## _wp11150

## Source details

**Canonical URL:** [_wp11150](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11150.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11150.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11150.pdf.json)

---

### Executive Summary: main conclusion
- The monetary policy implemented by the Central Bank of the Republic of Turkey (CBRT) helped soften the impact of the global financial crisis.
- Without key reforms—specifically, the adoption of an inflation targeting framework underpinned by a flexible exchange rate regime—the global financial crisis would have been associated with a much deeper economic contraction.
- Results are derived from counterfactual simulations of an estimated dynamic stochastic general equilibrium (DSGE) model that includes standard nominal and real rigidities and a financial accelerator mechanism in an open-economy framework.

### Key quantitative counterfactual findings
- Actual annual growth in 2009: –4.8 percent.
- Counterfactual without the countercyclical and discretionary interest rate cuts implemented by the CBRT:
  - Growth in 2009 would have decreased from –4.8 percent to –6.2 percent (simulation taking “no countercyclical responses and no discretionary policy loosening”).
- Counterfactual without discretionary interest rate cuts (only):
  - Growth in 2009 would have decreased from –4.8 percent to –5.9 percent, a difference of 1.1 percentage point.
- Counterfactual under a fixed exchange rate regime instead of the inflation targeting regime with a flexible exchange rate:
  - Growth in 2009 would have been –8.0 percent, a difference from the actual outcome of 3.2 percentage points.
- Policy action by the CBRT during the crisis:
  - The CBRT decreased policy rates by 1025 basis points over the November 2008 to November 2009 period.

### Context and crisis shocks
- Turkey experienced a year-over-year contraction of 14.7 percent during the first quarter of 2009; annual growth in 2009 was –4.8 percent.
- Pre-crisis performance: over the 26 quarters from Q1 2002 to mid-2008, Turkish year-over-year growth exceeded five percent and inflation declined markedly.
- Three crisis disturbances modeled:
  - export demand shock (collapse in foreign demand),
  - sudden stop shock (UIP/risk premium, reversal of capital inflows, depreciation),
  - (financial) uncertainty shock (net worth destruction, higher external finance premium).
- Financial uncertainty shock estimated persistence = 0.72.

### Transmission mechanisms emphasized
- Interest-rate channel:
  - Higher real rates increase saving, raise opportunity cost of investment, reduce capital accumulation and domestic demand, lowering inflation.
- Exchange-rate channel (open economy):
  - Higher nominal rates appreciate the real exchange rate, suppress net exports and aggregate demand.
- Financial accelerator channel:
  - Higher rates depress asset prices, deteriorate net worth, raise external finance premium, amplify investment and output contractions.
- A one standard deviation contractionary monetary policy shock:
  - Corresponds to a 70 basis point (quarterly) increase in the nominal interest rate (annual increase in the policy rate of about three percent).
  - Output gap dips below steady state by 70 basis points.
  - Year-over-year inflation rate troughs about 140 basis points below steady state after four periods.

### Model, estimation, and data
- Model: small open-economy New Keynesian DSGE with sticky prices, sticky wages, variable capital utilization, investment adjustment costs, habit persistence, time-varying inflation target, and a Bernanke-style financial accelerator.
- Agents: households, entrepreneurs, capital producers, retailers, government; rest of world interactions included.
- Estimation approach: Bayesian methods (Schorfheide-style, Smets and Wouters).
- Sample: quarterly data from Q1 2002 to Q2 2010.
- Observables (12): domestic policy and foreign interest rate levels; inflation rates of domestic GDP deflator, core consumer price, and foreign consumer price indices; growth rates of GDP, consumption, investment, exports, imports, foreign GDP; real exchange rate growth.
- Parameters: 43 parameters estimated; some calibrated as strict priors.
- Posterior sampling: Metropolis-Hastings with total draws = 500,000 and chains = two; Brooks and Gelman (1998) convergence criteria achieved.
- Model fit (one-sided Kalman filter): correlations — inflation > 0.95; interest rate > 0.99; consumption > 0.97; investment > 0.99. Exports show notable misfit.

### Calibration and selected steady-state targets
- Consumption-to-GDP ratio: 70 percent.
- Investment-to-GDP ratio: 20 percent.
- Government expenditures-to-GDP ratio: 10 percent.
- Exports-to-GDP ratio: 24 percent.
- Discount factor β = 0.9928 (implying an annual riskless real interest rate of approximately three percent).
- Entrepreneurial survival rate: 0.9728.
- Monitoring cost fraction: 0.15.
- Variance of idiosyncratic shock to entrepreneur production: 0.4.
- Target leverage ratio for 2007: two.
- Average EMBI spread targeted: 300 basis points.
- Steady-state markups (prices and wages): 1.15.

### Posterior estimates (selected)
- Calvo parameter (Domestic prices): Mean 0.306; Confidence interval 0.177 0.435.
- Calvo parameter (Import prices): Mean 0.554; Confidence interval 0.434 0.679.
- Calvo parameter (Wages): Mean 0.558; Confidence interval 0.387 0.739.
- Indexation (Domestic prices): Mean 0.460; Confidence interval 0.296 0.623.
- Indexation (Wages): Mean 0.562; Confidence interval 0.400 0.731.
- Habit formation: Mean 0.904; Confidence interval 0.850 0.958.
- Investment adjustment cost: Mean 3.625; Confidence interval 2.769 4.511.
- Interest rate smoothing: Mean 0.724; Confidence interval 0.639 0.808.
- Inflation response (monetary policy): Mean 1.537; Confidence interval 1.383 1.685.
- Output gap response (monetary policy): Mean 0.021; Confidence interval -0.022 0.064.
- Nominal exchange rate response (monetary policy): Mean 0.173; Confidence interval 0.101 0.242.
- Shock persistence examples:
  - Stationary technology: Mean 0.767; Confidence interval 0.593 0.939.
  - Unit root technology: Mean 0.497; Confidence interval 0.386 0.604.
  - Investment specific technology: Mean 0.916; Confidence interval 0.893 0.940.
- Investment specific technology volatility (Inverse gamma mean): 0.421; Confidence interval 0.334 0.504.

### Sensitivity analysis and model selection
- Baseline log data density: 802.482.
- Alternatives (18 specifications) universally dominated by baseline according to posterior odds ratios reported (examples: Financial accelerator removed log data density 779.963; Fixed exchange rate regime log data density 752.835).
- Excluding the financial accelerator is decisively rejected in favor of the baseline, underscoring the role of financial frictions.
- Fixed exchange rate regime is decisively rejected relative to the baseline monetary policy rule.

### Historical decompositions: contributions to output growth
- Shock grouping:
  - Group 1: monetary policy shocks.
  - Group 2: crisis shocks (foreign demand, financial uncertainty, UIP/sudden stop).
  - Group 3: remaining supply and demand shocks.
- Monetary policy shocks positively contributed to growth during the crisis.
  - Average growth contribution of monetary policy shocks during the crisis episode is about 1.1 percent.
  - Counterfactual without discretionary monetary policy: 2009 growth would have been –5.9 percent instead of –4.8 percent (monetary policy contribution ≈ 1.1 percent).
- Crisis shocks:
  - Financial uncertainty shock explains a large fraction of the downturn among the three crisis shocks; foreign demand shock has lingering negative effects.
  - Sudden stop (UIP/risk premium) shock had limited effect on growth in this specification (entrepreneurs borrow in domestic currency).

### Counterfactual experiments: setup
- Baseline: monetary policy under flexible exchange rate, estimated empirical interest rate rule with discretionary deviations.
- Four counterfactuals:
  1. No monetary policy shocks: all monetary policy shocks set to zero.
  2. No response to the output gap: output gap coefficient set to zero (estimated coefficient noted as 0.02); monetary policy shocks set to zero.
  3. Peg (fixed exchange rate): policy focused on stabilizing nominal exchange rate; no response to output gap; no discretionary deviations.
  4. Peg with heightened financial vulnerability: peg with leverage ratio calibrated to 3 (contrast to 2 in 2007).

### Counterfactual results: selected quantitative outcomes
- Turkey context:
  - Year-over-year contraction in 2009 Q1: 14.7 percent.
  - CBRT decreased policy rates by 1025 basis points over November 2008–November 2009.
- 2009 simulated outcomes (annual real GDP growth, in percent):
  - Baseline (actual): –4.8.
  - No monetary policy shocks: –5.9; Difference –1.1.
  - No response to output gap: –6.2; Difference –0.3 (additional) and Difference –1.4 (from baseline).
  - Fixed exchange rate regime (peg): –8.0; Difference from actual –1.8; Difference –3.2 (stated elsewhere as the impact of regime absence).
  - Peg with heightened financial vulnerability: –9.3; Difference –1.3; Difference –4.5.
- Peak-to-trough cumulative contractions (levels of GDP):
  - Baseline (actual): –14.4.
  - No monetary policy shocks: –15.4; Difference –1.0.
  - No response to output gap: –15.8; Difference –0.4.
  - Fixed exchange rate regime (peg): –19.6; Difference –3.8.
  - Peg with heightened financial vulnerability: –22.1; Difference –2.4.
- Table 5 excerpts (In percent) — selected rows:
  - Average (2008Q4—2009Q4): 51.98 1.18 0.32 2.22 1.69 5.42.
  - Average (2009Q1—2009Q4): 2.40 1.14 0.35 1.86 1.45 4.81.
  - Cumulative (2008Q4—2009Q4): 59.90 5.92 1.60 11.11 8.44 27.08.
  - Cumulative (2009Q1—2009Q4): 9.59 4.56 1.42 7.45 5.81 19.23.
- Counterfactual narrative:
  - The inflation targeting framework underpinned by a flexible exchange rate regime softened the crisis impact; absence implied substantially larger contractions (up to –8.0 percent under peg, –9.3 percent with peg and higher leverage).

### Policy implications and interpretation
- The combination of an inflation targeting framework and exchange rate flexibility increased resilience by:
  - Enabling discretionary and countercyclical monetary easing via short-term interest rate cuts.
  - Allowing exchange rate movements to absorb part of the external shock.
- Discretionary departures from the empirical interest rate rule played an important role; average contribution of monetary shocks to output growth ≈ 1.14–1.18 percent in crisis episodes.
- Low estimated output gap coefficient (0.02) implies a small systematic response of the policy rate to the output gap and a larger role for discretionary policy.
- Financial sector reforms lowering leverage (from 3.0 in 2000 to 2.0 in 2007) reduced vulnerability; higher leverage exacerbates contractions under a peg.
- Other policies (CBRT liquidity measures starting in Q4 2008, fiscal measures) also likely contributed to stabilization and may be partially captured by monetary policy shocks in the model.

*Source: Executive Summary and selected sections of _wp11150 (PDF chapter/section).*

### Executive Summary ......................................................................................................

### _wp11150 - Executive Summary

### Executive summary: main conclusion
- The monetary policy implemented by the Central Bank of the Republic of Turkey (CBRT) helped soften the impact of the global financial crisis.
- Without key reforms—specifically, the adoption of an inflation targeting framework underpinned by a flexible exchange rate regime—the global financial crisis would have been associated with a much deeper economic contraction.
- Results are derived from counterfactual simulations of an estimated dynamic stochastic general equilibrium (DSGE) model that includes standard nominal and real rigidities and a financial accelerator mechanism in an open-economy framework.

### Key quantitative counterfactual findings
- Actual annual growth in 2009: –4.8 percent.
- Counterfactual without the countercyclical and discretionary interest rate cuts implemented by the CBRT:
  - Growth in 2009 would have decreased from –4.8 percent to –6.2 percent (simulation taking “no countercyclical responses and no discretionary policy loosening”).
- Counterfactual without discretionary interest rate cuts (only):
  - Growth in 2009 would have decreased from –4.8 percent to –5.9 percent, a difference of 1.1 percentage point.
- Counterfactual under a fixed exchange rate regime instead of the inflation targeting regime with a flexible exchange rate:
  - Growth in 2009 would have been –8.0 percent, a difference from the actual outcome of 3.2 percentage points.
- Policy action by the CBRT during the crisis:
  - The CBRT decreased policy rates by 1025 basis points over the November 2008 to November 2009 period.

### Rationale and mechanisms highlighted
- Two post-2001 macroeconomic reforms are emphasized as key to resilience:
  - Abandoning heavily managed and fixed exchange rate regimes in favor of floating exchange rates.
  - Transitioning to, and in 2006 officially implementing, a full-fledged inflation targeting regime as the nominal anchor.
- The inflation targeting framework underpinned by exchange rate flexibility:
  - Allowed the CBRT to implement countercyclical and discretionary interest rate cuts.
  - Exchange rate flexibility acted as a shock absorber, improving output stabilization.
- The DSGE model used:
  - Small open-economy New Keynesian DSGE augmented with real and nominal frictions (sticky prices, sticky wages, variable capital utilization, investment adjustment costs, habit persistence).
  - Incorporates a financial accelerator mechanism à la Bernanke and others (1999) to capture financial market imperfections in an open-economy setup.

### Context: Turkish experience during the crisis
- Turkey was one of the hardest hit countries during the global financial crisis:
  - Real GDP contracted sharply for four quarters, reaching a year-over-year contraction of 14.7 percent during the first quarter of 2009.
  - This resulted in an annual growth rate of –4.8 percent for 2009.
- Pre-crisis performance:
  - Over the 26 quarters from Q1 2002 to mid-2008, the Turkish economy grew by over five percent (year-over-year), and inflation declined markedly.
- Shocks hitting Turkey during the crisis:
  - Collapse in foreign demand (severe hit to exports).
  - Distress across international capital markets (asset price collapse including the currency, increased spreads, sizeable capital outflows).
  - Heightened uncertainty suppressing consumption, investment, and credit extension.

### Comparative perspective and related literature
- The paper situates itself within the literature on small open-economy DSGE models augmented with nominal rigidities and financial accelerators (references to Mendoza (1991), Gali and Monacelli (2002), Bernanke and others (1999), Cespedes and others (2004), Gertler and others (2007), Elekdag and Tchakarov (2007), Smets and Wouters (2003, 2007), Lubik and Schorfheide (2007), Teo (2006), Christensen and Dib (2006), Elekdag, Justiniano, and Tchakarov (2006), Garcia-Cicco (2010)).
- The magnitude of the estimated growth contribution from monetary policy (1.1 percentage point in one specification) lies within the range reported by Christiano and others (2008) for advanced economies (0.75 percent for the United States and 1.27 percent for the Euro area).

### Policy implications (summary)
- The combination of an inflation targeting framework and exchange rate flexibility increased the resilience of the Turkish economy to the global financial shock by:
  - Enabling discretionary and countercyclical monetary easing via short-term interest rate cuts.
  - Allowing exchange rate movements to absorb part of the external shock.
- Absent these features, the model-based counterfactuals indicate substantially larger contractions in output during 2009.

*Source: Executive Summary of _wp11150*

### 2. However, rather

### _wp11150 - 2. However, rather

### The transmission of shocks
- During the global financial crisis, the Turkish economy was affected by three crisis disturbances modeled as:
  - an export demand shock,
  - a sudden stop shock, and
  - a (financial) uncertainty shock.
- The paper provides quantitative evidence appraising the relative growth contribution of these three shocks (as well as other structural shocks) during the recession which intensified in the first quarter of 2009.

### The export demand shock
- Representation:
  - A collapse in export (foreign) demand is represented by a decline in the export demand term.
- Transmission:
  - Domestic output equals the sum of consumption of domestically produced goods (household and entrepreneurial consumption), domestic investment goods, government expenditures, and exports.
  - A decline in foreign demand reduces exports and thus domestic output.

### The sudden stop shock
- Representation:
  - The UIP condition is augmented with an exogenous sudden stop shock (also referred to as an exchange rate shock or UIP shock).
  - Variables in UIP: domestic and international (gross) interest rates, the nominal exchange rate (Turkish lira per U.S. dollar—an increase represents a depreciation), and expectations operator.
- Economic role:
  - Captures reversal of capital inflows, large capital outflows, and sharp depreciation episodes that were observed during the crisis.

### The (financial) uncertainty shock
- Financial accelerator and external finance premium:
  - The real cost of capital equals the real interest rate augmented by the external finance premium.
  - The external finance premium depends on the leverage ratio (assets scaled by net worth).
  - Total assets depend on the price of equity, which is not sticky, making leverage sensitive to asset price fluctuations.
- Net worth shock:
  - The (financial) uncertainty shock is an exogenous process that directly impacts aggregate net worth and therefore the external finance premium.
  - Interpreted as a shock to the rate of destruction of entrepreneurial financial wealth or as capturing counterparty/Knightian uncertainty that impairs assets and disrupts the financial system.

### What role for monetary policy?
- Monetary policy rule (log-linear simplified form):
  - The central bank sets the nominal interest rate taking into account: inflation deviation from a time-varying inflation target, the output gap, the rate of exchange rate depreciation, and the previous period‘s interest rate (policy smoothing).
  - The monetary policy shock is the error term capturing discretionary deviations from the empirical interest rate rule.
- Time-varying inflation target:
  - The time-varying inflation target, π*_t, evolves according to a stochastic process and captures changes in the inflation target over time and structural changes in policy conduct.
- Discretionary easing during the crisis:
  - When the output gap is negative, strict adherence to the rule implies interest rate decreases proportional to a coefficient on the output gap.
  - Monetary authority may decrease interest rates by more than the rule implies; such deviations are captured by the monetary policy shock.
  - During the most intense episode of the global financial crisis, interest rates decreased by more than the empirical rule implied, helping soften the crisis impact.

### Estimation: data and parameters
- Estimation approach:
  - The log-linearized model is estimated using Bayesian methods (Schorfheide-style, Smets and Wouters).
- Sample and variables:
  - Quarterly data from the first quarter of 2002 to the second quarter of 2010.
  - Twelve matched variables: levels of the domestic policy and foreign interest rates; inflation rates of domestic GDP deflator, core consumer price, and foreign consumer price indices; growth rates of GDP, consumption, investment, exports, imports, foreign GDP, and the real exchange rate.
- Parameters:
  - Some parameters calibrated as strict priors; remaining parameters taken from literature summarized in Table 2.
  - 43 parameters estimated (degree of real and nominal rigidities, monetary policy stance, persistence and volatility of exogenous shocks).
  - Posterior means and 5th and 95th percentiles obtained via Metropolis-Hastings sampling.
- Model fit and robustness:
  - One-sided Kalman filter comparisons indicate generally satisfactory sample fit (Figure 3).
  - Robustness assessed via 18 alternative specifications; log data density and posterior odds strongly favor the baseline (Table 4).

### Model dynamics: the monetary transmission mechanism
- Channels of shock propagation for a one standard deviation monetary tightening:
  1. Interest-rate channel:
     - Higher real interest rates increase saving (postpone consumption) via the Euler equation and raise the opportunity cost of investment, decreasing capital accumulation; domestic demand and output decrease, putting downward pressure on inflation.
  2. Exchange-rate channel (open economy):
     - Higher nominal interest rate increases real interest rates and the real exchange rate (appreciation), suppressing net exports (expenditure switching) and reducing aggregate demand.
  3. Financial accelerator channel:
     - Higher interest rates depress asset prices (real price of capital), deteriorating net worth; weakened balance sheets increase the external finance premium, further raising the opportunity cost of investment and amplifying the contraction in investment and output.
- Uncertainty and impulse responses:
  - Bayesian impulse responses with 90 percent bands incorporate parameter uncertainty.
  - A one standard deviation contractionary monetary policy shock corresponds to a 70 basis point (quarterly) increase in the nominal interest rate—an annual increase in the policy rate of about three percent.
  - Impulse response magnitudes:
    - The output gap dips below steady state by 70 basis points.
    - The year-over-year inflation rate reaches a trough of about 140 basis points below steady state after four periods.
- Model structure:
  - The model includes 15 structural shocks (including the monetary policy shock); impulse responses for remaining shocks are detailed in the Appendix.

### Historical decompositions: contributions to output growth
- Shock grouping for decomposition:
  - Group 1: monetary policy shocks.
  - Group 2: crisis shocks (foreign demand, financial uncertainty, uncovered interest parity/sudden stop).
  - Group 3: remaining supply and demand shocks.
- Monetary policy contribution:
  - Monetary policy shocks positively contributed to output growth during the crisis.
  - Average growth contribution of monetary policy shocks during the crisis episode is about 1.1 percent.
  - Context: year-over-year real GDP contraction in Turkey in 2009 was –4.8 percent.
  - Counterfactual without discretionary monetary policy: model indicates 2009 growth would have been –5.9 percent instead (i.e., monetary policy contributed to soften the impact by about 1.1 percent).
- Crisis shocks:
  - Sudden stop (UIP/risk premium) shock had limited effect on growth during the crisis in this specification (entrepreneurs borrow in domestic currency).
  - Phases:
    - Around 2005 run-up: positive foreign demand shocks supported growth.
    - Late 2008 downturn: financial uncertainty shock initially negatively impacted growth, followed by foreign demand shock.
    - Recovery onset: decrease in financial uncertainty shocks led the recovery.
  - Financial uncertainty shock explains a large fraction of the downturn among the three crisis shocks; foreign demand shock has lingering negative effects due to weak Euro area growth.
- Other supply and demand shocks:
  - Prominent supply shocks: unit-root technology shock and investment-specific technology shock.
  - Limited role for cost-push (markup) and stationary technology shocks.
  - Demand shocks: government spending, preference, and time-varying inflation target; the time-varying inflation target has a negligible role, while government spending and preference shocks tend to offset each other.
  - Net effect of these remaining shocks acted as a drag on growth, particularly in the early phase of the global financial crisis.

### The role of monetary policy during the crisis: counterfactual experiments
- Counterfactual question:
  - If flexible exchange rate regime and active countercyclical monetary policy within an inflation targeting framework were not adopted, how much deeper would the recession have been?
- Main counterfactual result:
  - The counterfactual experiments indicate that the countercyclical and discretionary interest rate cuts implemented by the CBRT within an inflation targeting regime underpinned by a flexible exchange rate added at least 3.2 percentage points to the 2009 real GDP growth outturn.

*Source: _wp11150 - 2. However, rather*

### 2006. Against this backdrop, while not the focus of the paper, as a by-product of our

### _wp11150 - 2006. Against this backdrop, while not the focus of the paper, as a by-product of our

### Counterfactual simulations: setup and purpose
- Baseline: monetary policy under a flexible exchange rate regime, follows the estimated baseline interest rate rule which reacts to the output gap and allows for deviations from the rule (monetary policy shocks).
- Counterfactual logic: four cumulative counterfactual simulations are compared with the actual realization (baseline) to quantify the role of monetary policy, exchange rate regime, and financial vulnerability (via leverage) on output dynamics during the global financial crisis.
- Leverage calibration: based on cross-section of firms, average leverage ratio decreased to a value of 2 in 2007 from a value of 3 in 2000; counterfactuals alter the steady-state leverage ratio to assess reforms’ role.

### The four counterfactual experiments
- No monetary policy shocks:
  - All monetary policy shocks set to zero.
  - Asks: What would output growth have been if the CBRT did not implement any discretionary policy (deviations from the interest rate rule) during the crisis?
- No response to the output gap:
  - Output gap coefficient in the empirical interest rate rule set to zero (estimated coefficient noted later as 0.02).
  - Monetary policy shocks also set to zero (cumulative with previous).
  - Asks: What would output growth have been if the CBRT neither implemented discretionary policy nor took the output gap into consideration?
- Peg (fixed exchange rate):
  - CBRT implements a strict fixed exchange rate regime.
  - Monetary policy does not react to the output gap and no discretionary deviations from the rule; policy solely focuses on stabilizing the nominal exchange rate.
  - Asks: What would output growth have been under a fixed exchange rate regime?
- Peg with heightened financial vulnerability:
  - Same fixed exchange rate as above.
  - Leverage ratio calibrated to equal 3, in line with the value in Table 1 during 2000 (contrast with 2.0 in 2007).
  - Asks: What would output growth have been with a peg and a more vulnerable financial sector?

### Results from counterfactual simulations: main findings
- Overall statement:
  - The inflation targeting framework underpinned by a flexible exchange rate regime softened the impact of the global financial crisis.
- Three main results:
  - First — exchange rate flexibility and financial soundness:
    - Output growth declines the most under the fixed exchange rate regime.
    - The lack of exchange rate as a shock absorber reduces resiliency to crisis shocks.
    - The illustrative counterfactual with heightened financial vulnerabilities (leverage = 3) leads to an even sharper decline in output, highlighting the role of exchange rate flexibility and financial reforms that improve banking soundness.
  - Second — output gap weight in the interest rate rule:
    - The estimated output gap coefficient is 0.02.
    - Giving weight to the output gap has a limited role because the coefficient is quite low; implies a small systematic response of the policy rate to the output gap and a large discretionary (nonsystematic) response.
  - Third — discretionary departures from the interest rate rule:
    - Discretionary departures (monetary policy shocks) played an important role in softening the crisis impact.
    - The average contribution of the monetary shocks to output growth is around 1 percent (1.14 or 1.18 percent depending on episode), comparable to values in the literature (Christiano and others (2007): U.S. 0.75 percent, euro area 1.27 percent).
    - Cumulative growth contributions from monetary policy are stated to be reasonable and contextually informative on monetary policy’s role during the crisis.

### Quantitative growth implications (selected figures and comparisons)
- Turkey macro and policy context:
  - Turkey experienced a year-over-year contraction of 14.7 percent during the first quarter of 2009.
  - The CBRT decreased policy rates by 1025 basis points over the November 2008 to November 2009 period.
- 2009: most intense crisis year — simulated outcomes:
  - Actual growth rate for 2009: –4.8 percent.
  - Counterfactual without discretionary departures from the empirical interest rate rule:
    - One statement: growth would have been –5.9 percent, a difference of 1.1 percentage points from actual.
    - A later summary statement: without the countercyclical interest rate cuts, growth in 2009 would have decreased from –4.8 percent to –6.2 percent.
  - Counterfactual with fixed exchange rate (peg):
    - Growth implied: –8.0 percent, a difference from actual of 3.2 percentage points.
- Counterfactual contributions table references (descriptive):
  - Table 5 presents average or cumulative contributions to growth for two crisis episode definitions: 2008:Q4–2009:Q4 and 2009:Q1–2009:Q4.
  - Columns 1–5 in Table 5 indicate incremental contributions due to consecutive implementation of each policy; example provided: for 2009:Q1–2009:Q4, column 4 indicates reducing financial vulnerabilities added, on average, 1.45 percentage points to growth, and column 3 indicates an incremental contribution of 1.86 percentage points from adopting a flexible exchange rate regime.
- Alternative severity metrics (appendix):
  - Peak-to-trough output contraction differences using level of GDP.
  - “Area under the curve” metric: annualized average output loss relative to the baseline.
  - Appendix Table 1 quantifies these alternative measures and supports the conclusion that inflation targeting with a flexible exchange rate helped soften the crisis impact.

### Model and methodological notes
- Model: estimated structural DSGE model for Turkey over 2002–10, open-economy New Keynesian with nominal and real rigidities, stochastic trend, financial accelerator mechanism.
- Agents: households, producers (entrepreneurs, capital producers, retailers), government; interactions with rest of world included.
- Household preferences incorporate habit persistence; consumption is a CES index of domestic and imported goods with domestic share and elasticity parameters specified in model equations.
- External finance premium and borrowing premium specification follow Gertler and others (2007) logic; country borrowing premium depends on net foreign indebtedness and an exogenous AR(1) process.

### Policy implications (summary)
- The combination of countercyclical interest rate cuts (including discretionary departures from the baseline interest rate rule) and exchange rate flexibility provided important output stabilization during the global financial crisis.
- Financial sector reforms that reduced leverage (from 3.0 in 2000 to 2.0 in 2007) lowered financial vulnerability and contributed to greater resilience; a higher leverage calibration (3.0) in a peg scenario materially worsens output outcomes.
- While monetary policy under inflation targeting and exchange rate flexibility played a central role, other policies — notably CBRT liquidity measures starting in Q4 2008 (extending repo terms, restarting foreign exchange auctions, reducing reserve requirements on foreign exchange deposits) and fiscal policy measures — also likely contributed to stabilization and may be partially attributed to monetary policy shocks in the model.

*Source: _wp11150 - 2006. Against this backdrop, while not the focus of the paper, as a by-product of our*

### introduction of this risk-premium is needed in order to ensure a well-defined steady state in

### _wp11150 - introduction of this risk-premium is needed in order to ensure a well-defined steady state in

### Model structure and key mechanisms
- Representative households maximize intertemporal utility yielding an Euler equation and marginal utility expressions; foreign bond choice gives an uncovered interest parity (UIP) condition with an exogenous process interpreted as a risk premium (UIP) shock.
- Wage and price rigidities:
  - Households are monopolistic suppliers of differentiated labor and set wages, with inelastic supply of demanded labor.
  - Wages follow a Calvo-style updating with indexation to the time-varying inflation target, past inflation, and technology growth; degree of wage indexation denoted by a parameter with support between 0 and 1.
  - Retailers of domestic and imported goods follow Calvo (1983) price-setting with indexation and time-varying markups.
- Open-economy trade and pricing-to-market:
  - Wholesale law of one price holds between domestic currency wholesale price, foreign currency price, and nominal exchange rate.
  - Foreign demand for home tradable goods modeled following Gertler and others (2007); foreign demand and foreign price level assumed exogenous AR(1) processes.
- Financial accelerator and entrepreneurs:
  - Entrepreneurs are risk neutral, have finite expected horizon (survival probability is time-varying and stochastic), and face idiosyncratic shocks to production and capital effectiveness.
  - Entrepreneurial balance sheet: capital financed by net worth and nominal bonds (standard identity).
  - Costly state verification generates an external finance premium that varies inversely with entrepreneur net worth and depends on the real price of capital, providing a channel for asset-price effects on financing costs.
  - The model embeds a financial accelerator linking borrower financial positions, external finance premia, and capital demand.
- Capital producers and investment:
  - Capital accumulation includes capital adjustment costs; investment good is a CES aggregate of domestic and imported investment goods.
  - Only the parameter associated with adjustment cost curvature is identified in the log-linearized model.

### Shocks and interpretation (channels of the global financial crisis)
- UIP shock (risk premium shock) used to mimic a sudden stop shock that causes large capital outflows; used to capture the financial aspect of the global financial crisis.
- Net worth shock interpreted as a financial uncertainty shock (impairment or destruction of entrepreneurial financial wealth; counterparty/Knightian risk) that directly affects aggregate net worth and the external finance premium.
- Entrepreneurial survival probability is time-varying and subject to an exogenous shock; adverse shocks interpreted as impairments of entrepreneurs’ assets (heightened financial uncertainty).
- Investment-specific technology shock follows an AR(1) process.
- Foreign demand and foreign price level are exogenous AR(1) processes; shocks to foreign demand for domestic products capture the trade channel of the global financial crisis.

### Wage, price, and inflation dynamics
- Wage stickiness and indexation:
  - Non-optimizing households update wages by indexing to the current inflation rate target, previous period inflation, and technology growth.
  - Indexation prevalence and sticky wages/prices are emphasized as important for capturing Turkish inflation persistence.
- Phillips-curve relations:
  - The model yields log-linearized Phillips curves for domestic and imported goods inflation under Calvo with indexation for each sector.
- Monetary policy rule:
  - Log-linear interest rate rule responds to deviations of inflation from a time-varying inflation target, the output gap, exchange rate depreciation, and the lagged interest rate; includes an i.i.d. monetary policy shock.
  - Inflation target evolves stochastically.
  - The monetary policy shock captures deviations from the systematic rule—for example, interest rates decreasing by more than the rule-implied amount during intense episodes of the global financial crisis.

### Financial sector, entrepreneurs, and quantitative mechanisms
- Entrepreneur production and financing:
  - Entrepreneurs produce using labor and capital utilization; capital utilization affects depreciation (Greenwood et al., 1988).
  - Idiosyncratic productivity shock with E[...] = 0 affects both production and effective capital quantity.
- External finance premium and leverage:
  - External finance premium is an increasing function of the leverage ratio and depends on the real price of capital (asset price).
  - The entrepreneur’s marginal cost of funds equals the gross external premium times the gross real opportunity cost.
- Evolution of entrepreneurial net worth:
  - Net worth dynamics are a function of realized project returns, survival rates, and managerial wage; stochastic survival rate drives time-varying net worth.
- Financial accelerator summary:
  - The model provides an explicit mechanism whereby asset-price movements and shocks to net worth/survival amplify the effects of shocks on capital demand and aggregate output via external finance premia.

### Calibration and identification
- Calibrated steady-state targets and parameter choices:
  - Consumption-to-GDP ratio: 70 percent.
  - Investment-to-GDP ratio: 20 percent.
  - Government expenditures-to-GDP ratio: 10 percent.
  - Exports-to-GDP ratio: 24 percent.
  - Discount factor β set to 0.9928 (implying an annual riskless real interest rate of approximately three percent).
  - Entrepreneurial survival rate chosen at 0.9728.
  - Monitoring cost fraction set to 0.15.
  - Variance of shocks to entrepreneurial productivity set to 0.40.
  - Target leverage ratio of two for year 2007 used in calibration.
  - Average EMBI spread targeted at 300 basis points (over the 2004-2010 period, excluding immediate aftershocks of 2001).
  - Share of entrepreneurial labor set at 0.01.
  - Steady-state price and wage markups chosen to be 15 percent.
- Other calibrated parameters drawn from existing literature (Gertler and others (2007), Bernanke and others (1999), etc.); various elasticities of substitution and structural parameters set consistent with prior studies.
- Identification note:
  - Only the parameter associated with capital adjustment cost curvature is identified in the log-linearized model.

### Data, estimation sample, and Bayesian estimation approach
- Data and observed variables:
  - Quarterly data from the first quarter of 2002 to the second quarter of 2010.
  - Twelve observed/model-mapped variables matched: levels of domestic policy and foreign interest rates; inflation rates of domestic GDP deflator and core consumer price; foreign consumer price index inflation; growth rates of GDP, consumption, investment, exports, imports, foreign GDP; and the real exchange rate growth rate.
  - Observables vector mapped using model notation (growth rates of exports and the real exchange rate are explicitly included).
  - Interest rates divided by four to obtain quarterly periodic rates.
  - Time series demeaned by removing sample mean, except inflation and interest rates which are demeaned by subtracting steady-state values.
- Foreign variables:
  - Foreign aggregates constructed as weighted averages of United States and Euro area time series for real GDP, interest rate, and inflation rate.
  - Rationale: Euro area is destination of over fifty percent of Turkish exports; United States origin of the global financial crisis.
- Estimation methodology:
  - Log-linearized model estimated by Bayesian methods following Schorfheide (2000) and Smets and Wouters (2003, 2007).
  - Calibration treated as infinitely strict priors for selected parameters.
  - Remaining 43 parameters estimated (priors in Table 3) covering degrees of real and nominal rigidities, monetary policy stance, and persistence/volatility of exogenous shocks.
  - Prior distributions: beta for parameters bounded between 0 and 1; inverse gamma for positive-only parameters (standard deviations); normal for unbounded parameters.
- Sample choice and justification:
  - Estimation sample covers 2002–2010 to capture the transition of the CBRT to inflation targeting (implicit to explicit starting in 2006) and to avoid spurious inference from regime changes and structural breaks in longer series.
  - Note on sample length: sample is shorter than ideal but comparable to other studies (Del Negro and Schorfheide (2008) cited).

*Italic line: Source: _wp11150 - introduction of this risk-premium is needed in order to ensure a well-defined steady state in (PDF chapter/section).*

### 0.5 as in Teo (2009). Similarly, the indexation parameter is set to 0.5 as well, as in Adolfson

### _wp11150 - 0.5 as in Teo (2009). Similarly, the indexation parameter is set to 0.5 as well, as in Adolfson

### Parameter calibration and priors
- Indexation parameter set to 0.5 as in Teo (2009) and Adolfson and others (2007).
- Baseline monetary policy rule:
  - Interest rate persistence = 0.7 (in line with Elekdag and other (2006)).
  - Responsiveness to inflation = 1.4 (in line with the Taylor principle; slightly lower than in some other studies, e.g., Garcia-Cicco (2010)).
- Habit persistence parameter = 0.7 as in Adolfson and other (2007).
- Investment adjustment cost parameter described as relatively lower (no numeric change from source text).
- Shock persistence parameter baseline = 0.8 (contrast: Adolfson and other (2007) use 0.85; Elekdag and other (2006) and Garcia-Cicco (2010) use 0.5).
- Priors for most standard deviations of shocks: inverse gamma distribution typically centered around 0.03 with one degree of freedom.

### Estimation setup and posterior sampling
- Posterior distributions obtained via Metropolis-Hastings sampling.
- Sampling details:
  - Total draws = 500,000.
  - Chains = two independent chains.
  - Brooks and Gelman (1998) convergence criteria achieved.
- Posterior reporting includes means and 5th and 95th percentiles (as in Table 3 referenced).
- Kernel density estimates for posteriors and priors shown in Appendix Figure 1a through Appendix Figure 1c; these indicate the data are quite informative for most estimated parameters.

### Comparison with other studies and caveats
- Cross-study comparisons caveats:
  - Different countries studied (examples: Garcia-Cicco (2010) — Mexico; Elekdag and others (2006) — Korea; Teo (2009) — Taiwan; Adolfson and other (2008) — Sweden; closed-economy counterparts — United States and euro area in Christiano and others (2008)).
  - Differences in sample periods and choice of time series (this paper includes intense periods of the recent global financial crisis whereas most other studies do not).
  - Differences in model structure, notably the monetary policy rule specification.
- Model, sample period, and data differences should be recognized when comparing posterior estimates across studies.

### Key estimated parameters and interpretation
- Nominal rigidities:
  - Wage-Calvo parameter = 0.6 → implies wages adjusted on average every 2.5 quarters.
  - Domestic prices adjust on average every 1.5 quarters.
  - Degree of indexation parameters in the 0.5–0.6 range → Philips curve has significant backward-looking components.
  - These findings are close to Adolfson and other (2007) and Teo (2009).
- Real rigidities:
  - Habit formation = 0.9.
  - Investment adjustment costs = 3.6.
  - Comparisons: Garcia-Cicco (2010) finds habit formation = 0.8; Teo (2009) estimates investment adjustment cost = 3.2.
- Policy rule estimates:
  - Interest smoothing (interest rate persistence) = 0.7.
  - Responsiveness to inflation deviation from target = 1.5 (noted as in line with Taylor, 1993; close to 1.6 in Adolfson and others (2007)).
  - Responsiveness to nominal exchange rate depreciation described as smaller (no numeric value provided).
  - Responsiveness of policy rates to the output gap = 0.02.
  - Note: small systematic response to output gap; discretionary departures from rule used by CBRT during crisis episodes.

### Shock persistence and volatility
- Persistence parameters range:
  - Unit root technology shock persistence = 0.49.
  - Foreign demand shock persistence = 0.93.
  - 95th percentile for the foreign demand shock persistence = 0.96 (indicating absence of unit roots).
- General finding: persistence parameters in this open economy setup are typically below those in closed economy frameworks like Smets and Wouters (2003).
- Standard deviations:
  - Foreign interest rate shock = least volatile (no numeric SD provided).
  - Preference and investment shocks = notably variable (no numeric SDs provided).
  - Unit-root technology shock more volatile than stationary technology shock.
- Role of unit-root technology shocks: plays a much more prominent role in driving the business cycle, consistent with Aguiar and Gopinath (2007).

### Model fit (one-sided Kalman filter assessment)
- Model versus data correlations:
  - Inflation correlation > 0.95.
  - Interest rate correlation > 0.99.
  - Consumption correlation > 0.97.
  - Investment correlation > 0.99.
- Weaknesses in fit:
  - Exports: actual evolution markedly differs from model predictions, particularly during the global financial crisis and due to rapid diversification of Turkey’s export destinations.
  - Possible causes: unprecedented collapse in global trade during crisis; changing export destinations; unstable interaction between exports and the real exchange rate; related to the "exchange disconnected" puzzle (Devereux and Engel (2002), Duarte and Stockman (2005)).

### Sensitivity analysis and model selection
- Decision metric: posterior odds ratio; baseline model outperforms alternatives.
- Alternative specifications considered = 18.
  - In all but one case, results are decisively in favor of the baseline; in the remaining case the results very strongly favor the baseline.
- Interpretation guidance: Jeffreys (1961)
  - Ratio in 10–100 range → strong to very strong evidence.
  - Ratios above 100 → decisive evidence.
- Key sensitivity findings:
  - Excluding the financial accelerator mechanism → decisively rejected in favor of baseline → underscores importance of financial frictions for emerging markets (also discussed in Elekdag and others (2006) and Garcia-Cicco (2010)).
  - Baseline favored over models with low nominal and real rigidities → canonical RBC or New Keynesian models insufficient to fit data.
  - Structural shocks:
    - Technology (trend) shocks important (Aguilar and Gopinath (2007); Smets and Wouters (2003) — labor supply shocks; Justiniano and others (2007) — investment-technology shocks).
    - Demand shocks (preference and government spending) important; models without them are rejected.
  - Eliminating financial shocks (financial uncertainty and UIP shocks) → baseline decisively chosen.
  - Setting persistence parameter of time-varying inflation target to unity → baseline decisively favored.
  - Five alternative monetary policy rules considered (including dropping nominal depreciation term, adding change in output and inflation, cases without interest rate smoothing, strict inflation targeting, fixed exchange rate regime) → results at least very strongly in favor of baseline; fixed exchange rate regime particularly decisively rejected relative to baseline interest rate rule.
- Empirical interest rate rules do not represent exact central bank reaction functions; data prefer including an exchange rate term, which may proxy forward-looking policymaker behavior (consistent with Lubik and Shorfheide (2007), Elekdag and others (2006), Adolfson and others (2007)).

### Model dynamics and impulse responses (selected shocks)
- Shock grouping: monetary policy shock; crisis shocks; other supply and demand shocks.
- UIP/risk premium shocks:
  - Two choices: foreign interest rate shock (pinned to data) or exogenous UIP shock (risk premium shock); persistence coefficients similar; impulse responses similar (Appendix Figure 2a and 2b).
  - Under fixed exchange rate regime:
    - Shock to UIP → increase in domestic nominal interest rate → sticky prices imply increase in real interest rate → contraction.
    - Financial accelerator amplifies initial impact (asset prices fall → leverage rises → external finance premium increases → investment and output decline).
  - Under baseline (flexible exchange rate + financial accelerator):
    - Exchange rate depreciation offsets some adverse effects; real interest rates do not spike; real exchange rate depreciates more → softens shock impact.
  - Flexible exchange rates with financial accelerator shut off → even smaller perverse effect on output.
- Foreign demand shock (Appendix Figure 3a):
  - Decrease in exports → directly depresses aggregate demand.
  - Sudden nominal exchange rate depreciation under baseline mitigates impact via real exchange rate movements.
- Financial uncertainty shock (Appendix Figure 3b):
  - Erodes entrepreneur net worth → leverage ratio increases → external finance premium rises → investment and output decline.
  - Estimated persistence (autoregressive coefficient) = 0.72.
  - Balance sheet evolves gradually → protracted shock.
  - Shock can convert financial frictions from accelerator to attenuator due to endogenous counter-cyclical adjustments (Christiano and others (2010); Christensen and Dib (2008); Iacoviello (2005) distinctions noted).
  - Under fixed exchange rate (associated with lower real interest rate and small appreciation of real exchange rate), output contraction is less severe than under baseline.
- Trend (unit-root) technology shock (Appendix Figure 4a):
  - Permanently affects trend → output gap settles at permanently lower level.
  - Household consumption decreases; associated real exchange rate depreciation.
  - Under peg: nominal exchange rate cannot depreciate → decrease in inflation is greater → larger rise in real interest rate → larger decline in output.
- Stationary technology shock (Appendix Figure 4b):
  - Output gap widens while inflation increases (textbook negative correlation between changes in output and prices).
  - Households smooth consumption by borrowing internationally → exchange rate appreciates.
- Domestic markup shock (Appendix Figure 5a):
  - Remarkably similar to stationary technology shock.
  - Under baseline, external finance premium increases → output contraction slightly more severe with financial accelerator active.
  - Labor supply shock produces similar responses.
- Foreign cost-push shock (Appendix Figure 5b) (can be thought of as terms of trade / oil price shock):
  - Dynamics similar to foreign demand shock (terms of trade shocks resemble external income shocks).
- Government spending and preference shocks (Appendix Figure 6a and Appendix Figure 6b):
  - Contractionary demand shocks put downward pressure on real marginal costs and thereby inflation.
  - Choice of exchange rate regime importantly affects final outcomes for exchange and interest rates and the external finance premium.

### Growth counterfactuals (note)
- Main text discusses counterfactual scenarios using levels of real GDP.
- The same policy messages can be conveyed using demeaned, year-over-year growth rates under baseline (actual) and counterfactual simulations (figures/data referenced but not reproduced in source excerpt).

*Italic: Source — _wp11150 (excerpts provided from the referenced PDF chapter/section).*

### Appendix Figure 7. In sum, without the adoption of an inflation targeting framework

### _wp11150 - Appendix Figure 7. In sum, without the adoption of an inflation targeting framework

### Summary conclusion
- Without the adoption of an inflation targeting framework underpinned by a flexible exchange rate regime, the impact of the recent global financial crisis would have been substantially more severe.
- Growth rates under fixed exchange rate regimes appear quite high in the last few quarters of the sample mainly as a result of base effects; the level of economic activity is substantially depressed under the fixed exchange rate regimes, so year-over-year growth rates include base effects and should be carefully interpreted.

### Calibrated parameters (selected)
- Discount factor: 0.9928
- Consumption intra-temporal elasticity of substitution: 1
- Share of domestic goods in consumption: 0.75
- Investment intra-temporal elasticity of substitution: 0.25
- Share of domestic goods in investment: 0.77
- Inverse of the elasticity of work effort with respect to the real wage: 1
- Share of capital in production function: 0.4
- Elasticity of marginal depreciation with respect to utilization rate: 1
- Steady state markup rate for domestically produced goods: 1.15
- Steady state markup rate for imported goods: 1.15
- Steady state markup rate for wages: 1.15
- Share of entrepreneurial labor: 0.01
- Steady state external finance premium: 1.03
- Number of entrepreneurs who survive each period (at steady state): 0.9728
- Variance of idiosyncratic shock to entrepreneur production: 0.4
- Fraction of monitoring cost: 0.15
- Depreciation rate (at steady state): 0.035
- Elasticity of country risk premium with respect to net foreign debt: 0.01

### Posterior estimates (selected parameters with 5%–95% confidence intervals)
- Calvo parameter (Domestic prices): Mean 0.306; Confidence interval 0.177 0.435
- Calvo parameter (Import prices): Mean 0.554; Confidence interval 0.434 0.679
- Calvo parameter (Wages): Mean 0.558; Confidence interval 0.387 0.739
- Indexation (Domestic prices): Mean 0.460; Confidence interval 0.296 0.623
- Indexation (Import prices): Mean 0.475; Confidence interval 0.313 0.641
- Indexation (Wages): Mean 0.562; Confidence interval 0.400 0.731
- Export demand elasticity: Mean 0.182; Confidence interval 0.090 0.271
- Export demand inertia: Mean 0.883; Confidence interval 0.794 0.979
- Habit formation: Mean 0.904; Confidence interval 0.850 0.958
- Investment adjustment cost: Mean 3.625; Confidence interval 2.769 4.511
- Interest rate smoothing: Mean 0.724; Confidence interval 0.639 0.808
- Inflation response (monetary policy): Mean 1.537; Confidence interval 1.383 1.685
- Output gap response (monetary policy): Mean 0.021; Confidence interval -0.022 0.064
- Nominal exchange rate response (monetary policy): Mean 0.173; Confidence interval 0.101 0.242
- Shock persistence examples:
  - Stationary technology: Mean 0.767; Confidence interval 0.593 0.939
  - Unit root technology: Mean 0.497; Confidence interval 0.386 0.604
  - Investment specific technology: Mean 0.916; Confidence interval 0.893 0.940
- Shock volatility example:
  - Investment specific technology (Inverse gamma mean): 0.421; Confidence interval 0.334 0.504

### Sensitivity analysis (Table 4; model comparison)
- Baseline log data density: 802.482
- Alternatives (Log Data Density, Posterior Odds Ratio, Model Superior?)
  - 1 Financial accelerator: 779.963; 6.025E+09; No
  - 2 Low price stickiness: 787.606; 2.888E+06; No
  - 3 Low stickiness including wages: 726.449; 1.049E+33; No
  - 4 Low habit persistence: 719.789; 8.193E+35; No
  - 5 Low investment costs: 753.752; 1.456E+21; No
  - 6 Technology (all): 619.305; 3.573E+79; No
  - 7 Labor supply: 706.645; 4.185E+41; No
  - 8 Investment-specific technology: 664.178; 1.161E+60; No
  - 9 Preference: 664.068; 1.296E+60; No
  - 10 Government: 699.794; 3.951E+44; No
  - 11 Financial (uncertainty and UIP): 796.283; 4.924E+02; No
  - 12 Unit root inflation target: 794.754; 2.272E+03; No
  - 13 Add change in output and inflation: 798.223; 7.073E+01; No
  - 14 Baseline rule, but without ΔS rule: 796.913; 2.622E+02; No
  - 15 Same as (13), but without ΔS: 792.821; 1.570E+04; No
  - 16 No interest rate smoothing: 779.072; 1.469E+10; No
  - 17 Strict inflation targeting: 776.368; 2.193E+11; No
  - 18 Fixed exchange rate regime: 752.835; 3.644E+21; No

### Role of monetary policy and financial reforms (selected numbers from Table 5 and related text)
- Table 5 reports impacts (In percent) for scenarios including Cut in policy rates, Responsiveness to shocks, Flexible exchange rate, Reduced output gap, All factors combined.
- Average (2008Q4—2009Q4): 51.98 1.18 0.32 2.22 1.69 5.42
- Average (2009Q1—2009Q4): 2.40 1.14 0.35 1.86 1.45 4.81
- Cumulative (2008Q4—2009Q4): 59.90 5.92 1.60 11.11 8.44 27.08
- Cumulative (2009Q1—2009Q4): 9.59 4.56 1.42 7.45 5.81 19.23
- Comparative references:
  - Christiano and others (2008) United States (2001Q2-2002Q2): 40.75 (average) / 43.00 (cumulative)
  - Christiano and others (2008) Euro area (2001q4-2004q4): 131.27 (average?) / 1317.00 (cumulative?) [numbers reported in source]

### Counterfactual scenarios and severity measures (Table 6 and Appendix Table 1)
- In percent (calculations relative to actual 2009 annual real GDP growth):
  - Baseline (actual): –4.8
  - No monetary policy shocks: –5.9; Difference –1.1; Difference –1.1
  - No response to output gap: –6.2; Difference –0.3; Difference –1.4
  - Fixed exchange rate regime (peg): –8.0; Difference –1.8; Difference –3.2
  - Peg with heightened financial vulnerability: –9.3; Difference –1.3; Difference –4.5
- Peak-to-trough cumulative contractions (Appendix Table 1):
  - Baseline (actual): –14.4
  - No monetary policy shocks: –15.4; Difference –1.0; Difference –1.0
  - No response to output gap: –15.8; Difference –0.4; Difference –1.5
  - Fixed exchange rate regime (peg): –19.6; Difference –3.8; Difference –5.3
  - Peg with heightened financial vulnerability: –22.1; Difference –2.4; Difference –7.7
- Relative to baseline (cumulative, peak-trough and annual):
  - No monetary policy shocks: –5.7 (relative difference) and –5.7 (relative)
  - No response to output gap: –1.5 and –7.2
  - Fixed exchange rate regime (peg): –6.8 and –14
  - Peg with heightened financial vulnerability: –8.2 and –22.2

### Model validation and dynamic results (high-level)
- Model predictions versus data are presented for Output, Consumption, Investment, Inflation (core), Interest rate, Real exchange rate, Exports, Imports with filtered lines and actual data comparisons.
- Dynamic responses to a monetary policy shock show:
  - Output gap, Inflation (year-over-year), Nominal interest rate, Real exchange rate, External finance premium, Real interest rate, Imports, Real price of capital, Consumption, Exports, Net worth, Investment with baseline and “No financial accelerator” comparisons.
- Historical decompositions isolate contributions to demeaned year-over-year real GDP growth from:
  - Monetary policy; Crisis shocks (Foreign demand, UIP, Financial uncertainty); Other supply and demand shocks (Unit root technology, Other supply, Investment, Other demand).
- Counterfactual levels of real GDP across scenarios (Baseline; No monetary policy shocks; No response to output gap; Fixed exchange rate regime (peg); Peg with heightened financial vulnerability) show materially lower levels under peg scenarios (visualized in Figure 9 levels series).

*Source: _wp11150 - Appendix Figure 7. In sum, without the adoption of an inflation targeting framework*

### Appendix Figure 3a. Impulse Responses: Foreign Demand Shock

### Appendix Figure 3a. Impulse Responses: Foreign Demand Shock

### Description
- Source: Authors’ calculations.
- Footnote: Interest rates, inflation rates, and the external finance premium are shown as absolute deviations from their steady states, while the other variables are percentage deviations from their steady states.
- Figure shows impulse responses over horizons labeled 0 2 4 6 8 10 12 14 16 18 20 for a foreign demand shock.

### Variables (as labeled in the figure)
- Output gap
- Inflation (year-over-year)
- Nominal interest rate
- Real exchange rate (increase indicates appreciation)
- External finance premium
- Real interest rate

### Scenarios / Series (legend entries)
- Baseline
- Fixed exchange rate regime
- No financial accelerator

### Axis scales and tick values shown in the figure
- For Output gap (first panel): -4.50%, -4.00%, -3.50%, -3.00%, -2.50%, -2.00%, -1.50%, -1.00%, -0.50%, 0.00%
- For Inflation (year-over-year) (second panel): -5.00%, -4.00%, -3.00%, -2.00%, -1.00%, 0.00%, 1.00%, 2.00%
- For Nominal interest rate (third panel): 0.00%, 0.10%, 0.20%, 0.30%, 0.40%, 0.50%, 0.60%
- For Real exchange rate (fourth panel): -10.00%, -9.00%, -8.00%, -7.00%, -6.00%, -5.00%, -4.00%, -3.00%, -2.00%, -1.00%, 0.00%
- For External finance premium (fifth panel): -0.20%, 0.00%, 0.20%, 0.40%, 0.60%, 0.80%, 1.00%, 1.20%
- For Real interest rate (sixth panel): -0.20%, 0.00%, 0.20%, 0.40%, 0.60%, 0.80%, 1.00%, 1.20%, 1.40%, 1.60%

### Plot horizon
- Horizontal axis labels: 0 2 4 6 8 10 12 14 16 18 20

*Source: Authors’ calculations.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11150.pdf_
