy interest rate in November 2011.

### III. Stylized MTM channels (as depicted)
- Main channels referenced:
  - Money Channel
  - Interest Rate Channel
  - Asset Price Channel
  - Exchange Rate Channel
  - Credit Channel
  - Expectation Channel
- Endogenous feedback: output and inflation feed back into monetary policy, complicating empirical identification of exogenous policy shocks.

### IV. Stability and predictability of velocity and the money multiplier
- Using compiled measures of nominal GDP, broad money, and reserve money, the paper finds mixed evidence on stability and predictability (Table 1, Figure 3).
- Aggregate stylized findings:
  - "All EAC countries show declines in velocity and increases in money multiplier, consistent with increase in monetization or financial deepening experienced by all countries."
  - "Therefore, neither velocity nor multiplier is stable, 'constant' parameters, but both are predictable."
- Correlation statistics reported:
  - "Correlation coefficient between velocity and lagged velocity exceeds 0.6 for each country, while correlation coefficient between money multiplier and lagged money multiplier exceeds 0.65 for all but two EAC countries (Burundi and Uganda have correlation coefficients of 0.4 and 0.3, respectively)."
- Country-specific remarks:
  - "Rwanda has the most stable velocity in the EAC, followed by Kenya and Uganda, while Tanzania and Burundi show the largest decline in velocity (a decline of" [text continues in source beyond excerpt].

### V. Table 1 — Decomposition of the Money Multiplier and Velocity in the EAC (January 2000–December 2010)
- Decomposition of Velocity, January 2000- December 2010 (table lines as presented):
  - Burundi: -0.4 1.0 1.5 3175 -5.5 0.1 -0.8
  - Kenya: -0.1 0.9 1.0 47493.6 0.0 -0.7
  - Rwanda: 0.0 1.3 1.4 1883 -1.1 0.0 -0.9
  - Tanzania: -0.6 1.1 1.7 28710.4 0.0 -0.8
  - Uganda: -0.2 1.3 1.5 17766.1 -0.1 -0.9
  - Notes in table: "1 In percent of total variance of velocity." and "Variables defined in first differences of log levels. v, GDP and BM refer to income velocity of broad money, nominal GDP and broad money. Cov and Corr refer to covariance and correlation coefficient, respectively between variables listed in each column."
  - Average Monthly Growth and Variance columns are in the table but presented in the source with the above numeric lines.
- Decomposition of Money Multiplier, January 2000-December 2010 (table lines as presented):
  - Burundi: 0.3 1.5 1.5 38151 -88.0
  - Kenya: 0.3 1.0 0.9 12107 -18.9
  - Rwanda: 0.6 1.4 1.0 2599 -24.3
  - Tanzania: 0.3 1.7 1.6 1696 -11.4
  - Uganda: 0.3 1.5 1.6 1699 -15.0
  - Notes in table: "1 In percent of total variance of money multiplier." and "Variables are in first differences of log levels. m, BM, and RM refer to broad money multiplier, broad money and reserve money, respectively. Cov and Corr refer to covariance and correlation coefficient, respectively between variables listed in each column." Correlation column entry shown as:
    - Corr. BM,RM: 0.6 (Burundi), 0.3 (Kenya), 0.2 (Rwanda), 0.1 (Tanzania), 0.2 (Uganda)

### VI. Implications for RMP and monetary harmonization
- Because velocity and the money multiplier are not constant but are predictable, RMP implementation requires:
  - accommodating shifts in multipliers and velocity (ex-post adjustments or flexible RMP);
  - frequent use of high-frequency data and more regular Monetary Policy Committee meetings to fine tune stance.
- Heterogeneity across Partner States (stability/predictability differences) implies challenges for a one-size-fits-all common monetary policy; a common instrument must have similar effects across countries to be effective in a monetary union.

*Source: IMF Working Paper — "1. Decomposition of the Money Multiplier and Velocity in the EAC" (section and tables as provided in the source content).*

### 6.7 percent and 5.7 percent a year, respectively, based on annual data) and the most

### _wp1339 - 6.7 percent and 5.7 percent a year, respectively, based on annual data) and the most

### III. Channels of Monetary Transmission Mechanism
- Overview:
  - Central bankers need to know how changes in monetary policy instruments affect inflation and output, and the timing and size of such effects.
  - Traditional views: money channel (quantity) versus credit/interest rate channels (price). Nominal price or wage rigidity is assumed in most models to link money, interest rates, and output.

- A. Money Channel:
  - Assumes changes in reserve money are transmitted to broad money via the money multiplier.
  - Broad money used as an intermediate target in money-focused monetary policies.

- B. Interest Rate Channel:
  - Summarized in Keynesian IS-LM: expansionary monetary policy → fall in real interest rate → lower cost of capital → higher investment → higher aggregate demand and output.
  - Real spending is affected by real interest rates; policy controls short-term nominal rate.
  - Price stickiness and the term structure allow short-term policy rates to influence long-run real rates.

- C. Exchange Rate Channel:
  - Crucial in small, open economies; based on uncovered interest rate parity (UIP).
  - A cut in policy rate can depreciate domestic currency, boosting net exports and output.
  - UIP often augmented with a risk-premium term to account for foreign investor compensation.

- D. Credit Channel:
  - Based on asymmetric information; monetary contraction can raise external finance premium and lower loan supply.
  - Two types: bank-lending channel (monetary contraction reduces bank reserves/deposits and loan supply) and balance-sheet channel (monetary contraction reduces firm net worth and collateral values, raising external finance premium).

- E. Asset Price Channel:
  - Monetary shocks affect asset prices (equity, housing) via discount rates.
  - Tobin’s q: low equity prices reduce equity-financed investment.
  - Wealth effects: higher stock/ housing prices raise lifetime resources and consumption.

- F. Expectation Channel:
  - Fundamental in forward-looking economies; expectations of future policy rates affect medium/long-term rates and inflation expectations.
  - Inflation expectations influence real interest rates and wage/price setting, feeding into future inflation.

### IV. Review of the Empirical Literature on MTM in EAC
- General finding:
  - More studies on Kenya than other EAC countries.
  - MTM is strong in Kenya (mainly for prices) but generally weak in the rest of the EAC for output or price.

- A. Kenya (selected findings):
  - Cheng (2006): SVAR (1997–2005) — contractionary monetary policy (short-term interest rate shock) leads to initial price increase (price puzzle) then significant fall for about two years; output rises initially (output puzzle) then falls (not statistically significant). Interest rate shocks explain 30 percent of inflation variance and 10 percent of output variance. Exchange rate depreciates initially then appreciates for about two years.
  - Maturu, Maana, and Kisinguh (2010): Using M3 (2000–2010) — M3 shock has no effect on real output but raises prices for almost 18 months (statistically significant). Interest rate shock leads to falling prices (not statistically significant). M3 explains as much inflation variability as interest rate shock in Cheng.
  - Buigut (2009): Three-variable VAR (1984–2006) for Kenya, Tanzania, Uganda — interest rate transmission weak; shocks to interest rate have no statistically significant effect on inflation or output. Study limitations noted (small sample, structural shifts).
  - Buigut (2010): VECM (1979–2008) — price puzzle present; tightening reduces loan quantity and raises lending rates (positive bank-lending channel). No confidence bounds presented for impulse responses.
  - Sichei and Njenga (2010): Panel bank-level (2001–2008) — loan demand unresponsive to lending rates, banks contract loan supply after tightening → evidence for credit rationing. Suggests monetary policy works through quantities (credit).
  - Misati and others (2010): Financial innovation proxies (M2/M1, bank assets/GDP) weaken transmission of repo rate to output.

- B. Tanzania:
  - Buigut (2009): SVAR (1984–2005) — interest rate shocks weak and insignificant for output and inflation.
  - Montiel and others (2012): Monthly data (Jan 2002–Sep 2010) — reserve money statistically significant for price level in recursive VAR but not economically significant; in structural VAR, significance disappears. No output effects.
  - Adam and others (2012): Single-equation inflation study — broad money affects inflation short-run and long-run (statistically significant); parsimonious models may uncover significance ignored by VARs.

- C. Uganda:
  - Peiris (2005): Six-variable recursive VAR — shock to M2 increases price level; interest rate shock has no effects (no confidence intervals provided).
  - Saxegaard (2006): Threshold VAR (1990–2004) — money supply contraction significantly reduces CPI inflation; effect larger when involuntary excess liquidity of banks is low.
  - Mugume (2011): Structural VAR (1999–2009) — all channels ineffective; interest rate channel remains weak though some pass-through from treasury bill rate to lending rates.

- D. Rwanda and Burundi:
  - Studies analyze stability of money demand (Sayinzoga and Simson, 2006; Rutayisire, 2010) but not monetary transmission.
  - No published studies on monetary transmission in Burundi are identified.

### V. Empirical Methodology
- Main approach: Vector autoregression (VAR) models — recursive SVARs, Bayesian VARs (BVAR), and factor-augmented VARs (FAVAR).
- Rationale:
  - Recursive SVARs are widely used but can suffer from over-parameterization and misspecification.
  - BVARs (Bayesian estimation with Minnesota priors) shrink parameter space using prior information to address over-parameterization.
  - FAVARs use principal components to extract factors from many time series, reducing dimensionality while preserving information.

- A. Recursive Structural VAR:
  - Reduced form: equation (1) in the source.
  - Structural form: equation (2); relationship between reduced and structural form via equations (3).
  - Identification via Choleski decomposition (Sims, 1980) with ordering (from slow to fast response):
    - Ordering: output (y), price level (p), reserve money (m), policy/short-term interest rate (r), credit to private sector (c), exchange rate (e).
  - Key identification assumptions (preserved):
    - Output and price respond with a lag to reserve money innovations.
    - Reserve money shocks are considered monetary policy shocks (consistent with RMP).
    - Policy rate ordered after reserve money to allow signaling role.
    - Credit placed after policy rate to allow delayed bank response.
    - Exchange rate last, responding contemporaneously to macro fundamentals.

- B. Bayesian VAR:
  - Use Minnesota priors (Doan, Litterman, and Sims, 1984; Litterman, 1986).
  - Prior treats variables as random walks in levels; prior mean for coefficient on first own lag can be set (example: 0.9).
  - Prior variance-covariance matrix diagonal; more weight on recent own lags.
  - Posterior of parameter vector α is normal, facilitating computation without extensive sampling.

- C. Factor-Augmented VAR:
  - FAVAR extends VAR by including unobserved factors estimated via principal component analysis (Stock and Watson; Bernanke, Boivin, and Eliasz).
  - For EAC: first principal component of four exogenous variables used as single exogenous factor; first principal component of monetary indicators (credit, M1, M2, M3, NEER) used as an endogenous factor.
  - Static principal component approach used (Stock and Watson, 2002).

### VI. Data
- Data sources: IMF databases, national authorities, staff estimates.
- Frequency considerations:
  - Money, price level, asset prices: monthly data available.
  - GDP available quarterly for all EAC countries except Burundi.
  - Quarterly national accounts start dates: Kenya (2000 Q1), Rwanda (2006 Q1), Uganda (1999 Q4), Tanzania (2001 Q1).
  - Maximum of 40 quarterly observations per country (limits degrees of freedom).

- Strategies for missing GDP series:
  - Rwanda before 2006 Q1: apply post-2006 quarterly seasonality factors to annual real GDP to interpolate quarterly.
  - Burundi: use Rwanda’s quarterly seasonality factors on Burundi’s annual GDP, given similar production structure.
  - Monthly GDP estimates derived by interpolating quarterly GDP using a cubic spline, then seasonally adjusted with X-12 ARIMA.

- Benchmark model:
  - Period: January 2000 through December 2010.
  - Variables in log levels except interest rates in percent.
  - SVAR specification: six endogenous variables — real GDP, CPI, reserve money, short-term interest rate, credit to private sector, nominal effective exchange rate (NEER).
  - Four exogenous variables: global oil price index, global food price index, U.S. federal funds rate, U.S. industrial production.
  - FAVAR: first principal component of four exogenous variables; first principal component of credit and NEER plus M1, M2, M3 (M3 not used for Burundi).

### VII. Empirical Results (estimation design and inference)
- Estimation details:
  - VAR lag length chosen via standard criteria (Akaike, Shwartz, Hannan-Quinn, Final Prediction Error): maximum lag length of three.
  - Three lags sufficient to render serially uncorrelated VAR errors.
  - Using more lags increases over-parameterization and widens confidence bounds.

- Main inference points:
  - Effects of monetary policy in some EAC countries may last beyond three months cumulatively, but impulse responses tend to be short lived compared with advanced countries.
  - If impulse response confidence interval includes zero, monetary policy has no statistically significant effect on prices or output (i.e., MTM is weak).
  - Expanding horizons beyond six months generally widens impulse response confidence intervals, often producing economically or statistically insignificant results.
  - Choice of confidence interval affects conclusion: e.g., using a 90 percent interval increases Type II error relative to the 95 percent standard.

*Source: _wp1339 - 6.7 percent and 5.7 percent a year, respectively, based on annual data) and the most (IMF PDF content provided).*

### Appendix II reports the 95 percent confidence interval, showing that MTM is weak in the

### _wp1339 - Appendix II reports the 95 percent confidence interval, showing that MTM is weak in the

### Confidence intervals, heteroscedasticity, and data frequency
- Appendix II reports the 95 percent confidence interval, showing that MTM is weak in the EAC.
- The results reported below use a 90 percent confidence interval, showing a somewhat stronger MTM.
- Monthly data often displayed hetroscedasticity; adding many “intervention” dummies could reduce hetroscedasticity but also distorted results significantly.
- The authors decided not to pursue the saturated dummy/intervention-dummy approach further, noting difficulty in interpreting many dummies and concern about adding noise to the MTM.
- Using quarterly data produces significantly less heteroscedasticity owing to data averaging, but using quarterly data also loses significant degrees of freedom; this line of research was not pursued.

### Methodological details and lag selection
- For Burundi: one lag was selected by the Akaike, Schwartz, and Hannan-Quinn information criteria. However, the authors proceeded with 2 lags to avoid serial correlation of residuals at the first lag order.
- Shocks in impulse-response analysis are one standard deviation; vertical axes are percentages, horizontal axes are months, and the shaded areas denote the 90% confidence bands.

### A. Burundi — baseline SVAR and FAVAR comparisons
- Baseline recursive model:
  - Impulse responses of output and prices to a one standard deviation positive shock in reserve money (monetary loosening) are shown in the first row of Figure 4; responses to a positive shock in the policy rate (monetary tightening) are shown in the second row.
  - Because of the lack of a structured interbank market, the T-bill rate is used as the policy interest rate in the baseline VAR for Burundi.
  - Output responds positively and significantly to a shock in reserve money, albeit with a considerable lag that spans almost a year.
  - Output’s response to a shock in the T-bill rate is prompt (peaking in 6 months) and sustained over a longer time though not statistically significant.
- Nesting the recursive VAR into a FAVAR model:
  - Generates relatively muted impulse responses of output to shocks in both reserve money and the T-bill rate.
  - Generates a more pronounced response of prices to an interest rate shock.
  - This contrast perhaps illustrates the instability of MTM in Burundi.
- Additional observation on exchange rates:
  - As shown in Appendix II, output moves in tandem with exchange rates; an appreciation of the nominal effective exchange has an expansionary effect on the economy.
  - This contradicts the Keynesian argument; possible explanations include appreciation pressures generated by large inflows channeled into investment and, given Burundi’s large trade deficit, an appreciating exchange rate increasing the net worth of domestic assets and thereby their purchasing power.

### A. Burundi — testing the interest rate channel (Figure 5)
- Procedure:
  - The VAR is re-run with the T-bill rate exogenized (lagged values of the T-bill rate treated as exogenous) in a smaller VAR involving GDP, CPI, reserve money, credit, and NEER.
  - This generates a VAR identical to the original except that it blocks off any responses within the VAR that pass through the interest rate.
- Findings:
  - Activating the interest rate channel in this fashion morphs a negative impulse response of output to innovations in reserve money to a positive (and statistically significant) response.
  - The channel also has a positive impact on prices.
  - The considerable swing is only captured when using the T-bill rate as opposed to the indicative discount rate.
  - Similar dynamics are observable even when the credit and exchange rate channels are inactivated, demonstrating the robustness of the interest rate channel in Burundi.
  - Conclusion: Movements in interest rates amplify the impact of reserve money on output and the price level. Interest rate therefore appears to be a transmission channel, but the effect is not statistically significant, given the direct impact of interest rate on either output or the price level.

### B. Kenya — benchmark recursive SVAR results (Figure 6)
- The benchmark model with recursive identification displays impulse responses of GDP and CPI to a one standard deviation shock to reserve money and the repo rate.
- Key findings:
  - A positive shock to the policy rate has a significant and persistent effect on CPI which peaks 9 to 11 months after the shock.
  - A positive shock to reserve money has a positive impact on CPI that peaks later at 15 months, later than an interest rate shock.
  - Both types of shocks have no statistically significant effect on GDP.
- Note on data series:
  - The time series switches to reverse the repo rate from June 2009 onward. Re-running the VARs that end in May 2009 does not change the results.

*Source: _wp1339 - Appendix II reports the 95 percent confidence interval, showing that MTM is weak in the*

### Appendix II includes the complete set of impulse responses, allowing for a more detailed

### _wp1339 - Appendix II includes the complete set of impulse responses, allowing for a more detailed

### Kenya — impulse responses and channel tests
- Appendix II: impulse responses show that a shock to both reserve money and the policy rate has a significant impact on NEER and credit; a shock to NEER (unexpected nominal exchange rate appreciation) and a shock to credit (unexpected credit expansion) both have a significant impact on CPI.
- Since the two shocks are by construction orthogonal, results are interpretable as indirect evidence of an exchange rate and credit channel of monetary policy.
- Exogeneity–endogeneity exercise:
  - Comparing a 5-variable VAR (GDP, CPI, reserve money, interest rate, credit endogenous) with a 4-variable VAR (credit exogenous) shows that allowing credit to be endogenous increases the impact of a monetary policy shock on CPI (Figure 7).
  - Allowing NEER to be endogenous magnifies the impact of both reserve money and policy rate shocks on CPI, with a more pronounced effect for policy-rate shocks (Figure 8).
- Note: Shocks are one standard deviation; vertical axes are percentages, horizontal axes are months, and the shaded areas denote the 90% confidence bands.

### Rwanda — findings from SVAR and FAVAR
- A reserve money shock induces a positive and statistically significant output effect, but no price effect.
- A positive response of prices to a shock in the key repo rate is observed (a price puzzle) that lasts for quite some time.
- The FAVAR approach (Appendix III) confirms the positive output response to a monetary base shock and reduces the extent of the price puzzle versus the recursive specification.
- A shock to private sector credit:
  - Significant effect on output for the first 5 months (statistically significant).
  - Significant effect on the price level for the first 15 months (statistically significant).
- Reserve money significantly influences private sector credit (Figure 3 in Appendix II), suggesting an important credit channel.
- Exogeneity test: rerunning the VAR with private sector credit exogenized and comparing to baseline confirms a strong credit channel in Rwanda; credit channel seems stronger statistically for GDP response to reserve money than for CPI response.

### Tanzania — weak transmission and possible explanations
- Recursive VAR (2000–2010): positive shock to reserve money increases CPI in the first year-and-a-half, though not statistically significant.
- Longer sample (1993m1–2010m12) in baseline recursive structure: impact on CPI becomes highly significant.
- FAVAR: reserve money shock has statistically significant positive output effects.
- Positive shock to the interest rate increases CPI (price puzzle) in some specifications, but not statistically significant.
- Confidence intervals for impulse responses for output and price include zero, indicating weak monetary transmission.
- Possible causes of weak MTM:
  - Shallow financial markets and oligopolistic banking structure (Montiel and others, 2012).
  - Instability in velocity (our Section II.C) rather than in the money multiplier; shocks to reserve money transmit to money but transmission from money to prices or output is weak due to shifts in velocity.
- Policy note: removal of capital controls by 2015 (objective of the Tanzanian authorities) should strengthen the role of the exchange rate and interest rate channels.
- Note: Shocks are one standard deviation; vertical axes are percentages, horizontal axes are months, and the shaded areas denote the 90% confidence bands.

### Uganda — output responsive, prices less so; FAVAR differences
- Output responds significantly to positive shocks to reserve money over the short term; prices largely unaffected.
- Positive shock to policy rate (bank rate to commercial banks) has ambiguous effect on output but a persistent deflationary impact on prices.
- Estimating the VAR with five lags (vs three in baseline) generates a more pronounced CPI response to policy rate innovations.
- FAVAR: interest rate increases reduce inflation more than in simple SVAR or BVAR, suggesting the Central Bank of Uganda uses a larger information set when deciding interest rate changes.
- Sensitivity: fitting baseline recursive VAR over September 1995 to December 2010 increases magnitude and persistence of shocks, suggesting transmission channels have withered over time.
- Note: Shocks are one standard deviation; vertical axes are percentages, horizontal axes are months, and the shaded areas denote the 90% confidence bands.

### Results from variance decomposition
- Major findings:
  - Changes in output are more due to shocks to reserve money than to the interest rate; largest in Uganda and Rwanda.
  - Inflation is more due to shocks to the interest rate than shocks to reserve money; more pronounced in Tanzania and Uganda.
- Table 2 (percent of variance in GDP or CPI at 36 month horizon; each entry shows percent of variance in GDP or CPI accounted by shocks to reserve money and interest rate in the estimated baseline SVAR for each country at 36 month horizon):
  - Burundi: GDP 8.3, 7.6; CPI 4.5, 0.3
  - Kenya: GDP 5.4, 3.8; CPI 19.3, 12.8
  - Rwanda: GDP 12.7, 3.1; CPI 3.9, 4.4
  - Tanzania: GDP 0.7, 0.7; CPI 1.4, 11.1
  - Uganda: GDP 11.8, 2.7; CPI 0.3, 7.2

### Discussion and conclusions — synthesis of evidence
- Many EAC countries have shifted from quantity-based (monetary aggregates) to price-based (interest rate) monetary policy, with variable adherence to targeting rules.
- A flexible reserve money program has operational relevance in some EAC countries to account for shifts in velocity and multiplier.
- Key summarized empirical findings (evaluated at a 90 percent confidence interval):
  - An expansionary monetary policy (positive shock to reserve money) increases output significantly in Burundi, Rwanda, and Uganda but has no statistically significant effect on prices in any EAC country.
  - Monetary policy (reserve money shocks) has short lags in Uganda (statistically significant output effects for the first six months only) but long lags in Burundi and Rwanda (statistically significant output effects from 6 to 15 months).
  - An expansionary monetary policy (a negative shock to the policy rate) increases prices significantly in Kenya and Uganda and output in Burundi, Kenya, and Rwanda.
  - Monetary policy (policy rate shocks) has long lags for prices and output across countries (varying from 5 months in Burundi to some 36 months in Uganda); Rwanda has the shortest lag (3 to 5 months).
  - Channels differ across EAC: exchange rate and credit channels important in Kenya; credit in Rwanda; interest rate in Burundi.
  - Policy rate matters more for price evolution in countries with deeper financial markets and more competitive banking systems (Kenya and Uganda).
  - Interest rates typically decline in response to a positive shock to reserve money; statistically significant at the 90 percent confidence interval in Burundi, Kenya, and Uganda.
  - In some cases (Tanzania and Rwanda), shocks to reserve money and policy rate may move in directions that exert offsetting expansionary and contractionary impulses, producing statistically insignificant impacts on prices.
  - Simultaneous attempts to choose both prices and quantities inconsistently can weaken MTM and development of market-determined rates.

### Policy recommendations and challenges to strengthen MTM
- Ensure monetary targets and interest rate policy are consistent with each other; interest rates should be allowed to be endogenous when money is the operating target.
- Address a high share of currency in circulation in reserve money, which reduces central bank influence on cost conditions.
- Reduce constraints from a large informal economy; improve public awareness of the cost of holding idle money balances.
- Deepen financial markets and improve access to finance.
- Improve integration and depth of interbank FX and money markets to enhance exchange rate and interbank interest rate transmission.
- Promote competition in the banking sector to improve interest rate pass-through.
- Reduce high commercial bank excess reserves that weaken central bank regulation of reserve markets.
- Reconsider capital controls that can weaken exchange rate and interest rate channels.
- Improve high quality and high frequency data to reduce noise and aid timely policy responses.

### Transition issues for a future East African Monetary Union
- How to measure monetary policy stance across five countries with different instruments?
- What does tightening or loosening policy mean when an instrument tightens in one country but loosens in another?
- To what extent do impacts on inflation and output depend on specific transmission channels, and does heterogeneity of channels across countries matter if outcomes align with monetary principles?
- Would monetary policy be a blunt instrument if factors of production are not mobile across borders (i.e., without a fully functional common market)?
- What should be the pace of deepening domestic financial markets given disparities among member countries?
- Can countries with narrow financial markets piggy-back on those with deeper markets?
- More research using methodologies beyond VARs is needed to better understand MTM in the EAC.

*Source: _wp1339 - Appendix II includes the complete set of impulse responses, allowing for a more detailed (PDF chapter/section).*

### REFERENCES

### _wp1339 - REFERENCES

### Appendix I. Literature Review on Monetary Transmission Mechanism in the EAC

- Kenya — Cheng (2006)
  - Methodology: Five variable recursive and structural VARs (following Kim and Roubini, 2000). A monetary policy shock is identified as a shock to interest rate.
  - Data: Monthly data (January 1997– June 2005) in log levels for all variables except interest rate: real GDP, CPI, broad monetary aggregate (reserve money and M3), short-term interest rate (Repo rate, interbank rate) nominal effective exchange rate (NEER), oil price, U.S. federal funds rate, and U.S. commodity prices. The first five variables are endogenous; the last three are exogenous.
  - Main Findings:
    - A monetary contraction (a positive shock to interest rate) (i) does not have a statistically significant effect on output; (ii) lowers prices persistently after an initial increase in inflation; (iii) leads to an initial depreciation of the exchange rate followed by an appreciation that persists for almost two years and is statistically significant.
    - Shocks to interest rate account for 33 percent of the forecast error variance of inflation, 50 percent of forecast error variance of NEER, and 10 percent of forecast error variance of output.

- Kenya — Maturu, Maana, and Kisinguh (2010)
  - Methodology: Five variable recursive and structural VARs (following Kim and Roubini, 2000). Monetary policy shock is identified primarily as a shock to M3.
  - Data: Quarterly data (2000Q1–2010Q2) in log levels for all variables except interest rate: real GDP, CPI, monetary aggregates (M3, reserve money), interest rate (repo rate, interbank rate), NEER, oil price, commodity price index, and U.S. federal fund rate. The first five variables are endogenous; the last three are exogenous.
  - Main Findings:
    - A monetary contraction (a negative shock to M3) (i) decreases output and is marginally significant; (ii) leads to a lower price level which is statistically significant for almost four years; (iii) leads to an appreciation of the exchange rate though it is not statistically different from zero; and (iv) increases the interest rate but it is not statistically different from zero.
    - Shocks to M3 account for 30 percent of forecast error variance inflation but only 4 percent of forecast error variance of output.

- Kenya — Buigut (2010)
  - Methodology: A five-variable VECM. Monetary policy shock is identified as a shock to T-bill rate.
  - Data: Annual data (1979–2008) in log level except CPI in first difference of log: real GDP, real private sector credit, lending rate, T-bill rate.
  - Main Findings:
    - A positive shock to T-bill rate has positive but transitory effects on inflation (price puzzle), small negative impact on real GDP, and leads to a permanent fall in loan quantity while loan rates respond positively. Impulse responses are shown with no confidence bounds.

- Kenya — Misati, Lucas, Anne, and Shem (2010)
  - Methodology: Single equation methods: ARDL and 2SLS. Dependent variable: output gap. Monetary policy instrument is real repo rate.
  - Data: Monthly data (1996m1–2007m2): real repo rate, measures of financial innovation (ratio of bank asset to GDP, ratio of M3 to M1), and output gap.
  - Main Findings:
    - Coefficient on real interest rate is negative and statistically significant, indicating contractionary monetary effects.
    - Coefficient on the interaction of real interest rate and measures of financial innovation is positive and statistically significant, thus moderating effect of negative interest rate.

- Kenya — Sichei and Njenga (2010)
  - Methodology: Static panel data estimation: 3SLS.
  - Data: Annual data (2001–2008) in log levels for 37 banks: private credit, private deposit, total bank reserves, prudential and liquidity measures, and total capital ratio.
  - Main Findings:
    - (i) Strong evidence for the bank-lending channel through quantities rather than lending rate; (ii) credit of small, less capitalized and less liquid foreign-owned banks is more responsive to lending rate.

- Tanzania, Kenya, and Uganda — Buigut (2009)
  - Methodology: Three-variable recursive VAR. A monetary policy shock is identified as a shock to interest rate.
  - Data: Annual data (1984–2005) in log levels: real GDP, CPI inflation, and T-bill rate (discount rate for Tanzania).
  - Main Findings:
    - Weak interest rate channel: interest rate shock has insignificant effects on output and inflation in all three countries.

- All EAC countries — Baldini, Poplawski-Ribeiro (2011)
  - Methodology: Three-variable recursive VAR. Monetary policy shock is identified primarily as a shock to reserve money.
  - Data: Annual data (1980–2005) in log levels and first difference of logs: real GDP, CPI (GDP deflator), and reserve money (discount rate).
  - Main Findings:
    - A positive shock to reserve money has positive and negative effects on inflation, depending on the horizon in all EAC countries except for Burundi, which shows a positive effect. No confidence bounds are shown for impulse responses.
    - Shocks to reserve money account for 4 percent of inflation forecast error variance in Rwanda, 5 percent in Burundi, 13 percent in Uganda, 15 percent in Tanzania, and 31 percent in Kenya.

- Tanzania — Montiel and others (2012)
  - Methodology: VAR; Monthly data (December 2001– May 2010) in log levels; four VARs: two recursive (three- and six-variables) and two non-recursive (three- and six-variables). Ordering in three-variable recursive VAR: exchange rate, reserve money, and price level; ordering in the six-variable VAR: exchange rate, broad money, reserve money, loan rate, price level, and output.
  - Main Findings:
    - A positive shock to reserve money (an expansionary monetary policy) increases the price level in both recursive models; effects are statistically significant but not economically; no output effect in either VAR. In the non-recursive VAR, there are no statistically significant price or output effects.

- Uganda — Saxegaard (2006)
  - Methodology: Threshold VAR.
  - Data: Quarterly data (1990Q1–2004Q2) in log levels: deposit rate, lending rate, ratio of excess reserves to total deposits, private credit to GDP ratio, except real GDP for which output gap is used.
  - Main Findings:
    - (i) The presence of excess reserves lowers the negative effect of monetary contraction on inflation, thus weakening monetary transmission; (ii) excess reserves are high, implying low transmission in Uganda.

- Uganda — Mugume (2011)
  - Methodology: Five variable non-recursive VAR. Monetary policy shock is identified as a shock to interest rate (91 day T-bill rate).
  - Data: Quarterly data (1999q1–2009q1) in first difference of log except for interest rate which is in levels: real GDP, CPI, broad money, three-month T-bill rate (lending rate), nominal exchange rate, credit to private sector.
  - Main Findings:
    - A contractionary monetary policy drives output and inflation down. Only output effect is significant, lasting up to two quarters. Interest rate, credit, and exchange rate channels are weak. Innovation in M2 has no statistically significant effect on output and inflation.

- Regional/Other country findings
  - Peiris (2005): Six-variable recursive VAR; Monthly data (1993m6–2004m6) in log difference: international oil prices, coffee price, output gap, exchange rate, monetary aggregate (or interest rate) and consumer prices. Main finding: A 1 percent increase in M2 leads to a 0.2 percent rise in core inflation in three months. Interest rate has no effect.

### References (selection from the bibliography)
- Adam, Christopher, David Kwimbere, Wilfred Mbowe, and Stephen O’Connell, 2012, “Food Prices and Inflation in Tanzania,” International Growth Center WP/12/0459, http://www.theigc.org
- Baldini, A., and M. Poplawski-Ribeiro, 2011, "Fiscal and Monetary Determinants of Inflation in Low-Income Countries: Theory and Evidence from Sub-Saharan Africa," Journal of African Economies, Vol. 20, No. 3, pp. 419-462
- Bernanke, Ben S. and Mark Gertler, 1995, “Inside the Black Box: The Credit Channel of Monetary Policy Transmission,” Journal of Economic Perspectives, vol.9, no.4, pp.27-48
- Boivin, J., M. T. Kiley, and F. S. Mishkin, 2011, “How Has the Monetary Transmission Mechanism Evolved Over Time?,” in Handbook of Monetary Economics, ed. by B. M. Friedman, and M. Woodford, vol. 3, chap. 8, pp. 369–422.
- 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, International Monetary Fund
- Montiel, Peter, Christopher Adam, Wilfred Mbowe and Stephen O’Connell, 2012, “Financial Architecture and the Monetary Transmission Mechanism in Tanzania,” International Growth Center Working paper 12/0343
- Saxegaard, M., 2006, “Excess Liquidity and Effectiveness of Monetary Policy: Evidence from Sub-Saharan Africa,” IMF Working Papers 06/115, International Monetary Fund
- Sims, Christopher A., 1980, “Macroeconomics and Reality,” Econometrica, 48(1), 1–48

*Source: _wp1339 - REFERENCES (Appendix I and bibliography as provided).*

### Appendix II. Full Set of SVAR Impulse Responses

### _wp1339 - Appendix II. Full Set of SVAR Impulse Responses

### SVAR responses by country — figures and labeled series
- Figure 1. SVAR Responses for Burundi
  - Response panels and their exact labels:
    - Response of L_GDP_SA to L_GDP_SA (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPI_SA (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_RES (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to TB_R (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPS (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_NEER (y-axis ticks: -.005 .000 .005 .010; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_GDP_SA (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_CPI_SA (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_RES (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to TB_R (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_CPS (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_NEER (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_GDP_SA (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_CPI_SA (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_RES (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to TB_R (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_CPS (y-axis ticks: -.02 .00 .02 .04 .06 .08; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_NEER (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to L_GDP_SA (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to L_CPI_SA (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to L_RES (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to TB_R (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to L_CPS (y-axis ticks: -.8 -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of TB_R to L_NEER (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_GDP_SA (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_CPI_SA (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_RES (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to TB_R (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_CPS (y-axis ticks: -.04 -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_NEER (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_GDP_SA (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_CPI_SA (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_RES (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to TB_R (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_CPS (y-axis ticks: -.04 -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_NEER
  - Common footnote: Response to Cholesky One S.D. Innovations ± 2 S.E.

- Figure 2. SVAR Responses for Kenya
  - Response panels and exact labels (selected examples):
    - Response of L_GDP_SA to L_GDP_SA (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPI_NEW_SA (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_RES (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to POL_R (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPS (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_NEER (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_GDP_SA (y-axis ticks: -.02 .00 .02 .04; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of POL_R to L_GDP_SA (y-axis ticks: -0.5 0.0 0.5 1.0; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02; x-axis ticks: 5 10 15 20 25 30 35)
  - Common footnote: Response to Cholesky One S.D. Innovations ± 2 S.E.

- Figure 3. SVAR Responses for Rwanda
  - Response panels and labels include:
    - Response of L_GDP_SA to L_GDP_SA (y-axis ticks: -.005 .000 .005 .010 .015; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPI_SA (y-axis ticks: -.005 .000 .005 .010 .015; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_GDP_SA (y-axis ticks: -.01 .00 .01 .02; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_GDP_SA (y-axis ticks: -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of POL_R to L_GDP_SA (y-axis ticks: -.4 .0 .4 .8; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
  - Common footnote: Response to Cholesky One S.D. Innovations ± 2 S.E.

- Figure 4. SVAR Responses for Tanzania
  - Response panels and labels include:
    - Response of L_GDP_SA to L_GDP_SA (y-axis ticks: -.01 .00 .01; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_GDP_SA to L_CPI_SA (y-axis ticks: -.01 .00 .01; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_GDP_SA (y-axis ticks: -.004 .000 .004 .008; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_GDP_SA (y-axis ticks: -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of POL_R to L_GDP_SA (y-axis ticks: -1 0 1; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_GDP_SA (y-axis ticks: -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
  - Common footnote: Response to Cholesky One S.D. Innovations ± 2 S.E.

- Figure 5. SVAR Responses for Uganda
  - Response panels and labels include:
    - Response of L_GDP_SA to L_GDP_SA (y-axis ticks: -.004 .000 .004 .008 .012; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPI_SA to L_GDP_SA (y-axis ticks: -.01 .00 .01; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_RES to L_GDP_SA (y-axis ticks: -.02 .00 .02 .04 .06; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of POL_R to L_GDP_SA (y-axis ticks: -2 -1 0 1 2 3; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_CPS to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
    - Response of L_NEER to L_GDP_SA (y-axis ticks: -.02 -.01 .00 .01 .02 .03; x-axis ticks: 5 10 15 20 25 30 35)
  - Common footnote: Response to Cholesky One S.D. Innovations ± 2 S.E.

### Appendix III — FAVAR and BVAR comparison figures
- Figures compare impulse responses across SVAR, FAVAR, and BVAR variants for each country (Burundi, Kenya, Rwanda, Tanzania, Uganda).
  - Common plotted elements and exact axis tick labels observed:
    - Horizontal axis: Months (ticks shown as sequences: 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35)
    - Example vertical axis ranges and ticks from panels:
      - -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002 0.0025 0.003
      - -0.0025 -0.002 -0.0015 -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002
      - -0.003 -0.002 -0.001 0 0.001 0.002 0.003
      - -0.0008 -0.0006 -0.0004 -0.0002 0 0.0002 0.0004 0.0006 0.0008 0.001
      - -0.0015 -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035
      - -0.002 -0.0015 -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002
      - -0.0025 -0.002 -0.0015 -0.001 -0.0005 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035
      - -0.0035 -0.003 -0.0025 -0.002 -0.0015 -0.001 -0.0005 0 0.0005
    - Panels titled "ImpulseResponsesof CPI" and "ImpulseResponsesof GDP"
    - Shocks shown in panels labeled "Shock to Interest Rate" and "Shock to Reserve Money"
    - Legend entries shown as: SVAR, FAVAR, BVAR

- Figures are numbered:
  - Figure 1. Comparing Variants of VARs for Burundi
  - Figure 2. Comparing Variants of VARs for Kenya
  - Figure 3. Comparing Variants of VARs for Rwanda
  - Figure 4. Comparing Variants of VARs for Tanzania
  - Figure 5. Comparing Variants of VARs for Uganda

* _wp1339 - Appendix II. Full Set of SVAR Impulse Responses — https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1339.pdf_

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1339.pdf_