## wpiea2023209-print-pdf

## Source details

**Canonical URL:** [wpiea2023209-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023209-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023209-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023209-print-pdf.pdf.json)

---

### Key findings on recent financial conditions
- After loosening in 2020, the estimated FCI began tightening mid-2021, with funding constraints becoming more binding in Europe.
- Conditions tightened further in 2022 following the start of Russia’s war in Ukraine, driven by wider spreads and higher volatility (higher market price of risk).
- Monetary policy turned significantly tighter as the ECB and other central banks started hiking policy rates.
- As monetary policy tightened:
  - credit availability and costs tightened,
  - household and corporate lending rates rose,
  - stock prices fell,
  - government bond yields rose.
- Pace of tightening started slowing in late-2022 as retreating energy prices lowered market perceptions of risk and volatility.

### Scope and construction of the Financial Conditions Index (FCI)
- The FCI:
  - captures availability and affordability of financing,
  - estimates effects of changes in financial conditions on output, inflation, unemployment,
  - is constructed for a broad range of European countries (Euro area and non-Euro area), disaggregated by country and sector.
- Aggregation into five driver categories: Credit availability and costs; External conditions; Funding constraints; Policy stance; Price of risk.
- Main estimation method: supervised data-reduction via Partial Least Squares (PLS) estimation.
- Target supervising variable: quarterly year-on-year growth rate of total financial liabilities for main sectors (HH, NFC, GG); sectoral FCIs constructed for HH, NFC, GG; HH and NFC aggregated into private sector FCI for much of the analysis.

### Quantified macroeconomic impacts (sample 2000 to 2023)
- Over a three-year horizon, switching to a tighter FCI regime is estimated to:
  - lower real GDP by 2.2 percent,
  - lower inflation by 0.7 percentage points,
  - increase the unemployment rate by about 0.3 percentage points.
- Average cumulative tightening since mid-2021: about 2.5 standard deviations (magnitude varies across countries).
- Near-term drag on output: expected through 2025, with an average impact of around three quarters of a percentage point during 2023.
- Additional quantitative summaries reported:
  - Sacrifice ratio between 3 and 4.
  - A one-standard deviation increase in a worldwide financial stress indicator is associated with a reduction in level of output by 0.8 percent and an increase in the unemployment rate by 0.1 percentage point after a year (Hites et al., 2023).
  - After one year, a shift in regime reduces output by 0.9 percent and increases the unemployment rate by 0.06 percentage point (Hites et al., 2023).
- Sample evidence on unconditional average growth by FCI regime (mean annualized q/q real GDP growth, in percent):
  - Lag 1: Loose 3.5 / Neutral 2.0 / Tight 1.0
  - Lag 2: Loose 3.5 / Neutral 2.2 / Tight 1.0
  - Lag 3: Loose 3.3 / Neutral 2.0 / Tight 1.5
  - Lag 4: Loose 2.7 / Neutral 2.3 / Tight 1.8

### Cross-country and cross-sector heterogeneity
- Tightening is broad-based across sectors (households and corporations) and countries, but not uniform.
- Unlike prior cycles, government conditions also tightened in the recent cycle; increased government borrowing did not offset private tightening in 2022.
- Policy tightening cycle is synchronized across countries, affecting both Euro and non-Euro area economies, though the extent differs across countries and within the euro area.
- Divergences in estimated FCIs across countries imply heterogeneous macro effects of monetary policy tightening across Europe.
- Sectoral differences:
  - Households: tightening mainly from policy stance, lower credit availability, and higher lending rates; liabilities move more slowly (longer mortgage maturities) and react with longer lags.
  - Corporates: tightening more driven by price of risk (widening corporate spreads, rising volatility); corporate liabilities are more volatile and declined rapidly during 2022.
  - General government: historically countercyclical, but in 2022 government liabilities contracted in sync with private sector liabilities.

### Recent dynamics and drivers
- Historical episodes summarized:
  - 2003–2006: substantial loosening.
  - Global Financial Crisis: rapid tightening driven by decline in credit availability and rise in cost.
  - European sovereign debt crisis: wider sovereign spreads tightened FCs; later liquidity provision and government interventions loosened conditions.
  - Pandemic onset: exceptional policy support loosened conditions through end-2020 despite spike in price of risk; government funding increased and offset private deleveraging.
  - 2021–22: tightening beginning in 2021 and intensifying in early 2022; drivers shifted to lower availability and higher cost of credit and higher price of risk; policy stance tightened with rate hikes.
- Panel regression evidence (Italy, Germany, Spain, France; 363 observations) on price of risk:
  - HICP Inflation: 0.114*** (0.0392) and 0.139** (0.0377)
  - D.Policy stance: 0.802*** (0.105)
  - HICP Inflation # D.Policy stance: -0.0972*** (0.0253)
  - Adjusted R-squared: 0.90 and 0.92
  - Note: country and time fixed effects included; contributions in first differences; 4 quarter lags of inflation included. * p<0.1, ** p<0.05, *** p<0.01.
- Price of risk tends to rise with inflation, but tighter policy stance weakens inflation’s impact on price of risk.
- Near-term outlook: conditions likely remain tight; July 2023 euro area Bank Lending Survey indicated tighter credit standards and deteriorating bank funding.

### Policy-relevant takeaways and recommendations
- Combine policy levers: Given heterogeneous effects across countries, combining monetary, fiscal, and macroprudential policies can support macroeconomic stability.
- Macroprudential calibration: Calibrate macroprudential policies by sector (HH vs NFC) to contain future financial fragilities given sectoral divergence in FCIs.
- Temporary nature of tightening: Tighter monetary policy tightens FCs in the near term, but this effect is transitory—bringing inflation under control can reduce price of risk and support eventual normalization and loosening of FCs.
- Elevated government debt: High government debt ratios may constrain funding and contribute to tighter FCs.
- Policy challenge: Ensure continued credit provision to viable households and firms without undermining monetary normalization for price stability and fiscal consolidation to reduce public debt vulnerabilities; preserve buffers and prevent future market fragilities.
- Suggested research extensions: explore spillovers of the financial cycle, the impact of tightening financial cycle on vulnerabilities in the financial sector, their dynamics, and broader financial stability implications.

### Methodology — Partial Least Squares (PLS) and IPW causal identification
- PLS rationale and implementation:
  - PLS anchors data-reduction to a target variable (year-on-year growth rate of financial liabilities by sector) to construct FCIs; best specification includes two quarters’ lagged values for each indicator.
  - Advantages: handles large numbers of highly collinear variables; links explanatory variables to a standardized target aiding interpretation; enables sectoral FCIs and comparability across countries.
  - Decomposition used in PLS (notation preserved): X = T P′_X + E_X ; Y = U P′_Y + E_Y; algorithm computes P_X, P_Y, updates U, iterates until convergence; paper uses only the first PLS component for interpretability.
  - Observed liability behavior: NFCs account for largest share and are most dynamic; loans dominate household liabilities; households’ liabilities react with longer lags; corporate liabilities more volatile and declined rapidly in 2022.
- Addressing endogeneity — Inverse Probability Weighting (IPW) treatment-effects approach:
  - Endogeneity sources: policy endogeneity (policy responds to output/inflation) and market expectation effects (market indicators anticipate future developments).
  - Empirical observation: market expectation effects tend to dominate, causing loose FCIs to be on average associated with stronger growth absent adjustment.
  - Identification steps:
    - Discretize FCI into three regimes (tight, neutral, loose) with thresholds set by each country’s historical distribution so that on average 30 percent neutral, 40 percent loose, 30 percent tight.
    - Use multinomial logit to estimate propensity scores p̂(d_j | z_t) and reweight observations via IPW to correct overrepresentation of certain growth–FCI combinations.
    - Estimate outcome equations separately by treatment group and compute doubly-robust IPW/regression-adjusted estimates θ_{h j}.
  - Controls: lagged values, real effective exchange rate, output gap, government balances, current account balances as percent of GDP.
  - Validation: After IPW adjustments, differences in GDP growth, unemployment and inflation across FCI regimes vanish in period 0, supporting comparability across treatment groups.

### Drivers and representative variables used in FCIs (categories preserved)
- Credit availability and costs: Interest Rates of Loans to HHs; Interest Rates of Mortgages to HHs; Interest Rates of Consumer Credit & Other Lending to HHs; Rates on Outstanding Loans to NFCs; Commodity Index; 10-Year Government Bond Yield; Housing Prices; Lending Conditions; Stock Price Index; Brent Crude Oil*; Euro Area 10-Year Yield Curve Spot Rate*; Stock Trading Volume.
- External conditions: Bank Linkages Ratio; USD/EUR Exchange Rate; Germany: 2-Year Government Bond Yield; Germany: 2-Year Government Bond Yield Volatility; Germany: 10-Year Government Bond Yield; Nominal Effective Exchange Rate.
- Funding constraints: LTV Ratio; Government Debt Service; Non-performing Loans to Total Gross Loans; Interest Margin to Gross Income; Return on Assets; Return on Equity; Household Debt to GDP; Regulatory Capital to Risk-Weighted Assets; MSCI Financials Index; Liquid Assets to Short-Term Liabilities; General Government Debt Outstanding*; Nonfinancial Corporations Debt Outstanding*; Financial Corporations Debt Outstanding*; Household Debt Outstanding*; GG interest expense/revenue ratio.
- Policy stance: Households Deposit Rate; NFCs Deposit Rate; Policy Rate2/; Euro Area: Main Refinancing Rate3/; Euro Area: Euro Short-term Rate (€STR)3/; Euro Area: Shadow Short Rate Point Estimates3/; Wu-Xia Shadow ECB Rate3/; 2-Year Government Bond Yield; 1-Month Overnight Interest Rate Swap Close; Money Supply M1*; Money Supply M3*; 3-Month Yield Curve Spot Rate*; 2-Year Yield Curve Spot Rate*; Price of risk 5-Year CDS Premium; iBoxx EUR Non-Sovereigns BBB; iBoxx EUR Non-Sovereigns AAA; EUR Swap Annual 5-Year vs 6-Month; Germany: 10-Year Government Bond Yield Volatility; 2-Year Government Bond Yield Volatility; 10-Year Government Bond Yield Volatility; 2-Year Government Bond Spread; 10-Year Government Bond Spread; 3-Month Interbank Offer Rate; LIBOR-OIS Spread; Stock Price Volatility Index; 10-Year Interest Rate Swap; EURO STOXX 50 Volatility Index; EURO FTSE Volatility Index*; United States CBOE Volatility Index.
- Notes: 1/ Data availability may differ among countries; daily and monthly data are converted to quarterly. 2/ Used for non-EA countries. 3/ Categorized as Policy stance for EA countries, External conditions for non-EA countries. * Used for EA aggregate only.

### Annex IV — Individual country results (overview)
- FCI results provided for Euro Area and individual euro-area countries with panels showing:
  - FCI and Liability growth (rhs, reversed),
  - FCI Changes (contribution to quarter-over-quarter first difference) decomposed by the five drivers,
  - FCI level (contribution to unscaled FCI) by the five drivers,
  - Sectoral FCI panels: Nonfinancial Corporations, Households, General government.
- Countries covered include Austria, Belgium, Croatia, Cyprus, Czech Republic, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Slovakia, Slovenia, Spain.
- Presentation notes preserved: "FCI is not scaled. 2023Q2 and 2023Q3 are forecasts." and "Sources: Author's calculations."

*Source: wpiea2023209-print-pdf*

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

### Executive Summary

### Key findings on recent financial conditions
- After loosening in 2020, the estimated FCI began tightening mid-2021, with funding constraints becoming more binding in Europe.
- Conditions tightened further in 2022, after the start of Russia’s war in Ukraine, as spreads widened and volatility across asset classes increased, reflecting a higher market price of risk.
- The monetary policy stance turned significantly tighter as the ECB and other central banks started hiking policy rates.
- As monetary policy tightened:
  - credit availability and costs tightened,
  - household and corporate lending rates rose,
  - stock prices fell,
  - government bond yields rose.
- The pace of tightening in FCs started slowing down in late-2022 as retreating energy prices lowered market perceptions of risk and volatility.

### Scope and construction of the FCI
- The paper introduces a Financial Conditions Index (FCI) that:
  - captures the availability and affordability of financing,
  - estimates the effect of changes in financial conditions on economic activity (output, inflation, unemployment),
  - is constructed for a broad range of European (both Euro area and non-Euro area) countries, disaggregated by countries and sectors.
- The FCI is constructed consistent with the definition of financial conditions as the availability and affordability of financing.
- Methodology highlights:
  - Uses a supervised learning algorithm, the Partial Least Squares (PLS) estimation.
  - Aggregates indicators into five broad categories of drivers: Credit availability and costs; External conditions; Funding constraints; Policy stance; Price of risk.

### Quantified macroeconomic impacts (sample 2000 to 2023)
- Over a three-year horizon, tighter FCs are estimated to:
  - lower real GDP by 2.2 percent,
  - lower inflation by 0.7 percentage points,
  - increase the unemployment rate by about 0.3 percentage points.

### Cross-country and cross-sector heterogeneity
- The tightening in financial conditions is broad-based across sectors (households and corporations) and countries, but not uniform.
- Unlike prior financial cycles, government conditions have also tightened; increased government borrowing has not acted as an offsetting force in the recent cycle.
- The policy tightening cycle is synchronized across countries (affecting both Euro and non-Euro area European economies), though the extent of tightening differs across countries, including within the euro area.
- Divergences in estimated FCs across countries suggest heterogeneous effects of monetary policy tightening across Europe.

### Policy-relevant takeaways
- Multiple policy levers are beneficial: Given heterogeneous effects across countries, combining monetary, fiscal, and macroprudential policies can help pursue macroeconomic stability.
- Macroprudential calibration: Differences in FCs across sectors point to the need to calibrate macroprudential policies in line with sector-specific developments to contain future financial fragilities.
- Temporary effect of policy tightening: While tighter monetary policy results in tighter FCs in the near term, the analysis suggests this impact is temporary—bringing inflation under control eventually reduces price of risk, possibly supporting normalization and loosening of FCs later.
- Elevated government debt matters: Results suggest elevated government debt ratios may constrain funding, contributing to tighter FCs.
- Policy challenge: Policymakers should aim to ensure continued credit provision to viable households and firms without undermining monetary normalization for price stability and fiscal consolidation to reduce public debt vulnerabilities. Macro-financial policies should preserve adequate buffers and prevent future market fragilities.

### Research contributions and organization
- Contributions:
  - Provides a new methodology to construct an FCI aligned with the availability and affordability of financing.
  - Builds a comprehensive dataset and estimates FCIs for many European countries and by sector (HH, NFC, GG).
  - Estimates the causal impact of changes in the FCI on output, inflation, and unemployment, correcting for potential endogeneity.
- Paper structure (high-level):
  - Section 2: defines financial conditions, reviews literature, introduces conceptual framework of drivers.
  - Section 3: presents data and methodology for the FCI and the framework to assess impacts on macro variables.
  - Section 4: presents findings.
  - Section 5: discusses policy implications and extensions.

*wpiea2023209-print-pdf - Executive Summary*

### Box 1. Alternative Methodologies to Construct FCIs

### Box 1. Alternative Methodologies to Construct FCIs

### A. Macro-based FCIs with Regressions
- Methodology summary:
  - Weighted averages of financial indicators with weights from OLS regressions (Gauthier, Graham, and Liu, 2004; Monetary Conditions Index by the Bank of Canada).
  - IS-curve estimation to construct FCIs via regression methods (Goodhart and Hofmann (2001), Mayes and Viren (2001)).
  - Vector Autoregressions (VAR) mixing target variables (credit, inflation) with financial variables; aggregate Impulse Response Functions to create an FCI (Batini and Turnbull (2002), Gauthier, Graham, and Liu (2004), Swiston (2008)).
  - VAR-based FCIs aim to account for impacts of shocks and to purge endogeneity to retrieve a “pure” financial cycle.
- Advantages:
  - Easy to interpret when derived from a macro model (IS curve interpretation; VAR purges simultaneity depending on identification).
- Shortcomings and limitations:
  - (1) Limited set of variables: adding dozens of regressors in a VAR or OLS increases parametric noise and estimator variances.
  - (2) Severe multicollinearity issues common with financial variables.
  - Specification bias as financial systems grow more complex; may miss substantial parts of the financial cycle.
  - VAR identification challenges: good instruments are hard to find; reliance on arbitrary Cholesky decompositions offers relatively poor identification (see Stock and Watson (2001)).
  - Result: macro-model approach became less popular after the global financial crisis.

### B. Data-Reduction FCIs (PCA / Factor Models)
- Methodology summary:
  - Atheoretical, purely data-driven approach using PCA or factor models; first component/factor often interpreted as the FCI.
  - Examples:
    - Hatzius et al. (2010): U.S. FCI using 57 variables.
    - Chicago Fed: first PCA component on 105 financial variables (Brave and Kelly, 2017).
    - ECB: first three principal components on 24 variables (Angelopoulou, Balfoussia, and Gibson, 2012).
    - IMF GFSR: PCA-based FCIs over 29 jurisdictions (IMF GFSR, online annex 2.1, 2021).
  - Components commonly include policy rates, spreads, exchange rates, asset prices (equity, housing), and volatility measures.
- Advantages:
  - Agnostic “let data speak” approach; can capture a wide range of effects with large numbers of variables.
  - Linear data reduction is simple and widely available.
- Disadvantages and interpretability issues:
  - Aggregating many variables into one metric without a target/anchor complicates interpretation.
  - First components weight variables that explain the most common variance regardless of policy relevance.
  - Loadings are hard to interpret: they are linear combination coefficients embedding impacts of other variables, unlike regression coefficients with ceteris paribus interpretation.
  - Risk of overweighting volatile market variables in economies where bank funding dominates, leading to misrepresentation.

### C. Hybrid FCI Models: Supervised Data Reduction Methods
- Methodology summary and examples:
  - Aim: retain interpretability of macro-based FCIs while remaining data-driven.
  - Koop and Korobilis (2014): TVP-FAVAR (Time-Varying Parameters Factor Augmented VAR) — purges financial cycle from standard macro-cycle via VAR, estimates FCI through latent factor model with time-varying parameters; IMF used TVP-FAVAR (Elekdag et al., 2018) but noted few differences versus PCA despite higher complexity, so recent IMF GFSR versions use simple PCA (IMF 2021).
  - Goldman Sachs Index: two-step approach aggregating 6 categories with weights determined by impact on GDP growth over following 4 quarters; result: heavy weight on corporate and sovereign spreads (for the euro area such weight would make up 40 percent of the index).
  - US Fed researchers (Crump et al., 2021): Bayesian VAR with shrinkage to include 31 variables; captures many transmission channels but is complex, depends on choice of priors and Bayesian hyper-parameters.
  - Bank of France (Petronevich and Sahuc 2019): 18 financial series → first 6 PCA components → fit GARCH for each component → compute conditional volatility → use conditional volatility as weights to aggregate components into one FCI (more volatile factor → higher weight).
  - Bundesbank: ~70 variables → 6 sub-indicators via PCA → aggregate into single FCI with time-varying weights capturing time-varying correlation structure of sub-indicators.
- Advantages and caveats:
  - Two-step and Bayesian hybrid methods can recover higher-order factors and accommodate time variation.
  - Weighting by conditional volatility favors volatile market factors over more stable balance-sheet variables; it is not clear this is desirable since volatile variables may not represent financial conditions more adequately.
  - Complexity and reliance on non-obvious modelling choices (priors, hyper-parameters) require expertise.

### PLS-based FCIs (Supervised Data Reduction / Targeted Approaches)
- Methodology summary:
  - Partial Least Squares (PLS) resembles PCA in avoiding subjective selection but anchors estimated relationships to a target variable (Wold et al. 2001).
  - Chosen target variable here: growth rate of total financial liabilities for main sectors (HH, NFC, GG), quarterly year-on-year:
    - Rationale: (i) reflects availability and affordability of funding to households (HH), non-financial corporates (NFC), and general government (GG); (ii) encompasses loans and debt securities; (iii) is readily available from quarterly financial accounts consistent across countries.
  - FCIs constructed for each sector (HH, NFC, GG). HH and NFC FCIs aggregated into a private sector FCI used throughout the paper.
  - For each target, each PLS regression produces a regression score reflecting explanatory power. Best specification includes two quarters’ lagged values for each indicator.
- Observed sectoral liability composition and behavior:
  - NFCs account for largest share of financial liabilities and are the most dynamic (largest contribution to change in overall financial liabilities).
  - Loans largest for households; NFCs rely on equity and investment fund shares, loans, and debt securities.
  - Households’ liabilities move more slowly (longer mortgage maturities) and react to monetary policy with longer lags.
  - Corporate liabilities are more volatile and declined rapidly during 2022.
  - General government liabilities generally played a countercyclical role, though not in 2022 when government liabilities contracted in sync with private sector liabilities.
- PLS advantages highlighted:
  - PLS handles large numbers of highly collinear variables via linear projections.
  - PLS links explanatory variables to a standardized target variable, aiding economic interpretation: variables with highest loadings have best predictive power for financial liabilities.
  - Supervised PLS ensures comparability across FCIs by using an anchor variable standardized and consistent across countries (observations remain country-specific).
  - PLS enables sectoral FCIs using sector-specific financial liabilities.
  - Other studies: Bank of England (Kapetanios, Price, and Young, 2018) show PLS-based FCIs outperform PCA-based FCIs in forecasting monthly GDP for the United Kingdom; Duo and Wang (2016) find PLS-based FCIs for the United States outperform PCA-based FCIs in out-of-sample GDP forecasting setups; PLS can improve identification of credit supply shocks in SVARs.

### Assessing the Impact of FCIs on the Real Economy (Endogeneity and Causal Identification)
- Endogeneity sources affecting FCI–output relationship:
  - Policy endogeneity effects:
    - When output growth decelerates below potential and expected inflation falls, policy typically loosens (lower policy rates).
    - Inclusion of policy variables in FCI can make looser FCs appear associated with weak output (reverse sign).
    - Conversely, when growth accelerates and policy tightens, tighter FCs may appear associated with strong growth.
  - Market expectation effects:
    - FCIs include market indicators that move in anticipation of future developments (stock prices, spreads).
    - When growth accelerates, stock prices rise, making FCIs look looser — reverse causality from growth to financial conditions.
    - When growth decelerates, stock prices fall, making FCIs look tighter — again reverse causality.
- Empirical observation in sample:
  - Market expectation effects tend to dominate policy endogeneity effects; loose financial conditions are on average associated with stronger growth (see Table 1).
- Table 1 (Unconditional Averages of Growth in Different FCI Regimes)
  - Mean annualized q/q real GDP growth (in percent)
    - Loose FCI / Neutral FCI / Tight FCI
    - Lag 1: 3.5 / 2.0 / 1.0
    - Lag 2: 3.5 / 2.2 / 1.0
    - Lag 3: 3.3 / 2.0 / 1.5
    - Lag 4: 2.7 / 2.3 / 1.8
- Identification strategy used in paper:
  - Treatment effects approach: Inverse Probability Weighting (IPW).
  - Rationale: treatment effects approaches study causal effects without assuming a specific functional form; combination with regression adjustment produces robust estimators (technical presentation in Annex III).
  - Implementation steps:
    - Discretization of FCIs into three regimes: tight, neutral, loose.
      - Using historical distribution for each country over past 10–20 years, assign thresholds so well-known episodes (GFC, European sovereign debt crisis) map to tight; post-crisis accommodative periods to loose.
      - Thresholds set such that on average 30 percent of the time FCI is neutral, 40 percent loose, 30 percent tight. Numerical thresholds vary by country to match distribution.
    - Reweighting observations via IPW:
      - Because market expectation effects dominate, observations with strong growth and loose FCI are overrepresented; IPW reweights to avoid biased estimators, ensuring unbiased estimates of FCI impacts on output, unemployment, and inflation.
  - Note on control group: observations with high growth and tight FCI can define control group; unbiased impact is difference between average predicted outcome under reweighted treatment and control groups.

### Empirical Findings on Financial Conditions in Europe (Summary of Results)
- Historical movements and drivers:
  - 2003–2006: financial conditions loosened substantially.
  - Global Financial Crisis (GFC): conditions tightened rapidly driven by decline in credit availability and rise in cost.
  - European sovereign debt crisis: wider government bond spreads in euro area periphery and sovereign default concerns significantly impacted FCs; major central bank liquidity provision and government interventions later loosened conditions.
  - Pandemic onset: financial conditions loosened significantly due to exceptional policy support; price of risk spiked but policy/support led to loosened environment through end-2020; increase in government funding more than offset private deleveraging.
  - 2021–22: significant tightening in financial conditions beginning in 2021 and intensified with Russia’s war in Ukraine in early 2022; tightening drivers shifted to lower availability and higher cost of credit and higher price of risk (wider spreads and higher volatility); policy stance tightened as ECB and other central banks hiked rates.
- Recent dynamics:
  - Pace of tightening started to slow in late 2022 as lending rates and government/corporate yields began decelerating.
  - Price of risk tends to rise with inflation, but tighter monetary policy weakens inflation’s impact on price of risk: panel regression for Italy, Germany, Spain, France shows higher price of risk with higher inflation; the relationship is weakened with tighter policy stance.
    - Regression table excerpt (model summary):
      - Coefficients (standard errors in parentheses):
        - HICP Inflation: 0.114*** (0.0392) and 0.139** (0.0377)
        - D.Policy stance: 0.802*** (0.105)
        - HICP Inflation # D.Policy stance: -0.0972*** (0.0253)
      - Observations: 363
      - Adjusted R-squared: 0.90 and 0.92
      - Note: Country and time fixed effects included; contributions in first differences; 4 quarter lags of inflation included. * p<0.1, ** p<0.05, *** p<0.01.
- Cross-country and cross-sector divergence:
  - Cross-country dispersion in FCs increased from 2022; tightening cycle broad-based across Euro and non-Euro European economies but heterogeneity rose.
  - For some countries (Italy, Spain), policy stance driver estimated to have had larger impact on FCs during 2022–2023.
  - Cross-sector: both households (HH) and nonfinancial corporations (NFCs) faced tighter FCs but driven by different factors:
    - Households: tightening mainly driven by tighter policy stance and lower credit availability and higher costs (higher lending rates).
    - Corporates: price of risk (widening corporate spreads and rising volatility) more prominent.
- Near-term outlook and comparisons:
  - In near term, conditions likely remain tight, consistent with survey data (July 2023 euro area Bank Lending Survey indicated tighter credit standards and deteriorating bank funding).
  - Comparison with other FCIs:
    - Methodological differences produce differences in volatility and episodes captured (e.g., Bank of France’s FCI more volatile; Goldman Sachs Index weights corporate spreads heavily (almost 40 percent for euro area)).
    - IMF GFSR FCI (price and volatility variables, not quantities) exhibited sharper easing in late 2022.
    - The FCI in this paper is anchored to quarterly financial accounts and incorporates market and non-market indicators, aiming to identify longer-term financial cycles rather than short-term gyrations.
- Quantified macroeconomic effects:
  - Almost all euro area countries’ FCIs have been tightening with countries accounting for 91 percent of euro area GDP having entered a significantly tight regime.
  - Average estimated macro impacts (based on estimates across euro area and selected emerging European countries since 2000):
    - Switching to a tight regime could lower output by 2.2 percent and raise the unemployment rate by 0.3 percentage points over three years.
  - Normalization and magnitudes:
    - Indices normalized by means and standard deviations.
    - Cumulative tightening since mid-2021 is on average about 2.5 standard deviations (magnitude varies across countries).
  - Near-term drag:
    - Current tightening expected to put a drag on output through 2025, with an average impact of around three quarters of a percentage point during 2023.
  - Inflation impact:
    - Tightening FCIs can reduce inflation on average by 0.3 percentage points (text truncates here in source).

*Italic: Source — Box 1. Alternative Methodologies to Construct FCIs, wpiea2023209-print-pdf.*

### 0.7 percentage points after 3 years. Our estimates show

### wpiea2023209-print-pdf - 0.7 percentage points after 3 years. Our estimates show

### Policy implications and main conclusions
- The paper presents a novel Financial Conditions Index (FCI) designed to assess the costs, conditions, and availability of funds to the economy and investigates relationships between fluctuations in financial conditions and output, inflation, and unemployment.
- The FCI is computed for a broad set of European countries, including both Euro area and non-Euro area countries, and is broken down by individual countries and economic sectors.
- Key quantitative summary:
  - Sacrifice ratio between 3 and 4.
  - A one-standard deviation increase in a worldwide financial stress indicator is associated with a reduction in the level of output by 0.8 percent and an increase in the unemployment rate by 0.1 percentage point after a year (Hites et al., 2023).
  - After one year, a shift in regime reduces output by 0.9 percent and increases the unemployment rate by 0.06 percentage point (Hites et al., 2023).
  - Based on historical data from 2000 to 2023Q3, a shift from a neutral to a tight FCI regime (which happens around 30 percent of the time) is estimated to:
    - Lower real GDP by 2.2 percent over a three-year horizon.
    - Lower inflation by 0.7 percentage points over a three-year horizon.
    - Increase the unemployment rate by about 0.3 percentage points over a three-year horizon.
  - During 2021–22, more than 90 percent of the countries in the sample have entered a tight regime by historical standards.

### Findings relevant for policy design
- Divergence in estimated FCIs across countries highlights heterogeneous impacts of monetary policy tightening in Europe, and underscores benefits of considering macroeconomic stability when setting fiscal and macroprudential policies.
- Differences across Household (HH) and Non-Financial Corporate (NFC) financial conditions emphasize careful calibration of macroprudential policies by sector to mitigate risk of creating future fragilities.
- Tighter monetary policy initially leads to tighter FCs, but this tightening is transitory: as policies rein in inflation, the price of risk can decrease over time, potentially resulting in looser FCs.
- High government debt ratios could potentially limit available financing, which may result in more restrictive FCs.
- Possible research extensions: explore spillovers of the financial cycle, impact of a tightening financial cycle on vulnerabilities in the financial sector, their dynamics, and broader financial stability implications.

### Methodology — Partial Least Squares (PLS)
- Motivation:
  - PLS (Partial Least Squares Regression - PLSR) models covariance between two data sets, Y and X, and is well suited for many collinear variables, noisy data, and incomplete observations.
  - PLS constructs weights, scores, and loadings as linear combinations of the datasets to address collinearity that reduces OLS efficiency.
- Decomposition (as presented in the paper):
  - X = T P′_X + E_X
  - Y = U P′_Y + E_Y
  - T and U are scores (latent structure); P′_X and P′_Y are loadings; E_X and E_Y are error terms.
- Algorithm summary (procedural steps preserved in wording):
  - Compute loadings of X as P_X = X′T(T′T)−1
  - Compute loadings of Y as P_Y = Y′T(T′T)−1
  - Update Y score as U = Y P_Y/(P′_Y P_Y)−1
  - Repeat until convergence of Y scores (U1,...,Uk) where k is number of iterations
  - After convergence, orthogonal residual E_X = X − T P′_X
- Number of components:
  - The paper uses only the first PLS component to construct FCIs for interpretability; alternative FCIs can be produced by changing the supervising variable Y.
- Implementation note:
  - Main specification: target variable Y is the year-on-year growth rate of financial liabilities of the private sector (households and firms). The same growth rate is also considered for household, firm, government sectors. X matrix includes variables listed in Annex I. Variables are detrended and normalized to be stationary.
  - The Python implementation used for the paper is publicly available (link provided in source).

### Methodology — Inverse Probability Weighting (IPW) and causal estimation
- Framework:
  - Outcome variable y_t (e.g., GDP growth, inflation, unemployment) is related to D_t, a discretized version of changes in the FCI taking values d_L, d_C, d_T.
  - z_t is a vector of predetermined covariates relevant for predicting both D_t and y_t.
  - A multinomial logit is used for treatment assignment: Pr(D_t = d_j | z_t; j = {d_j, d_c}) = Λ(z_t γ_j + ε_t).
  - Propensity score: p̂(d_j | z_t) = Λ(z_t γ̂_j).
- Outcome equation estimated separately by treatment group at horizon h:
  - y_{t+h|j} = z_t Γ_j + e_{t j}
- Doubly-robust causal estimate (Regression Adjustment, Inverse Probability Weighting estimator):
  - θ_{h j} = E[ŷ_{t+h|j} ( 1{D_t = d_j} / p̂(z_t) − 1{D_t = d_0} / (1 − p̂(z_t)) )]
  - Intuition: The logit model explains likelihood of D_t = d_j conditional on z_t; more weight is given to ‘surprise’ observations where the predicted probability p̂(d_j | z_t) is low.
- Controls and outcomes:
  - Outcome variables: real GDP, year-on-year core inflation, unemployment rates.
  - Control variables: all lagged values, real effective exchange rate, output gap, government balances, current account balances as percent of GDP.
- Validation:
  - Success of IPW assessed by checking impulse response functions: after IPW adjustments, differences in GDP growth, unemployment and inflation across FCI regimes vanish in period 0, supporting comparability across treatment groups.

### Annex I — Drivers and representative variables used in FCIs (categories and examples)
- Credit availability and costs:
  - Interest Rates of Loans to HHs; Interest Rates of Mortgages to HHs; Interest Rates of Consumer Credit & Other Lending to HHs; Rates on Outstanding Loans to NFCs; Commodity Index; 10-Year Government Bond Yield; Housing Prices; Lending Conditions; Stock Price Index; Brent Crude Oil*; Euro Area 10-Year Yield Curve Spot Rate*; Stock Trading Volume.
- External conditions:
  - Bank Linkages Ratio; USD/EUR Exchange Rate; Germany: 2-Year Government Bond Yield; Germany: 2-Year Government Bond Yield Volatility; Germany: 10-Year Government Bond Yield; Nominal Effective Exchange Rate.
- Funding constraints:
  - LTV Ratio; Government Debt Service; Non-performing Loans to Total Gross Loans; Interest Margin to Gross Income; Return on Assets; Return on Equity; Household Debt to GDP; Regulatory Capital to Risk-Weighted Assets; MSCI Financials Index; Liquid Assets to Short-Term Liabilities; General Government Debt Outstanding*; Nonfinancial Corporations Debt Outstanding*; Financial Corporations Debt Outstanding*; Household Debt Outstanding*; GG interest expense/revenue ratio.
- Policy stance:
  - Households Deposit Rate; NFCs Deposit Rate; Policy Rate2/; Euro Area: Main Refinancing Rate3/; Euro Area: Euro Short-term Rate (€STR)3/; Euro Area: Shadow Short Rate Point Estimates3/; Wu-Xia Shadow ECB Rate3/; 2-Year Government Bond Yield; 1-Month Overnight Interest Rate Swap Close; Money Supply M1*; Money Supply M3*; 3-Month Yield Curve Spot Rate*; 2-Year Yield Curve Spot Rate*; Price of risk 5-Year CDS Premium; iBoxx EUR Non-Sovereigns BBB; iBoxx EUR Non-Sovereigns AAA; EUR Swap Annual 5-Year vs 6-Month; Germany: 10-Year Government Bond Yield Volatility; 2-Year Government Bond Yield Volatility; 10-Year Government Bond Yield Volatility; 2-Year Government Bond Spread; 10-Year Government Bond Spread; 3-Month Interbank Offer Rate; LIBOR-OIS Spread; Stock Price Volatility Index; 10-Year Interest Rate Swap; EURO STOXX 50 Volatility Index; EURO FTSE Volatility Index*; United States CBOE Volatility Index.
- Notes:
  - 1/ Data availability may differ among countries; daily and monthly data are converted to quarterly.
  - 2/ Used for non-EA countries.
  - 3/ Categorized as Policy stance for EA countries, External conditions for non-EA countries.
  - * Used for EA aggregate only.

*Source: wpiea2023209-print-pdf*

### Annex IV. Individual Country Results

### Annex IV. Individual Country Results

### Overview
- Presents Financial Conditions Index (FCI) results for the Euro Area and individual Euro Area countries.
- FCI is reported alongside "Liability growth (rhs, reversed)" in country-level panels.
- Component contributions shown: "Credit availability and costs", "External conditions", "Funding constraints", "Policy stance", "Price of risk".
- Sectoral decomposition presented by index for "Household", "Nonfinancial corporations", and "General government".
- Notes repeated across panels: "FCI is not scaled. 2023Q2 and 2023Q3 are forecasts." and "Sources: Author's calculations."

### Country-level FCI panels and comparisons
- Euro Area: FCI and Liability Growth; FCI Changes (contributions to quarter-over-quarter first difference); FCI level (contributions to unscaled FCI); sectoral FCI by index.
- Individual country panels (each with the same breakdowns as Euro Area) include:
  - Austria
  - Belgium
  - Croatia
  - Cyprus
  - Czech Republic
  - Estonia
  - Finland
  - France
  - Germany
  - Greece
  - Hungary
  - Ireland
  - Italy
  - Latvia
  - Lithuania
  - Luxembourg
  - Malta
  - Netherlands
  - Poland
  - Portugal
  - Slovakia
  - Slovenia
  - Spain

### Component contributions and FCI dynamics (method and presentation)
- For each country the report provides:
  - Time series panel of FCI and liability growth (Index; percent).
  - Quarter-over-quarter first-difference decomposition titled "FCI Changes (Contribution to quarter-over-quarter first difference of FCI)" with components: Credit availability and costs; External conditions; Funding constraints; Policy stance; Price of risk.
  - Contribution-to-level panel titled "FCI level (Contribution to unscaled FCI)" with the same components.
  - Sectoral breakdown panels:
    - "FCI Changes — Nonfinancial Corporations (Contribution to quarter-over-quarter first difference of FCI)"
    - "FCI Changes — Households (Contribution to quarter-over-quarter first difference of FCI)"
    - "FCI by Sectors (Index)" showing Household, Nonfinancial corporations, General government.
- Recurrent label "Tightening" appears on panels indicating directionality interpretation (tightening = tighter financial conditions).
- Forecast horizon indicated for panels: "Projection" and explicit note that "2023Q2 and 2023Q3 are forecasts."

### What the annex conveys (interpretation of provided figures and panels)
- The annex systematically documents FCI historical trajectories, recent changes, and projected short-run movements (2023Q2 and 2023Q3) for the Euro Area and listed countries.
- The decomposition panels identify the relative contributions of five channels—credit availability and costs, external conditions, funding constraints, policy stance, price of risk—to quarter-over-quarter FCI changes and to FCI levels.
- Sectoral panels allow assessment of differential FCI impacts across households, nonfinancial corporations, and general government for each jurisdiction.

*Source: Annex IV. Individual Country Results — IMF Working Papers, "Financial Conditions in Europe"; Sources: Author's calculations.*

### References

### References

### Literature on Financial Conditions Indexes and Financial Stress
- Adrian, T, F. Duarte, T. Iyer, 2023, “The Market Price of Risk and Macro-Financial Dynamics,” CEPR Press Discussion Paper No. 17777.
- Ahir, H., G. Dell’Ariccia, D. Furceri, C. Papageorgiou, H. Qi, “Financial Stress and Economic Activity—Evidence from a New Worldwide Index,” IMF Working Paper, forthcoming.
- Aikman, D., J. Bridges, S. Burgess, R. Galletly, I. Levina, C. O’Neill, and A. Varadi, 2018, “Measuring Risks to UK Financial Stability,” Staff Working Paper No. 738 (London: Bank of England).
- Angelopoulou, E., H. Balfoussia, and H. D. Gibson, 2012, “Building a Financial Conditions Index for the Euro Area and Selected Euro Area Countries: What Does It Tell Us about the Crisis?” Working Paper Series No. 1541 (Frankfurt: European Central Bank).
- Arregui, Nicolas, Selim Elekdag, Gaston Gelos, Romain Lafarguette, and Dulani Seneviratne, 2018, “Can Countries Manage Their Financial Conditions Amid Globalization?” IMF WP 2018/015.
- Braves, S., and D. L. Kelly, 2017, “Introducing the Chicago Fed’s New Adjusted National Financial Conditions Index,” Chicago Fed Letter No. 386, (Chicago: The Federal Reserve Bank of Chicago).
- Gauthier, C., C. Graham, and Y. Liu, 2004, “Financial Conditions Indexes for Canada,” Staff Working Papers 2004–22, (Ottawa, Bank of Canada)
- Gómez, E., A. Murcia, and N. Zamudio, 2011, “Financial Conditions Index: Early and Leading Indicator for Colombia,” Ensayos sobre Política Económica, Vol. 29, No. 66, pp. 174–221 (Bogotá: Banco de la República).
- Hatzius, J., P. Hooper, F. S. Mishkin, K. L. Schoenholtz, and M. W. Watson, 2010, “Financial Conditions Indexes: A Fresh Look after the Financial Crisis,” NBER Working Paper No. 16150, Cambridge, MA: National Bureau of Economic Research.
- Hatzius, J., S. Jari Stehn, 2018, “The Case for a Financial Condition Index”, Goldman Sachs Economics Research.
- Kapetanios, G., S. Price, and G. Young, 2018, “A UK Financial Conditions Index Using Targeted Data Reduction: Forecasting and Structural Identification,” Econometrics and Statistics, Vol 7, pp. 1–17 (Philadelphia, PA: Elsevier B.V.)
- Khudrakpam, J. K., R. Kavediya, and J. M. Anthony, 2017, “Estimating Financial Conditions Index for India,” Journal of Emerging Market Finance, Vol. 16, pp. 61–89 (Chennai: Institute for Financial Management and Research).
- Klein, N., N. Gumata, and E. Ndou, 2012, “A Financial Conditions Index for South Africa,” IMF Working Papers 12/196 (Washington: International Monetary Fund).
- Koop, G., and D. Korobilis, 2014, “A New Index of Financial Conditions,” European Economic Review, Vol. 71, pp. 101–16 (Philadelphia, PA: Elsevier B.V.).
- Nicoletti, G., K. M. Wacker, and D. Lodge, 2014, “Measuring Financial Conditions in Major Non-Euro Area Economies,” Working Paper Series No. 1743 (Frankfurt: European Central Bank).
- Petronevich, A., and J. G. Sahuc, 2019, “A New Banque de France Financial Conditions Index for the Euro Area,” Banque de France Bulletin, Vol. 223, Article 1.
- Swiston, M. A. J., 2008, “A U.S. Financial Conditions Index: Putting Credit Where Credit Is Due,” IMF Working Papers 08/161 (Washington: International Monetary Fund).
- Zheng, G. and W. Yu, 2014, “Financial Conditions Index’s Construction and its Application on Financial Monitoring and Economic Forecasting,” Procedia Computer Science, Vol. 31, pp. 32–39 (Philadelphia, PA: Elsevier B.V.).

### Empirical Macroeconomics, Monetary Policy, and Macrofincial Dynamics
- Acemoglu, Daron, Suresh Naidu, Pascual Restrepo, and James A. Robinson, 2019, “Democracy Does Cause Growth,” Journal of Political Economy, Vol. 127, Issue 1, pp. 47–100 (Chicago: University of Chicago Press).
- Blanchard, Oliver, Christopher Erceg and Jesper Linde, 2016, “Jump-starting the Euro Area Recovery: Would a Rise in Core Fiscal Spending Help the Periphery?” NBER Macroeconomics Annual 2016
- Cardarelli, Roberto, Selim Elekdag, and Subir Lall, 2011. “Financial Stress and Economic Downturns,” Journal of Financial Stability, 7(2): 78-97
- Christiano, J. Lawrence, Martin Eichenbaum, and Charles L. Evans, 2005, Nominal Rigidities and the Dynamic Effects of a Shock to Monetary Policy,” Journal of Political Economy Vol. 113, No. 1
- Coibion, Oliver, 2012, “Are the Effects of Monetary Policy Shocks Big or Small?” American Economic Journal: Macroeconomics Vol 4, No. 2, April 2012.
- Crump. R. K., S. Eusepi, D Giannone, E. Qian, and A. M. Sbordone, 2021, “A Large Bayesian VAR of the United States Economy,” Staff Report No. 976 (New York: Federal Reserve Bank of New York).
- Gertler, Mark and Peter Karadi, 2015, “Monetary Surprises, Credit Costs, and Economic Activity,” American Economic Journal: Macroeconomics, 7(1):44-76
- Goodhard, C. A. E., and B. Hofmann, 2001, Asset Prices, Financial Conditions and The Transmission of Monetary Policy, Proceedings, March (San Francisco: Federal Reserve Bank of San Francisco).
- International Monetary Fund, 2017, “Financial Conditions and Growth at Risk.” Global Financial Stability Report. Chapter 3: October, Washington, DC.
- International Monetary Fund, 2021, “Global Financial Stability Report, Online Appendix 2.1 (Washington).
- Romer, Christina and David Romer, 2014, “A New Measure of Monetary Shocks: Derivation and Implications,” American Economic Review.
- Stock, J. H., and M. W. Watson, 2001, “Vector Autoregressions,” Journal of Economic Perspectives, Vol.15, No. 4, pp. 101–15 (Pittsburgh: American Economic Association).

### Econometric Methods, Program Evaluation, and Statistical Techniques
- Angrist, Joshua, 1998, “Estimating the Labor Market Impact of Voluntary Military Service Using Social Security Data on Military Applicants,” Econometrica, Vol. 66, Issue 2, pp. 249–288 (Bethesda, MD: Econometrica, Inc.).
- Angrist, Joshua D., Òscar Jordà, and Guido M. Kuersteiner, 2018, “Semiparametric Estimates of Monetary Policy Effects: String Theory Revisited,” Journal of Business and Economic Statistics, Vol. 36, Issue 3, pp. 371–387 (Alexandria: American Statistical Association).
- Ashenfelter, Orley, 1978, “Estimating the Effect of Training Programs on Earnings,” Review of Economics and Statistics, Vol. 60, pp. 47–57 (Cambridge: MIT Press).
- Card, David., and Daniel Sullivan, 1988, “Measuring the Effect of Subsidized Training Programs on - Movements In and Out of Employment,” Econometrica, Vol. 56, Issue 3, pp. 497-530 (Bethesda, MD: Econometrica, Inc.).
- Delaigle, A., and P. Hall, 2012, “Methodology and Theory for Partial Least Squares Applied to Functional Data,” The Annals of Statistics, Vol. 40, No. 1, pp. 322–352.
- Duo, Q., and Q. Wang, 2016, “Predictive Macro-Impacts of PLS-based Financial Conditions Indices: An Application to the USA,” School of Oriental and African Studies,Department of Economics Working Paper Series No. 201(London, University of London).
- Heckman, James J., Hidehiko Ichimura, and Petra E. Todd, 1998, “Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Program,” Review of Economic Studies, Vol. 64, Issue 4, pp. 605–654 (Cambridge: MIT Press).
- Hirano, Keisuke, Guido W. Imbens, and Geert Ridder, 2003, “Efficient Estimation of Average Treatment Effects Using the Estimated Propensity Score,” Econometrica, Vol. 71, Issue 4, pp. 1161–1189, (Bethesda, MD: Econometrica, Inc.).
- Imbens, Guido W, 2004, “Nonparametric Estimation of Average Treatment Effects under Exogeneity: A Review,” Review of Economics and Statistics, Vol. 86, Issue 1, pp. 4–29 (Cambridge: MIT Press).
- Imbens, Guido W., and Jeffrey M. Wooldridge, 2009, “Recent Developments in the Econometrics of Program Evaluation,” Journal of Economic Literature, Vol. 47, Issue 1, pp.5–86.
- Jordà, Òscar and Alan M. Taylor, 2016, “The Time for Austerity: Estimating the Average Treatment Effect of Fiscal Policy,” Economic Journal, Vol. 126, Issue 590, pp. 219–255 (Oxford: Oxford University Press).
- LaLonde, Robert J., 1986, “Evaluating the Econometric Evaluations of Training Programs with Experimental Data,” American Economic Review, Vol. 76, pp. 604–620 (Nashville: American Economic Association).
- Wold, S., M. Sjostrom, and L. Eriksson, 2001, “PLS-regression: a Basic Tool of Chemometrics,” Chemometrics and Intelligent Laboratory Systems, Vol. 58, Issue 2, pp. 109–30 (Philadelphia, PA: Elsevier B.V.).
- Wooldridge, Jeffrey M, 2010, Econometric Analysis of Cross Section and Panel Data. (Cambridge: MIT Press).

### Country- and Region-specific Studies and Other Topics
- Arena, M., Di Bella G., Cuevas A., Gracia B., Nguyen, V. and Pienkowski, A., 2020, “It is Only Natural: Europe’s Low Interest Rates” IMF Working Paper, 2020/116
- Batini, N. and K. Turnbull, 2002, “A Dynamic Monetary Conditions Index for the UK,” Journal of Policy Modeling, Vol 24, Issue 3, pp. 257–81 (Philadelphia, PA: Elsevier B.V.).
- Caselli, Francesca, and Philippe Wingender, 2021, “Heterogeneous Effects of Fiscal Rules: The Maastricht Fiscal Criterion and the Counterfactual Distribution of Government Deficits,” European Economic Review, Vol. 136, 103748 (Philadelphia: Elsevier B.V.).
- Duprey, T. and B. Klaus, 2015, "Dating systemic financial stress episodes in the EU countries", ECB Working Paper Series, No 1873.
- Kapetanios, G., S. Price, and G. Young, 2018, “A UK Financial Conditions Index Using Targeted Data Reduction: Forecasting and Structural Identification,” Econometrics and Statistics, Vol 7, pp. 1–17 (Philadelphia, PA: Elsevier B.V.)
- Mayes, D. G., and M. Viren, 2001, “Financial Conditions Indexes,” Bank of Finland Discussion Papers No. 17/2001 (Helsinki: Bank of Finland).
- Petronevich, A., and J. G. Sahuc, 2019, “A New Banque de France Financial Conditions Index for the Euro Area,” Banque de France Bulletin, Vol. 223, Article 1.
- Zheng, G. and W. Yu, 2014, “Financial Conditions Index’s Construction and its Application on Financial Monitoring and Economic Forecasting,” Procedia Computer Science, Vol. 31, pp. 32–39 (Philadelphia, PA: Elsevier B.V.).

*Financial Conditions in Europe: Dynamics, Drivers, and Macroeconomic Implications Working Paper No. WP/2023/209*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023209-print-pdf.pdf_