## _wp12243

## Source details

**Canonical URL:** [_wp12243](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12243.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12243.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12243.pdf.json)

---

### Introduction and scope
- Objective: estimate the neutral real interest rate (NRIR) and assess the monetary stance for ten Latin American countries that use the policy interest rate as the main tool.
- Countries analyzed: Chile, Brazil, Colombia, Mexico, Peru, Uruguay, Dominican Republic, Guatemala, Costa Rica, Paraguay.
- NRIR definition: short-run (operational) neutral real interest rate depicting stable inflation with a closed output gap (over the medium-term—the horizon relevant for monetary policy decisions).
- Real values are deflated by one-year-ahead inflation expectations unless otherwise stated.

### Methodology and analytical approach
- Multiple methodologies applied to estimate NRIR; the estimated NRIR is used to compute the interest rate gap (difference between the actual policy rate and the neutral rate, both in real terms).
- Comparison of monetary stance to the output gap to inspect inter-linkages between monetary policy and economic activity.
- Assessment of macro-prudential policies (MaPPs) effects on NRIR levels and the stance of monetary policy, with focus on Brazil and Peru.
- Claim of novelty: first cross-country study on NRIR developments and monetary stance in Latin America using multiple methodologies.

### Methods summarized
- Static methods
  - Consumption-based CAPM (closed-economy, consumption-smoothing models; Euler equation approach with and without habit persistence).
  - Uncovered Interest Parity (UIP) condition.
- Dynamic methods
  - HP filters, linear de-trending, moving averages.
  - Kalman-filter based small-scale macro models (Laubach and Williams (2003) style).
  - Implicit common stochastic trend between short- and long-term nominal interest rates (term spread approaches, Kalman filter).
  - Dynamic Taylor rule and expected-inflation augmented Taylor rule (Kalman filter state-space estimation).
  - General equilibrium semi-structural (S-I Macro) model (saving-investment equilibrium with IS and Phillips equations); NRIR and potential GDP estimated via Kalman filter.
- Key technical settings and calibrations (selected)
  - HP filter end-of-sample correction: append about 18 months of projections (ARIMA).
  - General equilibrium model smoothness restrictions:
    - σ_η^2 / σ_g^2 = 5
    - σ_φ^2 / σ_v^2 = 1/5
  - Monthly data customary λ = 14400 for trend smoothness.

### Data description (Section IV)
- Static UIP estimations: medium-term inflation and interest rate projections from April 2012 World Economic Outlook.
- Country risk premium proxied by J.P. Morgan’s Emerging Bond Index (EMBI).
- Expected exchange rate depreciation/appreciation: May 2012 Foreign Exchange Consensus Forecasts.
- Consumption-based static estimations: medium-term projections of per-capita GDP potential growth from April 2012 WEO.
- Dynamic estimations: seasonally adjusted monthly economic and financial data for January 2000 to end-2013 (data permitting); quarterly/annual data interpolated when monthly not available.
- End-of-sample bias correction: sample extended post May 2012 using projected data; ARIMA post estimation dynamic forecasting used to calculate 2012M6-13M12 for multiple series.
- Where interest rate observations are spotty, a 12-month moving average for interest rates is used (typically 90-day Treasury bill for short-term and 10-year Treasury bond for long-term).
- Interest rate sources: Haver Analytics and national sources; proxies or interpolations used where policy rates missing.
- Inflation targets: official central bank inflation targets since 2000; if IT adopted later, average of actual annual inflation rate in sample used as target for that year.
- One-year-ahead inflation expectations: one-year-ahead WEO forecasts.
- Output gap estimates: April 2012 World Economic Outlook or IMF internal country desks’ estimations.
- Caveat: Less financially integrated economies in sample (Costa Rica, Dominican Republic, Guatemala, Paraguay, and to some extent Uruguay) have thinner financial markets and scarce long-term interest rate data; dynamic estimates for these countries should be interpreted with caution and a sub-sample of methodologies is used.

### Key empirical findings and stylized facts
- NRIR point estimates:
  - Each country’s NRIR point estimates are usually clustered within a 200 basis points band—particularly for the more developed Latin American economies.
  - More economically and financially developed economies exhibit lower levels of NRIR; Brazil is noted as an exception.
- Time trends:
  - A documented downward trend in the NRIR for all countries in the sample during recent years.
  - Contributing factors: stronger domestic economic fundamentals (lower exchange rate risk and inflation risk premiums, fiscal consolidation) and easing global financial conditions.
  - Near-record low global interest rates following the 2008 global financial crisis affected NRIRs.
- Monetary stance (using data up to May 2012, with end-August 2012 updates referenced)
  - For most countries, the monetary stance is currently appropriate—close to neutral, in line with closing output gaps.
  - End-August data point to monetary easing in Brazil and Mexico given their negative output gaps.
  - Costa Rica, Dominican Republic, Guatemala, and Paraguay have somewhat accommodative monetary policy despite closing output gaps; estimated interest rate gaps might not accurately reflect the monetary stance in these countries due to:
    - Weaker monetary transmission mechanisms;
    - A monetary framework still under development;
    - Segmented short-term funding markets causing the policy rate to not accurately reflect financing conditions in all markets.
- Interest rate gap relationships:
  - The interest rate gap and the output gap are strongly and positively correlated.
  - The estimated interest rate gap (sign and magnitude) is correlated with future GDP growth rates for most countries; periods of accommodative monetary policy (negative interest rate gap) are followed (typically within 9 months) by strong economic expansions.
  - Magnitude matters: negative interest rate gaps approaching zero are followed by a slowdown in economic growth.
  - Comparison with inflation gaps: central banks typically undertake restrictive monetary policies if inflation exceeds the target (and vice-versa). Uruguay and Mexico are exceptions due to particularly persistent inflation rates above target for the whole sample period.

### NRIR: Summary results from different methodologies (percent) — selected reported values
- Brazil: Uncovered Interest Parity 4.5; Consumption-based CAPM 4.5; HP Filter 4.8; Implicit Common Stochastic Trend 5.4; Dynamic Taylor Rule 5.7; Expected-Inflation Augmented Taylor Rule 5.5; General Equilibrium Model 5.5; Average 5.1
- Chile: 1.3; 2.9; 2.0; 2.1; 2.3; 2.2; 1.2; Average 2.0
- Colombia: 2.5; 4.4; 1.9; 1.8; 1.6; 1.7; 2.1; Average 2.3
- Mexico: 2.0; 4.2; 1.7; 1.3; 1.3; 1.3; 2.9; Average 2.1
- Peru: 2.3; 5.0; 1.3; 1.5; 1.8; 1.0; 1.3; Average 2.0
- Uruguay: 3.6; 3.3; 1.3; 2.1; 5.3; -7.2; 3.8 (note: sub-sample of methodologies used due to data limitations)
- Costa Rica: 2.6; 4.1; - - - - ; 3.7; Average 3.5 (sub-sample)
- Dominican Republic: 3.2; 4.2; 1.7; 2.7; 3.8; 3.1; 3.9; Average 3.2
- Guatemala: 2.3; 3.2; - - - ; 2.0; 3.7; Average 2.8 (sub-sample)
- Paraguay: 2.0; 3.8; 1.0; 1.3; 2.2; 2.2; 3.2; Average 2.2
- Descriptive graphical conventions: red dots denote end-August 2012 real policy rate (deflated by expected inflation); rectangles represent values between the 70th and 30th percentile.

### Interpretation and policy-relevant observations
- Cross-country NRIR differences may reflect stronger fundamentals and higher levels of capital account openness and financial development.
- Data limitations and thin financial markets can bias dynamic estimates downward and widen estimate ranges for certain countries.
- For highly dollarized economies, those with fiscal dominance, weak monetary transmission, segmented funding markets, and concentrated banking sectors, NRIR estimates may not fully capture domestic financing conditions—complement NRIR estimates with a financial/monetary condition index to add information on domestic financial conditions.

### Macro-prudential policies (MaPPs) and NRIR
- Preliminary evidence (Brazil and Peru) suggests MaPPs can raise the interest rate gap without altering the policy rate by reducing the NRIR through the credit channel.
- Conjectured mechanism: MaPPs restrain credit growth, reduce demand for loanable funds, contract the output gap, and thereby lower the NRIR.
- Policy implications:
  - In overheating driven by domestic shocks, MaPPs can complement conventional monetary tightening.
  - For external shocks that attract capital flows (e.g., positive terms of trade), MaPPs could substitute for conventional interest rate tightening by directly tightening the credit channel without further increasing capital inflows.
- Research needs: quantify MaPPs’ impact on credit, the output gap, and NRIR; determine whether effects are temporary or permanent.

### Box 1 — Brazil: drivers and observations
- Identified drivers for a high Brazilian NRIR
  - Public debt: around 65 percent of GDP in gross terms.
  - Share of domestic public debt indexed to SELIC: about half.
  - SELIC movements:
    - SELIC rose by 3¾ percentage points between March 2010 and June 2011.
    - SELIC cuts since mid-2011: 4½ percentage points.
  - GDP growth: 2.7 percent in 2011 (with GDP essentially flat in the second half of 2011).
  - Other contributors: low domestic savings, weak creditor rights, lack of full central bank independence (mixed evidence), widespread financial indexation, subsidized lending, higher inflation target, minimum remuneration requirements in saving accounts.
- Empirical/model observations relevant to Brazil
  - NRIR has declined considerably over time but remains high by international standards.
  - Static consumption-CAPM with habit persistence yields plausible NRIR estimates; without habit persistence estimates are implausibly large.
  - Monetary stance: Brazil’s stance has recently turned accommodative in response to a growth slowdown.
  - Transmission caveats: effectiveness reduced by high public sector demand for credit, insufficient exchange rate flexibility, excessive bank concentration, high financial dollarization, low financial intermediation, excess liquidity, segmented short-term funding markets.

### Box 2 — Credit channel strength (LA6) and empirical evidence
- Credit dynamics and trends
  - Carry-trade driven capital inflows increased domestic currency deposits and credit growth.
  - Brazilian private credit increased from 26 percent of GDP in 2000 to about 50 percent in early 2012 (boosted by a higher credit multiplier—M2-to-M0 ratio) despite a marked sterilization effort (declining M0 to net international reserves ratio).
  - Higher credit is observed in rising M2-to-M1 ratio.
- OLS regression evidence (rows reported in country order: Brazil, Chile, Colombia, Mexico, Peru, Uruguay)
  - Monetary policy rate coefficients (as reported): 0.07- 0.23- 0.01- 0.26 ***0.22 *0.32 ***
  - Sterilization: - 13.06 ***0.05 ***0.011.29 ***- 15.18 ***1.06 ***
  - Money multiplier: 3.90 ***0.010.82- 1.27 ***- 0.81 ***14.40 ***
  - REER: 0.29 ***0.11 **0.21 ***- 0.06 *0.58 ***- 0.28 ***
  - Financial account/GDP: - 6.11 **- 9.39 ***- 5.80 *- 2.84- 1.95 **- 20.77 ***
  - Constant and model fit: Adjusted R2: 0.920.340.540.600.670.90; Prob(F-statistic): 0.000000.000000.000000.000000.000000.00000
  - Significance notation: *,**, and *** indicate the 10 percent, 5 percent, and 1 percent level of statistical significance.
- Interpretation of regressions
  - For Peru and Uruguay the monetary policy rate coefficient is positive and statistically significant.
  - For Mexico the monetary policy rate coefficient is negative and statistically significant.
  - Preliminary evidence suggests that in a majority of the countries considered, interest rate hikes either do not reduce credit to the private sector or, in some cases, even increase it.
  - Important endogeneity issues remain; deeper econometric assessment needed.
- MaPPs, sterilization, and quasi-fiscal costs
  - Sterilization of capital inflows increases quasi-fiscal costs; capital outflows do not increase quasi-fiscal deficit.
- General conclusions
  - MaPPs can complement monetary policy for domestic shocks and substitute for it for external shocks.
  - Multiple methodologies produce similar NRIR estimates (most clustered within two percentage points) and show a downward trend in NRIR.
  - Central banks should rely on well-informed subjective judgment about unobservables such as the output gap and NRIR.

### Appendix I — Recent MaPPs in Brazil and Peru (selected entries)
- Timeframe covered: 2006–2012.
- Brazil (selected measures and dates)
  - Apr-06: Quantitative easing — widen definition of securities allowed for repo auctions.
  - May-06: Quantitative easing — increase repo maturity up to three months; placements of 20-year nominal sovereign bonds.
  - Oct-08 / Feb-10 / Dec-10 / Dec-11 / Jul-12: Varying reserve requirement adjustments and uses.
  - Feb-10: Limits on net open currency positions: Long position: 75%; Short position: 15% of capital (Feb-10); Nov-10 revised: Long position: 60%; Short position: 15% of capital. Historical increases preserved exactly (Dec-06 10.5→12; Mar-07→13.5; Jun-07→15; Feb-08→16; Jan-08→17; Apr-08→20; Oct-09→22; Jan-10→24; Jun-10→26; Jul-10→28; Sep-10→30).
  - Jun-10: Limits to foreign currency purchase of domestic pension funds: Daily and weekly limits: 0.85 percent and 1.95 percent of Assets Under Management.
  - Foreign exchange interventions (selected transactions): Jan-06 Sale of US$ 364 million; Feb-06 Purchase of US$59 million; May-06 Purchase of US$0.5 million; Jun-06 Purchase of US$2.5 million; Jul-06 Purchase of US$600 million; Aug-06 Purchase of US$1.41 billion; Sep-06 Purchase of US$166 million; Nov-06 Purchase of US$696 million; Dec-06 Purchase of US$610 million.
- Peru (selected measures and dates)
  - Apr-06 / May-06: Quantitative easing measures similar to Brazil (widen repo securities; increase repo maturity up to three months; placements of 20-year nominal sovereign bonds).
  - Series of legal reserve requirement changes (selected chronology preserved exactly): Apr-08 7→8 percent; May-08 8→8.5 percent; Jul-08 8.5→9 percent; Dec-08 9→7.5 percent; Mar-09 7.5→6 percent; Jul-10 6→7 percent; Aug-10 7→8 percent; Sep-10 8→8.5 percent; Oct-10 8.5→9 percent; Oct-10 9→9.5 percent.
  - Series of marginal/special reserve ratio changes (selected chronology preserved exactly): Apr-08 15→20 percent; Jul-08 20→25 percent; Aug-10 25→12 percent; Sep-10 12→15 percent; Oct-10 15→25 percent; May-12 25→30 percent; Jan-07 30→0 percent; Jul-08 0→9 percent; Oct-08 9→0 percent; Mar-08 30→40 percent; May-08 40→45 percent; Aug-08 45→49 percent; Oct-08 49→35 percent; Dec-08 35→30 percent; Jun-10 30→35 percent; Aug-10 35→45 percent; Sep-10 45→50 percent; Oct-10 50→55 percent.
  - May-12: New special reserve requirement for holders of long-term instruments that exceed two and a half times the effective capital of the financial entity.
  - 2009–2011: Restrictions/taxes to limit foreign access and short-term FX volatility (2009 ban on foreign investors' purchases of central bank bills; 2010 increased fee on foreign purchases of central bank liquidity draining instruments to 400 basis points; Jan-10: 30 percent tax on foreign investors' profits from short term currency futures).
  - Jan-11 / Oct-11: Limits on net derivative position of financial institutions: Jan-11 introduce limit of either 40 percent of assets or S/.400 million, whichever is highest; Oct-11 tighten to 30 percent of assets or S/.350 million, whichever is highest.
  - Dec-08: Countercyclical/dynamic provisioning with activation rule when GDP growth for the last 30 months is 5 percent or higher.

### Appendix II — Data sources and methodological notes (selected)
- Inflation target: National Authorities and Haver Analytics.
- Policy rate, short-term and long-term rates, core inflation, oil inflation: ARIMA post estimation dynamic forecasting for 2012M6-13M12 where needed; seasonal adjustment with X-12 ARIMA additive or multiplicative seasonal adjustment.
- EMBI: JP Morgan through Bloomberg LP. For PRY the EMBI of a country with similar sovereign rating was used.
- Commodity price index: Commodity Research Bureau through Haver Analytics; country-specific commodity index constructed based on trade shares.
- Country-specific series notes (selected exact labels): BRA: Selic Target Rate. CHL: Monetary policy rate. COL: BDLR Intervention Rate. CRI: Policy rate (changes in deposit rate 30-90 days used to interpolate). DOM: Overnight rate (interbank loan rate used as proxy). PER: Reference rate (changes in overnight deposit interest rate used to interpolate). PRY: Money Market rate substitution and EMBI proxy. MEX: CETES 91 days, EMED. URY: ITLUP 90 days and ITLUP 1080 days.
- Miscellaneous: Historical data interpolated where possible; a 12-month moving average used for certain interest rate series.

*Source: _wp12243*

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

### References

### Introduction and scope
- Objective: estimate the neutral real interest rate (NRIR) and assess the monetary stance for ten Latin American countries that use the policy interest rate as the main tool.
- Countries analyzed: Chile, Brazil, Colombia, Mexico, Peru, Uruguay, Dominican Republic, Guatemala, Costa Rica, Paraguay.
- NRIR definition: short-run (operational) neutral real interest rate depicting stable inflation with a closed output gap (over the medium-term—the horizon relevant for monetary policy decisions). Footnotes reference Wicksell (1898), Archibald and Hunter (2001), Bernhardsen and Gerdrup (2007).
- Real values are deflated by one-year-ahead inflation expectations unless otherwise stated.

### Methodology and analytical approach
- Multiple methodologies commonly used in the literature are applied to estimate NRIR for the ten countries.
- The estimated NRIR is used to compute the interest rate gap: difference between the actual policy rate and the neutral rate (both in real terms).
- Comparison of monetary stance to the output gap is performed to inspect inter-linkages between monetary policy and economic activity.
- Assessment of macro-prudential policies (MaPPs) effects on NRIR levels and the stance of monetary policy, with focus on Brazil and Peru.
- Paper claims novelty: first cross-country study on NRIR developments and monetary stance in Latin America using multiple methodologies.

### Key empirical findings and stylized facts
- NRIR estimates:
  - Each country’s NRIR point estimates are usually clustered within a 200 basis points band—particularly for the more developed Latin American economies.
  - More economically and financially developed economies exhibit lower levels of NRIR; Brazil is noted as an exception.
- Time trends:
  - A documented downward trend in the NRIR for all countries in the sample during recent years.
  - Contributing factors: stronger domestic economic fundamentals (lower exchange rate risk and inflation risk premiums, fiscal consolidation) and easing global financial conditions.
  - Near-record low global interest rates following the 2008 global financial crisis affected NRIRs.
- Monetary stance (using data up to May 2012):
  - For most countries, the monetary stance is currently appropriate—close to neutral, in line with closing output gaps.
  - More recent data (at end-August) point to monetary easing in Brazil and Mexico given their negative output gaps.
  - Costa Rica, Dominican Republic, Guatemala, and Paraguay have somewhat accommodative monetary policy despite closing output gaps; however, estimated interest rate gaps might not accurately reflect the monetary stance in these countries due to:
    - Weaker monetary transmission mechanisms;
    - A monetary framework still under development;
    - Segmented short-term funding markets causing the policy rate to not accurately reflect financing conditions in all markets.
- Interest rate gap relationships:
  - The interest rate gap and the output gap are strongly and positively correlated.
  - Correlation could indicate central banks respond counter-cyclically to business cycle fluctuations.
  - Periods of relaxing monetary policy (decreasing interest rate gaps) are followed by shrinking (negative) output gaps, and vice versa.
  - Estimated interest rate gap (sign and magnitude) is correlated with future GDP growth rates for most countries; periods of accommodative monetary policy (negative interest rate gap) are followed (typically within 9 months) by strong economic expansions.
  - Magnitude matters: examples where a negative interest rate gap approaches zero are followed by a slowdown in economic growth.
  - Comparison with inflation gaps: central banks typically undertake restrictive monetary policies if inflation exceeds the target (and vice-versa). Uruguay and Mexico are exceptions due to particularly persistent inflation rates above target for the whole sample period.

### Macro-prudential policies (MaPPs) and the NRIR
- Preliminary evidence indicates Brazil and Peru successfully tightened their monetary stance (raised the interest rate gap) via MaPPs without altering their policy rate on several occasions (2006, 2008, and 2010).
- Conjecture: increase in the interest rate gap was achieved by reducing the NRIR, possibly through contracting the output gap (quantifying and rigorously analyzing these effects is left for future research).
- Mechanism: NRIR appears affected by the credit channel. Recent carry-trade driven capital inflows increased domestic currency deposits and credit growth; MaPPs seem to have lowered the NRIR by mitigating the expansionary effect of the credit channel on GDP by containing demand for loanable funds.
- Policy implication: In overheating situations, MaPPs could be complementary to conventional monetary policy. For external shocks that attract capital flows (e.g., positive terms of trade shocks), MaPPs could act as a substitute to conventional interest rate policy by directly tightening the credit channel without further increasing capital inflows.

### Contribution and organization
- Paper contribution: first cross-country perspective on NRIR developments and monetary stance in Latin America using multiple methodologies; existing literature usually focuses on single-country studies (mostly Brazil, Chile, and Colombia) and limited methodologies.
- Organization: Section II reviews literature and methodology pros/cons; Section III describes the approaches used to estimate NRIR.

*Source: _wp12243 - References*

### Section IV delves into the data set briefly. In Section V we present the results, as well as the

### _wp12243 - Section IV delves into the data set briefly. In Section V we present the results, as well as the

### Existing literature
- Most studies estimate the NRIR in advanced economies and typically concentrate on one country.
- Few studies estimate the NRIR for emerging economies; Latin America studies largely focus on Chile, Colombia and Brazil.
- References cited as examples (listed in the source) include Laubach and Williams (2003); Bernhardsen and Gerdrup (2007); ECB (2004); Bjorksten and Karagedikli (2003); Lam and Tkacz (2004); Adolfson et al. (2011); Ogunc and Batmaz (2011); Calderon and Gallego (2002); Fuentes and Gredig (2007); Minella et al. (2002); Portugal and Barcellos (2009); Duarte (2010); Perrelli (2012); Pereda (2010); Humala and Rodriguez (2009); Castillo et al. (2006); Gonzalez et al. (2010, 2012); Torres (2007).

### Methodological approaches (overview)
- No single best method for estimating the neutral real interest rate (NRIR); a broad array of alternative methods is presented to provide a range of magnitudes.
- Methods summarized in the paper:
  - Static methods
    - Consumption-based CAPM (closed-economy, consumption-smoothing models; Euler equation approach with and without habit persistence following Cochrane (2001) and Campbell and Cochrane (1999)).
    - Uncovered Interest Parity (UIP) condition.
  - Dynamic methods
    - Simple statistical/filtering techniques: HP filters, linear de-trending, moving averages (not structurally grounded; sensitive to sample period and end-of-sample bias).
    - Kalman-filter based small-scale macro models (Laubach and Williams (2003) style): reduced-form IS and backward-looking Phillips curve variants.
    - Implicit common stochastic trend between short- and long-term nominal interest rates (term spread approaches, Kalman filter estimation).
    - Dynamic Taylor rule and expected-inflation augmented Taylor rule (Kalman filter state-space estimation).
    - General equilibrium semi-structural (S-I Macro) model in the spirit of Laubach and Williams (2003) – saving-investment equilibrium with IS and Phillips equations; NRIR and potential GDP estimated via Kalman filter.

### Static methodologies — technical details and calibration
- Consumption-smoothing models
  - Euler equation representation includes the real interest rate r_t, intertemporal discount factor β, utility function u(.), expectation operator E(.), consumption c and per-capita potential GDP y; resource constraint and CRRA assumption lead to a log-linearized form:
    - ln(r_t) = ln(β) + γ * Δ ln(y_{t+1}) - (γ^2 / 2) Var(Δ ln(y_{t+1}))
  - Habit persistence variant: surplus consumption ratio S_t and habit parameter x_t; Euler equation leads to:
    - ln(r_t) = ln(β) + γ g - (1/2) γ (σ^2) (1/φ1 - ...)
  - Parameters γ and β are treated as plausible free parameters; φ is calibrated using risk aversion and discount factor parameters for a given level of potential GDP.
- Uncovered Interest Parity (UIP)
  - Nominal domestic and international rates linked via expected nominal depreciation and country risk premium:
    - i_t = i_t* + E(Δe_t) + ρ
  - Expected nominal depreciation decomposed into rate of depreciation of the real exchange rate and domestic-international inflation differential:
    - E(Δe_t) = Δq_t + (π_t - π_t*)
  - Assumptions for static UIP estimation:
    - International nominal interest and inflation rates set to 4 and 2 percent, respectively.
    - Country-specific risk premium proxied by J.P. Morgan’s EMBI spreads.
    - Expected depreciation rates from medium-term Consensus Forecasts; a range reported as ±1 percentage point around the mean estimate.

### Dynamic methodologies — technical details
- HP Filters
  - Standard Hodrick-Prescott (HP) filter applied to real short-term interest rate series (typically 3-month Treasury bill deflated by 12-month-ahead inflation expectations).
  - To minimize end-of-sample bias, about 18 months of projections are appended to the sample (ARIMA procedure details in Appendix II).
- Implicit common stochastic trend (term spread / four-equation dynamic system)
  - Four-equation system links nominal short-term rate r_t, long-term rate R_t, unobserved NRIR r_t*, and term premium α_t; NRIR follows a random walk; term premium is AR(1).
  - Estimated using a Kalman filter; simultaneous shifts in both long- and short-term rates (after cyclical adjustment) interpreted as shifts in the NRIR.
- Dynamic Taylor Rule
  - Monetary policy rate responds to inflation deviation from target and output gap; when both deviations are zero, the constant equals the nominal neutral rate.
  - System estimated by Kalman filter with NRIR modeled as a random walk; notation includes r_t (90-day), r_t* (neutral nominal), π_t, π_t* (inflation target), and y_gap.
- Expected-Inflation Augmented Taylor Rule
  - Variant where neutral interest rate is in real terms and an equation for the long-term nominal interest rate with one-year ahead inflation expectations is introduced; long-term rate equals short-term rate plus premium α.
- General Equilibrium Model (S-I Macro Model)
  - IS equation: output gap as function of its lags, lagged deviations of actual real monetary policy rate from NRIR, and control variables x1,t (cyclical real exchange rate deviations); disturbances assumed zero-mean white noise with variance σ_ε^2.
  - Phillips curve: inflation deviations from target explained by lagged inflation, lagged output gap, and inflation controls x2,t (cyclical real exchange rate and oil/commodity prices); disturbance variance σ_u^2.
  - NRIR and potential GDP estimated with Kalman filter; NRIR follows a random walk with residual variance σ_v^2; potential GDP growth g follows a random walk with variance σ_g^2.
  - Smoothness restrictions imposed in maximum likelihood estimation:
    - σ_η^2 / σ_g^2 = 5
    - σ_φ^2 / σ_v^2 = 1/5
  - Monthly data customary λ = 14400 for trend smoothness.

### Data description (Section IV)
- Static UIP estimations use medium-term inflation and interest rate projections from April 2012 World Economic Outlook.
- Country risk premium proxied by J.P. Morgan’s Emerging Bond Index (EMBI).
- Expected exchange rate depreciation/appreciation from May 2012 Foreign Exchange Consensus Forecasts.
- Consumption-based static estimations use medium-term projections of per-capita GDP potential growth from April 2012 World Economic Outlook.
- Dynamic estimations use seasonally adjusted monthly economic and financial data for January 2000 to end-2013 (data permitting); quarterly/annual data interpolated when monthly not available.
- To correct end-of-sample bias, sample period extended post May 2012 using projected data.
- Where interest rate observations are spotty, a 12-month moving average for interest rates is used (typically 90-day Treasury bill for short-term and 10-year Treasury bond for long-term).
- Interest rate sources: Haver Analytics and national sources; proxies or interpolations used where policy rates missing (details in Appendix II).
- Inflation targets: official central bank inflation targets since 2000; if IT adopted later, average of actual annual inflation rate in sample used as target for that year.
- One-year-ahead inflation expectations: one-year-ahead WEO forecasts (similar results if expectations implied by indexed bonds used when available).
- Output gap estimates: April 2012 World Economic Outlook or IMF internal country desks’ estimations (statistical filters or production function approaches).
- Caveat: Less financially integrated economies in sample (Costa Rica, Dominican Republic, Guatemala, Paraguay, and to some extent Uruguay) have thinner financial markets and scarce long-term interest rate data; dynamic estimates for these countries should be interpreted with caution and a sub-sample of methodologies is used.

### Results (Section V) — summary findings and key statistics
- Point estimates across methodologies are clustered for each country, typically within 200 basis points, and consistent with country-specific studies.
- Dynamic estimates are somewhat lower than static estimates in less financially open economies—possibly reflecting limitations in financial data and thinner yield curves.
- Less financially developed economies exhibit a larger range of estimates; only methodologies yielding reasonable results given data reliability are reported.
- Cross-country patterns:
  - NRIR typically lower in (i) more economically and financially developed economies, and (ii) countries with a longer inflation-targeting (IT) history; other country-specific factors also matter.
  - Notable exception: Brazil, where NRIR is among the highest in the region (referred to as the Brazilian interest rate puzzle).
  - For highly dollarized countries and those with large fiscal dominance, weaker monetary transmission, segmented short-term funding markets, and banking sector concentration, the estimated NRIR may not fully capture domestic financing conditions.
- Tabulated summary (authors' calculations) — NRIR: Summary Results from Different Methodologies (percent) — selected values as reported:
  - Brazil: Uncovered Interest Parity 4.5; Consumption-based CAPM 4.5; HP Filter 4.8; Implicit Common Stochastic Trend 5.4; Dynamic Taylor Rule 5.7; Expected-Inflation Augmented Taylor Rule 5.5; General Equilibrium Model 5.5; Average 5.1
  - Chile: 1.3; 2.9; 2.0; 2.1; 2.3; 2.2; 1.2; Average 2.0
  - Colombia: 2.5; 4.4; 1.9; 1.8; 1.6; 1.7; 2.1; Average 2.3
  - Mexico: 2.0; 4.2; 1.7; 1.3; 1.3; 1.3; 2.9; Average 2.1
  - Peru: 2.3; 5.0; 1.3; 1.5; 1.8; 1.0; 1.3; Average 2.0
  - Uruguay: 3.6; 3.3; 1.3; 2.1; 5.3; -7.2; 3.8 (note: sub-sample of methodologies used due to data limitations)
  - Costa Rica: 2.6; 4.1; - - - - ; 3.7; Average 3.5 (sub-sample)
  - Dominican Republic: 3.2; 4.2; 1.7; 2.7; 3.8; 3.1; 3.9; Average 3.2
  - Guatemala: 2.3; 3.2; - - - ; 2.0; 3.7; Average 2.8 (sub-sample)
  - Paraguay: 2.0; 3.8; 1.0; 1.3; 2.2; 2.2; 3.2; Average 2.2
- Graphical summary (descriptive): red dots denote end-August 2012 real policy rate (deflated by expected inflation); rectangles represent values between the 70th and 30th percentile.

### Interpretation and policy-relevant observations
- Cross-country NRIR differences may reflect stronger fundamentals and higher levels of capital account openness and financial development (see Archibald and Hunter, 2001).
- Data limitations and thin financial markets can bias dynamic estimates downward and widen estimate ranges for certain countries.
- For highly dollarized economies, those with fiscal dominance, weak monetary transmission, segmented funding markets, and concentrated banking sectors, NRIR estimates may not fully capture domestic financing conditions—complementing NRIRs with a financial/monetary condition index is recommended to add information on domestic financial conditions.

*Source: Authors' calculations and text from the provided IMF content unit.*

### Box 1. Why is Brazil’s Neutral Real Interest Rate so High?

### Box 1. Why is Brazil’s Neutral Real Interest Rate so High?

### Identified drivers for a high Brazilian NRIR
- Fiscal considerations
  - Brazilian public debt is around 65 percent of GDP in gross terms.
  - About half of the domestic public debt is indexed to the SELIC, creating endogeneity between the policy rate (SELIC) and public debt and restricting monetary policy degrees of freedom; this feeds back into a higher SELIC, and thus a higher NRIR (World Bank, 2006).
  - Rogoff (2005) argues Brazil incurs a significant default risk premium due to its inflationary history; empirical reinforcement by World Bank (2006).
- Low domestic savings
  - Brazil’s low domestic savings, and thus investment, is cited as a reason for a higher NRIR (Fraga, 2005; Miranda and Muinhos; 2003; Hausmann, 2008; and Segura, 2012).
  - Segura (2012) finds that low domestic savings cannot adequately explain the cross-country discrepancy.
- Institutional factors
  - Weak creditor rights and contractual enforcement have been cited as possible explanations (Arida et al., 2004; Rogoff, 2005).
  - Lack of full central bank independence is also proposed as an explanation, though Nahon and Meuer (2009) find no changes in central bank credibility due to recent changes in its Board of Directors.
- Widespread financial indexation
  - Extensive indexation of financial contracts to the overnight interest rate has produced unusually short-duration financial contracts, institutional and psychological inertia, and path-dependency that is difficult to dislodge (World Bank, 2006).
  - This legacy of indexation has made inflation less responsive to interest rate changes.
- Other Brazil-specific factors
  - Subsidized lending has resulted in credit market segmentation, pushing up market-determined interest rates.
  - Additional potential contributors include an inflation target higher than in other emerging markets and minimum remuneration requirements in saving accounts (see Segura (2012) and Central Bank of Brazil (2012) for details).

### Empirical and model-based observations relevant to Brazil (from wider LA6 analysis)
- Overview
  - The NRIR has declined considerably over time but remains high by international standards.
  - Static consumption-CAPM with habit persistence yields NRIR estimates; without habit persistence estimates are implausibly large.
- Interest rate policy and recent historical movements
  - SELIC rose by 3¾ percentage points between March 2010 and June 2011.
  - GDP growth was at 2.7 percent in 2011, with GDP essentially flat in the second half of 2011.
  - SELIC cuts of 4½ percentage points since mid-2011 have placed the interest rate gap again in negative territory (ceteris paribus implying stronger economic growth based on the model).
  - About half of domestic public debt being indexed to SELIC reinforces higher NRIR via fiscal-monetary interactions (see fiscal considerations above).
- Monetary stance and transmission caveats
  - Brazil’s stance has recently turned accommodative in response to a growth slowdown.
  - Effectiveness of monetary policy may be reduced by factors that raise effective market interest rates for the private sector (high public sector demand for credit, insufficient exchange rate flexibility, excessive bank concentration, high financial dollarization, low financial intermediation).
  - Weak monetary transmission (e.g., excess liquidity, segmented short-term funding markets) can make policy rates less reflective of financing conditions across markets.

### Macro-prudential policies (MaPPs) as a complement/substitute to interest rate policy (implications for NRIR)
- Use and objectives
  - MaPPs covered include changing reserve requirements, imposing limits on currency mismatches or loan-to-value ratios, imposing specific asset risk weights, etc.
  - Brazil and Peru relied mostly on restrictive MaPPs; Peru was the most active.
- Observed effects (preliminary)
  - Graphical inspection suggests Brazil and Peru successfully tightened their monetary stances (raised interest rate gaps) via MaPPs without altering their policy rate on several occasions (2006, 2008, and 2010) by lowering the NRIR.
  - Mechanism conjectured: MaPPs contracted the output gap by restraining credit growth, reducing demand for loanable funds and lowering the NRIR.
- Complementarity versus substitution
  - If overheating arises from a domestic shock, MaPPs can be complementary to conventional monetary tightening by containing credit-driven expansion that might otherwise be amplified by capital inflows.
  - If the shock is external (e.g., positive terms of trade attracting capital flows), MaPPs could be superior to conventional interest rate tightening because they tighten the credit channel directly without further increasing capital inflows.
- Research needs and caveats
  - More research is needed to quantify MaPPs’ impact on credit, the output gap, and NRIR, and to determine whether effects are temporary or permanent.
  - Preliminary evidence suggests the credit channel may be relevant in LA6, but deeper analysis is required.

### Key numeric facts and time-specific indicators preserved exactly as reported
- Public debt: around 65 percent of GDP in gross terms.
- Share of domestic public debt indexed to SELIC: about half.
- SELIC change: rose by 3¾ percentage points between March 2010 and June 2011.
- SELIC cuts since mid-2011: 4½ percentage points.
- GDP growth: 2.7 percent in 2011 (with GDP essentially flat in the second half of 2011).

*Source: Box 1. Why is Brazil’s Neutral Real Interest Rate so High? (excerpt)._wp12243 - Box 1. Why is Brazil’s Neutral Real Interest Rate so High?*

### Box 2. How Strong is the Credit Channel in Latin America?

### Box 2. How Strong is the Credit Channel in Latin America?

### Credit dynamics and recent trends
- LA6 economies experienced a surge in (carry-trade driven) capital inflows, increasing domestic currency deposits and thus credit growth.
- Brazilian private credit increased from 26 percent of GDP in 2000 to about 50 percent in early 2012 (boosted by a higher credit multiplier—M2-to-M0 ratio) despite a marked sterilization effort (reflected in the declining M0 to net international reserves ratio).
- Higher credit is also observed in the rising M2-to-M1 ratio, as more long-term deposits facilitate greater credit growth by the banking system.

### Empirical evidence on the credit channel (OLS regressions for LA6)
- After controlling for the money multiplier (M2/M0), the degree of sterilization (M0/NIR), the real effective exchange rate (REER), and the capital and financial account balance (as a percentage of GDP), the coefficient of the monetary policy rate in determining the credit-to-GDP ratio is oftentimes positive; statistical significance varies by country.
- Regression coefficient rows (country order: Brazil, Chile, Colombia, Mexico, Peru, Uruguay) as reported:
  - Monetary policy rate: 0.07- 0.23- 0.01- 0.26 ***0.22 *0.32 ***
  - Sterilization: - 13.06 ***0.05 ***0.011.29 ***- 15.18 ***1.06 ***
  - Money multiplier: 3.90 ***0.010.82- 1.27 ***- 0.81 ***14.40 ***
  - REER: 0.29 ***0.11 **0.21 ***- 0.06 *0.58 ***- 0.28 ***
  - Financial account/GDP: - 6.11 **- 9.39 ***- 5.80 *- 2.84- 1.95 **- 20.77 ***
  - Constant: 1.1645.95 ***- 11.4827.15 ***- 23.38 ***7.75
  - Adjusted R2: 0.920.340.540.600.670.90
  - Prob(F-statistic): 0.000000.000000.000000.000000.000000.00000
- Significance notation: *,**, and *** indicate the 10 percent, 5 percent, and 1 percent level of statistical significance.
- Interpretation:
  - For Peru and Uruguay the monetary policy rate coefficient is positive and statistically significant.
  - For Mexico the monetary policy rate coefficient is negative and statistically significant.
  - Preliminary evidence suggests that in a majority of the countries considered, interest rate hikes either do not reduce credit to the private sector or, in some cases, even increase it.
- Caveats: Important endogeneity issues remain and a deeper econometric assessment is needed to confirm robustness and to explore dependence on open economy features (e.g., exchange rate regimes).

### MaPPs, sterilization, and quasi-fiscal costs
- MaPPs (macroprudential policies) seemed to have been employed to mitigate the expansionary effect of the credit channel, also resulting in a lower NRIR.
- MaPPs could affect the output gap via changes in demand for loanable funds (credit channel).
- Sterilization symmetry and fiscal cost:
  - If capital inflows are sterilized (purchasing reserves with domestic paper) the quasi-fiscal cost increases.
  - Capital outflows do not increase the quasi-fiscal deficit, as they entail selling reserves while purchasing domestic assets by issuing domestic currency.

### Policy implications and conclusions (selected)
- MaPPs can:
  - Complement conventional monetary policy when an economy faces domestic shocks.
  - Substitute for interest rate policy in the case of external shocks.
  - Affect the interest rate gap through the NRIR even when the policy rate remains unchanged.
- Benefits: Monetary policy combined with MaPPs can be a stronger mechanism to smooth business cycles.
- Research needs: Further work is required to:
  - Understand the mechanics of MaPPs.
  - Quantify their impact on credit, the output gap, and NRIR.
  - Investigate whether their effects are temporary or permanent.
- On NRIR and policy making:
  - Multiple methodologies produce similar NRIR estimates (most clustered within two percentage points) and show a downward trend in NRIR.
  - No single methodology yields “the” correct NRIR; central banks must therefore rely on well-informed subjective judgment about unobservables such as the output gap and NRIR.
  - The paper’s NRIR-based interest rate gap analysis finds:
    - The current monetary stance is appropriately neutral in most sample countries, in line with closing output gaps.
    - The stance remains stimulative in Brazil and Mexico given their negative output gaps.
    - Costa Rica, Dominican Republic, Guatemala, and Paraguay still have a somewhat accommodative monetary policy despite closing output gaps.
    - The interest rate gap is correlated with future economic growth and inflation.

*Source: Authors' calculations.*

### Appendix I: Recent Macroprudential Measures in Brazil and Peru

### Appendix I: Recent Macroprudential Measures in Brazil and Peru

### Overview
- Timeframe covered: 2006–2012.
- Focus: recent macroprudential instruments and changes in Brazil and Peru, including quantitative easing (repo operations), reserve requirement adjustments, differentiated reserve regimes, sector-dependent asset risk weights, limits on currency and derivative positions, countercyclical/dynamic provisioning, and foreign exchange interventions.
- Objective categories repeatedly cited: to improve access to short-term funding; to widen the yield curve; to manage capital flows; to build/accumulate a buffer stock of international reserves; to increase availability of international liquidity for local banks; to avoid short-term capital inflows translating into unsustainable credit expansion; to increase or decrease liquidity in the financial system; to increase the risk-sensitivity of capital requirements; to reduce exchange rate volatility; to mitigate procyclicality.

### Brazil — Key measures, dates, and stated reasons
- Apr-06: Quantitative easing — Widening the definition of securities allowed for repo auctions.
  - Policy change: Widen the definition of securities allowable for temporary repos and allowed private sector non-financial securities in local currency to be used for direct repos.
  - Reason: To improve access to short-term funding for financial entities.
- May-06: Quantitative easing — Increasing instruments maturity.
  - Policy change: Increase the maturity on repos to up to three months and make placements of 20-year nominal sovereign bonds for the first time.
  - Reason: To widen the yield curve.
- Feb-10: Sector-dependent asset risk weights / Loan-to-value and maturity-dependent asset risk weights.
  - Policy change: Higher risk weights for certain automobile and personal loans; higher risk weights introduced for certain categories of automobile and personal loans with longer maturies and higher loan-to-value ratios.
  - Reason: To increase the risk-sensitivity of capital requirements in a scenario of rapid growth in credit to these sectors.
- Nov-11: Recalibration of consumer loan capital requirements.
  - Policy change: A recalibration lowered the capital requirements for consumer loans according to their maturity, removing the loan-to-value ratio criteria.
- Oct-08 / Feb-10 / Dec-10 / Dec-11 / Jul-12: Varying reserve requirements and use of reserves.
  - Examples and policy uses:
    - Oct-08: Decrease in required bank reserves, with liquidity channeled to smaller financial institutions.
    - Feb-10: Increase in required bank reserves.
    - Dec-10: Increase in required bank reserves.
    - Dec-11: Large banks may acquire small banks using reserve requirements on time deposits—initially a temporary measure taken in October 2008; it allows large banks to use the non-remunerated part of the required reserves on time deposits to acquire small bank assets.
    - Jul-12: Cut the "additional" bank-reserve requirements on cash deposits to 6 percent from 12 percent and raise the requirement for credit directed to the farm sector to 34 percent from 28 percent.
  - Reason: To increase or decrease liquidity in the financial system.
- Feb-10: Limits on net open currency positions (selected values).
  - Policy values: Long position: 75%; Short position: 15% of capital (Feb-10); Nov-10 revised: Long position: 60%; Short position: 15% of capital.
  - Additional historical series of limit increases (selected entries preserved exactly): Dec-06 Increase the limit from 10.5 percent to 12 percent; Mar-07 Increase the limit to 13.5 percent; Jun-07 Increase the limit to 15 percent; Feb-08 Increase the limit to 16 percent; Jan-08 Increase the limit to 17 percent; Apr-08 Increase the limit to 20 percent; Oct-09 Increase the limit to 22 percent; Jan-10 Increase the limit to 24 percent; Jun-10 Increase the limit to 26 percent; Jul-10 Increase the limit to 28 percent; Sep-10 Increase the limit to 30 percent.
  - Reason: Reduce the volatility of the domestic currency.
- Jun-10: Limits to foreign currency purchase of domestic pension funds.
  - Policy values: Daily and weekly limits: 0.85 percent and 1.95 percent of Assets Under Management.
  - Reason: Reduce the volatility of the domestic currency.
- Foreign exchange interventions (selected transactions, Peru table includes similar FX rows for Brazil context):
  - Jan-06 Sale of US$ 364 million.
  - Feb-06 Purchase of US$59 million.
  - May-06 Purchase of US$0.5 million.
  - Jun-06 Purchase of US$2.5 million.
  - Jul-06 Purchase of US$600 million.
  - Aug-06 Purchase of US$1.41 billion.
  - Sep-06 Purchase of US$166 million.
  - Nov-06 Purchase of US$696 million.
  - Dec-06 Purchase of US$610 million.

### Peru — Key measures, dates, and stated reasons
- Reserve requirements to long-term external liabilities (in foreign currency, more than two years).
  - Stated reason: To manage capital flows and build a accumulate a buffer stock of international reserves.
- Differentiated reserved requirements for foreign currency: general regime, marginal requirement for foreign currency deposits.
  - Stated reason: To increase availability of international liquidity for local banks. These measures are also intended to avoid short term capital inflows and liquidity to translate into an unsustainable expansion of credit.
- Changes in legal minimum reserve requirements and changes in marginal reserve requirements for domestic currency deposits.
  - Stated reason: As a complement to policy changes; to manage capital flows.
- Apr-06: Quantitative easing — Widening the definition of securities allowed for repo auctions.
  - Policy change: Widen the definition of securities allowable for temporary repos and allowed private sector non-financial securities in local currency to be used for direct repos.
  - Reason: To improve access to short-term funding for financial entities.
- May-06: Quantitative easing — Increasing instruments maturity.
  - Policy change: Increase the maturity on repos to up to three months and make placements of 20-year nominal sovereign bonds for the first time.
  - Reason: To widen the yield curve.
- Series of changes to legal reserve requirements (selected chronology preserved exactly):
  - Apr-08 Increase the legal reserve requirements from 7 to 8 percent.
  - May-08 Increase the ratio to 8.5 percent.
  - Jul-08 Increase the ratio to 9 percent.
  - Dec-08 Decrease the ratio to 7.5 percent.
  - Mar-09 Decrease the ratio to 6 percent.
  - Jul-10 Increase the ratio to 7 percent.
  - Aug-10 Increase the ratio to 8 percent.
  - Sep-10 Increase the ratio to 8.5 percent.
  - Oct-10 Increase the ratio to 9 percent.
  - Oct-10 Increase the ratio to 9.5 percent.
- Series of changes to marginal or special reserve ratios (selected chronology preserved exactly):
  - Apr-08 Increase ratio from 15 to 20 percent.
  - Jul-08 Increase ratio to 25 percent.
  - Aug-10 Decrease ratio to 12 percent.
  - Sep-10 Increase ratio to 15 percent.
  - Oct-10 Increase ratio to 25 percent.
  - May-12 Increase ratio to 30 percent.
  - Jan-07 Decrease the ratio to zero percent from 30 percent.
  - Jul-08 Increase the ratio to 9 percent.
  - Oct-08 Decrease the ratio to zero percent.
  - Mar-08 Increase the reserve ratio from 30 percent to 40 percent.
  - May-08 Increase the ratio to 45 percent.
  - Aug-08 Increase the ratio to 49 percent.
  - Oct-08 Decrease the ratio to 35 percent.
  - Dec-08 Decrease the to 30 percent.
  - Jun-10 Increase the ratio to 35 percent.
  - Aug-10 Increase the ratio to 45 percent.
  - Sep-10 Increase the ratio to 50 percent.
  - Oct-10 Increase the ratio to 55 percent.
- May-12: New reserve requirement (special rule).
  - Policy change: Introduce a new special reserve requirement for holders of long-term instruments, such as bonds, excluding sol-denominated mortgage bonds, that exceed two and a half times the effective capital of the financial entity.
  - Reason: To moderate credit growth and prevent future demand pressures on the economy.
- 2009–2011: Restrictions and taxes to limit foreign access and short-term FX volatility.
  - 2009: Ban on foreign investors' purchases of central bank bills.
  - 2010: Increased fee on foreign purchases of central bank liquidity draining instruments to 400 basis points.
  - Jan-10: Government imposed a 30 percent tax on foreign investors' profits from short term currency futures.
  - Stated reason: Avoid volatility of the domestic currency.
- Jan-11 / Oct-11: Limits on net derivative position of financial institutions.
  - Jan-11: Introduce a limit of either 40 percent of assets or S/.400 million, whichever is the highest.
  - Oct-11: Tighten the limit to 30 percent of assets or S/.350 million, whichever is the highest.
  - Stated reason: To mitigate short-term capital flows and exchange rate volatility by extending the maturity of foreign debt of domestic banks, thus increasing domestic banks’ resilience to sudden stops.
- Dec-08: Countercyclical/dynamic provisioning.
  - Policy change: Introduce generic provision to banking and non-banking loans (microfinance). There are three activation rules, the most important one is when GDP growth for the last 30 months is 5 percent or higher.
  - Reason: To reduce the procyclicality of the banking business.
- Measures on reserve requirements to domestic currency liabilities of domestic banks with non-residents and to short-term external liabilities (in foreign currency, up to two years).
  - Stated reasons: To mitigate short-term capital flows and exchange rate volatility by reducing the use of bank’s deposits in domestic currency as a vehicle to take long positions in Soles; to manage foreign exchange risks; to mitigate short-term capital flows and exchange rate volatility by extending the maturity of foreign debt of domestic bank, thus increasing domestic bank’s resilience to sudden stops.
- Other measures: Restrict foreign investors' access to central bank instruments.
  - Stated reason: To mitigate short-term capital flows and exchange rate volatility.

### Additional notes on data, indicators, and sources (as referenced in appendix)
- Inflation data: Haver Analytics; for the period 2012M6-2013M12 projections from the 2012 April WEO were interpolated; seasonal adjustment with X-12 ARIMA additive seasonal adjustment.
- 12-month ahead inflation expectations: 2012 April WEO; data interpolated by obtaining the 12-month ahead inflation projection of the corresponding year; for 2012M6-13M12 the last published value was repeated.
- International reserves and exchange rate: International Financial Statistics.
- Real effective exchange rate: Information Notice System.
- Country-specific notes on series definitions and proxies (selected examples preserved exactly):
  - BRA: Selic Target Rate.
  - CHL: Monetary policy rate.
  - COL: BDLR Intervention Rate.
  - CRI: Policy rate; changes in the deposit rate (30-90 days) was used to interpolate historical data.
  - DOM: Overnight rate; interbank loan rate was used as a proxy for historical data.
  - PER: Reference rate, changes in the overnight deposit interest rate was used to interpolate historical data.

*Sources: IMF (2011c, 2012d); Mihalijek and Sybelyte (2011); Rossini, Quispe and Rodriquez (2011); Tovar, Garcia-Escribano, and Vera-Martin (2012); and authors' research based on national sources.*

### Appendix II: Data Sources and Description

### Appendix II: Data Sources and Description

### Variable sources and comments
- Inflation target: National Authorities and Haver Analytics
- Policy rate: In order to calculate 2012M6-13M12 an ARIMA post estimation dynamic forecasting was used.
- Emerging Markets Bond Indices: JP Morgan through Bloomberg LP. For PRY the EMBI of a country with similar sovereign rating was used.
- M0, M1, M2: (source not separately listed in table)
- Core inflation: In order to calculate 2012M6-13M12 an ARIMA post estimation dynamic forecasting was used. Data was seasonally adjusted using an X-12 ARIMA additive seasonal adjustment.
- Oil inflation: (comment entries show repeated methodology) In order to calculate 2012M6-13M12 an ARIMA post estimation dynamic forecasting was used. Data was seasonally adjusted using an X-12 ARIMA additive seasonal adjustment.
- Short-term interest rate: In order to calculate 2012M6-13M12 an ARIMA post estimation dynamic forecasting was used.
- Long term rate: In order to calculate 2012M6-13M12 an ARIMA post estimation dynamic forecasting was used.
- Commodity price index: Commodity Research Bureau through Haver Analytics. A commodity index was constructed for each country based on the country’s trade shares for each commodity.
- Credit to the private sector; Real and Nominal GDP; Public consumption: 2012 April WEO and author’s calculations.
- Public Debt: 2012 April WEO.
- Potential GDP: 2012 April WEO.
- Population: 2012 April WEO for all countries.

### Country-specific rate and series notes (as listed)
- BRA: Swaps Reference Rate: Daily: 90 days. BRA: 10-year bond, Datastream. Historical data were interpolated, if possible.
- CHL: PDBC 90 days, Haver Analytics. CHL: BCP 5 year, Haver Analytics. A 12-month moving average was used.
- COL: 90 day certificate of deposit, Haver Analytics. COL: 8 year bond, Bloomberg LP.
- CRI: 90-180 day Central Directo por plazos de vencimiento, Central Bank. CRI: 5 year BEM offer rate, National Authorities.
- PER: Peru 3-month bond. PER: Peru 8-year bond.
- PRY: Money Market rate, International Financial Statistics. PRY: Money Market rate substitution: For PRY the EMBI of a country with similar sovereign rating was used. PRY: Lending rate (foreign currency), IFS.
- DOM: Interbank loan rate, Haver Analytics. DOM: Certificado de inversión especial, 7 años, National Authorities.
- GTM: Rate on open-market operations, National Authorities. GTM: 10 year yield, primary market, National Authorities.
- MEX: CETES 91 days, EMED. MEX: 9 year bond, Bloomberg LP.
- URY: ITLUP 90 days, Bolsa Electronica de Valores del UruGTMy. URY: ITLUP 1080 days, Bolsa Electronica de Valores del UruGTMy.

### Seasonal adjustment, interpolation, and forecasting methods
- ARIMA post estimation dynamic forecasting was applied to calculate 2012M6-13M12 for multiple series.
- Data was seasonally adjusted using an X-12 ARIMA additive seasonal adjustment for some series.
- Data was seasonally adjusted using an X-12 ARIMA multiplicative seasonal adjustment for other series.
- Data was seasonally adjusted using an X-12 ARIMA seasonal adjustment (generic phrasing appears in table).
- Yearly actual and projected data was interpolated.
- Historical data were interpolated, if possible.
- A 12-month moving average was used for certain series (noted for CHL).

### Miscellaneous methodological notes
- A commodity index was constructed for each country based on the country’s trade shares for each commodity (World Integrated Trade Solutions, 2011).
- Yearly/quarterly actual and projected data was interpolated for some series.
- Several country-specific series draw on National Authorities, Haver Analytics, Bloomberg LP, Datastream, International Financial Statistics, EMED, and Bolsa Electronica de Valores del Uruguay.

*Appendix II: Data Sources and Description — _wp12243*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12243.pdf_