## _wp1336

## Source details

**Canonical URL:** [_wp1336](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1336.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1336.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1336.pdf.json)

---

### I. Introduction — research question and main empirical findings
- Research question:
  - How do efforts to reduce public debt levels affect economic activity, specifically the short-term cost of corporate credit (syndicated loan spreads)?
- Framing:
  - Keynesian view: cutting spending or raising taxes reduces aggregate demand and economic activity in the short run.
  - Expectation/credibility view: fiscal consolidations can be expansionary in the short run by reducing sovereign default risk and restoring credibility, lowering interest rates and creating wealth effects.
  - This paper examines whether and how banks price in the potential effects of fiscal consolidation when lending, focusing on near-contemporaneous/short-term relations between fiscal consolidation and loan spreads.
- Summary of main empirical findings:
  - Fiscal consolidations are positively associated with syndicated loan spreads on average.
  - Loan spreads are higher for small firms, domestic firms, and firms without access to international markets through ADRs.
  - The adverse relation between spreads and fiscal consolidation is mitigated substantially when consolidations are large.
  - Consolidations accompanied by accommodating monetary policy, flexible exchange rates, currency devaluations, and stable governments show no short-run adverse effect on corporate borrowing costs.
  - Corporate tax hikes are major drivers of higher spreads; social benefit cuts and VAT hikes are associated with lower spreads.
  - Sectoral analysis: tax hikes increase spreads in directly affected sectors; spending cuts do not show significant adverse effects for related sectors.
- Policy implication highlights:
  - Long-term gains of debt reduction may come with short-term pain in corporate financing conditions, borne disproportionately by small firms, domestic firms, and firms with limited financing sources.
  - Large consolidations that convey credibility and consolidations implemented with accommodating macroeconomic policy and stable governments mitigate adverse effects on spreads.

### II. Channels linking fiscal consolidation and corporate credit costs
- Demand-side (Keynesian) channel:
  - Spending cuts or tax increases reduce aggregate demand, lower corporate profitability, and raise borrowing costs, particularly for “marginal” borrowers facing asymmetric information and limited access to finance.
- Supply-side (credibility) channel:
  - Large fiscal consolidations reduce sovereign default risk premium, mitigate crowding-out, increase available funds for private debt issuance, and lower loan spreads.
- Expectations channel:
  - Large and credible fiscal adjustments raise expectations of successful adjustment, potentially lowering loan spreads.
- Wealth effects:
  - If spending cuts are permanent, forward-looking agents may anticipate decreased future tax burdens and adjust consumption/investment accordingly.
- Composition matters:
  - Direct tax increases or cuts in politically sensitive areas may signal stronger credibility; cuts in public investment or indirect-tax–focused packages may be perceived as weak.

### III. Data, sample, and measurement
- Data sources:
  - Syndicated loan characteristics: Dealogic.
  - Firm financials: Worldscope.
  - Macro-level data: OECD, IMF, World Bank.
  - Sovereign ratings: Moody’s.
  - Government debt: Reinhart and Rogoff (2011).
  - Country risk indices: PRS Group’s ICRG.
- Sample and coverage:
  - Final dataset: 4,967 observations covering 16 countries over the period 1990–2011.
  - 16 advanced economies: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, United Kingdom.
  - Total consolidation programs considered between 1990 and 2011: 116 consolidation programs.
  - Number of large consolidations (consolidations exceeding 1.5 percent of GDP): 30.
- Fiscal consolidation measure:
  - Action-based measure recording announced and implemented government “actions” aimed at reducing public debt (expressed in percent of GDP at the time of implementation), following Devries et al. (2011) and the Romer and Romer (2010) logic.
  - Rationale: treats actions-based measure as exogenous and avoids biases of cyclically-adjusted primary balance (CAPB).
- Key variable definitions (selected):
  - LogSpread: natural logarithm of syndicated loan spread, inclusive of all fees (spread in basis points relative to benchmark).
  - FiscalConsolidation: fiscal consolidation measures announced and implemented aimed to reduce public debt, in percent of GDP.
  - FiscalConsolidation > 1.5% GDP: dummy = 1 if FiscalConsolidation exceeds 1.5 percent of GDP.
  - TaxHikes and SpendingCuts: components of consolidation, in percent of GDP.
- Controls included:
  - Loan-level: loan maturity, log loan size, loan purpose, loan currency, loan type, InvestmentGrade, DomesticBank.
  - Firm-level: Leverage, Profitability, Tangibility, GrowthOpportunities, RatedBorrower.
  - Country-level: SovereignRating (Moody’s converted Aaa = 21, etc.), SovereignDebt (percent of GDP), ExchangeRate change, InterestRate (level and change), LogRealGDP, RealGDPGrowth, StockMarketTurnover, TradeOpenness, Inflation, GovernmentStability, Law&Order, LackofCorruption, GovernmentSize, InstitutionalInvestorCreditRating.

### IV. Descriptive statistics and notable sample moments
- Sample size and aggregation:
  - Observations: 4,967; Observations with Fiscal Consolidation: 1,528.
- Fiscal consolidation statistics:
  - FiscalConsolidation: mean 0.30, Std.Dev. 0.60, Min 0.0, Max 4.5.
  - Average fiscal consolidation: around 0.3 percent of GDP.
  - Maximum fiscal consolidation observed: 4.5 (Italy in 1993).
  - Note: Ireland’s 4.74 percent of GDP in 2009 noted but dropped in regressions due to data gaps.
- Tax and spending components:
  - TaxHikes: mean 0.10, Std.Dev. 0.30, Min 0.0, Max 2.4.
  - SpendingCuts: mean 0.20, Std.Dev. 0.40, Min 0.0, Max 3.7.
  - Average spending cuts: 0.2 percent of GDP; average tax hikes: 0.1 percent of GDP.
- Loan and firm summary:
  - Average loan spread: around 134 basis points (Spread mean 133.6, Std.Dev. 136.4).
  - LogSpread: mean 4.41, Std.Dev. 1.00.
  - Investment-grade loans: around 70 percent.
  - Loans in major currencies: around 80 percent.
  - Borrowers with a credit rating: around 30 percent.
  - Loans extended by domestic banks: around 30 percent.
  - Average loan maturity: 4 years.
  - Average loan size: around US$ 790 million (DealValue mean 790.5; Std.Dev. 2,005.2; Min 0.2; Max 45,000.0).
  - Largest observed loan: US$ 45 billion (InBev to buyout Anheuser-Busch in 2008) — maturity around 3 years and spread of 200 basis points above LIBOR, inclusive of all fees.
- Country-level moments:
  - SovereignRating: mean 20.5, Std.Dev. 0.8, Min 14.0, Max 21.0.
  - SovereignDebt: mean 67.03, Std.Dev. 37.44, Min 4.5, Max 233.1.
  - ExchangeRate, change: mean 1.9, Std.Dev. 8.1, Min -25.3, Max 20.2.
  - InterestRate, level: mean 4.0, Std.Dev. 2.6, Min 0.0, Max 15.2.
  - TradeOpenness: mean 64.7, Std.Dev. 32.6.
  - Inflation: mean 2.0, Std.Dev. 1.4.
  - Law&Order: mean 91.4, Std.Dev. 9.8.

### V. Baseline empirical results
- Main baseline result:
  - Syndicated loan spreads are positively associated with fiscal consolidations on average, indicating lenders price in recessionary effects of fiscal consolidations more than anticipated improvements in future economic performance or reduced sovereign default risk.
- Key quantitative magnitudes (Table 3, Column 3 evaluated at average spread):
  - An additional 1 percent of GDP in fiscal consolidation is associated with an 11 percent increase in spreads, corresponding to around 15 basis points.
  - For a typical loan (average loan size around US$ 800 million and average term of 4 years), this corresponds to around US$ 5 million in additional borrowing costs over the life of the loan.
  - If the consolidation is large (> 1.5 percent of GDP), the increase in spreads is only 2 basis points; additional financing costs are negligible when consolidations are large.
- Selected baseline regression coefficients (Table 3, selected columns):
  - FiscalConsolidation: Column (2) 0.1080*** [0.0322]; Column (4) 0.1200*** [0.0297].
  - FiscalConsolidation > 1.5% GDP: Column (1) -0.0933** [0.0378]; Column (2) -0.1160*** [0.0341].
  - InterestRate, change: coefficients around -0.0005** to -0.0006***.
  - Law&Order: significant negative coefficients (e.g., -0.0056*** [0.0019]).
  - InvestmentGrade: approx. -1.2010*** [0.0560].
  - LogDealValue: approx. -0.0879*** [0.0132].
  - Size: approx. -0.0380*** [0.0105].
  - Leverage: approx. 0.3850*** [0.0967].
  - GrowthOpportunities: approx. -0.0549*** [0.0148].
  - Tangibility: approx. -0.1330** [0.0453].
- Robustness:
  - Results robust to dropping one country at a time.

### VI. Firm and loan heterogeneity — who bears the short-run cost?
- Small vs large firms (Table 4, Panel A):
  - For small firms:
    - FiscalConsolidation coefficient: 0.0597** [0.0268].
    - For a 1 percent of GDP consolidation: 17 percent increase in spreads for small firms → average spread 132 basis points ⇒ 23 basis points increase.
    - Average deal value for small-firm sample: around US$ 600 million → this increase ≈ US$ 1.4 million additional financing costs in a year.
    - If consolidation is large (> 1.5 percent of GDP), increase in spreads is 4 percent ⇒ 6 basis points.
  - For large firms: adverse effect not observed or substantially smaller.
- Firms without bond ratings and without ADRs (Panels B and C):
  - Firms without ratings:
    - Average spread 135 basis points; 12 percent ⇒ 16 basis points for a 1 percent consolidation.
    - Large consolidation → less than 8 percent ⇒ 2 basis points.
  - Firms without ADRs:
    - Average spread 139 basis points; 14 percent ⇒ 19 basis points.
    - Large consolidation → less than 8 percent ⇒ 7 basis points.
- External finance dependence (Panel D):
  - Firms that do not use external markets often:
    - Average spread 137 basis points; 19 percent ⇒ 25 basis points for 1 percent consolidation.
    - Large consolidation → less than 8 percent ⇒ 11 basis points.
- Domestic vs multinational firms (Panel E):
  - Domestic firms (no foreign sales and no foreign income):
    - Average spread 135 basis points; 17 percent ⇒ 23 basis points for 1 percent consolidation.
    - Large consolidation → less than 8 percent ⇒ 1 basis point.
  - Multinational firms: generally insulated from consolidation-related changes in borrowing costs.
- Overall implication:
  - Short-run burden of fiscal consolidations falls disproportionately on small, domestic, and financing-constrained firms unless consolidations are large and credible.

### VII. Country conditions and interactions (Table 5)
- Sovereign debt levels:
  - When sovereign debt is low: fiscal consolidations not relevant for loan spreads.
  - When sovereign debt is high: loan spreads increase with consolidations unless consolidations are large.
- Government size:
  - Consolidation positively associated with spreads for both large and small governments; offsetting effect of large consolidations present only when government size is large.
- Monetary policy stance (Taylor-rule residuals; Columns 5–6):
  - Under loose/accommodative monetary policy and large consolidations, the overall effect of large consolidations is negative (sum of coefficients = -0.126 in Column 6).
  - Under tight monetary policy, the sum of coefficients is 0.003 in Column 5.
  - Interpretation: accommodative monetary policy helps shield the economy from short-term pain of large consolidations.
- Exchange rate regime and currency depreciation:
  - Flexible exchange-rate regimes: fiscal consolidations do not have a significant relation with cost of credit.
  - Fixed regimes: fiscal consolidations positively associated with spreads unless consolidations are large enough to reverse adverse effects.
  - Countries that allow currency depreciation: fiscal consolidations do not have a significant relationship with corporate credit cost.
- Political stability:
  - Large consolidations reduce cost of credit when lenders believe the government is stable (sum of coefficients negative and significant for stable governments; near zero for unstable governments).
- Overall message:
  - Large consolidations accompanied by flexible macroeconomic policy frameworks and political stability shield corporations (especially small and financing-constrained firms) from adverse effects on cost of credit.

### VIII. Composition of consolidation packages (Tables 6–7)
- Aggregate measures:
  - SpendingCuts and TaxHikes (as percent of GDP): neither is statistically significant on aggregate (coefficients positive but not significant in some specifications).
- Specific tax and spending measures:
  - CorporateTax hikes:
    - Large and highly significant positive coefficient (e.g., 0.5450*** [0.1060]; 0.5660*** [0.1390]) — corporate tax hikes are major driver of the positive association between loan spreads and consolidations.
  - ValueAddedTax (VAT) hikes:
    - Negative association with loan spreads (e.g., -0.2030*** [0.0542]; -0.1530** [0.0676]).
  - SocialBenefitsCut:
    - Negative association with loan spreads (e.g., -0.1200** [0.0545]; -0.1010* [0.0598]).
  - WageCut and PensionCut:
    - Some negative or mixed coefficients (e.g., WageCut -0.1280** [0.0628]; PensionCut 0.1540* [0.0926]).
- Sector-specific interactions (Table 7):
  - Tax hikes raise spreads in sectors directly affected:
    - ConsumptionTax * RetailSector: 0.2300*** [0.0714].
    - BankTax * BankingSector: 0.2710** [0.0946].
    - HealthcareTax * HealthcareSector: 0.7460*** [0.1980].
  - Spending cuts in a sector do not show similar adverse effects on loan spreads for that sector.
- Interpretation:
  - Corporate tax hikes appear particularly contractionary for corporate credit costs; VAT hikes and social-benefits cuts are associated with lower spreads, consistent with credibility or profitability channels.

### IX. Robustness checks and alternative specifications (Tables 8–9)
- Sample-selection concerns (Heckman selection, Table 8):
  - First-stage instrument: external finance dependence (Rajan and Zingales).
  - FiscalConsolidation coefficients in Heckman second-step close to baseline and significant (e.g., Column (3): 0.1060*** [0.0325]).
  - Lambda (inverse Mill’s ratio) not significant → evidence not driven by sample selection bias.
- Alternative samples:
  - Firms that issued at least one loan: coefficients on fiscal consolidation and large consolidation comparable to baseline.
- Alternative specifications (Table 9):
  - Lagged fiscal consolidation and lagged real GDP growth included; results similar (e.g., RealGDPGrowth, lagged: -0.0366** [0.0148]).
  - Replacement of Moody’s rating with Institutional Investor rating: fiscal consolidation coefficients remain close to baseline.
  - Alternative large-consolidation definitions (e.g., > 1 percent of GDP in two consecutive years as additional criterion): positive and significant fiscal consolidation effect persists; large-consolidation variable comparable in sign but may lose statistical significance at conventional levels.
- CAPB-based measures:
  - CAPB-based consolidation measures show no significant relation with spreads (potential endogeneity).
  - IV approach instrumenting CAPB-based measure with action-based measure yields a positive association between spreads and fiscal consolidations unless consolidations are large.
- Overall robustness conclusion:
  - Core result — fiscal consolidations raise syndicated loan spreads on average, with mitigation for large consolidations and interactions with macro framework and political stability — is robust to multiple specifications and checks.

### X. Policy implications and conclusions
- Short-run trade-offs:
  - Long-term benefits of debt reduction may be accompanied by short-term tightening of corporate financing conditions.
  - The short-term burden falls disproportionately on small firms, domestic firms, and firms with limited financing channels.
- Design considerations for fiscal consolidation:
  - Ensure consolidations are large enough and credible to convey improvement in economic outlook and sovereign solvency, thereby mitigating adverse credit effects.
  - Complement consolidation packages with safeguards targeted at the most vulnerable firms if credit conditions are tight.
  - Implement consolidations alongside accommodative macroeconomic policies (monetary accommodation, exchange-rate flexibility) and under stable governments to avoid unfavorable developments in corporate credit markets.
- Overall interpretation:
  - Lenders price in potential short-term recessionary effects of fiscal consolidations, especially when consolidations appear noncredible; expectations and credibility channels operate to lower borrowing costs when consolidations are large and supported by accommodating policies and stable governments.

*Source: Excerpt from IMF working paper _wp1336 (introduction through methodology, results, and appendix tables; dataset: 4,967 observations across 16 countries over 1990–2011).*

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

### _wp1336 - References

### I. INTRODUCTION
- Legacy of the global financial crisis of 2007-08: increase in government debt levels in many advanced economies; debt-to-GDP ratio reached a historical peak as governments used public funds to bail out financial systems and provide stimuli.
- Fiscal consolidation is a major policy issue to bring debt levels back to sustainable levels.
- Key research question: How do efforts to reduce public debt levels affect economic activity, specifically the short-term cost of corporate credit?
- Competing views:
  - Keynesian view: cutting spending or raising taxes reduces aggregate demand and economic activity in the short run.
  - Expectation and credibility view: fiscal consolidations can be expansionary in the short run by reducing sovereign default risk and restoring credibility, lowering interest rates and creating wealth effects.
- This paper’s focus: whether and how banks price in the potential effects of fiscal consolidation when lending; emphasis on the near contemporaneous or short-term relationship between fiscal consolidation and loan spreads.
- Summary of main empirical findings (detailed later):
  - Fiscal consolidations are positively associated with spreads in the syndicated loan market.
  - Loan spreads are higher for small firms, domestic firms, and firms without access to international markets through ADRs.
  - The adverse relation between spreads and fiscal consolidation is mitigated substantially when consolidations are large.
  - Consolidations accompanied by accommodating monetary policy, flexible exchange rates, currency devaluations, and stable governments do not suffer short-run adverse effects on corporate credit costs.
  - Corporate tax hikes are major drivers of higher spreads; social benefit cuts and VAT hikes are associated with lower spreads.
  - Sectoral analysis: tax hikes increase spreads in directly affected sectors (e.g., bank tax hikes -> banking sector; health care tax hikes -> health care sector); spending cuts do not show significant adverse effects for related sectors.
- Policy implication highlights:
  - Long-term gains of debt reduction may come with short-term pain in corporate financing conditions, borne disproportionately by small firms, domestic firms, and firms with limited financing sources.
  - Policymakers should carefully design fiscal consolidation measures and consider safeguards for the most vulnerable, especially if credit conditions are already tight.
  - Large consolidations that convey credibility and consolidations implemented with accommodating macroeconomic policy and stable governments mitigate adverse effects on spreads.
- Positioning relative to literature:
  - Existing studies using aggregate corporate bond data show mixed results (e.g., Ardagna (2009) negative relationship; Laubach (2009) insignificant relationship).
  - This study is the first, to the authors’ knowledge, to use syndicated loan micro-level data matched with firm and loan characteristics and measures of fiscal consolidation based on government actions announced and implemented to reduce public debt (Romer and Romer (2010) approach).

### II. BACKGROUND ON THE POTENTIAL CHANNELS OF TRANSMISSION
- Channels linking fiscal consolidation to the cost of credit:
  - Demand-side (Keynesian): spending cuts or tax increases reduce aggregate demand, lower corporate profitability, and raise borrowing costs, particularly for “marginal” borrowers subject to asymmetric information and limited access to finance.
  - Supply-side (credibility channel): large fiscal consolidations reduce sovereign default risk premium, mitigate crowding-out, increase available funds for private debt issuance, and lower loan spreads.
  - Expectations channel: large and credible fiscal adjustments raise expectations of successful adjustment, potentially leading to expansionary effects and lower loan spreads.
  - Wealth effects channel (discussed in literature): if spending cuts are permanent, forward-looking agents anticipate decreased future tax burdens and could respond via consumption/investment changes (reference to Alesina and Ardagna, 1998).
- Importance of consolidation composition:
  - Measures involving direct tax increases or cuts in politically-sensitive areas (wages, welfare) may signal stronger credibility.
  - Cuts in less politically sensitive areas (public investment) or revenue measures focusing on indirect taxes may be perceived as weak or prone to creative accounting, reducing credibility.
- Net effect ambiguity: consolidation can reduce default risk premium (lower spreads) or, if not credible or growth-harming, increase borrowing costs (higher spreads).

### III. DATA AND METHODOLOGY
- Data sources:
  - Syndicated loan characteristics: Dealogic.
  - Firm-level financial and accounting data: Worldscope; loan-level data matched with firm-level data manually.
  - Macro-level data: OECD, IMF, World Bank.
  - Sovereign ratings: Moody’s.
  - Government debt data: Reinhart and Rogoff (2011).
  - Country risk indices (government stability, law and order, corruption): PRS Group’s International Country Risk Guide (ICRG).
- Sample and coverage:
  - Final dataset: 4,967 observations covering 16 countries over the period 1990-2011.
- Measurement approach:
  - Fiscal consolidation measures based on government actions announced and implemented to reduce public debt levels (sovereign-debt–reduction-driven measures), following Romer and Romer (2010) logic.
  - Controls include the level and change in policy interest rates when examining relationships between fiscal consolidation and corporate loan spreads.
- Empirical strategy:
  - Loan-level panel dataset exploiting firm and loan characteristics to identify heterogeneous effects across firm size, domestic vs. international orientation, and sectoral exposure to specific tax or spending measures.
  - Examination of country- and policy-framework interactions (monetary stance, exchange rate flexibility, devaluations, government stability) to assess conditional effects of consolidations.

### IV. KEY RESULTS AND INTERPRETATIONS
- Aggregate and heterogeneous effects:
  - Fiscal consolidations are associated with higher syndicated loan spreads in the short run.
  - Adverse effects concentrated among:
    - Small firms.
    - Domestic firms.
    - Firms with limited access to alternative funding (e.g., without ADR issuance).
  - Large consolidations mitigate or reverse adverse short-run effects on spreads, consistent with credibility and expectations channels.
- Role of policy context:
  - Consolidations implemented with flexible/exogenous macro frameworks (accommodating monetary policy, flexible exchange rates, currency devaluations) and by stable governments show no short-run adverse effect on corporate borrowing costs.
- Composition of consolidation measures:
  - Corporate tax hikes: associated with higher loan spreads (interpreted as lenders pricing in decreased corporate profitability).
  - Social benefit cuts: associated with lower loan spreads.
  - VAT hikes: associated with lower loan spreads.
  - Sectoral specificity: tax hikes raise spreads in directly affected sectors (e.g., bank tax hikes -> banking sector; health care tax hikes -> health care sector); spending cuts examined (transportation, health care, defense) show no significant adverse effect on related sectors’ loan spreads.
- Interpretation and robustness:
  - Findings support a view that lenders price in expected short-term growth and profitability effects and sovereign-creditworthiness effects when extending syndicated loans.
  - Results are consistent with asymmetric-information and procyclicality of lending standards: firms more likely to be credit-constrained face larger spread increases.
  - Because the sample is composed of syndicated loans, average effects for the broader population of firms (typically smaller and more constrained) may be larger than those reported.

### V. POLICY IMPLICATIONS
- Short-run trade-offs:
  - Long-term benefits of debt reduction may be accompanied by short-term tightening of corporate financing conditions.
  - The short-term burden falls disproportionately on small firms, domestic firms, and firms with limited financing channels.
- Design considerations for fiscal consolidation:
  - Ensure consolidations are large enough and credible to convey improvement in economic outlook and sovereign solvency to mitigate adverse credit effects.
  - Complement consolidation packages with safeguards targeted at the most vulnerable firms if credit conditions are tight.
  - Implement consolidations alongside accommodative macroeconomic policies (monetary accommodation, exchange-rate flexibility) and under stable governments to avoid unfavorable developments in corporate credit markets.

*Source: Excerpt from IMF working paper (introduction through methodology and background) covering fiscal consolidation impacts on corporate loan spreads; final dataset: 4,967 observations across 16 countries over 1990-2011.*

### Appendix Table 1. The number of observations for each country and the number of observations

### Appendix Table 1. The number of observations for each country and the number of observations

### Fiscal consolidation measure — Methodology
- Fiscal consolidation measures follow Devries et al. (2011), building on Ramey and Shapiro (1998), Ramey (2011), and Romer and Romer (2010).
- The measure records announced and implemented “actions” aimed at reducing public debt rather than actual budgetary outcomes.
- Measures include spending cuts or tax hikes and are expressed in percent of GDP at the time of implementation.
- Rationale for actions-based measure:
  - Avoids biases of cyclically-adjusted primary balance (CAPB) which can misattribute consolidations to growth shocks or omit discretionary consolidations during recessions.
  - The resulting measure of fiscal consolidation is treated as exogenous (Romer and Romer, 2010).

### Consolidation episodes
- Sample: 16 advanced economies — Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, United Kingdom.
- Total consolidation programs considered between 1990 and 2011: 116 consolidation programs.
- Consolidations exceeding 1.5 percent of GDP are categorized as large (shown in bold in Table 1 in the source). Number of large consolidations: 30.
- Temporal patterns: the 1990s and 2010-11 are more active than the early 2000s.

### Details of fiscal consolidation packages
- Fiscal consolidation disaggregated into tax-hike and spending-cut components (percent of GDP).
- Tax-hike components considered: corporate tax, personal income tax, VAT, property tax, health care tax.
- Spending-cut components considered: government capital expenditure cuts, cuts in social security contribution and social benefits, wage cuts, personnel cuts, pension cuts, health care cuts, public investment cuts.
- Binary indicators: each measure takes value 1 if included in the package, 0 otherwise.
- Sector-specific tax and spending measures considered (binary indicators):
  - Taxes affecting banking sector (e.g., taxes on deposits).
  - Taxes affecting retail sector (e.g., consumption taxes).
  - Taxes affecting health care sector (e.g., contributions, premiums).
  - Spending cuts affecting transportation, health care, and defense sectors.

### Loan and firm characteristics (variables controlled in regressions)
- Dependent variable: natural logarithm of syndicated loan spreads, inclusive of all fees, measured as the difference between loan rate and relevant benchmark rate (usually LIBOR).
- Loan-level controls include:
  - Loan maturity.
  - Natural logarithm of loan size.
  - Loan purpose (working capital, refinancing, general purposes, etc.).
  - Loan currency.
  - Loan type (term credit, revolving credit, etc.).
  - Investment grade indicator.
  - Indicator for loans given by a domestic bank.
- Firm-level controls include:
  - Leverage (ratio of total debt to total assets).
  - Profitability (net income over total assets).
  - Tangibility (tangible assets over total assets).
  - Growth opportunities (ratio of market value of assets to book value of assets).
  - Indicator whether borrower is rated.

### Country characteristics (controls)
- Country-level controls include:
  - Sovereign rating (Moody’s), converted to numeric scale where Aaa = 21, Aa1 = 20, Aa2 = 19, etc.
  - Total government debt (general-government-debt-to-GDP ratio or central government debt where general-government unavailable).
  - Change in exchange rate and interest rate, level of interest rate.
  - Natural logarithm of real GDP, real GDP growth.
  - Stock market turnover, trade openness, inflation.
  - Government stability, degree of law and order, level of corruption (PRS Group’s ICRG database).
- ICRG indices: higher values correspond to more stable government, better law and order, and less corruption.

### Empirical specification and identification
- Regression equation (conceptual form):
  - LoanSpread = β1 Fiscal + β2 Country + β3 Loan + β4 Firm + fixed effects (country, industry, year) + clustered robust SEs by country.
- Fiscal variable in baseline: contemporaneous total fiscal consolidation in percent of GDP.
- Alternative specifications: tax-based versus spending-based consolidations; package details (taxes on consumption, corporate income, personal income, property, health care; cuts in transfers, government consumption, public investment, social security contributions).
- Identification arguments for causal interpretation:
  1. Fiscal consolidation measure is exogenous (records discretionary actions to reduce debt).
  2. Loan-level data reduce reverse causality risk from aggregate credit conditions to fiscal decisions.
  3. Falsification test: sector-specific consolidation measures should affect spreads for firms in targeted sectors but not for firms in other sectors (results support this for tax hikes; spending cuts generally do not have significant effects on loan spreads).
- Caveat: latent common factors may still drive both consolidation decisions and loan spreads despite extensive controls.

### Empirical findings — Descriptive statistics (sample moments and notable observations)
- Average fiscal consolidation: around 0.3 percent of GDP.
- Maximum fiscal consolidation observed: 4.5 (Italy in 1993).
  - Note: one larger episode exists — Ireland’s consolidation amounting to 4.74 percent of GDP in 2009 — but was dropped in regressions due to data gaps.
- Spending cuts vs. tax hikes:
  - Average spending cuts: 0.2 percent of GDP.
  - Average tax hikes: 0.1 percent of GDP.
  - Maximum spending cut: 3.7 percent of GDP (Finland in 1993).
  - Maximum tax hike: 2.4 percent of GDP (Italy in 1995).
- Sovereign credit ratings in sample:
  - Range between Baa1 and Aaa.
  - Average rating stated as Aaa.
  - Lowest value observed: Baa1 (Ireland in 2010).
  - Note: Ireland reduced to Ba1 and Portugal to Ba2 in 2011, but dataset lacks full loan-spread observations for those cases in 2011.
- Government debt:
  - Average government debt: around 67 percent of GDP.
  - Minimum around 5 percent of GDP (Australia in 2007 and 2008).
  - Maximum above 200 percent of GDP (Japan after 2009).
- Interest rates and exchange rates:
  - Exchange rates change on average by 2 percent.
  - Interest rates change on average around 15 percent.
  - Low short-term interest rates observed: Japan after mid-1990s; Sweden in 2009.
  - High short-term interest rates (>15 percent) observed for Italy, Spain, Portugal, and Australia from mid-1980s to early 1990s.
- Trade openness and inflation:
  - Average trade openness: 65 percent of GDP.
  - Average inflation: around 2 percent (Japan experienced deflation starting in the early 1990s).
- Governance indices:
  - Lowest government stability observed for Italy and Spain in late 1980s/early 1990s, and for several European countries during 2010-11 eurozone crisis.
  - Law and order and lack of corruption indices mostly high.
- Loan sample characteristics:
  - Average loan spread: around 134 basis points.
  - Investment-grade loans: around 70 percent of loans.
  - Loans issued in a major currency (Euro, Japanese Yen, British Pound, U.S. Dollar): around 80 percent.
  - Borrowers with a credit rating: around 30 percent.
  - Loans extended by domestic banks: around 30 percent.
  - Average loan maturity: 4 years.
  - Average loan size: around US$ 790 million.
  - Loan size range: US$ 250 thousand to US$ 45 billion.
  - Largest observed loan: US$ 45 billion (InBev to buyout Anheuser-Busch in 2008) — maturity around 3 years and spread of 200 basis points above LIBOR, inclusive of all fees.
- Firm characteristics: sample contains both large and small firms, a mix of unlevered and levered firms, firms with low and high growth opportunities, and firms with and without tangible assets.

### Baseline regressions (intro)
- Analysis begins by estimating Equation (1) with fiscal consolidation measured as percent of GDP based on announced and implemented government actions to reduce sovereign debt.
- Main variable of interest: contemporaneous total fiscal consolidation in percent of GDP; alternative regressions use disaggregated consolidation measures.

*Source: Appendix Table 1 and accompanying text from the PDF chapter.*

### Section III.A. Table 3 presents the results. In Column 1, we include only the country-, loan- and

### _wp1336 - Section III.A. Table 3 presents the results. In Column 1, we include only the country-, loan- and

### Baseline results (Table 3)
- Syndicated loan spreads are positively associated with fiscal consolidations (Column 2), indicating lenders price in recessionary effects of fiscal consolidations more than anticipated improvements in future economic performance or reduced sovereign default risk.
- Introducing an interaction for large fiscal consolidations (indicator = 1 if consolidation/GDP > 1.5 percent) (Column 3) reduces the positive relation between loan spreads and consolidations considerably.
- Economic magnitudes (Table 3, Column 3, evaluated at the average spread of 134 basis points):
  - An additional 1 percent of GDP in fiscal consolidation is associated with an 11 percent increase in spreads, corresponding to around 15 basis points.
  - For a typical loan (average loan size around US$ 800 million and average term of 4 years), this corresponds to around US$ 5 million in additional borrowing costs over the life of the loan.
  - If the consolidation is large (> 1.5 percent of GDP), the increase in spreads is only 2 basis points; additional financing costs are negligible when consolidations are large.
- The differential effect for large consolidations is interpreted as evidence that both:
  - Keynesian recessionary-effects channel (raising spreads), and
  - Expectations/credibility channel (reducing spreads) are operative; large consolidations mitigate lenders’ emphasis on recessionary effects by signaling reduced sovereign default risk and stronger permanency.
- Results are robust to dropping one country at a time (robustness check; coefficients remain broadly the same).

### Loan-, firm-, and country-level determinants (Table 3 additional findings)
- Country-level:
  - Spreads decrease with increasing interest rates (coefficient very small).
  - Spreads decrease with real GDP growth.
  - Spreads decrease with the degree of law and order.
- Loan characteristics associated with lower spreads:
  - Investment grade (lower default risk).
  - Larger deal size (fixed origination costs spread over larger deals).
- Loan characteristic associated with higher spreads:
  - Denomination in foreign currency (higher risk premium).
- Firm characteristics:
  - Lower spreads associated with firm size, growth opportunities, and tangibility.
  - Higher spreads associated with leverage.

### Firm and loan heterogeneity (Section C; Table 4)
- Small vs. large firms (Panel A):
  - Positive association between loan spreads and fiscal consolidations observed for small firms; large firms are not affected.
  - When distinguishing consolidations above 1.5 percent of GDP, adverse relation is reduced for large consolidations.
  - Magnitudes (Table 4, Panel A, Column 2):
    - For small firms, increasing fiscal consolidation by 1 percent of GDP increases loan spreads by 17 percent.
    - Average spread in small-firm sample: 132 basis points → 17 percent ⇒ 23 basis-point increase.
    - Average deal value in small-firm sample: around 600 million dollars → this increase ≈ US$ 1.4 million additional financing costs in a year.
    - If consolidation is large (> 1.5 percent of GDP), increase in spreads is 4 percent ⇒ 6 basis points.
- Firms without bond ratings and without ADRs (Panels B and C):
  - Firms without bond ratings and without ADRs experience adverse borrowing-cost effects from fiscal consolidations.
  - Large consolidations substantially reduce this effect for firms without bond ratings; reduction not statistically significant for firms without ADRs.
  - Subsample mean spreads and 1 percent consolidation effects:
    - Firms without ratings: average spread 135 basis points; 12 percent ⇒ 16 basis points; large consolidation → less than 8 percent (2 basis points).
    - Firms without ADRs: average spread 139 basis points; 14 percent ⇒ 19 basis points; large consolidation → less than 8 percent (7 basis points).
- Firms that do not use external markets often (Panel D):
  - These firms experience increases in loan spreads with fiscal consolidations; effect reduced for large consolidations.
  - Average spread for firms that do not use external markets often: 137 basis points; 19 percent ⇒ 25 basis points; large consolidation → less than 8 percent (11 basis points).
- Domestic vs multinational firms (Panel E):
  - Domestic firms (no foreign sales and no foreign income): borrowing terms adversely associated with fiscal consolidations unless consolidations are large.
    - Average spread for domestic firms: 135 basis points; 17 percent ⇒ 23 basis points; large consolidation → less than 8 percent (1 basis point).
  - Multinational firms (nonzero foreign sales and/or foreign income) do not observe changes in borrowing costs during fiscal consolidations.
- Overall implication:
  - Effects of fiscal consolidations are asymmetric across firms: small, domestic, and financing-constrained firms face higher borrowing costs unless consolidations are large and credible; large, multinational firms with alternative financing sources are generally insulated.

### Country conditions and interactions (Section D; Table 5)
- Sovereign debt levels:
  - When sovereign debt is low: fiscal consolidations are not relevant for loan spreads.
  - When sovereign debt is high: loan spreads increase with consolidations unless consolidations are large.
- Government size (total expenditures/GDP):
  - Consolidation positively associated with spreads for both large and small governments; offsetting effect of large consolidations present only when the government is large.
- Monetary policy stance (Taylor-rule residuals; Columns 5-6):
  - Under loose (accommodative) monetary policy and large consolidations, the overall effect of large consolidations is negative (sum of coefficients = -0.126 in Column 6).
  - Under tight monetary policy, the sum of coefficients is 0.003 in Column 5.
  - Interpretation: accommodative monetary policy helps shield the economy from short-term pain of large consolidations.
- Exchange rate regime (Columns 7-8; Reinhart and Rogoff classification):
  - Flexible regimes: fiscal consolidations do not have a significant relation with cost of credit.
  - Fixed regimes: fiscal consolidations positively associated with spreads unless consolidations are large enough to reverse adverse effects.
- Currency depreciation (Columns 9-10):
  - Countries without currency depreciation: loan spreads increase with consolidations (mitigated for large consolidations).
  - Countries that let their currency depreciate: fiscal consolidations do not have a significant relationship with corporate credit cost.
- Political stability (Columns 11-12):
  - Large consolidations create positive expectations and reduce cost of credit when lenders believe the government is stable.
  - Sum of coefficients in Column 12 (stable government) is negative and significant; Column 11 (unstable) sum ≈ zero.
- Overall: large consolidations accompanied by flexible macroeconomic policy frameworks and political stability shield corporations from adverse effects on cost of credit; most relevant for small and financing-constrained firms and domestic firms.

### Composition of fiscal consolidation packages (Section E; Tables 6–7)
- Aggregate measures:
  - Spending cuts and tax hikes as percent of GDP (Columns 1–3, Table 6): neither has a statistically significant association with cost of corporate credit (coefficients positive but not significant).
- Specific measures (Columns 4–6, Table 6):
  - Corporate tax hikes: coefficient large and highly significant — major driver of positive association between loan spreads and consolidations.
  - Social benefit cuts: associated with lower loan spreads (lenders reflect improved credibility or projected increase in corporate profitability).
  - VAT hikes: negative association with loan spreads.
- Sector-specific interactions (Table 7):
  - Loan spreads increase in sectors that face tax hikes; spending cuts in a sector do not have similar adverse effects.
  - Random-match robustness checks: interaction variables generated by random matching of measures and sectors are not significant (iterations available upon request), supporting cross-sector exposure interpretation.
- Interpretation relative to literature:
  - Overall evidence leans toward expenditure-based consolidations being less contractionary than tax-based measures; corporate tax hikes appear particularly influential for loan spreads.

### Alternative specifications and robustness (Section F; Tables 8–9)
- Sample-selection concerns:
  - Heckman (1979) selection model:
    - First-stage instrument: external finance dependence (Rajan and Zingales methodology).
    - Findings (Table 8, Columns 1–4): coefficient on fiscal consolidation positive and significant; close to baseline magnitude.
    - Lambda (inverse Mill’s ratio) not significant → results not driven by sample selection bias.
  - Alternative universe (firms that issued at least one loan in sample period) (Table 8, Columns 5–8):
    - Coefficients on fiscal consolidation and large consolidation comparable to baseline; lambda not significant.
- Alternative specifications (Table 9, Columns 1–3):
  - Add lagged fiscal consolidation measure; add lagged real GDP growth; replace Moody’s sovereign rating with Institutional Investor rating.
  - Coefficients on fiscal consolidation remain close in magnitude to baseline.
- Large-consolidation definition robustness (Table 9, Column 4):
  - Extend definition to also include > 1 percent of GDP in two consecutive years (in addition to > 1.5 percent in a year).
  - Positive and statistically significant coefficient on fiscal consolidation persists; large consolidation variable same sign and comparable magnitude but no longer statistically significant at conventional levels.
- CAPB-based consolidations (cyclically-adjusted primary balance; Table 9, Columns 5–6 and 9–10):
  - No significant relation between CAPB-based consolidation measure and spreads (potential endogeneity of CAPB).
  - IV approach: instrument CAPB-based measure with action-based measure (Guajardo et al. approach) (Table 9, Columns 7–8 and 11–12):
    - Positive association between spreads and fiscal consolidations unless consolidations are large enough to signal credibility and create positive expectations.

### Conclusions (Section V)
- Using a new dataset covering 22 years in 16 advanced economies and syndicated loan market data, fiscal consolidations are associated with higher corporate loan spreads on average, particularly for:
  - Small firms, domestic firms, and firms with limited access to alternative financing.
- This adverse effect is substantially mitigated when fiscal consolidations are large (> 1.5 percent of GDP) and when large consolidations are accompanied by:
  - Flexible macroeconomic policy frameworks (e.g., accommodative monetary policy, flexible exchange rates, currency depreciation), and
  - Political stability (credible governments able to carry out consolidations).
- Among consolidation measures, corporate tax hikes are positively related to loan spreads; sectoral tax hikes increase spreads in affected sectors, while spending cuts generally do not raise sectoral loan spreads and can lower spreads when they signal credibility or improve corporate profitability expectations.
- Overall interpretation:
  - Lenders price in potential short-term recessionary effects of fiscal consolidations, especially when consolidations appear noncredible; expectations and credibility channels operate to lower borrowing costs when consolidations are large and supported by accommodating policies and stable governments.

*Source: _wp1336 - Section III.A. Table 3 presents the results. In Column 1, we include only the country-, loan- and firm-level controls.*

### References

### _wp1336 - References

### Key Data and Variable Definitions
- Main Variables of Interest:
  - LogSpread: Natural logarithm of syndicated loan spread, inclusive of all fees, measured as the difference between the rate charged on the loan and the relevant benchmark (in basis points), where the benchmark is the interbank lending rate agreed upon by the lenders in the syndicate — Dealogic.
  - FiscalConsolidation: Fiscal consolidation measures announced and implemented with an aim to reduce the public debt level and expressed in percent of GDP — Devries et al. (2011) and authors' calculations based on IMF and OECD reports and national budgets.
  - FiscalConsolidation > 1.5% GDP: Dummy variable that takes the value of 1 if FiscalConsolidation exceeds 1.5 percent of GDP — Devries et al. (2011) and authors' calculations based on IMF and OECD reports and national budgets.
  - TaxHikes: Fiscal consolidation accomplished through increases in revenues (in percent of GDP) — Devries et al. (2011) and authors' calculations based on IMF and OECD reports and national budgets.
  - SpendingCuts: Fiscal consolidation accomplished through decreases in expenditures (in percent of GDP) — Devries et al. (2011) and authors' calculations based on IMF and OECD reports and national budgets.

- Country Variables:
  - SovereignRating: Indicator for sovereign default risk ranging from Aaa to C, converted to numerical values by assigning the value of 21 to the highest rating (Aaa) and decreasing the assigned value by 1 for each notch down — Moody's.
  - SovereignDebt: General government debt outstanding (central, if general is not available), in percent of GDP — Reinhart and Rogoff (2011) and OECD.
  - ExchangeRate: Real effective exchange rate (basket of currencies per local currency), index — IMF International Financial Statistics.
  - InterestRateShort: short interest rate — OECD.
  - LogRealGDP: Natural logarithm of real GDP — World Bank Development Indicators.
  - RealGDPGrowth: Annual change in real GDP — World Bank Development Indicators.
  - StockMarketTurnover: Total number of shares traded during the period divided by the average market capitalization for the period (average of the end-of-period values for the current period and the previous period) — World Bank Development Indicators.
  - TradeOpenness: Sum of exports and imports, in percent of GDP — IMF International Financial Statistics.
  - Inflation: Annual change in the consumer price index — IMF International Financial Statistics.
  - GovernmentStability: Index with higher values indicating more stable governments — PRS Group's International Country Risk Guide.
  - Law&Order: Index with higher values indicating stronger and impartial legal systems and better popular observance of the law — PRS Group's International Country Risk Guide.
  - LackofCorruption: Index with higher values indicating less corrupt systems — PRS Group's International Country Risk Guide.
  - GovernmentSize: Total government revenues to GDP — OECD.
  - InstitutionalInvestorCreditRating: Index with higher values indicating lower sovereign default probability — Institutional Investor.

- Loan Variables:
  - InvestmentGrade: Dummy = 1 if the deal is investment grade — Dealogic.
  - RatedBorrower: Dummy = 1 if the borrower is rated — Dealogic.
  - LogDealValue: Natural logarithm of the deal value — Dealogic.
  - DealMaturity: Maturity of the loan in years — Dealogic.
  - MajorCurrencyDenomination: Dummy = 1 if the loan is denominated in euros, Japanese yens, British pounds, or U.S. dollars — Dealogic.
  - ForeignCurrencyDenomination: Dummy = 1 if the loan is denominated in a currency different from the local currency used in the borrower's country — Dealogic.
  - DomesticBank: Dummy = 1 if the lender is located in the same country as the borrower — Dealogic.

- Firm Variables:
  - Size: Natural logarithm of total assets — WorldScope.
  - Leverage: Ratio of total debt to total assets — WorldScope.
  - Profitability: Ratio of net income to total assets — WorldScope.
  - GrowthOpportunities: Tobin's Q (ratio of market value of assets to book value of assets) — WorldScope.
  - Tangibility: Ratio of tangible assets to total assets — WorldScope.

- Appendix Table: Appendix Table 1. Variable Definitions and Data Sources.

### Country Observations (sample counts)
- Australia: Observations 213; Observations with Fiscal Consolidation 31; Year Observations 1990 99; Observations with Fiscal Consolidation 6
- Austria: Observations 18; Observations with Fiscal Consolidation 3; Year Observations 1991 60; Observations with Fiscal Consolidation 22
- Belgium: Observations 68; Observations with Fiscal Consolidation 14; Year Observations 1992 104; Observations with Fiscal Consolidation 26
- Canada: Observations 469; Observations with Fiscal Consolidation 94; Year Observations 1993 128; Observations with Fiscal Consolidation 36
- Denmark: Observations 56; Observations with Fiscal Consolidation 2; Year Observations 1994 172; Observations with Fiscal Consolidation 118
- Finland: Observations 98; Observations with Fiscal Consolidation 31; Year Observations 1995 237; Observations with Fiscal Consolidation 174
- France: Observations 528; Observations with Fiscal Consolidation 86; Year Observations 1996 204; Observations with Fiscal Consolidation 131
- Germany: Observations 422; Observations with Fiscal Consolidation 268; Year Observations 1997 103; Observations with Fiscal Consolidation 96
- Ireland: Observations 34; Observations with Fiscal Consolidation -; Year Observations 1998 131; Observations with Fiscal Consolidation 80
- Italy: Observations 270; Observations with Fiscal Consolidation 185; Year Observations 1999 175; Observations with Fiscal Consolidation 76
- Japan: Observations 611; Observations with Fiscal Consolidation 176; Year Observations 2000 297; Observations with Fiscal Consolidation 25
- Netherlands: Observations 233; Observations with Fiscal Consolidation 65; Year Observations 2001 305; Observations with Fiscal Consolidation 2
- Portugal: Observations 36; Observations with Fiscal Consolidation 16; Year Observations 2002 190; Observations with Fiscal Consolidation 2
- Spain: Observations 416; Observations with Fiscal Consolidation 125; Year Observations 2003 347; Observations with Fiscal Consolidation 75
- Sweden: Observations 177; Observations with Fiscal Consolidation 53; Year Observations 2004 359; Observations with Fiscal Consolidation 103
- United Kingdom: Observations 1,318; Observations with Fiscal Consolidation 379; Year Observations 2005 456; Observations with Fiscal Consolidation 77
- Additional year-level rows listed: 2006 305 with 78; 2007 503 with 131; 2008 167 with -; 2009 163 with -; 2010 269 with 106; 2011 193 with 164
- Total: Observations 4,967; Observations with Fiscal Consolidation 1,528

*Source: _wp1336 - References (Appendix Table 1 and Country Observations).*

### Appendix Table 2. Distribution of Observations by Country and Year

### _wp1336 - Appendix Table 2. Distribution of Observations by Country and Year

### Distribution of Observations by Country and Year
- Country-year cells report fiscal consolidation episodes (in percent of GDP) across 1990–2011 for sample countries (selected entries shown in source):
  - Australia: 1990 0.25, 1991 0.50, 1992 0.62, 1993 0.71, 1994 0.37, 1995 0.40
  - Austria: 1990 2.41, 1991 1.56
  - Belgium: 1990 0.60, 1991 1.79, 1992 0.92, 1993 1.15, 1994 1.00, 1995 0.91
  - Canada: 1990 0.86, 1991 0.40, 1992 0.23, 1993 0.37, 1994 0.49, 1995 0.99, 1996 0.97, 1997 0.48
  - Finland: 1990 0.91, 1991 3.71, 1992 3.46, 1993 2.91, 1994 1.47, 1995 0.23
  - France: 1990 0.25, 1991 0.28, 1992 1.33, 1993 0.50
  - Germany: 1990 1.11, 1991 0.46, 1992 0.11, 1993 0.91, 1994 1.08, 1995 1.60, 1996 0.30, 1997 0.70
  - Italy: 1990 2.77, 1991 3.50, 1992 4.49, 1993 1.43, 1994 4.20, 1995 0.34, 1996 1.82, 1997 0.68
  - Japan: 1990 1.43, 1991 0.48
  - Netherlands: 1990 0.87, 1991 0.74, 1992 0.12
  - Portugal: 1990 0.50
  - Spain: 1990 0.70, 1991 1.10, 1992 1.60, 1993 0.74, 1994 1.30, 1995 1.20
  - Sweden: 1990 1.81, 1991 0.78, 1992 3.50, 1993 2.00, 1994 1.50, 1995 1.00
  - United Kingdom: 1990 0.83, 1991 0.28, 1992 0.30, 1993 0.69, 1994 0.31, 1995 0.22
- Later years (2001–2011) entries (selected):
  - Austria: 2001 1.02, 2002 0.55, 2003 0.73
  - Belgium: 2001 0.43, 2002 0.79
  - Finland: 2001 0.23, 2002 0.94
  - France: 2001 2.11
  - Germany: 2001 0.74, 2002 0.40, 2003 0.50, 2004 0.90, 2005 0.43
  - Ireland: 2001 4.74, 2002 2.60, 2003 3.70
  - Italy: 2001 1.30, 2002 1.00, 2003 1.39, 2004 1.61, 2005 0.39, 2006 0.78
  - Japan: 2001 0.48, 2002 0.64, 2003 0.28, 2004 0.72, 2005 0.15
  - Netherlands: 2001 1.70, 2002 0.50, 2003 0.34
  - Portugal: 2001 1.60, 2002 0.60, 2003 1.65, 2004 1.40, 2005 2.26, 2006 4.10
  - Spain: 2001 1.90, 2002 2.20
  - United Kingdom: 2001 0.60, 2002 1.90
  - Observations: see descriptive statistics section for total observations.

### Descriptive Statistics (Table 2)
- Observations: 4,967
- Main variables:
  - LogSpread: Mean 4.41, Std.Dev. 1.00, Min 0.17, Max 7.2
  - Spread (basis points): Mean 133.6, Std.Dev. 136.4, Min 1.1, Max 400.0
  - FiscalConsolidation: Mean 0.30, Std.Dev. 0.60, Min 0.0, Max 4.5
  - FiscalConsolidation > 1.5% GDP: Mean 0.10, Std.Dev. 0.20, Min 0.0, Max 1.0
  - TaxHikes: Mean 0.10, Std.Dev. 0.30, Min 0.0, Max 2.4
  - SpendingCuts: Mean 0.20, Std.Dev. 0.40, Min 0.0, Max 3.7
- Country variables (selected):
  - SovereignRating: Mean 20.5, Std.Dev. 0.8, Min 14.0, Max 21.0
  - SovereignDebt: Mean 67.03, Std.Dev. 37.44, Min 4.5, Max 233.1
  - ExchangeRate, change: Mean 1.9, Std.Dev. 8.1, Min -25.3, Max 20.2
  - InterestRate, change: Mean 5.3, Std.Dev. 57.5, Min -85.3, Max 740.3
  - InterestRate, level: Mean 4.0, Std.Dev. 2.6, Min 0.0, Max 15.2
  - LogRealGDP: Mean 27.9, Std.Dev. 0.8, Min 25.2, Max 29.0
  - RealGDPGrowth: Mean 2.2, Std.Dev. 1.9, Min -8.4, Max 10.7
  - TradeOpenness: Mean 64.7, Std.Dev. 32.6, Min 0.2, Max 197.3
  - Inflation: Mean 2.0, Std.Dev. 1.4, Min -1.4, Max 10.4
  - Law&Order: Mean 91.4, Std.Dev. 9.8, Min 0.0, Max 100.0
  - LackofCorruption: Mean 74.0, Std.Dev. 15.2, Min 33.3, Max 100.0
- Loan variables (selected):
  - InvestmentGrade: Mean 0.70, Std.Dev. 0.40, Min 0.0, Max 1.0
  - RatedBorrower: Mean 0.30, Std.Dev. 0.40, Min 0.0, Max 1.0
  - LogDealValue: Mean 5.5, Std.Dev. 1.6, Min -1.4, Max 10.7
  - DealValue (millions of USD): Mean 790.5, Std.Dev. 2,005.2, Min 0.2, Max 45,000.0
  - MajorCurrencyDenomination: Mean 0.80, Std.Dev. 0.40, Min 0.0, Max 1.0
  - ForeignCurrencyDenomination: Mean 0.60, Std.Dev. 0.50, Min 0.0, Max 1.0
  - DomesticBank: Mean 0.30, Std.Dev. 0.40, Min 0.0, Max 1.0
- Firm variables (selected):
  - Size: Mean 17.0, Std.Dev. 2.9, Min 7.3, Max 26.0
  - Leverage: Mean 0.30, Std.Dev. 0.20, Min 0.0, Max 1.9
  - Profitability: Mean 0.10, Std.Dev. 0.10, Min -1.5, Max 0.6
  - GrowthOpportunities: Mean 0.90, Std.Dev. 0.80, Min 0.0, Max 17.8
  - Tangibility: Mean 0.30, Std.Dev. 0.30, Min 0.0, Max 1.0

### Baseline Regression Results (Table 3)
- Dependent variable: natural logarithm of syndicated loan spread. Observations: 4,967; R-squared: 0.61 (columns vary between 0.61 and 0.611).
- Key coefficients (selected; robust SEs in brackets; significance markers preserved as in source):
  - FiscalConsolidation:
    - Column (1): 0.0287* [0.0156]
    - Column (2): 0.1080*** [0.0322]
    - Column (4): 0.1200*** [0.0297]
  - FiscalConsolidation > 1.5% GDP:
    - Column (1): -0.0933** [0.0378]
    - Column (2): -0.1160*** [0.0341]
  - InterestRate, change:
    - Columns show coefficients around -0.0005** to -0.0006*** (e.g., -0.0005** [0.0002])
  - Law&Order:
    - Significant negative coefficients across specifications (e.g., -0.0056*** [0.0019])
  - InvestmentGrade:
    - Strong negative and highly significant everywhere (approx. -1.2010*** [0.0560])
  - LogDealValue:
    - Negative and highly significant (approx. -0.0879*** [0.0132])
  - Size:
    - Negative and highly significant (approx. -0.0380*** [0.0105])
  - Leverage:
    - Positive and highly significant (approx. 0.3850*** [0.0967])
  - GrowthOpportunities and Tangibility:
    - GrowthOpportunities: negative and highly significant (approx. -0.0549*** [0.0148])
    - Tangibility: negative and significant (approx. -0.1330** [0.0453])

### Which Firms Face Higher Borrowing Costs around Fiscal Consolidations? (Table 4 — Panel A: Firm Size)
- Sample split by firm size (Small vs Large). Observations and R-squared:
  - Small: Observations 2,463; R-squared 0.71
  - Large: Observations 2,504; R-squared 0.58
- FiscalConsolidation:
  - Small firms: 0.0597** [0.0268]
  - Large firms: 0.1590*** [0.0528]
- FiscalConsolidation > 1.5% GDP:
  - Small firms: -0.1150** [0.0523]
- Notable other coefficients:
  - InvestmentGrade: large negative (e.g., -1.0630*** [0.0395] for small; -1.3550*** [0.0868] for large)
  - LogDealValue: negative and significant for both subsamples (e.g., -0.0661*** [0.00922] small; -0.0962*** [0.0208] large)
  - Leverage: positive and significant for both, stronger for large (e.g., 0.2770*** small; 0.5740** large)

### Heterogeneity by Credit Rating, Access to Global Markets, External Finance Dependence (Table 4 — Panels B–D)
- Panel B: Credit Rating (Not rated vs Rated)
  - Not rated (Observations 3,659; R-squared 0.64): FiscalConsolidation 0.0276 [0.0228] (not significant), FiscalConsolidation > 1.5% GDP -0.1020** [0.0367]
  - Rated (Observations 1,308; R-squared 0.63): FiscalConsolidation 0.1150** [0.0442]
- Panel C: Access to Global Markets (No ADR vs ADR)
  - No ADR (Observations 2,688; R-squared 0.66): FiscalConsolidation 0.0570* [0.0279]
  - ADR (Observations 2,279; R-squared 0.62): FiscalConsolidation 0.1270* [0.0693]
- Panel D: External Finance Dependence (Low vs High)
  - Low Dependence (Observations 2,485; R-squared 0.69): FiscalConsolidation 0.0962** [0.0341]
  - High Dependence (Observations 2,482; R-squared 0.58): FiscalConsolidation 0.1700*** [0.0563]
- Selected ancillary findings:
  - ForeignCurrencyDenomination positive and significant across subsamples (e.g., 0.1180*** [0.0306] No ADR; 0.1530*** [0.0387] ADR)
  - Size remains negative and significant in most splits (e.g., -0.0478*** [0.0095] No ADR)

### Cross-Border Diversification of Firm Activities (Table 4 — Panel E)
- Domestic vs Multinational:
  - Domestic (Observations 3,150; R-squared 0.64): FiscalConsolidation 0.0232 [0.0225]
  - Multinational (Observations 1,817; R-squared 0.61): FiscalConsolidation 0.1580*** [0.0322]
  - FiscalConsolidation > 1.5% GDP: Domestic -0.1520*** [0.0287]; Multinational -0.0029 [0.0793]
- Additional notable coefficients in this panel:
  - Leverage positive and significant in both groups (e.g., 0.2420** [0.0954] Domestic; 0.6960** [0.2770] Multinational)
  - Size negative and significant in both groups (e.g., -0.0306** [0.0113] Domestic; -0.0662*** [0.0153] Multinational)

### Country Conditions Matter (Table 5 — Selected splits)
- Table 5 splits sample by median country-condition measures (High vs Low or analogous splits). Selected results (columns and variables in source):
  - FiscalConsolidation coefficients vary across 12 splits (examples):
    - Column (1) High Sovereign Debt: FiscalConsolidation 0.2270** [0.0780]
    - Column (2) Low Sovereign Debt: FiscalConsolidation 0.0118 [0.0828]
    - Column (3) Large Government Size: FiscalConsolidation 0.1100*** [0.0320]
    - Column (4) Small Government Size: FiscalConsolidation 0.1800** [0.0621]
    - Column (5) Tight Monetary Policy: FiscalConsolidation 0.1410*** [0.0388]
    - Column (6) Loose Monetary Policy: FiscalConsolidation 0.1070 [0.0622]
  - FiscalConsolidation > 1.5% GDP often negative and significant in several splits (e.g., -0.2130*** [0.0691] in column 1)
- Loan and firm variable patterns across splits:
  - InvestmentGrade: consistently large negative and highly significant across splits (e.g., -1.2190*** [0.1310] to -1.0920*** [0.0609])
  - ForeignCurrencyDenomination: consistently positive and frequently significant (e.g., 0.1510** [0.0511] to 0.1800*** [0.0328])
  - Size: consistently negative and often significant (e.g., -0.0503** [0.0183] to -0.0294** [0.0134])
  - Leverage: consistently positive and significant (e.g., 0.5310*** [0.1040] to 0.3680*** [0.0849])
- Observations and R-squared for splits: Observations vary across columns (e.g., 1,903 to 3,323); R-squared range approx. 0.60–0.67 in reported splits.

### Tax Hikes versus Spending Cuts (Table 6)
- Dependent variable: log syndicated loan spread. Observations: 4,967; R-squared: approx. 0.61 across columns.
- Coefficients (selected):
  - TaxHikes: Column (1) 0.0405 [0.0342]; Column (2) 0.0241 [0.0482]
  - SpendingCuts: Column (1) 0.0431 [0.0311]; Column (2) 0.0329 [0.0436]
  - CorporateTax: Column (1) 0.5450*** [0.1060]; Column (2) 0.5660*** [0.1390]
  - ValueAddedTax: Column (1) -0.2030*** [0.0542]; Column (2) -0.1530** [0.0676]
  - SocialBenefitsCut: Column (1) -0.1200** [0.0545]; Column (2) -0.1010* [0.0598]
  - WageCut: Column (1) -0.1280** [0.0628]; Column (2) -0.0703 [0.0718]
  - UnemploymentBenefitCut: Column (1) 0.1320* [0.0707]; Column (2) 0.0079 [0.0835]
  - PensionCut: Column (1) 0.1540* [0.0926]; Column (2) 0.0806 [0.1030]
- Notes: country-, loan-, and firm-level controls included (coefficients omitted for space in table).

### Differential Effect of Fiscal Consolidation on Sectors (Table 7)
- Selected sector-interaction results (observations 4,967; R-squared approx. 0.61 across columns):
  - ConsumptionTax * RetailSector: 0.2300*** [0.0714]
  - BankTax * BankingSector: 0.2710** [0.0946]
  - HealthcareTax * HealthcareSector: 0.7460*** [0.1980]
  - DefenseCuts: 0.0896* [0.0482]
  - Other sector interactions reported with standard errors; significance preserved as in source.

### Heckman Selection Model Results (Table 8)
- Two-step Heckman selection results reported. Columns 1–4 use Worldscope sample as all firms; columns 5–8 restrict to firms that issued at least one loan.
- Key coefficients (second-step dependent variable = log spread; first-step dependent variable = dummy if spread nonmissing):
  - FiscalConsolidation (second-step examples):
    - Column (1): 0.0275* [0.0158]
    - Column (3): 0.1060*** [0.0325]
    - Column (5): 0.0276* [0.0157]
    - Column (7): 0.1060*** [0.0322]
  - FiscalConsolidation > 1.5% GDP: -0.0930** [0.0379] (columns 1 and 3)
  - ExternalFinanceDependence (first-step): 2.2553*** [0.2263] (columns 1 and 2), 1.8973*** [0.1531] (columns 3 and 4)
  - Lambda (selection correction term): -0.0370 [0.0254] to -0.0408 [0.0501] (not significant as reported)
- Sample sizes:
  - First-step observations large (e.g., 196,157 in some columns; 20,382 in others)
  - Second-step observations smaller subsets (e.g., 196,157 second-step in some columns; 20,382 in others per table layout)

### Robustness (Table 9)
- Multiple robustness checks (observations 4,967; R-squared 0.61 across specifications).
- Selected coefficients:
  - FiscalConsolidation:
    - Column (1): 0.0258* [0.0128]
    - Column (2): 0.0279* [0.0155]
    - Column (3): 0.0449** [0.0199]
    - Column (4): 0.1020** [0.0415]
  - FiscalConsolidation, lagged: 0.0331 [0.0336] (column with lag)
  - RealGDPGrowth, lagged: -0.0366** [0.0148]
  - InstitutionalInvestorCreditRating: 0.0063 [0.0054]
  - FiscalConsolidation * Large: -0.0819* [0.0449]
  - CAPB-based FiscalConsolidation specifications:
    - FiscalConsolidation, CAPB-based: examples include 0.0527* [0.0305], 0.1390** [0.0583] in some columns
    - FiscalConsolidation, CAPB-based > 1.5% GDP: -0.0347 [0.0320], -0.1210** [0.0495], -0.1520*** [0.0299] in various columns
- Notes on specification variants:
  - Column-specific approaches: inclusion of lagged variables, alternative sovereign rating index (Institutional Investor), alternative definition of large consolidation (Alesina and Ardagna (2010)), CAPB-based identification, instrumental variables estimation where CAPB-based measure is instrumented by action-based measure, and matching alternative episodes.
  - Country-, loan-, and firm-level control sets as in Table 3 included in all robustness regressions.

*Italic: Source: _wp1336 - Appendix Table 2. Distribution of Observations by Country and Year (extracted from provided PDF content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1336.pdf_
