Fiscal Consolidation and the Cost of Credit
Evidence from Syndicated Loans†
Author: Ms. Senay Agca and Ms. Deniz O Iganhttps://orcid.org/0000-0002-9501-1807

Contributor Notes

Authors’ E-Mail Addresses: sagca@gwu.edu; digan@imf.org
Publication Date:
01 Feb 2013
eISBN:
9781475564440
Language:
English
Keywords:
WP; real GDP; credit rating; economic activity; external finance; monetary policy; firm growth opportunity; loan characteristic; firm leverage; firm profitability; firms in the Worldscope database; firm in a country
Download PDF (1.4 MB)

We examine how the cost of corporate credit varies around fiscal consolidations aimed at reducing government debt. Using a new dataset on fiscal consolidations and syndicated corporate loan data, we find that loan spreads increase with fiscal consolidations, especially for small firms, domestic firms, and for firms with limited alternative financing sources. These adverse effects are mitigated substantially if consolidations are large, and can be avoided if consolidations are also accompanied with more adaptable macroeconomic policies and implemented by a stable government. These findings suggest that lenders price the short-term recessionary effects in loans but large consolidations can reduce or undo the increase in spreads, especially under favorable country conditions, by signaling credibility and creating expansionary expectations.

Abstract

We examine how the cost of corporate credit varies around fiscal consolidations aimed at reducing government debt. Using a new dataset on fiscal consolidations and syndicated corporate loan data, we find that loan spreads increase with fiscal consolidations, especially for small firms, domestic firms, and for firms with limited alternative financing sources. These adverse effects are mitigated substantially if consolidations are large, and can be avoided if consolidations are also accompanied with more adaptable macroeconomic policies and implemented by a stable government. These findings suggest that lenders price the short-term recessionary effects in loans but large consolidations can reduce or undo the increase in spreads, especially under favorable country conditions, by signaling credibility and creating expansionary expectations.

I. Introduction

One legacy of the global financial crisis of 2007-08 is the increase in government debt levels in many advanced economies. The debt-to-GDP ratio has reached a historical peak in the aftermath of the crisis as governments used public funds to bail out the troubled financial systems and to provide stimuli to their ailing economies. As a consequence, fiscal consolidation needed to bring debt levels back to sustainable levels is currently a major policy issue and is likely to remain so in the coming years.1

A particularly important question is how efforts to reduce public debt levels affect economic activity. Many studies aim to provide answers to this question by focusing on the aggregate effects of fiscal consolidations on economic performance (see, among others, Giavazzi and Pagano, 1990, 1996; Alesina and Perotti, 1995, 1997; Alesina and Ardagna, 1998, 2010; Clinton et al., 2010; Leigh et al., 2010; and Cottarelli and Jaramillo, 2012). On the long-run effects, there is consensus that public debt reduction has benefits for economic performance. Lower public sector borrowing requirements translate into lower real interest rates, reducing crowding out and spurring private investment. On the short-run effects, the consensus disappears. On the one hand, the Keynesian view asserts that cutting spending or raising taxes put brakes on aggregate demand and, hence, reduces economic activity in the short run. On the other hand, the “expectation and credibility” view argues that fiscal consolidations can be expansionary even in the short run because reduced sovereign default risk and restored credibility lowers interest rates, creating wealth effects, and improves the economy’s prospects.

In this paper, we focus on how the cost of corporate credit varies in the short term with fiscal consolidations that are aimed at reducing government debt.2 Particularly, we seek to answer whether and how banks price in the potential effects of fiscal consolidation when lending. One of the main differences between the “short-run pain” argument (for example, Guajardo et al., 2011) and the “expansionary fiscal consolidations” view (for example, Alesina and Perotti, 1997) is whether the anticipated favorable impact of fiscal consolidation on the economy will materialize in the short term or not. In this regard, a priori, the sign of the relationship between fiscal consolidations and the cost of credit is ambiguous. If lenders price in the reduced default risk premium and the expected improvement in the long-run growth outlook, then the cost of credit would go down, encouraging investment and consumption. If lenders, instead, are more concerned about Keynesian effects or do not find the efforts credible, the cost of credit would increase, reflecting expectations of lower corporate profitability and reduced ability to make timely repayments.3

Using a loan-level panel dataset and detailed fiscal consolidation measures based on government actions to reduce public debt levels in16 advanced economies over the period 1990-2011, we find that fiscal consolidations are positively associated with spreads in the syndicated loan market. Specifically, loan spreads are higher for small firms, domestic firms, and firms that have limited access to alternative funding such as those without access to international markets through issuance of American depositary receipts (ADRs).

The adverse relation between spreads and fiscal consolidation is mitigated substantially when consolidations are large. This suggests that, when consolidations are not large enough, lenders worry about possible short-term recession due to fiscal consolidation and/or do not find the efforts of consolidation credible in achieving the desired improvement in the economy. As a result, they raise the cost of credit rather than price in the potential benefit of debt reduction on default risk. The expectation of better economic performance in the future and reduced sovereign default risk weigh favorably in pricing loans when consolidations are large, and reduce the adverse effects of consolidation considerably. The fact that the findings are largely driven by small firms, domestic firms, and by firms with limited financing options further supports this interpretation because these are the firms that are subject to a higher degree of asymmetric information (and, hence, suffer more from procyclicality of lending standards) or have a higher degree of exposure to domestic economic policies.

Following Ardagna (2009), Leigh et al. (2010), Guajardo et al. (2011), and Alesina et al. (2012), we also explore whether country conditions matter. Our results indicate that consolidations accompanied by more adaptable policy frameworks featuring accommodating monetary policy stance, flexible exchange rates, and currency devaluations, and implemented by stable governments do not suffer from the short-run adverse effects on the cost of corporate credit.

Examining the details of fiscal consolidation packages, we find that corporate tax hikes are one of the major drivers of higher spreads, whereas social benefit cuts and value-added tax (VAT) hikes are associated with lower spreads. Thus, when pricing loans, lenders seem to account for the possibility of decreasing corporate profitability due to corporate tax hikes. In contrast, social benefit cuts, which are politically sensitive and have a direct effect on corporate profitability to the extent that corporations bear some of the costs through contributions, and VAT hikes, which are less prone to creative accounting, are considered favorable.

We also exploit the variation across sectors in the sense that some measures may have a disproportionate effect on certain sectors. For example, consumption-related tax hikes should affect the firms in the retail sector more adversely than they do other firms. The existence of a significant relation between fiscal consolidations and the cost of credit in sectors that are more directly affected compared to others gives some support to a causal link running from the fiscal consolidation measures to the cost of credit. We consider hikes in consumption, bank, and health care taxes, and spending cuts in transportation, health care, and defense. While we do not find any significant adverse effect of spending cuts on the cost of credit for the related sectors, we find that loan spreads are higher for firms that are in the sectors that face tax hikes, e.g., banking sector with bank tax hikes and health care sector with health care tax hikes. Thus, lenders appear to price in the potential decrease in the profitability of firms in the sectors that experience tax hikes due to fiscal consolidation.

The literature on corporate borrowing costs in relation to fiscal consolidations is rather limited. The existing studies use country-level aggregate data on corporate bonds and the findings tend to vary from one study to another: for instance, Ardagna (2009) reports a negative relationship between fiscal consolidations and corporate bond yields while Laubach (2009) finds an insignificant relationship.4 To the best of our knowledge, ours is the first study that examines how the cost of corporate credit varies with fiscal consolidations by focusing on syndicated loan market and using a detailed micro-level dataset that exploits information on firm and loan characteristics. By taking advantage of the richness in data regarding variation across firms, we can shed light on how fiscal consolidation affects credit conditions of different corporations. Furthermore, we use measures of fiscal consolidation that are based on government actions that are announced and implemented to reduce the public debt level, and, hence, are exogenous to economic developments, as articulated in Romer and Romer (2010).

Overall, our results have important economic and policy implications. In particular, the long-term gains of debt reduction may come at short-term pain in corporate financing conditions as lenders price in the possible negative impact of fiscal consolidation on growth, creating upward pressure on corporate borrowing costs. Such pain is likely to be borne by small firms, domestic firms, and by firms that have limited financing sources. Also the effects are likely to be stronger for the average firm in a country than those reported in this paper. This is because our sample is comprised of syndicated loans, and the average firm in a country is likely to be smaller and more constrained in terms of alternative financing options than those that have access to syndicated loan markets. Given the plausibility of such an adverse effect in corporate debt markets, policymakers may need to carefully contemplate the design of the measures in fiscal consolidation packages and perhaps complement the package with safeguards for the most vulnerable, especially if credit conditions are already tight, to minimize the damage in the short run. In particular, if consolidations are large enough to convey credibility of these efforts in improving the economic outlook and reducing sovereign default risk, then the adverse effects of consolidation on spreads are mitigated substantially. Furthermore, these unfavorable developments in corporate credit markets can be avoided if consolidations are accompanied by accommodating macroeconomic policy and implemented by stable governments.

The rest of the paper is organized as follows. Section II describes the channels through which fiscal consolidation may have a bearing on corporate borrowing costs. Section III explains the data and the empirical approach. In Section IV, we present and discuss the results on the relation between cost of credit and debt-reduction driven fiscal consolidations. Section V concludes.

II. Background on the Potential Channels of Transmission

There are several channels that could link fiscal consolidation to the cost of credit. The related literature has looked at the potential impact of fiscal consolidation on the economy from various aspects, namely, the demand side, the supply side, and expectations formation. We summarize the major channels identified in this literature, with an eye on the implications for the short-run relationship between consolidations and cost of credit.

According to the Keynesian view, cutting spending or raising taxes put brakes on aggregate demand and, hence, reduces economic activity in the short run. The projected decrease in the economic activity, as a result, will push up the borrowing costs, especially for those firms that are likely to be affected the most from the recessionary effects of fiscal consolidations. Hence, one would expect to see an increase in the cost of corporate credit around the announcement and after the immediate implementation of fiscal consolidations. Such an effect would be more pronounced for the “marginal” borrowers, who are more subject to asymmetric information and, hence, have limited access to finance, given the procyclicality of lending standards.

The credibility channel, on the other hand, predicts lower borrowing costs and is based on a supply-side argument (Alesina and Perotti, 1997). Large fiscal consolidations, particularly in a high-debt country, reduce the default risk premium as the consolidation enhances the credibility of the government in meeting its obligations. The reduction in the default premium could be reflected as reduced spreads in the corporate loan market. In addition, the expectation that public borrowing requirements will decline as a result of fiscal consolidation could mitigate the crowding-out effect of public debt, increase the available funds for issuance of private debt, and lower loan spreads as a result.

The magnitude of the impact could vary with the measures actually taken to accomplish fiscal consolidation. If the consolidation features direct tax increases and expenditure reduction through cuts focusing on more politically-sensitive areas, such as government wages and welfare spending, the improvement in credibility may be more pronounced. Alternatively, cuts in government spending in less politically-sensitive areas such as public investment and revenue-raising measures focusing on indirect taxes may lead to the perception that the desire to reduce the deficit and bring down the public debt level is not strong. Such a perception may also feed suspicions of “creative accounting,” further weakening the credibility of the fiscal consolidation package.

The expectations channel provides another, credibility-related explanation for “expansionary fiscal contractions” (Giavazzi and Pagano, 1990). In this interpretation, large fiscal adjustments can be expansionary because expectations of such adjustments being successful are higher (compared to less ambitious adjustments). This channel can be combined with the credibility channel discussed above. If fiscal adjustment is expected to be successful, then loan spreads should decrease. Thus, improvement in loan terms when fiscal consolidations are large can be interpreted as evidence in support of both the credibility and expectations channels.5

Overall, tightening of fiscal policy with an objective to reduce budget deficits and lower public debt levels can have opposing effects. On the one hand, such fiscal consolidations reduce the default risk premium that arises from sovereign debt and lead to expectations of better economic performance by reducing sovereign debt levels. This improvement could be transmitted to the corporate debt market through a lower cost of credit. On the other hand, if fiscal consolidations are not considered credible in reducing deficit and improving economic outlook, or if the growth prospects of corporations and their ability to make timely repayments worsen due to falling aggregate demand, borrowing conditions can deteriorate. We examine these possibilities in the following sections using corporate syndicated loan data and a detailed dataset on sovereign-debt–reduction-driven fiscal consolidations.

III. Data and Methodology

A. Data

The data used in this study come from several sources. Syndicated loan characteristics are from Dealogic. Firm-level financial and accounting data are from Worldscope. Loan-level data are matched with firm-level data manually. The majority of the macro-level data are from the OECD, IMF and the World Bank. Sovereign ratings are from Moody’s. Government debt data are from Reinhart and Rogoff (2011). Indices summarizing information on government stability, law and order, and corruption come from the PRS Group’s International Country Risk Guide (ICRG). The final dataset has 4,967 observations, covering 16 countries over the period 1990-2011. The list of variables, along with the definition and the data source for each variable, is in Appendix Table 1. The number of observations for each country and the number of observations in each year are given in Appendix Table 2.

Fiscal Consolidation Measure

Methodology

Fiscal consolidation measures we use are constructed following the methodology described in Devries et al. (2011), which in turn builds upon the approach depicted in Ramey and Shapiro (1998), Ramey (2011), and Romer and Romer (2010). The measure of fiscal consolidation is based on “actions” rather than the actual budgetary outcomes, which would be affected by numerous factors determining the pace of economic growth. More precisely, only the measures announced and implemented with an aim to reduce the public debt level are recorded as a fiscal consolidation. These measures could be spending cuts or tax hikes, and are expressed in percent of GDP at the time of their implementation.

This choice of focus on actions rather than actual outcomes reflects a major criticism directed at the usage of cyclically-adjusted primary balance (CAPB), in percent of GDP, as the measure of fiscal consolidation. In particular, CAPB may misestimate the cyclical adjustment component and, more importantly, may record a significant consolidation when there is a positive income growth shock or omit fiscal consolidation efforts to reduce debt levels when a recession hits and some of these efforts get offset by countercyclical stimulus measures. Therefore, as discussed in Devries et al. (2011) and Guajardo et al. (2011), the use of CAPB as the measure of fiscal consolidation may lead to a bias towards finding a positive association between GDP growth and the size of fiscal consolidation.6 In contrast, the definition we use to identify fiscal consolidation focuses on discretionary decisions to cut budget deficits and reduce public debt levels, independently of the developments in economic activity. Thus, our measure of fiscal consolidation is exogenous, as further discussed in Romer and Romer (2010).

Consolidation episodes

The fiscal consolidation measure is calculated for 16 advanced economies (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, and United Kingdom) using information from Devries et al. (2011), IMF Article IV Staff Reports, OECD Country Reports, and national budgets. At the end, 116 consolidation programs between 1990 and 2011 are considered. Table 1 shows the list of fiscal consolidation episodes. There have been efforts to reduce public debt levels in all of the 16 countries during our sample period. Some countries have been more active in these efforts than others (for instance, comparing Italy to Ireland). Also, the 1990s and 2010-11 have been more active than the early 2000s in terms of fiscal consolidation episodes. We also distinguish between small and large consolidations. Table 1 shows consolidations exceeding 1.5 percent of GDP, which are deemed to be large, in bold font. Of the 116 consolidations identified, 30 were large according to this cut-off point.

Table 1.

Fiscal Consolidation Episodes in the Sample Countries, 1990-2011

(in percent of GDP)

article image
Sources: Devries et al. (2011); authors’ calculations based on IMF and OECD reports and national budgets.Note: Large fiscal consolidations, defined as those greater than 1.5 percent of GDP, are shown in bold.
Details of fiscal consolidation packages

In addition to overall fiscal consolidation as a percentage of GDP, we distinguish between tax hike and spending cut components as a percentage of GDP. We also look at the details of each fiscal consolidation package. For tax hikes, we consider increases in corporate tax, personal income tax, VAT, property tax, and health care tax. For spending cuts, we consider government capital expenditure cuts, cuts in social security contribution and social benefits, wage cuts, personnel cuts, pension cuts, health care cuts and public investment cuts. These measures take a value of one if a fiscal consolidation package includes these measures, and are zero otherwise.

Some consolidation measures can have a disproportionate effect on a number of sectors, such as the effect of consumption taxes on the retail sector. To consider these effects, we look at tax hikes related to (i) the banking sector such as increase in taxes on deposits, (ii) the retail sector such as consumption taxes, and (iii) the health care sector such as hikes in contributions to health care and increases in premiums for public health insurance. Among spending cuts, we look at spending cuts that affect the transportation sector, the health care sector, and the defense sector. Like the other fiscal consolidation details, these measures take a value of one if they are part of the consolidation package, and are zero otherwise.

Loan and Firm Characteristics

As a measure of cost of corporate credit, we use the natural logarithm of syndicated loan spreads, inclusive of all fees, measured by the difference between the rate charged on the loan and the relevant benchmark rate.7 We include a large number of loan-level characteristics that are related to the cost of credit. Specifically, we include loan maturity, the natural logarithm of loan size, the purpose of the loan such as working capital, refinancing, general purposes, etc., the currency in which the loan is denominated, the type of the loan such as term credit, revolving credit, etc., whether the deal is investment grade or not, and whether the loan is given by a domestic bank or not.

A corporation’s cost of credit is also related to a number of firm characteristics that determine the credit risk of a firm. In this regard, consistent with the related literature (see, for example, Chava et al., 2009), we include firm leverage (the ratio of total debt to total assets), profitability (net income over total assets), tangibility (tangible assets over total assets), growth opportunities (the ratio of market value of assets to book value of assets) and whether the borrower is rated or not.

Country Characteristics

In examining how cost of corporate credit varies with fiscal consolidations, it is crucial to control for country-level factors that have a bearing on both fiscal consolidations and corporate credit risk in order to reduce the risk of spurious correlations between fiscal consolidations and corporate loan spreads. We control for country-level variables found to be relevant in the fiscal consolidation literature (see, among others, Ardagna, 2009, and Guajardo et al., 2011, for more details). These controls are sovereign rating of countries (from Moody’s), total government debt (from Reinhart and Rogoff, 2011, for the period between 1990 and 2010 and from OECD for 2011)8, change in exchange rate and interest rate, level of interest rate, natural logarithm of real GDP, real GDP growth, stock market turnover, trade openness and inflation (from the IMF and World Bank datasets), government stability, degree of law and order, and level of corruption (from the PRS Group’s ICRG database).

For sovereign bond ratings, we convert each letter rating to a numeric one as follows: the highest rating, Aaa, takes the value of 21 and the value declines for each downward notch in rating, i.e., Aa1 corresponds to 20, Aa2 corresponds to 19, and so on and so forth. For total government debt, we include the general-government-debt-to-GDP ratio as a measure of the sovereign debt level. If general government debt data are not available, we use central government debt instead. The measure of government stability from the ICRG database assesses a government’s ability to carry out its declared programs and its ability to stay in office. The index consists of government unity, legislative strength, and popular support subcomponents. The measure of law and order is an assessment of the strength of the legal system and popular observance of the law. The measure of corruption from the ICRG database includes two subcomponents: financial corruption that makes it difficult to conduct business effectively (e.g., bribes) and political corruption such as “quid pro quo” and suspiciously close ties between government and business. Higher values of these measures correspond to a more stable government, to a more orderly society, and to less corruption (“lack of corruption”).

B. Methodology

Our empirical specification treats the yield spread on a loan as the dependent variable. Reflecting the richness of the dataset, the specification controls for a battery of loan, firm, and country characteristics. Specifically, we estimate the following regression equation:

LoanSpreadijkt=β1⁢Fiscalkt+β2⁢Countrykt+β3⁢Loanjt+β4⁢Firmit+ck+nĭ+yt+ɛijkt(1)

where LoanSpread is the difference between the rate charged for firm i (headquartered in country k) on loan j (issued in year t) and the benchmark rate as determined by the lender(s). Fiscal is the amount of fiscal consolidation in percent of GDP (or other measures related to fiscal consolidation such as spending cuts and tax hikes as percentage of GDP or details of the fiscal consolidation package) in country k in year t, Country is the set of country-level controls, Loan is the set of loan-level controls, and Firm is the set of firm-level controls. In addition to these time-varying controls, fixed effects for each country, industry, and year are included. Robust standard errors are clustered by country.

Our main variable of interest in the baseline is the contemporaneous total fiscal consolidation in percent of GDP. In separate regressions, we substitute this variable with more detailed measures of the fiscal consolidation effort in question. In particular, we distinguish between tax-based versus spending-based fiscal consolidations as well as the details of fiscal consolidation packages such as taxes on consumption, corporate income, personal income, property, and health care, and cuts in government transfers (social benefits, unemployment benefits), government consumption (health care, pension, personnel), public investment, and social security contributions.

There are three main reasons as to why any association we find between our measures of fiscal consolidation and loan spreads is likely to indicate an “effect” of fiscal consolidation on the cost of corporate credit, rather than the other way around. First, the fiscal consolidation measure we use is exogenous (Romer and Romer, 2010; Devries et al., 2011) in the sense that it records discretionary changes in fiscal policy aimed at reducing government debt and, therefore, it is a response to past conditions rather than the current ones. Second, the use of loan-level data introduces a certain degree of separation such that it is more difficult to argue that the spread on a particular loan would lead the decision to engage in fiscal consolidation compared to the case where an aggregate measure of cost of credit is used. Third, any difference in results for different sectors when considering consolidation measures that are applicable to certain sectors could be interpreted as a “falsification test”.9 If the introduction of the measure coincides with a general increase in spreads in the syndicated loan market attributable to other factors, we would expect to see spreads increase in all sectors. As the results below show, this is not the case with regard to tax hikes: tax hikes in a given industry are associated with an increase in spreads for the firms operating in that industry but do not have a statistically significant relationship with spreads for the firms in other industries. Spending cuts in general do not have a significant effect on loan spreads.

That said, we should note that even though we control for a battery of firm, loan and country conditions, it is still possible that there can be a latent factor driving both the fiscal consolidation decisions and the corporate loan spreads. We include an extensive list of controls to alleviate this concern but we cannot totally eliminate it.

IV. Empirical Findings

A. Descriptive Statistics

Table 2 shows the summary statistics for the dataset on which Equation (1) is estimated. Average fiscal consolidation amounts to around 0.3 percent of GDP with a maximum of 4.5 observed for Italy in 1993.10 Tax hikes and spending cuts have comparable distributions in the sample, but spending cuts are relatively larger, with an average of 0.2 percent of GDP compared to 0.1 percent of GDP for tax hikes. Spending cuts are also more severe at the maximum than tax hikes, with a spending cut of 3.7 percent of GDP observed for Finland in 1993, and a tax hike of 2.4 percent of GDP observed for Italy in 1995.

Table 2.

Descriptive Statistics

article image
Sources: Dealogic, WorldScope, Reinhart and Rogoff (2011), PRS Group’s ICRG, IMF, World Bank, OECD, national sources; authors’ calculations.

With respect to macroeconomic characteristics, the sample exhibits a fair degree of variation. Sovereign credit ratings vary between Baa1 and Aaa, with an average of Aaa. The lowest value, Baa1, is observed for Ireland in 2010.11 There is considerable variation in sovereign debt levels.

The average government debt is around 67 percent of GDP, with the minimum around 5 percent observed for Australia in 2007 and 2008, and the maximum above 200 percent of GDP observed for Japan after 2009. There is also wide variation in short-term interest rates, where lower than 1 percent rates are observed for Japan after mid-1990s and in Sweden in 2009, and above 15 percent is observed for Italy, Spain, Portugal, and Australia from the mid-1980s to the early 1990s. On average, exchange rates change by 2 percent and interest rates change are around 15 percent. Stock market turnover variation indicates that both liquid and illiquid markets are represented in the sample. Given that the sample consists of advanced economies, average real GDP is high. Trade openness on average is 65 percent of GDP. The average inflation in the sample is around 2 percent (Japan has deflation starting in the early 1990s). As for the country-level governance variables, we consider the indices for government stability, law and order, and lack of corruption. Lowest levels of government stability are observed for Italy and Spain in the late 1980s and the early 1990s, and for a number of European countries such as Ireland, France, Italy and Spain during the eurozone crisis in 2010-11. Law and order and lack of corruption indices mostly have high values, reflecting better functioning institutions in advanced economies.

In terms of loan characteristics, the average loan spread is around 134 basis points. Around 70 percent of the loans are investment grade, 80 percent are issued in a major currency (Euro, Japanese Yen, British Pound, or the U.S. Dollar), 30 percent are given to borrowers with a credit rating, and 30 percent are extended by domestic banks. In other words, majority of loans are investment grade, in major currencies, and given to nonrated borrowers. The average loan in the sample matures in 4 years. The size of a loan is around US$ 790 million on average, varying between US$ 250 thousand and US$ 45 billion. The largest amount of US$ 45 billion was launched by the Belgian brewer InBev to buyout Anheuser-Busch in 2008. The loan had a maturity of around 3 years and spread of 200 basis points above LIBOR, inclusive of all fees. This example demonstrates the ability of large corporations to issue jumbo loans during adverse economic conditions such as the recent financial crisis. Regarding firm characteristics, we observe that the sample has both large and small firms. Similarly, the sample consists of both unlevered and levered firms, firms with low and high growth opportunities, and firms with and without tangible assets. Thus, there is reasonable variation in the sample across firm characteristics as well.

B. Baseline Regressions

We start the analysis by examining how syndicated loan spreads vary with fiscal consolidations as in Equation (1), where fiscal consolidation is measured as a percentage of GDP based on the announced and implemented government actions to reduce sovereign debt as discussed in Section III.A. Table 3 presents the results. In Column 1, we include only the country-, loan- and firm-level controls. In Column 2, we introduce the measure of fiscal consolidation in percentage of GDP in addition to all the controls in Column 1. We also allow for a differential effect of large fiscal consolidations, defined as those exceeding 1.5 percent of GDP, in Column 3.12

Table 3.

Baseline Regression Results

article image
Notes: Dependent variable is the natural logarithm of the syndicated loan spread. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.

Syndicated loan spreads are positively associated with fiscal consolidations, as shown in Column 2, suggesting that lenders price in the recessionary effects of fiscal consolidations in loans, whereas the expectation of better economic performance in the future or reduced default risk in sovereign debt does not seem to be weighed in as strongly.

It has been argued in the literature that large fiscal consolidations project positive expectations about future economic output and also are credible in reducing sovereign default risk (see, for instance, Giavazzi and Pagano, 1990, and Alesina and Perotti, 1997). To capture the potential differential effect of large fiscal consolidations, we introduce the interaction of our fiscal consolidation measure with an indicator variable that takes the value of one if the ratio of fiscal consolidation to GDP is above 1.5 percent.13 The result is in Table 3, Column 3. If fiscal consolidations are above 1.5 percent of GDP, the positive relation between loan spreads and consolidations is indeed reduced considerably.

In addition to being statistically significant, these baseline results are also economically meaningful. Evaluated at the average spread of 134 basis points, the coefficients in Table 3, Column 3 imply that an additional 1 percent of GDP in fiscal consolidation is associated with an 11 percent increase in spreads, which is in the magnitude of around 15 basis points. For a typical loan in the sample (defined by the average loan size of around US$ 800 million and the average term of 4 years), this increase in spreads corresponds to around US$ 5 million in additional borrowing costs over the life of the loan. If the consolidation is large, however, the increase in spreads is only 2 basis points. Thus, additional financing costs are negligible when consolidations are large.

This finding on the differential effect of large fiscal consolidations can be interpreted as evidence that both Keynesian recessionary effects channel and expectations channel are in play. Lenders’ emphasis on recessionary effects and/or lower credibility of fiscal consolidation in creating a better economic outlook is mitigated substantially if fiscal consolidations are large enough to reduce default risk in sovereign debt and bring about positive expectations about economic performance in the future. This finding is also consistent with the notion that large consolidations may be considered to be less likely to be reversed in the future. Hence, they tend to have a stronger permanent component than small consolidations, which can be viewed as more temporary. Further evidence that expectations matter and there is more than just the aggregate demand effect comes from the fact that we do control for real GDP growth, firm profitability, and firm growth opportunities in our baseline and still find the effects of fiscal consolidation to be positive on average.

In Table 3, Column 4, we also introduce an interaction of fiscal consolidation and large fiscal consolidation variables with an indicator variable that takes the value of one for 2010-11 to capture any change in the relationship due to the recent sovereign debt crisis in Europe. The interaction variables are not significant and the coefficients on fiscal consolidation variables are comparable to those reported in Column 3. These results suggest that the relation between loan spreads and fiscal consolidation measures is not different during the eurozone crisis.

Regarding firm and loan characteristics, and other country-level variables, Table 3 shows that both loan- and firm-level variables are major determinants of loan spreads. Among country-level factors, loan spreads decrease with increasing interest rates, in line with Longstaff and Schwartz (1995) and Leland and Toft (1996), but the coefficient is very small. Spreads also decrease with real GDP growth and the degree of law and order. These findings indicate that lenders price in the rapid growth performance in an economy while pricing loans and also reduce spreads where the legal system is functioning well. One explanation for the lack of a significant association for the other country variables is that the regressions already include country and time effects. Among loan characteristics, being investment grade, the size of the deal, and foreign currency denomination have a statistically significant relationship with loan spreads. Investment grade loans have relatively lower default risk, which is also reflected in loan pricing. Larger loans are associated with lower spreads. Since there are some fixed costs in loan originations, these fixed costs are likely to be lower for larger loan deals. Finally, loans denominated in a currency other than the borrower’s own currency carry a larger risk premium, which is reflected in the higher spread on these loans. Among firm characteristics, size, growth opportunities, and tangibility are associated with lower loan spreads, whereas leverage is positively related to spreads. Large firms have less information asymmetry and more collateral. Therefore, they get lower rates from lenders. Our results also suggest that firms with growth opportunities are considered to be lower-default-risk customers by lenders. Firms with more tangibility have more to offer as collateral and thus have lower spreads. Lastly, increasing debt levels in a corporate capital structure increases the default risk of a firm, which is reflected as higher loan prices for firms with high leverage.

C. Firm and Loan Characteristics

The baseline regressions have demonstrated a positive association between loan spreads and fiscal consolidations, which is mitigated for large consolidations. We next examine whether this finding is driven by certain firm and loan characteristics. In this regard, we differentiate between small and large firms, firms with bond ratings and without bond ratings, firms that issue ADRs and those that do not, firms that go to external markets more often and those that do not, and finally firms that have revenues from abroad and those that do not. The results are in Table 4.

Table 4.

Which Firms Face Higher Borrowing Costs around Fiscal Consolidations?

Notes: Dependent variable is the natural logarithm of the syndicated loan spread. The sample is split by firm size. Small (Large) firms are those with assets below (above) the median in a given country. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.
Notes: Dependent variable is the natural logarithm of the syndicated loan spread. The sample is split by credit rating availability. (Not) rated firms are those with (no) credit rating published. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.
Notes: Dependent variable is the natural logarithm of the syndicated loan spread. The sample is split by access to global markets. Firms with no ADR issuance are considered to have no or limited access to finance outside their country. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.
Notes: Dependent variable is the natural logarithm of the syndicated loan spread. The sample is split by external finance dependence. Firms operating in industries that have been argued to have more need for external finance, as assigned by Rajan and Zingales (1998), are considered to have high dependence. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.
Notes: Dependent variable is the natural logarithm of the syndicated loan spread. The sample is split by cross-border diversification in firm activities. A firm is considered to be multinational if its foreign sales and foreign income are nonzero. All regressions are estimated using ordinary least squares. Country, industry, and year fixed effects are included. Also included are indicator variables for the purpose and type of loan. Robust standard errors are clustered at the country level. Statistically significant coefficients are shown in bold. ***, **, and * denote significance at the 1, 5, and 10 percent level, respectively.