Firm characteristics and bank loan distribution: Who borrows in Austria? (OeNB Bulletin Q1/25)

Bernhard Hirsch, Aleksandra Riedl, Stefan Trappl ¹

JEL classification: G21, G32, E44, D22, L25

Keywords: bank loan distribution, nonfinancial companies, firm size and financing, Austria

This paper provides the first detailed analysis of the characteristics of Austrian nonfinancial companies that hold bank loans and investigates how these loans are distributed among borrowers. Utilizing a novel and comprehensive dataset, we are able to offer a thorough overview of bank loan allocation within the Austrian corporate sector, highlighting patterns of credit reliance across firms of varying sizes, asset structures and profitability levels. These findings provide valuable insights into the role of bank loans in corporate financing and have critical implications for policy design and implementation.

Understanding the distribution of bank loans among nonfinancial firms is essential for both macroprudential and monetary policy. From a macroprudential perspective, analyzing which firms hold bank loans and the concentration of loan volumes helps identify potential systemic risks to financial stability. Larger firms with significant loan volumes pose heightened risks, as financial distress in these firms could have widespread economic repercussions (Bernanke et al., 1996). Furthermore, the high concentration of bank loans among a few major borrowers amplifies this risk, as defaults by these firms could trigger systemic shocks, destabilizing banks and the broader economy (Rao et al., 2015). For monetary policy, the reliance of firms on bank financing sheds light on the economy’s sensitivity to fluctuations in interest rates and credit conditions. When a substantial proportion of firms depends on bank loans, changes in credit availability or borrowing costs can significantly impact business operations, potentially amplifying economic cycles (Bernanke and Gertler, 1995; Kashyap and Stein, 2000). This understanding enables policymakers to better evaluate the scope and immediacy of policy transmission across the corporate sector.

Despite the importance of these issues, there has been little evidence on how bank loans are distributed across Austrian companies or on the role of bank loans within their debt structures. Studies using firm-level data (e.g. Beer and Waschiczek, 2019) provide valuable insights into the relationship between equity and debt financing, but generally, i.e. for almost all firms, lack detailed information on firms’ liabilities, including whether firms hold bank loans. Consequently, it has not been possible to conduct a comprehensive analysis of the importance of bank loans for Austrian nonfinancial companies at the firm level, leaving a gap in understanding the full scope and structure of corporate bank debt reliance in Austria.

Our study fills this gap by using a unique dataset, the Integrated Firm-level Database (IFLD), compiled by the central bank of Austria (OeNB). This database allows us to combine loan-level data from AnaCredit ² , which contains information on bank loans granted by Austrian credit institutions, with firm-level financial statement data from companies required to disclose their annual financial statements. This combination enables us to distinguish between firms with and without bank loans and to analyze how the prevalence and distribution of bank loans relate to key firm characteristics. Furthermore, the dataset fulfills a key prerequisite for analyzing the distribution of bank loans among Austrian firms: the precise definition of the target population. We define this population as all nonfinancial companies that have to disclose their annual financial statements (henceforth, firms), including, among others, all public limited companies (AG) and private limited companies (GmbH). The IFLD includes all firms within this group. For analyses requiring balance sheet information, however, the sample excludes firms without such data. Nevertheless, defining the target population allows us to assess the representativeness of the sample and ensure its alignment with the broader Austrian corporate sector.

The paper is organized as follows: Section 1 reviews key insights from the literature on Austrian firms’ funding structures, while section 2 presents our dataset. Section 3 provides a descriptive examination of the prevalence of bank loans among firms, and section 4 investigates the characteristics of firms that rely on bank financing. A logit regression highlights key differences between firms with and without bank loans, without aiming to establish causality. Section 5 analyses the concentration of bank loan amounts and their role within firms’ liabilities. A box section provides an in-depth look at debt structures for a subset of firms. Finally, section 6 summarizes our findings and discusses their implications for monetary and macroprudential policy.

1 Setting the scene – three stylized facts from the literature

In this section, we review the literature on the funding structure of Austrian nonfinancial companies, focusing on the importance of bank loans. Research on this topic is limited, but we highlight three key findings that serve as a foundation for our analysis.
First, Austrian nonfinancial companies are mainly debt- rather than equity-financed, i.e. the equity ratio is below 50%. This pattern is consistent across data sources (Breyer et al., 2021; Beer and Waschiczek, 2019; Elsinger et al., 2016; Wiesinger, 2015). Aggregated data from the financial accounts of Austria indicate that debt capital exceeds equity capital, with Elsinger et al. (2016) reporting a mean equity ratio of 44.9% in 2014. ³ Similarly, micro-level data from the SABINA database reveal a median equity ratio of 34% in 2016 (Beer and Waschiczek, 2019), indicating that debt is the primary financing source for at least half of Austrian nonfinancial companies. Notably, in a European context, Austria’s equity ratios rank below average (Kristofik and Medzihorsky, 2022; Elsinger et al., 2016).

Second, equity ratios of nonfinancial companies are very heterogenous across firms and industries (Breyer et al., 2021; Beer and Waschiczek, 2019). For example, firms with a high proportion of tangible assets tend to have lower equity ratios, perhaps because tangible assets are easier to value and use as collateral for secured debt. This phenomenon may also explain industry-level differences: firms in sectors like accommodation, energy and construction – where tangible assets are abundant – typically exhibit lower equity ratios.

Third, in an international comparison, bank loans appear to be an important source of financing for Austrian companies in terms of their aggregate volume and their prevalence. According to financial accounts data, bank loans constitute a major share of the corporate sector’s debt finance (Elsinger et al., 2016; Wiesinger, 2015). In 2014, Austria had the third-highest share of bank loans in total assets (20.4%) among 14 European countries (Elsinger et al., 2016), a figure that rose to 20.9% by Q4 2023, making Austria the leader in this category. This share is more than double the EU-20 average (9.9%), underscoring the importance of bank loans for Austrian firms. While bank loans are clearly critical for the corporate sector as a whole, little is known about how they are distributed across companies or their role within firms’ total liabilities. One exception is the proportion of companies with bank loans. Lawless et al. (2015), using the Survey on the Access to Finance of Enterprises (SAFE) ⁴ data from 2010–2013, report that 39.8% of Austrian small and medium-sized enterprises (SMEs) had used bank loans, the fourth-highest share among 16 European countries and slightly above the euro area average of 39%. ⁵ Although current figures are unavailable due to changes in SAFE survey questions after 2015, related questions continue to highlight the importance of bank loans for Austrian companies.

2 Data

Our analysis utilizes data from the Integrated Firm-level Database (IFLD) compiled by the Austrian central bank (OeNB), focusing on 2021 as it is the most recent and complete dataset. The IFLD is well-suited for analyzing the bank loan distribution among Austrian nonfinancial companies due to two key features. First, it includes multiple data sources, which is critical for a comprehensive view of corporate financing. For our analysis, we rely on two main datasets: (1) firm-level balance sheet data and (2) loan-level data detailing loans from Austrian banks to domestic companies. Previous studies on Austrian company financing have generally used financial statement data alone (e.g. SABINA dataset), which lack detailed liability information needed to identify all bank-financed firms. By combining balance sheet and loan-level data, the IFLD allows us to study both companies with and without bank loans, facilitating an in-depth analysis of loan distribution and associated factors across the corporate sector.

The second key feature of the IFLD is that it includes all Austrian companies listed in the commercial register (Firmenbuch), excluding sole proprietorships, providing a broad and representative basis for analysis. ⁶ This coverage ensures that the dataset comprehensively captures the target population, which we define as all (1) nonfinancial companies that (2) are obliged to disclose their annual financial statements (N=194,263 companies). While the IFLD covers all firms within this defined population, the sample used in our analysis may exclude certain firms lacking complete balance sheet information, particularly for analyses requiring detailed financial data. However, the precise definition of the target population allows us to evaluate the representativeness of our sample and its alignment with the broader Austrian corporate sector, minimizing the risk of biased outcomes.

In our analysis, we distinguish between two categories of bank loans which are based on the financing purpose: revolving loans and nonrevolving loans. Revolving loans, such as overdrafts and credit lines, offer firms flexible access to short-term liquidity, typically to cover working capital needs. In contrast, nonrevolving loans, such as lump-sum loans and financial leases, are used for long-term investments and are disbursed as a single payout. This distinction between loan types is crucial for understanding firms’ financing strategies and their reliance on bank loans for different operational needs. For a detailed description of the dataset, methodology and supplementary tables, refer to annexes I and II.

3 The prevalence of bank loans among firms

Using the full target population defined in the data chapter – comprising all Austrian nonfinancial companies required to disclose their annual financial statements (henceforth “firms”) – table 1 presents key statistics, revealing that 35.9% of these firms held bank loans with Austrian banks as of the end of 2021. This share aligns closely with findings by Lawless et al. (2015), who reported 39.9% for Austrian firms from 2010 to 2013, and by Hooks (2003), who noted 35% for US firms.

Interestingly, bank-financed firms hold a disproportionately high share of (aggregate) assets. While only 35.9% of firms have bank loans, these firms control 56% of total assets (not reported in table 1), highlighting their greater economic weight. ⁷ Still, if only 35.9% of firms have bank loans, how do the others finance themselves? Box 1 delves into this, examining alternative financing methods among a smaller subset of firms.

Although 35.9% may seem low given the dominance of bank financing reflected in aggregated data, this indicates that a substantial loan volume is concentrated among those firms that do have bank loans. As table 1 shows, the average loan amount among bank-financed firms is EUR 2.45 million. Interestingly, the median is only EUR 374,000. The resulting mean-to-median ratio (6.6) reflects a high concentration of loan volumes among bank-financed firms. This can be attributed to the fact that firms are very different in terms of size, with few very large ones holding far higher amounts. Appendix table A2.1 further illustrates the positive relationship between firm size and loan volume, which we explore in detail in section 5.

When considering the share of bank-financed companies broken down by loan type in table 1, one can see that more firms have nonrevolving loans, where the average outstanding loan amount is also significantly higher. This is not surprising, since nonrevolving loans typically finance long-term investments and are thus larger in volume than revolving loans, which are usually taken out to finance day-to-day business transactions.

The left-hand panel in chart 1 shows that out of the 28.1% of firms holding nonrevolving loans, almost half (13.2%) also have revolving loans and that only 7.8% of all firms rely solely on revolving loans. While the difference between the type of loan holders (i.e. 28.1% versus 21.0%) is not large, the total outstanding amount differs greatly between the two broad loan types, with nonrevolving loans predominating by far. Hence, a large proportion of the total volume granted to firms by Austrian banks is distributed among relatively few borrowers, as shown in chart 1. Out of all firms, 28.1% (14.9%+13.2%, left-hand panel) hold 83.8% of the total loan volume as of the end of 2021 (right hand panel).

Finally, returning to table 1, we present the average number of loans and bank connections per firm. The median bank-financed firm has two bank loans. This figure reduces to one loan for the median firm if we consider the subgroup of bank-financed firms with revolving loans and the subgroup with nonrevolving loans. Again, the conditional mean is higher compared to the median, indicating that some bank-financed firms have a far higher number of bank loans. Interestingly, regarding the number of bank connections, the median bank-financed firm has only one banking relationship, which also applies to the two subgroups of loan holders. As reported by Kosekova et al. (2023), the number of banking relationships varies significantly across euro area countries. According to their analysis, Austria ranks in the mid-range, together with countries like Germany and France, while the Netherlands and Ireland are closer to the one-bank model.

4 The characteristics of firms that have bank loans

Despite the absence of a specific theory outlining which firms opt for bank loans, we integrate the insights from three distinct literature streams to analyze the characteristics of Austrian bank-financed firms: capital structure theory, credit constraints and the diversification of debt structures. We also reference empirical data from the Survey on Access to Finance (SAFE).

The capital structure literature focuses on the trade-offs between debt and equity financing, with various theories often presenting conflicting predictions about firm behavior (Myers, 2001). An extensive empirical study by Frank and Goyal (2009) identifies key predictors of firm leverage, showing that larger firms with significant tangible assets tend to have higher leverage, while profitability negatively impacts debt levels. Similar findings are echoed in Rajan and Zingales (1995). The positive relationship between firm size and leverage stems from the lower default risk that larger firms face. According to the trade-off theory, firms evaluate the benefits of debt, like tax advantages, against costs such as bankruptcy risk. Since larger, more diversified firms have a lower default risk, they are expected to have higher leverage. A similar rationale applies to firms with a large amount of tangible assets, as these are safer than intangible assets, reducing business risk and encouraging higher debt levels (Kraus and Litzenberger, 1973). Finally, the negative impact from profits aligns with the pecking order theory, which posits that more profitable firms use less external financing due to greater internal funds, leading to lower leverage (Myers, 1984). Although this literature stream focuses on the debt-equity choice rather than the stock of bank loans, it is presumable that higher-leveraged firms are more likely to secure bank loans. Thus, size, profitability and asset composition are potential critical factors influencing the likelihood of seeking bank loans.

In addition to capital structure, literature on credit constraints highlights that smaller and younger firms are more likely to struggle with accessing bank financing (Holton et al., 2014). This aligns with Mac an Bhaird et al. (2016), who identify discouraged firms – those avoiding credit applications due to perceived rejection risks – as typically smaller and younger. Devos et al. (2012) also note that firms with zero debt are often younger and smaller, reinforcing the link between age, size and access to financing. Younger firms may face challenges in securing loans due to limited operating history and the risks associated with new ventures. Additionally, larger firms often enjoy greater financial stability and a more diversified revenue stream, which reduces perceived risk for lenders. Thus, if we assume that some Austrian firms face credit constraints, we should observe that younger and smaller firms are less likely to have bank loans.

The third stream of literature addresses the diversification of debt structures . Research by Lawless et al. (2015) suggests that older and larger firms exhibit more diversified financial structures, which increases the likelihood of having bank loans as part of their financing portfolio.

Finally, results from the SAFE – where firms are asked if they had taken out or renewed a loan in the past six months – reveal that bank borrowing increases with firm size (Kwaak et al., 2021). While the survey does not directly ask about the stock of bank loans, it likely correlates positively with borrowing, suggesting that larger firms are also more likely to hold bank loans.

The empirical model

Given these literature insights, we will empirically examine the association between the specific firm characteristics of size, tangibles, profits, and age, and the likelihood of having a bank loan. We employ a logit model (e.g. Wooldridge, 2019), where the dependent variable is a binary dummy variable, coded as 1 for firms with bank loans (i.e. $D_{i}^{Loan} = 1\ \text{if}\ {ONA}_{i} > 0)$ and 0 for those without loans. In addition to firm characteristics, we include control variables such as the legal form of the firm, the NACE industry classification and geographical region to account for variations in the data. We report our results in table 2.

We start with specification 1, which does not differentiate between loan types. As predicted by the trade-off theory and in line with the insights from the literature on credit constraints and diversification, firm size enters with a positive sign. The results show that medium-sized firms, i.e. those belonging to the second and third quartile in terms of their total assets, have a 26 percentage points higher probability of securing a bank loan compared to firms with total assets of less than EUR 130,000. This increase in probability is even higher (+36 percentage points) for large firms with assets of more than EUR 2 million, clearly highlighting a strong relationship between size and access to bank financing.

A comparably strong connection is found between tangible assets and bank loans. The likelihood of securing loans increases by 17 percentage points for firms with positive tangible assets making up less than 35% of their total assets, compared to firms with no tangible assets (i.e. 30% of all firms). This trend is even more pronounced (+36 percentage points) for firms with tangibles comprising 35% or more of total assets. This finding aligns with the trade-off theory. Another key reason why tangible assets are so critical for bank financing, perhaps even more so than the leverage implications suggested by the trade-off theory, is that these assets can be pledged as collateral, significantly lowering lender risk. This makes it easier and more cost-effective for firms with high tangible assets to access credit.

The estimated relationship between profitability and bank loan prevalence aligns with the pecking order theory. However, this relationship is not as strong as with the previous two variables. Firms with zero or positive returns making up less than 15% of their assets – which applies to around 40% of all companies – are 1.5 percentage points less likely to secure loans compared to those firms with no or negative profits. This relationship becomes more pronounced (–11.8 percentage points) for firms with a return on assets of 15% or more.

Interestingly, the relationship between firm age and bank loan prevalence varies significantly across the age distribution. As predicted by the literature on credit constraints, age has a positive effect on loan prevalence, but only for firms in the early stages of their life cycle. Compared to newly founded firms (up to 2 years old), those aged 3 to 6 years are significantly more likely to have a bank loan, though the impact is relatively small. However, as firms grow older, they become progressively less likely to have bank loans compared to newly established firms. This finding suggests that credit constraints may primarily affect firms at the beginning of their life cycle, while older firms are likely to accumulate retained earnings over time, reducing their reliance on external financing. With more internal funds available, these firms may prefer equity financing to avoid debt obligations, aligning with the pecking order theory. Additionally, as firms age, they often build stronger reputations and networks, potentially granting them access to alternative financing sources such as trade credit, bond issuance or private equity.

In specifications (2) and (3) we differentiate between nonrevolving and revolving loans. We want to highlight three key differences between these two models. First, tangibles play a much greater role for nonrevolving loans (2), as evidenced by a higher marginal effect for firms in the high tangibles category (compared to model 3). This might suggest that banks place a high value on collateral for nonrevolving loans due to their larger loan amounts and longer-term commitments, mitigating risk with tangible security. Second, the observed relationship between firm age and holding bank loans differs for revolving credit. While the positive relationship for nonrevolving loans reverses as firms age, this is not the case for revolving credit. Firms of all ages appear to have a slightly higher likelihood of holding revolving credit compared to newly founded firms. This could indicate that newly established companies face challenges in accessing short-term liquidity.

Finally, the relationship between profitability and bank loan holdings is less pronounced for nonrevolving loans and even points to the opposite direction for firms with medium profitability (i.e.: [0%,15%]). These results still suggest that higher profitability reduces the need for bank loans, as firms can rely on retained earnings to fund their operations. However, for larger investment loans (nonrevolving), this effect could be offset by banks actively selecting more profitable firms, which they consider more reliable in repaying substantial, long-term debt. This makes them preferable candidates for these investment-focused loans. This creates a dual dynamic: While profitable firms generally have less need for external financing, banks still favor them for larger loan amounts due to the perceived financial security they offer. In contrast, for revolving loans aimed at working capital, high profitability correlates with a lower likelihood of seeking a loan, as firms can manage short-term needs internally, reducing the demand for external working capital financing.

5 The distribution of loan volumes among bank-financed firms

This section focuses on the subset of firms that have bank loans. As discussed in the introduction, understanding the concentration of bank loan volumes is critical from a financial stability perspective. Chart 2 displays a Lorenz curve illustrating this concentration among bank-financed firms. The blue line represents perfect equality, where each firm holds an identical loan volume, while the red line shows the actual distribution of bank loans across Austrian firms as of the end of 2021. The results reveal a high degree of concentration: 10% of firms hold 76% of the total loan volume, and just 1% of firms account for 43% of the total.

While these findings are vital for financial stability, given that high concentration heightens systemic vulnerability to economic shocks, they are not entirely surprising. Two factors explain this distribution: first, the nature of firm size distribution in Austria, and second, the correlation between firm size and bank loan volumes for bank-financed firms.

Starting with firm size distribution, chart 3 provides the Lorenz curve for firm size, measured by total assets. The green line represents the distribution for all firms (including those without bank loans), while the red line shows the size distribution for bank-financed firms. Two aspects are particularly noteworthy. First, firm size is highly unequal in both groups, as evidenced by the wide gap from the equality line (blue). Second, the selection of only bank-financed firms slightly reduces this inequality, though the overall concentration remains pronounced. Notably, 10% of bank-financed firms hold 85% of the total assets among these firms.

Chart 4 highlights the strong correlation between firm size and loan volume for bank-financed firms. This relationship, noted briefly in section 3, alongside the high inequality in firm size, explains why a few firms hold significant volumes of outstanding bank loans.

Interestingly, however, when measured as a share of firms’ total liabilities, bank loans are distributed more uniformly across firm sizes. This is illustrated by chart 5 showing the loan-to-debt ratio for bank-financed firms, broken down by firm size percentile, and differentiates between revolving (blue bars) and nonrevolving (red bars) loans. For each percentile, we calculate the average share of revolving and nonrevolving loans in firms’ total liabilities. The loan-to-debt ratio generally remains stable between 50% and 60% across different firm sizes, except in the top percentiles, where the ratio drops. ⁹ Due to data limitations on debt structure (outlined in section 2), further analysis is challenging, though it is plausible that larger firms substitute bank loans with bonds. This assumption is supported by our dataset. Among the 59 firms with bond data available (i.e. non-missing and non-zero entries), the vast majority are concentrated at the upper end of the size distribution. Specifically, 97% of these firms are within the top 10% of all firms in terms of total assets, and an impressive 86% are positioned within the top 3%.

While further research is needed to fully understand why the loan-to-debt ratio remains almost constant across firm sizes, it is notable that similar findings are observed in other contexts. For example, Colla et al. (2020) analyze over 4,500 US public firms from 2002 to 2018 and find that bank loans account for approximately 59% of total debt for firms with bank debt, aligning closely with the stable ratio observed here.

6 Conclusions

This study offers the first comprehensive analysis of the bank loan distribution among Austrian nonfinancial firms using data from the Integrated Firm-level Database (IFLD). Our findings provide a nuanced view of how bank loans are allocated within the Austrian corporate sector, with implications for monetary and macroprudential policy. Three key insights emerge from our analysis.

First, we find that 35.9% of Austrian nonfinancial firms hold bank loans, a figure that may appear unexpectedly low, especially considering the widespread belief that Austrian companies traditionally rely on bank financing ¹⁰ and the high prevalence of such financing in aggregate data. This figure, however, can be explained by several factors. Many firms use alternative funding sources, such as trade credit and intercompany loans, as observed in a subset of firms with detailed liability disclosures. Additionally, our findings suggest that some firms may favor internal financing, such as retained earnings, over taking on debt, providing support for this preference within our analysis. Furthermore, certain sectors, especially those less capital-intensive, have probably only limited financing needs. This is underscored by the positive association we observe between tangible assets and loan prevalence: Firms with higher tangible assets are more likely to hold bank loans, while 30% of firms have no tangible assets at all.

From a monetary policy perspective, the fact that over one-third of Austrian firms have bank loans indicates that interest rate changes have a substantial impact on the corporate sector, via the cost channel. Notably, these bank-financed firms control 56% of total corporate assets, underscoring the broader economic significance of this group and highlighting their sensitivity to interest rate changes, especially as most loans are variable rate (OeNB, 2024). Understanding interfirm connections is equally vital for assessing monetary transmission; for instance, intercompany loans – which nearly 80% of firms hold – suggest that monetary policy effects may extend beyond direct bank-financed firms. Given that this insight is drawn from a small subset of firms with detailed liability disclosures, gathering more granular data on firms’ debt structures would enhance policy assessments.

Second, we observe a strong correlation between firm size, tangible assets and the likelihood of obtaining bank loans, likely reflecting banks’ preference for larger, asset-rich firms with strong collateral. This pattern supports financial stability, as these firms are typically more diversified across markets and regions, making them resilient to sector-specific or regional economic shocks. At the same time, however, firms with lower profitability, which may pose higher credit risks, have greater bank loan reliance. This could increase systemic vulnerabilities, especially if economic downturns affect these firms more severely.

Third, our study reveals a high concentration of loan volumes among a small subset of large firms. Specifically, 10% of bank-financed firms account for 76% of total loan volume, and the top 1% hold 43%. This concentration is primarily driven by Austria’s skewed firm size distribution and the strong correlation between size and loan volume. While such concentration within large, financially stable firms is reassuring for financial stability, it also raises systemic risk concerns. Should financial distress occur within this small group, it could disrupt the banking sector and the broader economy. Monitoring these large borrowers is crucial, and further research could assess their default risk and its implications for overall loan exposure. The IFLD’s granularity provides a strong foundation for future studies on financial stability.

An additional area for future research is the interconnectedness of firms. If firms are closely linked, distress in one could rapidly impact others, potentially creating a contagion effect that destabilizes multiple sectors. Understanding these connections would enable policymakers to anticipate and mitigate systemic risks, thereby strengthening the resilience of the financial system. The IFLD, with its detailed data on the ownership structure of companies with respect to each other, offers a valuable resource for exploring these interconnections.

Another avenue for future research is to explore the time dimension of this data, which will soon become possible as the OeNB plans to extend the database to include additional years of observation. Our current analysis focuses on 2021, a year shaped by COVID-19. While we do not expect our general findings on the distribution of bank loans to change significantly, this could be verified with a panel dataset. Panel data would also allow for deeper financial stability analyses, such as tracking how firms’ reliance on bank loans evolves over time, how external shocks like economic downturns affect the financial health of bank-financed firms, and whether changes in key characteristics – such as profitability or leverage – increase their vulnerabilities. These insights would provide a more dynamic perspective, offering valuable contributions to future research and policy design.

7 References

Bachmann, E., M. Hameter, T. Kemetmüller, C. Leitner, P. Reisinger and S. Brachtl. 2021. Progression der Kreditrisikoanalyse durch AnaCredit und die Granulare Kreditdatenerhebung in Österreich. In: Statistiken – Daten & Analysen Q4/21. OeNB. 49–60.

Beer, C. and W. Waschiczek. 2019. Equity ratios of Austrian nonfinancial corporations – evidence from balance sheet data. Monetary Policy & the Economy Q3/19. OeNB. 25–41.

Bernanke, B. and M. Gertler. 1995. Inside the Black Box: The Credit Channel of Monetary Policy Transmission. In: Journal of Economic Perspectives, 9(4). 27–48.

Bernanke, B., M. Gertler and S. Gilchrist. 1996. The Financial Accelerator and the Flight to Quality. In: The Review of Economics and Statistics 78(1). 1–15.

Breyer, P., E. Endlich, D. Huber, D. Oswald, C. Prenner, L. Reiss, M. Schneider and W. Waschiczek. 2021. Corporate equity finance in Austria – impediments and possible improvements. Monetary Policy & the Economy Q3/21. OeNB. 39–57.

Colla, P., F. Ippolito and K. Li. 2020. Debt structure. In: Annual Review of Financial Economics 12(1). 193–215

Devos, E., U. Dhillon, M. Jagannathan and S. Krishnamurthy. 2012. Why are firms unlevered? In: Journal of Corporate Finance 18(3). 664–682.

Elsinger, H., R. Köck, M. Kropp and W. Waschiczek. 2016. Corporate financing in Austria in the run-up to capital markets union. In: Financial Stability Report 31. OeNB. 96–119.

ECB. 2024. Survey on the Access to Finance of Enterprises. Methodological Information on the Survey and User Guide for the Anonymised Micro Dataset, April.

European Commission. 2023. Survey on the Access to Finance of Enterprises (SAFE). Analytical Report 2023.

Frank, M. and V. Goyal. 2009. Capital Structure Decisions: Which Factors Are Reliably Important? In: Financial Management. 38(1). 1–37.

Hirsch, B., T. Kemetmüller and M. Lingo. 2020. AnaCredit und die Granulare Kreditdatenerhebung (GKE) in Österreich. Statistiken – Daten & Analysen Q1/20. OeNB. 21–26.

Holton, S., M. Lawless and F. McCann. 2014. Firm credit in the euro area: a tale of three crises. In: Applied Economics 46(2). 190–211.

Kashyap, A. K. and J. C. Stein. 2000. What Do a Million Observations on Banks Say About the Transmission of Monetary Policy? In: American Economic Review 90(3). 407–428.

Kemetmüller, T., C. Leitner and M. Sakka. 2024. Insolvenzprognose nichtfinanzieller Unternehmen anhand von Jahresabschlusskennzahlen unter Anwendung verschiedener Machine Learning Modelle. Statistiken – Daten & Analysen H2/24. OeNB. 73–92.

Kosekova, K., A. Maddaloni, M. Papoutsi and F. Schivardi. 2023. Firm-bank relationships: a cross-country comparison. ECB Working Paper 2826.

Kraus, A. and R. H. Litzenberger. 1973. A State-Preference Model of Optimal Financial Leverage. In: The Journal of Finance 28(4). 911–922.

Kristofik, P. and J. Medzihorsky. 2022. Capital structure of non-financial companies in EU countries – macroeconomic view. In: WSB Journal of Business and Finance 56(1). 8–16.

Kwaak, T., M. Clarke, J. Kok, J. Snijders, S. Tollenaar, S. Galesloot and M. Sadilek. 2021. Survey on the access to finance of enterprises (SAFE). Analytical Report. European Commission.

Lawless, M., B. O’Connell and C. O’Toole. 2015. Financial structure and diversification of European firms. In: Applied Economics.47(23), 2379–2398.

Mac an Bhaird, C., J. S. Vidal and B. Lucey. 2016. Discouraged borrowers: Evidence for Eurozone SMEs. In: Journal of International Financial Markets, Institutions and Money 44. 46–55.

Myers, S. 1984. The Capital Structure Puzzle. In: The Journal of Finance, 39(3), 574–592.

Myers, S. 2001. Capital Structure. In: Journal of Economic Perspectives. 15(2). 81–102.

OeNB. 2024. Financial Stability Report 47.

Rao, P., H. Yue and J. Zhu. 2015. An investigation of credit borrower concentration. In: Journal of Banking & Finance. 54. 208–221.

Rajan, R. and L. Zingales. 1995. What Do We Know about Capital Structure? Some Evidence from International Data. In: The Journal of Finance 50(5). 1421–1460.

Wiesinger, S. 2015. Innersektorale Kredite und Handelskredite als Substitut klassischer Bankkredite? Statistiken – Daten & Analysen Q4/15. OeNB. 35–48.

Wooldridge, J. M. 2019. Introductory Econometrics: A Modern Approach (7th edition). Boston: Cengage Learning.

8 Annex I: detailed data description

Balance sheet information – firm-level data

Table A1.1 provides an overview of the nonfinancial companies included in the IFLD, divided according to their legal form. In 2021, 250,093 nonfinancial companies were registered, the majority being GmbHs. Our target population – a subset of this group – comprises 194,263 companies. These companies have limited liability and are therefore obliged to disclose their annual financial statements to the Austrian commercial register, ensuring transparency for investors, customers and other stakeholders. This allows us to monitor their finances, which makes them our target population. This subset, representing 78% of all nonfinancial companies, forms a solid basis for our analysis, as limited liability companies are typically more economically significant than firms with unlimited liability (such as KGs or OGs), which tend to be far smaller. ¹¹

The IFLD balance sheet data originate mainly from the commercial register, to which the OeNB has access and from which about 95% of firm entries originate, with the SABINA database supplementing where needed, particularly for larger companies that only submit financial statements as PDFs. ¹² Data from different sources are never mixed within individual companies. Overall, balance sheet data are available for 90% of the target population (174,659 firms), as some companies do not disclose despite the legal requirement. For the analysis, we further refine our sample by excluding observations with inconsistent or implausible data ¹³ . The final sample used for our analysis represents between 87% and 100% of the target population. While calculations such as the share of firms with bank loans are based on the full 100%, analyses of how bank-financed firms differ in their balance sheet characteristics rely on 87% of the population. Each analysis clearly states the number of observations it is based on.

Data availability varies by balance sheet item, especially regarding liabilities. While asset-side coverage is robust, the Austrian Business Code imposes less stringent disclosure requirements on small companies ¹⁴ , limiting the details available on liabilities. Only around 4,000 companies (<3% for which balance sheet data are available) report “liabilities to credit institutions,” underscoring the need to integrate loan-level data to examine bank loan distribution. In addition to balance sheet information, the IFLD dataset includes other firm characteristics relevant to our analysis, such as employee numbers, turnover and industry classification (NACE). ¹⁵

Bank loans – loan-level data

Information on bank loans provided in the IFLD loan-level dataset stems from the national AnaCredit (Analytical Credit) database ¹⁶ compiled by the OeNB. It is a central credit register which contains confidential data on bank credits granted by Austrian credit institutions to legal entities. Data collection started in September 2018 with a monthly reporting frequency. In this paper, we employ data as of December 31, 2021. Subject to reporting is any bank credit to a legal entity if the total commitment amount (i.e. sum of outstanding loan amount and unused credit lines) at the creditor-debtor level is greater than or equal to EUR 25,000 at month-end.

Bank credit is recorded in the form of different financing instruments. We consider six types of financing instruments to be bank loans and classify them according to their revolving nature ¹⁷ :

1. revolving loans: (a) overdraft, (b) credit card debt, and (c) revolving credit other than overdraft and credit card debt;

2. nonrevolving loans: (d) financial leases, (e) nonrevolving credit lines and (f) all other loans that are of a nonrevolving nature and are disbursed in a single payout (lump-sum credits).

We distinguish between these two loan types as they serve different financing purposes. Revolving loans allow firms to adjust their bank exposure up to a credit limit, providing short-term liquidity for working capital needs. In contrast, nonrevolving loans are typically used to finance long-term investments.

For our analysis, we aggregate the relevant bank-level information to obtain a firm-level variable reflecting the total amount of bank loans held by nonfinancial companies. We define the firm-level variable Outstanding Nominal Amount ${ONA}_{i}$ as the sum of all outstanding nominal amounts of loan instruments $j$ held by company $i$ at any Austrian bank $b$ as of the end of 2021, i.e.

${ONA}_{i} = \sum_{b}^{}{\sum_{j}^{}{ONA}_{ijb}\ }$.

Any firm with ${ONA}_{i} > 0$ will be considered a bank-financed firm. ¹⁸

Note that only Austrian credit institutions are included in the national AnaCredit database, so firms with loans exclusively from foreign banks are not considered bank-financed in our analysis. However, aggregate data indicate that Austrian companies predominantly borrow from domestic banks ¹⁹ .

Note also that the AnaCredit reporting threshold of EUR 25,000 has two implications for our analysis. First, estimates such as loan prevalence (share of bank-financed firms) and loan concentration (share of loan amount held by top-x% firms) might be slightly higher if loans below the threshold were included. Second, because the threshold applies per creditor-debtor relationship rather than per debtor, firms with multiple small loans below EUR 25,000 may not be classified as bank-financed even if their total loans exceed the threshold (introducing a downward bias). These cases are rare, however. Comparing AnaCredit’s total loan amount with loan data covering the total exposure (FinStab-V ²⁰ ) shows that over 99% of the loan volume from Austrian banks to nonfinancial companies is represented in AnaCredit, indicating that the bias is negligible.

Table A1.2 shows the number of bank-financed firms and the total loan amount. Among all companies in the commercial register, there are 80,891 bank-financed companies with EUR 176 billion in loans. This decreases to 69,751 firms and EUR 171 billion when focusing only on our target population. Although limiting to firms required to disclose their annual financial statements reduces the number of bank-financed companies by 14%, the impact on loan volume is only minimal (under 3%).

Note that the reported loan volume does not fully reflect the amount due to the bank, as loans jointly signed by multiple entities with joint liability are attributed in full to each if each debtor is fully liable. Consequently, summing these amounts may slightly overestimate the total from the bank’s perspective, though the impact is minimal and does not affect our conclusions. Only 2.5% of companies with loans have co-debtors, representing 2.3% of the total reported volume.

9 Annex II: detailed descriptive statistics