## _wp04166

## Source details

**Canonical URL:** [_wp04166](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04166.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04166.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04166.pdf.json)

---

### I. Introduction — regime change and policy preview
- Peso floated in September 1999 following the Asian, Russian, and Brazilian crises.
- Formal “inflation targeting” announced in October 2000.
- Monetary stance increasingly transmitted through the central interest rate (repo rate) of the Banco de la República (BR), managed by the Board of Directors, within a framework of “lombard rates.”
- Trinity framework components consolidated: (i) a flexible exchange rate; (ii) inflation targeting; and (iii) a monetary policy rule.
- Annual inflation stabilized at around 7 percent, completing four consecutive years of single-digit inflation in Colombia.
- Historical CPI inflation averaged 22 percent a year over the previous three decades (Dornbusch and Fischer, 1991).
- Real exchange rate depreciated by about 15–20 percent in real effective terms over the period 1999–2002.
- Central bank provided “last-resort” money during the 1998 housing crisis; estimated quasi-fiscal resources needed to fund this financial crisis: 4–6 percent of GDP over the period 1998–2007.
- Constitutional mandate of the BR: hierarchical — pursue low and stable inflation while aligning with the government development plan (high growth and low unemployment). BR has instrumental independence and is committed to disinflation over the medium term on the basis of sustainable real growth.
- Paper’s policy recommendations preview:
  - Adopt “operational inflation target ranges” given single-digit inflation over four years.
  - Strengthen the scheme of foreign exchange “options” to confront international capital market turbulence.
- Impact assessment: reductions in the reference rates of the Central Bank of Colombia will only be significant if household expenses rise as mortgages are refinanced at lower interest rates.
- Structure of the paper:
  - Section II: macroeconomic assessment (inflation, unemployment, growth) and EFF program outcomes.
  - Section III: monetary and exchange rate issues.
  - Section IV: comparison of BR and the Federal Reserve System (Fed).
  - Section V: Taylor rules in an open-economy framework with preliminary estimates for Colombia.
  - Section VI: concluding remarks.

### II. Macroeconomic performance: inflation, growth, unemployment, and EFF outcomes
- Macroeconomic “suffering” index defined as (Inflation + Unemployment) – (Economic Growth).
- Historical averages (Banco de la República, Table 1):
  - 1967–1974: Inflation (CPI-Average) = 12.1; Unemployment (Main Cities) = 9.9; Growth (Real-GDP) = 6.2; Index = 15.8
  - 1975–1981: Inflation = 24.7; Unemployment = 9.5; Growth = 4.5; Index = 29.6
  - 1982–1989: Inflation = 22.6; Unemployment = 11.7; Growth = 3.5; Index = 30.9
  - 1990–1997: Inflation = 24.0; Unemployment = 10.1; Growth = 4.0; Index = 30.0
  - 1998–2002: Inflation = 10.6; Unemployment = 18.1; Growth = 0.5; Index = 28.3
- Key findings:
  - Average annual inflation hovered around 22–24 percent over 1975–97, falling to 10.6 percent during 1998–2002.
  - Disinflation accompanied by weakening domestic aggregate demand since 1997 and structural deterioration due to high indebtedness and internal conflict.
  - Unemployment averaged 18 percent in urban areas (15 percent nation-wide) in recent years.
  - Estimates of NAIRU for Colombia around 10 percent.
  - Sum of inflation and unemployment averaged 34 percent during 1975–97; declined to 29 percent in 1998–2002.
  - “Suffering index” has been around 28–31 percent over the last three decades, about twice the 1967–74 average.
- EFF program (1999–2002) targets and results (percent of GDP or annual percentage change):
  - Consolidated fiscal deficit (-):
    - 1999: Target = -6.0; Result = -5.5; Overperformance (+) = +0.5
    - 2000: Target = -3.6; Result = -3.4; Overperformance (+) = +0.2
    - 2001: Target = -2.9; Result = -3.2; Overperformance (+) = -0.3
    - 2002: Target = -2.6; Result = -4.0; Overperformance (+) = -1.4
  - External deficit (-):
    - 1999: Target = -1.3; Result = 0.6; Overperformance (+) = +0.7
    - 2000: Target = -2.4; Result = 0.5; Overperformance (+) = +1.9
    - 2001: Target = -1.8; Result = -1.9; Overperformance (+) = -0.1
    - 2002: Target = -3.2; Result = -1.7; Overperformance (+) = +1.5
  - Real growth (annual percentage change):
    - 1999: Target = -3.5; Result = -4.2; Overperformance (+) = -0.7
    - 2000: Target = 3.0; Result = 2.7; Overperformance (+) = -0.3
    - 2001: Target = 3.8; Result = 1.6; Overperformance (+) = -2.2
    - 2002: Target = 3.0; Result = 1.7; Overperformance (+) = -1.3
  - Inflation (annual percentage change):
    - 1999: Target = 15.0; Result = 9.2; Overperformance (+) = +5.8
    - 2000: Target = 10.0; Result = 8.8; Overperformance (+) = +1.2
    - 2001: Target = 8.0; Result = 7.7; Overperformance (+) = +0.3
    - 2002: Target = 6.0; Result = 7.0; Overperformance (+) = -1.0
- Interpretation:
  - Fiscal deficit slightly overperformed in 1999–2000; 2002 saw larger shortfall versus target.
  - Current account significantly overperformed in 1999–2000 due to weaker local demand and improved terms of trade (terms of trade increased by 20 percent over 1999–2000).
  - Real growth underperformed program targets across 1999–2002.
  - Cumulative disinflation almost 7 percentage points in two years (47 percent of the original 1999 target); 2002 missed headline inflation target due to weather-related crop-price increases while nonfood inflation closed at 5.3 percent.

- Policy recommendation on targets:
  - Move from point-targets to range-targets given outcome uncertainty at single-digit inflation and exposure to supply shocks.
  - BR adopted range targets of 5–6 percent for 2003 and 3.5–5.5 percent for 2004, with deviations of +/- 2 percent allowed under IMF agreement since 2001.
  - Caution against excessive disinflation as inflation converges to the BR’s long-term value of 3 percent per-annum.

### III. Monetary policy design, instruments, and foreign exchange options
- Four main changes after adopting inflation targeting (October 2000):
  - (a) Announcing multi-annual inflation targets to guide expected inflation and expected nominal depreciation of the peso versus the dollar.
  - (b) Global assessment of macroeconomic variables with focus on real and financial sector solvency; credit transmission assessed via the repo rate; liquidity access enhanced by linking “lombard rates” with the “penalty rate” for discount window access.
  - (c) Signaling via interest rates: BR concentrates on transmitting stance through the repo rate, guiding the interbank rate; uses ceiling and floor lombard rates to reduce discount-window stigma.
  - (d) Foreign Exchange Options: use of “put” options to increase NIR and “call” options to decrease NIR; volatility-control options used to smooth exchange-rate swings.

- Foreign exchange options activity and statistics (Table 3, amounts in US$ millions, NIR in US$ millions, percentages preserved):
  - I. “Put” Options to Buy NIR
    - Colombia (1999–2002)
      - Trigger rule: Spot < Spot(MA20Days)
      - Amount offered per auction: 30–200
      - Cumulative amount purchased: 1,400
      - NIR: 10,840
      - Amount purchased/NIR (percent): 11.3
      - NIR/amortization due (percent): 1.0
    - Mexico (1995–2001)
      - Trigger rule: Spot < Spot(MA20Days)
      - Amount offered per auction: 250
      - Cumulative amount purchased: 12,000
      - NIR: 34,000
      - Amount purchased/NIR (percent): 35.0
      - NIR/amortization due (percent): 1.2
  - II. “Call” Options to Sell NIR
    - Colombia (1999–2003)
      - Trigger rule: Spot > Spot(MA20Days)
      - Amount offered per auction: 200 (or up to 1 billion)
  - III. Options to Control Volatility
    - Colombia (1999–2002)
      - Trigger rule (percent): Spot deviates by more than 4 percent from Spot(MA20Days)
      - Amount offered per auction: 180
      - Cumulative amount purchased: 414
      - Exercised options/NIR (percent): 3.8
    - Mexico (1995–2001)
      - Trigger rule (percent): Spot 2 percent more depreciated than Spot t-1
      - Amount offered per auction: 200
      - Cumulative amount purchased: 1,950
      - Exercised options/NIR (percent): 5.7
- Operational details and historical notes:
  - Since late 1999 “put” options allowed BR to buy between US$30–200 million per month; cumulative exercised = US$1,400 million over 1999–2002 (11 percent of NIR at end-2002).
  - “Call” options announced in 2001 and first used in February 2003 when the exchange rate was depreciating nearly 30 percent annually; BR offered US$200 million in the first month out of a package of up to US$1 billion.
  - Volatility control mechanism set at +/- 4 percent of the 20-day moving average; triggered conditions in 2002 permitted the financial system to buy US$414 million from the BR.

### IV. Transmission mechanism of monetary policy — Colombian case and international comparison
- Necessary condition: BR repo rate Granger-causes market rates (Julio, 2001).
- Three elements to analyze impact of a 1 percentage point reduction in interest rate:
  - (i) Effect on credit stock of the financial system
    - Credit stock: nearly 26 percent of GDP by end-2002.
    - A 1 percent reduction in the average interest rate could represent about 0.27 percent of GDP.
    - Banco de la República estimate for 1996–99 showed an impact of about this size if mortgage rates are included.
  - (ii) Mortgage debt restructuring
    - In the 1990s mortgages linked to floating rates with premiums fluctuating around 8–10 percent.
    - Housing Law 546 of 1999 fixed ex-ante the real interest rate for the life of the mortgage, limiting pass-through from market rate reductions.
    - Pre-payments now allowed without financial penalties; creation of a securitization market and long-term public bonds (now extending to 7–10 years) should help recontracting.
  - (iii) Profitability of new investment projects
    - Difficult to quantify ex ante; credit mechanism crucial for marginal impact; limited fiscal room for increasing public investment due to structural budget limitations.
- U.S. comparison (2000–2002):
  - Fed reduced Federal Funds Rate on 12 occasions; market rates fell about 6 percent.
  - Estimated reduction in mortgage rates: about 2 percent in real terms.
    - Typical 30-year mortgage nominal rate: from 8.7 percent down to 6.5 percent.
    - CPI inflation in the United States averaged 3 percent per annum.
  - Consumption and growth effects:
    - Increase in consumption of about US$150–200 per month for the average household.
    - Aggregate effect: increase of about 0.7 percent of GDP in two years.
    - Estimated U.S. GDP growth: 1 percent in 2001 and 2.4 percent in 2002.
    - Fiscal impulse estimated in the range of 1–1.5 percent of GDP contributed to recovery.
  - Wealth effects: a 10 percent increase in housing prices over 1982–99 induced an increase in household consumption of about 0.6 percent of GDP (local studies).

- Lessons:
  - Reductions in market interest rates spur household consumption especially when mortgage credits are affected and refinancing is feasible.
  - In Colombia, fixed long-term real mortgage interest rates (Court-mandated) hamper transmission; promoting competition to encourage mortgage recontracting is crucial.
  - Without mortgage recontracting and related reforms, average households may not benefit from real interest rate reductions by the central bank.

### V. BR–Fed institutional and operational comparison; reaction functions and Taylor-rule evidence
- Institutional contrasts:
  - Objectives:
    - BR: Hierarchical — Inflation Control and Coordination with Government Macro Policies.
    - Fed: Dual — Inflation Control and Generation of Employment.
  - Board composition:
    - BR: Seven members (including the Minister of Finance).
    - Fed: Twelve for the FOMC (and seven for the “Discount Window”).
  - Strategy and instruments:
    - BR: Explicit “Inflation Targeting”; instruments include Reference rates, Lombard Rates, Repo/Reverse Repo, Options: “puts” and “calls”.
    - Fed: Implicit inflation targeting; instruments include Fed Funds Rate, Discount Window, banking reserves; discount-window reform for liquidity support at about 100–150 basis points above the FFR starting 2003.
  - Transparency and operational cadence differences noted (meeting frequency, minutes/transcriptions, staff roles).

- Interest-rate stance indicators:
  - Nominal interest rate relative to nominal GDP growth used as a proxy for stance: if nominal interest rate exceeds nominal GDP growth, stance inferred as tight.
  - Colombia: 1996 and 1998–99 tight; 2001–2002 monetary policy neutral by this measure.
  - Real repo-rate history:
    - Repo rate of BR reduced from 20 percent to 8 percent during 1995–98, increased to 12 percent in mid-1998, then relaxed; real repo-rate close to zero since late 2001.
    - U.S. Federal Funds Rate declined to 1.25 percent by end-2002; in real terms it turned negative since late 2001.

- Taylor-rule and reaction-function evidence (preserve formulas and estimation results):
  - Fed-related generalized Taylor rule formulation:
    - k ≡ r* - (gπ - 1) π*.
    - Estimations for 1987–96: k = 0.63; gπ = 1.78; gy = 0.82.
  - Table 5 rule variants (reported forms):
    - Basic Taylor rule: i = r* - 0.5 π* + 1.5 π + 0.5 y.
    - Generalized Taylor rule (estimation): i = 0.63 + 1.7 π + 0.8 y + ε.
    - Optimal Taylor rule (estimation): i = 2.21 + 2.8 π + 1.6 y + ε.
    - Optimal dynamic (estimation): i = 2.21 + 2.8 π + 1.8 y + ε.
    - Optimal lagged (estimation): i = 2.21 + 2.5 π + 1.6 y + ε.
  - Interpretation: “Observed” Fed reaction parameters implied “slow” actions; literature suggests caution against rapid policy swings that could increase volatility.

- Banco de la República open-economy reaction-function estimates (preserve equations and estimation outputs):
  - Suggested extension: include uncovered interest-rate parity condition instead of direct real exchange rate targeting.
  - Sample estimations (dependent variable: interbank interest rate):
    - Period 1989–2002 (Quarterly):
      - A. Contemporaneous values:
        - i = 23.9 – 0.04 (M - M*) + 0.05 (i* + e) + 1.67 y
        - Significance levels: (99%) (40%) (30%) (85%)
        - R2 = 0.72; Dw = 2.1; AR(1) = 0.84
      - B. Lagged and contemporaneous values:
        - i = 16.6 - Σ 0.03 (M - M*) + Σ 0.39 (i* + e) + Σ 3.69 y
        - Significance levels: (99%) (1%) (99%) (99%)
        - R2 = 0.77; Dw = 1.95; AR(1) = 0.45
    - Period 1998–2002 (Monthly):
      - C. Contemporaneous values:
        - i = 4.0 + 0.64 (π - π*) + 0.13 (i* + e) – 0.11 y
        - Significance levels: (20%) (78%) (66%) (74%)
        - R2 = 0.86; Dw = 2.29; AR(1) = 0.53; AR(2) = 0.41
      - D. Lagged and contemporaneous values:
        - i = 5.3 - Σ 0.69 (π - π*) + Σ 0.18 (i* + e) + Σ 0.09 y
        - Significance levels: (48%) (47%) (27%) (7%)
        - R2 = 0.84; Dw = 2.31; AR(1) = 0.46; AR(2) = 0.44
  - Notes on results:
    - Contemporaneous specifications explain about 72 percent of interbank rate variation with strong autoregressive persistence in inflation.
    - Including lags improved fit (case B), with the (uncovered) interest-rate gap of last three quarters exerting about 39 basis points effect on the interbank rate and a lagged product gap coefficient near 3.7.
    - For 1998–2002, inflation-gap effects exceeded monetary-gap effects (0.64 vs. 0.03) but were only marginally significant; overall evidence is mixed and sample lengths limit definitive inference.

### VI. Fiscal constraints, structural reforms, and policy recommendations
- Fiscal and structural findings:
  - Structural fiscal deficits running at 3 percent of GDP resulted in pressure to issue local public debt and pushed up real domestic interest rates.
  - Financial sector favored local treasury securities yielding 8 percent in real terms with no risk.
  - In 2002, financial credit grew in real terms for the first time in almost three years.
  - Fiscal agenda progress by late 2002 included: second generation of the pension reform; labor reform; state downsizing law; tax reform imposing universal coverage for the VAT. Fiscal responsibility law approved in 2003.
  - Public debt in relation to GDP is “now about 55 percent.”
- Policy recommendations:
  - Break down fiscal dependency to help credit flow to private investment; otherwise monetary policy effectiveness and growth will be hampered.
  - Strengthen mechanisms to consolidate inflation targeting, including maintaining flotation of the peso.
  - Continue efforts to lower public debt in relation to GDP.
  - Consider strengthening the scheme of foreign exchange options to better confront turbulence in international capital markets.
  - Recognize that reductions in reference rates of the Central Bank of Colombia will have a significant effect only if resources are freed for more spending, e.g., through refinancing of mortgages at lower rates.
  - Prefer introducing the (uncovered) interest-rate parity condition in reaction functions for emerging markets rather than direct real exchange rate targeting, because capital flows affect NIR and monetary aggregates.
- Research and operational recommendations:
  - It is premature to postulate a clear reaction function of the repo rate with respect to inflation and output gaps or the (uncovered) interest-rate-parity condition for 1998–2002; expect econometric estimates to improve as the historical experiment lengthens and as the (uncovered) interest-parity condition is incorporated.
  - Explore use of forecasting values to compute gaps, but note forecasting values have been inconsistent over short periods and thus have not been a practical direct input to the BR reaction function to date.
  - Distinguish inflation targeting modeled as an “instrument rule” versus a “targeting rule” (Svensson, 2002).

*Source: _wp04166 (IMF working paper excerpt).*

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

### _wp04166 - References

### I. INTRODUCTION
- Peso floated in September 1999 following the Asian, Russian, and Brazilian crises.
- Formal “inflation targeting” announced in October 2000.
- Monetary stance increasingly transmitted through the central interest rate (repo rate) of the Banco de la República (BR), managed by the Board of Directors, within a framework of “lombard rates.”
- Trinity framework components consolidated: (i) a flexible exchange rate; (ii) inflation targeting; and (iii) a monetary policy rule.
- Annual inflation stabilized at around 7 percent, completing four consecutive years of single-digit inflation in Colombia.
- Historical CPI inflation averaged 22 percent a year over the previous three decades (Dornbusch and Fischer, 1991).
- Real exchange rate depreciated by about 15–20 percent in real effective terms over the period 1999–2002.
- Central bank provided “last-resort” money during the 1998 housing crisis; estimated quasi-fiscal resources needed to fund this financial crisis: 4–6 percent of GDP over the period 1998–2007.
- Constitutional mandate of the BR: hierarchical — pursue low and stable inflation while aligning with the government development plan (high growth and low unemployment). BR has instrumental independence and is committed to disinflation over the medium term on the basis of sustainable real growth.
- Paper’s policy recommendations preview:
  - Adopt “operational inflation target ranges” given single-digit inflation over four years.
  - Strengthen the scheme of foreign exchange “options” to confront international capital market turbulence.
- Impact assessment: reductions in the reference rates of the Central Bank of Colombia will only be significant if household expenses rise as mortgages are refinanced at lower interest rates.
- Structure of the paper:
  - Section II: macroeconomic assessment (inflation, unemployment, growth) and EFF program outcomes.
  - Section III: monetary and exchange rate issues.
  - Section IV: comparison of BR and the Federal Reserve System (Fed).
  - Section V: Taylor rules in an open-economy framework with preliminary estimates for Colombia.
  - Section VI: concluding remarks.

### II. MACROECONOMIC PERFORMANCE OF COLOMBIAN ECONOMY
A. Inflation, Growth, and Unemployment
- Defined “macroeconomic suffering” index: (Inflation + Unemployment) – (Economic Growth) (variation of Okun’s immiserising index).
- Historical averages (Table 1, Source: Banco de la República):
  - 1967–1974: Inflation (CPI-Average) = 12.1; Unemployment (Main Cities) = 9.9; Growth (Real-GDP) = 6.2; Index = 15.8
  - 1975–1981: Inflation = 24.7; Unemployment = 9.5; Growth = 4.5; Index = 29.6
  - 1982–1989: Inflation = 22.6; Unemployment = 11.7; Growth = 3.5; Index = 30.9
  - 1990–1997: Inflation = 24.0; Unemployment = 10.1; Growth = 4.0; Index = 30.0
  - 1998–2002: Inflation = 10.6; Unemployment = 18.1; Growth = 0.5; Index = 28.3
- Key findings:
  - Average annual inflation hovered around 22–24 percent over 1975–97, falling to 10.6 percent during 1998–2002.
  - Disinflation accompanied by weakening domestic aggregate demand since 1997 and structural deterioration due to high indebtedness and internal conflict.
  - Unemployment averaged 18 percent in urban areas (15 percent nation-wide) in recent years.
  - Estimates of NAIRU for Colombia around 10 percent.
  - Sum of inflation and unemployment averaged 34 percent during 1975–97; declined to 29 percent in 1998–2002.
  - “Suffering index” has been around 28–31 percent over the last three decades, about twice the 1967–74 average.
- Policy implication:
  - Opportunistic disinflation presents an opportunity to reduce financial and wage indexation to support aggregate demand recovery with stable, low inflation.

B. The EFF Program with the IMF, 1999–2002
- Table 2 summarizes targets and results under the Extended Fund Facility (EFF) program for 1999–2002 (values in percent of GDP or annual percentage change as presented):
  - Consolidated fiscal deficit (-):
    - 1999: Target = -6.0; Result = -5.5; Overperformance (+) = +0.5
    - 2000: Target = -3.6; Result = -3.4; Overperformance (+) = +0.2
    - 2001: Target = -2.9; Result = -3.2; Overperformance (+) = -0.3
    - 2002: Target = -2.6; Result = -4.0; Overperformance (+) = -1.4
  - External deficit (-):
    - 1999: Target = -1.3; Result = 0.6; Overperformance (+) = +0.7
    - 2000: Target = -2.4; Result = 0.5; Overperformance (+) = +1.9
    - 2001: Target = -1.8; Result = -1.9; Overperformance (+) = -0.1
    - 2002: Target = -3.2; Result = -1.7; Overperformance (+) = +1.5
  - Real growth (annual percentage change):
    - 1999: Target = -3.5; Result = -4.2; Overperformance (+) = -0.7
    - 2000: Target = 3.0; Result = 2.7; Overperformance (+) = -0.3
    - 2001: Target = 3.8; Result = 1.6; Overperformance (+) = -2.2
    - 2002: Target = 3.0; Result = 1.7; Overperformance (+) = -1.3
  - Inflation (annual percentage change):
    - 1999: Target = 15.0; Result = 9.2; Overperformance (+) = +5.8
    - 2000: Target = 10.0; Result = 8.8; Overperformance (+) = +1.2
    - 2001: Target = 8.0; Result = 7.7; Overperformance (+) = +0.3
    - 2002: Target = 6.0; Result = 7.0; Overperformance (+) = -1.0
- Interpretation:
  - Consolidated fiscal deficit showed slight overperformance in 1999–2000 (0.2–0.5 percent of GDP).
  - Current account overperformance significant in 1999–2000 (0.7–1.9 percent of GDP) due to weaker local aggregate demand and improved terms of trade (terms of trade increased by 20 percent over 1999–2000).
  - Real growth underperformed program targets in 1999–2002 (e.g., shortfalls of 0.7 percent in 1999 and 0.3 percent in 2000).
  - Inflation overperformed in 1999 (by 5.8 percentage points) and 2000 (by 1.2 percentage points); cumulative disinflation almost 7 percentage points in two years (47 percent of the original 1999 target).
  - 2002 missed the 6 percent inflation target by 1 percentage point (7.0 percent result) mainly due to weather-related crop-price increases; nonfood inflation closed at 5.3 percent, below headline target.
- Policy recommendation on targets:
  - Move from point-targets to range-targets due to higher outcome uncertainty at single-digit inflation in presence of supply shocks.
  - BR adopted range targets of 5–6 percent for 2003 and 3.5–5.5 percent for 2004, with deviations of +/- 2 percent allowed under IMF agreement since 2001.
  - Caution against excessive disinflation as inflation converges to the BR’s long-term value of 3 percent per-annum.

### III. THE NEW MONETARY POLICY AND EXCHANGE RATE FLEXIBILITY
A. Elements of the New Monetary Policy
- Four main changes following formal adoption of inflation targeting (October 2000):
  (a) Announcing multi-annual inflation targets to guide expected inflation and expected nominal depreciation of the peso versus the dollar.
  (b) Global assessment of macroeconomic variables with special attention to real sector and financial sector solvency; credit transmission assessed via the repo rate; liquidity access enhanced by linking “lombard rates” with the “penalty rate” for discount window access.
  (c) Signaling via interest rates: since late 2000 BR concentrates on transmitting monetary stance through the repo rate, which guides the interbank rate. Repo rate and lombard structure preferred instrument for communicating policy stance.
    - Comparisons: Fed uses Federal Funds Rate (FFR) and discount window; Fed ceased announcing monetary aggregate forecasts starting June 2000 (Board of Governors (2000 p. 2) and Meyer (2001a p.7–9)).
    - Central Bank of Chile similarly signals via repo rate after adopting inflation targeting and floating the peso.
    - Monetary stance can be altered without moving repo rate by switching the central bank’s net asset position (e.g., repo vs reverse repo relevance at year-end vs January).
    - BR uses ceiling lombard rate (cost of accessing unlimited resources) and floor lombard rate (remuneration for excess deposits) as an alternative to discount window to reduce stigma.
    - Fed announced in October 2002 transformation of its discount window into a liquidity support mechanism allowing solvent institutions access at a modest penalty rate of about 100–150 basis points above the FFR, beginning in 2003.
    - BR created an Internal Operational Committee in early 2001 to handle operational issues, compensating movements in Treasury accounts.
  (d) Foreign Exchange Options:
    - “Put” options to increase net international reserves (NIR); “call” options to decrease NIR.
    - Since late 1999 “put” options allowed BR to buy between US$30–200 million per month (about 1 to 5 percent of the market) when spot was more appreciated than the 20-day moving average; cumulative exercised = US$1,400 million over 1999–2002 (11 percent of NIR at end-2002). This is lower than Mexico’s accumulated 35 percent of NIR since 1995 (Mexico suspended in August 2001).
    - “Call” options announced in 2001 and first used in February 2003 when the exchange rate was depreciating nearly 30 percent annually; BR offered US$200 million in the first month out of a package of up to US$1 billion, exercisable when spot was more depreciated than the 20-day limit-moving around the spot rate.
    - Volatility control mechanism: set at +/- 4 percent of the 20-day moving average of the spot rate. Triggered first in 2001 (conditions not met) and activated in 2002 when volatility surpassed 4 percent. In 2002, the triggering conditions permitted the financial system to buy US$414 million from the BR (in several sessions).

*Source: _wp04166 - References.*

### 3.8 percent of the NIR at end-2002.

### _wp04166 - 3.8 percent of the NIR at end-2002.

### Foreign Exchange Options (Table 3)
- I. “Put” Options to Buy Net International Reserves (NIR)
  - Colombia (1999–2002)
    - Trigger rule: Spot < Spot(MA20Days)
    - Amount offered per auction: 30–200
    - Cumulative amount purchased: 1,400
    - NIR: 10,840
    - Amount purchased/NIR (percent): 11.3
    - NIR/amortization due (percent): 1.0
  - Mexico (1995–2001)
    - Trigger rule: Spot < Spot(MA20Days)
    - Amount offered per auction: 250
    - Cumulative amount purchased: 12,000
    - NIR: 34,000
    - Amount purchased/NIR (percent): 35.0
    - NIR/amortization due (percent): 1.2

- II. “Call” Options to Sell NIR
  - Colombia (1999–2003)
    - Trigger rule: Spot > Spot(MA20Days)
    - Amount offered per auction: 200 (or up to 1 billion)
  - Mexico (1995–2001)
    - Trigger rule: --
    - Amount offered per auction: --

- III. Options to Control Volatility
  - Colombia (1999–2002)
    - Trigger rule (percent): Spot deviates by more than 4 percent from Spot(MA20Days)
    - Amount offered per auction: 180
    - Cumulative amount purchased: 414
    - Exercised options/NIR (percent): 3.8
  - Mexico (1995–2001)
    - Trigger rule (percent): Spot 2 percent more depreciated than Spot t-1
    - Amount offered per auction: 200
    - Cumulative amount purchased: 1,950
    - Exercised options/NIR (percent): 5.7

- Source attribution within table: Author.

### Fiscal Dependency and Structural Reforms (Colombia)
- Findings
  - Structural fiscal deficits running at 3 percent of GDP resulted in high pressure to issue local public debt and driven up real domestic interest rates.
  - Financial sector increased investments mainly in local treasury securities, yielding 8 percent in real terms with no risk.
  - In 2002, financial credit grew in real terms for the first time in almost three years.
- Policy actions and progress
  - Since the IMF program launch in 1999, fiscal agenda identified; late 2002 progress included:
    - Second generation of the pension reform
    - Labor reform
    - State downsizing law
    - Tax reform imposing universal coverage for the VAT
  - Fiscal responsibility law approved in 2003.
- Recommendation
  - Break down fiscal dependency to help credit flow to private investment; otherwise monetary policy effectiveness and growth will be hampered.

### Central Bank Signaling: From “Quantities” to “Prices”
- Historical operational shift
  - Early 1990s BR followed Bundesbank-like model: monetary aggregates as intermediate targets (monetary nominalism).
  - 1995–99: BR moved toward interest rates as main instrument; exchange rate system shifted from “crawling peg” to “crawling band” and finally to flotation of the peso in September 1999.
  - Board of the BR moved toward inflation targeting by late 2000; viewed as a breakthrough after floating the peso.
- Proposed long-run framework to consolidate:
  - (i) a flexible exchange rate system;
  - (ii) an inflation target scheme;
  - (iii) a monetary policy rule based on signaling the stance through the repo interest rate of the BR.

### Transmission Mechanism of Monetary Policy — Case of Colombia
- Necessary condition established: BR repo rate Granger-causes market rates (Julio, 2001).
- To gauge aggregate demand effects, need net position of households (net debtors vs net savers).
- Three elements to analyze impact of a 1 percentage point reduction in interest rate:
  - (i) Effect on credit stock of the financial system
    - Credit stock: nearly 26 percent of GDP by end-2002.
    - A 1 percent reduction in the average interest rate could represent about 0.27 percent of GDP.
    - Banco de la República estimate showed an impact of about this size if mortgage rates are included (period 1996–99).
  - (ii) Mortgage debt restructuring
    - In the 1990s mortgages linked to floating rates with premiums fluctuating around 8–10 percent.
    - Housing Law 546 of 1999 fixed ex-ante the real interest rate for the life of the mortgage, limiting pass-through from market rate reductions.
    - Pre-payments now allowed without financial penalties; creation of a securitization market and long-term public bonds (now extending to 7–10 years) should help recontracting.
  - (iii) Profitability of new investment projects
    - Difficult to quantify ex ante; credit mechanism crucial for marginal impact; limited fiscal room for increasing public investment due to structural budget limitations.

### Transmission Mechanism — Case of the United States
- 2000–2002: Fed reduced Federal Funds Rate on 12 occasions; market rates fell about 6 percent.
- Estimated reduction in mortgage rates: about 2 percent in real terms (Bloomberg, 2001).
  - Typical 30-year mortgage nominal rate: from 8.7 percent down to 6.5 percent.
  - CPI inflation in the United States averaged 3 percent per annum.
- Consumption and growth effects
  - Increase in consumption of about US$150–200 per month for the average household.
  - Aggregate effect: increase of about 0.7 percent of GDP in two years.
  - Estimated U.S. GDP growth: 1 percent in 2001 and 2.4 percent in 2002.
  - Fiscal impulse estimated in the range of 1–1.5 percent of GDP contributed to recovery.
- Wealth effects
  - Local studies: a 10 percent increase in housing prices over 1982–99 induced an increase in household consumption of about 0.6 percent of GDP.

### Lessons on Interest Rate Reductions and Household Consumption
- Reductions in market interest rates spur household consumption especially when mortgage credits are affected.
- In the United States the refinancing mechanism is dynamic, enabling households to benefit.
- In Colombia, fixed long-term real mortgage interest rates (Court-mandated) hamper transmission; promoting competition in the financial system to encourage mortgage recontracting is crucial.
- Without such reforms, average households may not benefit from real interest rate reductions by the central bank.

### BR–Fed Comparison: Institutional and Operational Features
- Objectives
  - BR: Hierarchical — Inflation Control and Coordination with Government Macro Policies.
  - Fed: Dual — Inflation Control and Generation of Employment.
- Board composition
  - BR: Seven members (including the Minister of Finance).
  - Fed: Twelve for the FOMC (and seven for the “Discount Window”).
- Strategy and instruments
  - BR: “Inflation Targeting” (Explicit); central instruments: Reference rates, Limiting rates; secondary: Aggregates; support: Treasury; use of Repo and Reverse Repo; Lombard Rates-Discount Window; Monetary Corridors / Reference Lines (announced); Semi-Automatic compensation with Treasury accounts; Options: “puts” and “calls”.
  - Fed: “Inflation Targeting” (Implicit); Repo (Fed Funds Rate), Discount Window; Banking Reserves (un-announced); Automatic compensation; Flotation with intervention thru Treasury.
- Operational procedures and transparency
  - BR: Weekly Board meetings (various subjects), Monthly (Inflation); Board announcements immediate; minutes provided under certain conditions; transcriptions provided after three years in relevant areas; reports to public and Congress; staff active with recommendations; simple model of transmission mechanism.
  - Fed: Every month and a half (Eight Times a year) Board meetings; announcements immediate; minutes available two months later (excluding sensitive topics); transcriptions confidential for five years; reports to Congress twice a year; staff passive (avoiding recommendations); integrated national/international macro-models.
- Anti-inflationary policy
  - BR formally announced inflation targeting in October 2000.
  - Fed operates with a “disguised inflation target” around 3 percent per year, focusing on core inflation.

### Board Structure and Decision-Making (BR)
- Board composition details
  - The Board includes:
    - The Minister of Finance (the chairman)
    - Five full-time Directors chosen by the President (two may be replaced every four years; maximum tenure 12 years)
    - The General Manager, elected by the Directors for periods of four years, up to a maximum of 12 years
  - Each member has one vote; no veto power by the Chairman or tie-breaking vote by the General Manager.
  - The Head of the Office of Bank Supervision has voice but no vote.
- Practical dynamics
  - Board cannot be convened without Minister of Finance or Deputy; provisions prevent delays over two weeks; any member can call extraordinary meetings.
  - Joint technical work by Central Bank and Ministry staffs.

### Reaction Functions and Taylor Rules
- Reading the stance of monetary policy
  - Nominal interest rate relative to nominal GDP growth used as a proxy: if nominal interest rate exceeds nominal GDP growth, stance inferred as tight.
  - Colombia: 1996 and 1998–99 showed tight conditions; 2001–2002 monetary policy was neutral by this measure.
- Real interest rate evolution (Figure 1, 1995–2002)
  - Repo rate of BR reduced from 20 percent to 8 percent during 1995–98, then increased to 12 percent in mid-1998, then relaxed; real repo-rate close to zero since late 2001.
  - U.S. Federal Funds Rate (FFR) nominally fluctuated between 4–6 percent and declined to 1.25 percent by end-2002; in real terms hovered around 2–4 percent during 1998–2000 and fell to negative figures since late 2001.
- Taylor rules overview
  - Basic Taylor rule characterized by:
    - (1) presenting a premium with respect to the long-term real interest rate;
    - (2) increasing whenever inflation surpasses the long-term target set up by the FOMC;
    - (3) increasing whenever real output surpasses potential output.
  - The basic Taylor rule postulated long-term real interest rate and long-term inflation rate converged to 2 percent (r* = π* = 2 percent), with short-term deviations calling for responses (e.g., increases of the FFR of about 1.5 percent under the classical formulation).

*Source: _wp04166 - 3.8 percent of the NIR at end-2002.*

### 0.5 percent, respectively.

### _wp04166 - 0.5 percent, respectively.

### Fed reaction functions and Taylor-rule variants
- Generalized Taylor rule formulation:
  - k ≡ r* - (gπ - 1) π*.
  - Estimations for 1987–96: k = 0.63; gπ = 1.78; gy = 0.82.
- Table 5 reported rule variants and estimations:
  - Basic Taylor rule: i = r* - 0.5 π* + 1.5 π + 0.5 y.
  - Generalized Taylor rule (estimation): i = 0.63 + 1.7 π + 0.8 y + ε.
  - Optimal Taylor rule (estimation): i = 2.21 + 2.8 π + 1.6 y + ε.
  - Optimal dynamic (estimation): i = 2.21 + 2.8 π + 1.8 y + ε.
  - Optimal lagged (estimation): i = 2.21 + 2.5 π + 1.6 y + ε.
- Interpretation and literature:
  - “Observed” Fed reaction parameters implied “slow” actions; Fed appeared timid (Rudebusch, 2001).
  - One explanation: concerns about excess volatility from rapid Fed actions; supports acting firmly but in a slow fashion because changes of direction can harm economic stability.

### Case of Banco de la República: modeling open-economy reaction functions
- Extensions to Taylor rule for emerging markets:
  - Inclusion of real exchange rate qt and lagged qt-1 considered, but argued to be of scant practical use when qt changes reflect productivity or are temporary.
  - Recommendation: introduce the (uncovered) interest rate parity condition instead of real exchange rate targeting, since capital flows affect net international reserves (NIR) and monetary aggregates.
- Two sample periods analyzed:
  - 1989–2002 (quarterly): targeting of monetary aggregates and a fixed exchange rate system (crawling peg and crawling bands).
  - 1998–2002 (monthly): flotation of the peso, inflation targeting, and signaling through the repo rate.
- Equation (2) specification for 1989–2002: i = r* + gm (M - M*) + gi (i* + e) + gy y.
- Empirical estimations (Dependent Variable: Interbank Interest Rate):
  - Period 1989–2002 (Quarterly):
    - A. Contemporaneous values:
      - i = 23.9 – 0.04 (M - M*) + 0.05 (i* + e) + 1.67 y
      - Significance levels in parentheses: (99%) (40%) (30%) (85%)
      - R2 = 0.72; Dw = 2.1; AR(1) = 0.84
    - B. Lagged and contemporaneous values:
      - i = 16.6 - Σ 0.03 (M - M*) + Σ 0.39 (i* + e) + Σ 3.69 y  (sums over j=0..2 or t-j as specified)
      - Significance levels in parentheses: (99%) (1%) (99%) (99%)
      - R2 = 0.77; Dw = 1.95; AR(1) = 0.45
  - Period 1998–2002 (Monthly):
    - C. Contemporaneous values:
      - i = 4.0 + 0.64 (π - π*) + 0.13 (i* + e) – 0.11 y
      - Significance levels in parentheses: (20%) (78%) (66%) (74%)
      - R2 = 0.86; Dw = 2.29; AR(1) = 0.53; AR(2) = 0.41
    - D. Lagged and contemporaneous values:
      - i = 5.3 - Σ 0.69 (π - π*) + Σ 0.18 (i* + e) + Σ 0.09 y
      - Significance levels in parentheses: (48%) (47%) (27%) (7%)
      - R2 = 0.84; Dw = 2.31; AR(1) = 0.46; AR(2) = 0.44
- Notes on estimation quality:
  - Contemporaneous-only specifications explained about 72 percent of interbank rate variation, with strong autoregressive persistence in inflation.
  - In case A, only the output gap played a marginal role; monetary and interest rate gaps were not significant.
  - Measurement challenges: frequent changes in monetary targets/definitions; sensitivity of output gap to decomposition method; gaps I(0) versus interest rates I(1).
  - Including lagged values improved estimates: in case B the (uncovered) interest rate gap of the last three quarters had a significant effect of about 39 basic points on the interbank rate; lagged product gap coefficient ~ 3.7 (log difference between observed and potential real GDP); monetary aggregate gap remained insignificant.
  - For 1998–2002, results were mixed: inflation gap effect larger than monetary gap (0.64 vs. 0.03) but inflation gap significant only at the margin (78 percent). Introducing lags did not materially improve fit.

### Policy implications, constraints, and recommended strategies for Colombia
- Inflation and policy framework outcomes:
  - The trinity framework advocated: (i) a flexible exchange rate; (ii) inflation targeting; and (iii) a monetary policy rule.
  - Under this framework inflation fell to 6 percent by 2003, completing five consecutive years of single-digit inflation in Colombia.
- Fiscal constraints and risks:
  - Fiscal dependency threatens inflation-targeting strategy.
  - High fiscal deficits “hovering around 3 percent of GDP” push real interest rates up and may compromise long-term stability of the central bank reaction function.
  - Fiscal pressure extends beyond seigniorage when capital markets depend critically on public bond prices; central bank may face pressure to sustain bond values.
  - Public debt in relation to GDP is “now about 55 percent,” creating an imperative to further coordinate monetary and fiscal policies.
- Recommended policy measures:
  - Strengthen mechanisms to consolidate inflation targeting, including maintaining flotation of the peso.
  - Continue efforts to lower public debt in relation to GDP.
  - Consider strengthening the scheme of foreign exchange options to better confront turbulence in international capital markets.
  - Recognize that reductions in reference rates of the Central Bank of Colombia will have a significant effect only if resources are freed for more spending, e.g., through refinancing of mortgages at lower rates.
  - Prefer introducing the (uncovered) interest rate parity condition in reaction functions for emerging markets rather than direct real exchange rate targeting, because capital flows affect NIR and monetary aggregates.
- Research and operational recommendations:
  - It is premature to postulate a clear reaction function of the repo rate with respect to inflation and output gaps or the (uncovered) interest-rate-parity condition for 1998–2002; expect econometric estimates to improve as the historical experiment lengthens and as the (uncovered) interest-parity condition is incorporated.
  - Explore use of forecasting values to compute gaps, but note forecasting values have been inconsistent over short periods and thus have not been a practical direct input to the BR reaction function to date.
  - Distinguish inflation targeting modeled as an “instrument rule” versus a “targeting rule” (Svensson, 2002).

*Source: Excerpt from the IMF working paper content unit _wp04166 - 0.5 percent, respectively.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04166.pdf_