Insolvency frameworks could help address issues linked to high ineffective debt. They may first contribute to reducing the adverse effects of high private debt on economic activity by freeing up resources caught in unproductive activities and contribute to mitigating macro-financial spillovers from zombie firms to non-zombie ones (Albuquerque and Iyer 2023). Moreover, they can lower costs linked to bankruptcies. They should also diminish, ex ante, potential concerns regarding credit supply and demand in the event of insolvencies. Research has also shown that bankruptcy reform can help an economy to recover more quickly during a recession (Claessens and Klapper 2002).
Better insolvency regimes can also contribute to reducing corporate bond spreads, expand GDP, and increase employment (Simmons et al. 2016).
Still, it should be kept in mind that while modernising insolvency law helps address problems linked to high unproductive debt, it may not be sufficient on its own. Indeed, other factors play a key role (judicial infrastructure, regulatory and tax policies, among others) and these may take more time to change.
The use of detailed loan data enables us to better identify the impact of insolvency regimes, controlling for creditor and debtor factors
We investigate the link between insolvency regimes and non-performing loans (NPLs) at European banks across the globe from 2015 to 2020. This period corresponds to peaks in the number of insolvency reforms passed both in developed and developing countries (see Figure 1).
Figure 1 Number of reforms passed per year
Source: Doing Business, authors’ calculations.
Our primary objective is to estimate the impact of implementing reforms that enhance insolvency regimes on the NPL level of banks. We leverage the granularity of our dataset (lending bank by borrowing firm’s country by firm size) to investigate the differential between insolvency reforms across borrower and creditor categories, while also evaluating how the orientation of a country’s insolvency regime – whether creditor- or debtor-oriented – affects the efficacy of these reforms.
To estimate the relationship between insolvency reform implementation and bank NPL levels, a fixed effects estimator is used. Due to the granularity of our data from the European Banking Authority's Transparency Exercises, we account for unobserved time-invariant heterogeneity related to banks, debtor countries, firm size, and their interactions. More detail is available in our recent paper (Bricongne and Dufouleur 2024).
In a nutshell, we exploit the high detail of the data to analyse to what extent the difference of NPL ratios of say, BNP Paribas vis-à-vis big firms in Germany and Romania can be explained by different insolvency regimes and reforms in these countries, controlling for other factors. This is possible as foreign loans represent a sizeable fraction of total credit in many European countries (see Figure 2).
Figure 2 Breakdown of loans per origin country, 2020 (% of total loans)
Source: European Banking Authority (EBA) transparency exercises, authors’ calculations.
Following the NPL determinant literature, our analysis incorporates macroeconomic, institutional, and bank-specific factors to isolate the impact of insolvency reforms. For macroeconomic variables, we consider GDP growth, the inflation rate, and the unemployment rate. Thanks to the granularity of our dependent variable data, we are also able to control for time-varying factors associated with both the bank (including, for example, management and governance, which is seldom taken into account) and its origin country.
Non-performing loan reforms have an impact mainly on the dynamics of NPLs rather than on their levels
The analysis encompasses regressions between insolvency reform implementation and both NPL rates and dynamics.
The variable of interest, Strength, reflects the magnitude of insolvency reforms that were implemented between the years t-4 and t-1, since these reforms may take time to generate their effects (we also make robustness checks on the length of this moving period).
Concerning NPL rates in levels, despite a significance of Strength at the 5% threshold, the results do not reveal a robust connection between the adoption of insolvency reforms and bank NPL rates.
Conversely, a notably significant and negative relation emerges between Strength and the NPL midpoint growth rates, indicating that the implementation of insolvency reforms in the previous four years goes hand in hand with an accelerated resolution of bank NPLs. This finding holds true across all model specifications, with significance levels reaching 1%.
In summary, the findings underscore that bank NPL rates show no significant impact from insolvency reforms enacted in the past four years (excluding the present year). Nonetheless, the implementation of such reforms significantly correlates with faster NPL resolution.
Debtor-oriented reforms seem to have more effect on non-performing loan dynamics
Still, all insolvency reforms are not equivalent. In other words, what would be the optimal insolvency regime orientation to implement, and does it depend on context? This question refers to the creditor- versus debtor-oriented regimes debate concerning NPLs. Both orientations exhibit efficiencies and inefficiencies in reducing NPLs, and present conflicting symmetric mechanisms. On the one hand, increasing creditor protection equips creditors with enhanced tools for credit recovery; however, it may also diminish banks’ risk exposure, leading to decreased borrower screening. On the other hand, improving debtor protection increases credit demand from lower-quality borrowers while giving viable firms a means of restoring their financial health.
Using further the detail of our data, we carry out regressions on different samples depending on the type of debtor (small and medium-sized enterprises, or SMEs, versus non-SMEs), the type of creditor (small, medium, and large banks), and debtor country characteristics (high versus low NPL countries and insolvency regime types). We show that insolvency regime reforms are efficient at speeding up the resolution of NPLs, especially during financial distress. This effect is particularly true for big firms and big banks, in a debtor country with an already high NPL level. This result is driven by debtor-oriented reforms, more precisely, reforms that aim to facilitate business continuity.
Our findings also reveal that such reforms are more efficient in countries with a non-debtor- and creditor-friendly insolvency regime. Conversely, we find that creditor-oriented reforms present a perverse effect, as they are associated with higher NPL levels.
These findings are interesting to keep in mind to implement insolvency reforms, whether in the EU context or outside.
References
Albuquerque, B and R Iyer (2023), “The rise of the walking dead: Zombie firms around the world”, VoxEU.org, 28 August.
Becker, B (2019), “The EU’s insolvency reform: Right direction, not enough, and important issues left behind”, VoxEU.org, 27 June.
Bricongne, J-C and M Dufouleur (2024), “Impact of Insolvency Regimes on NPLs: Two Birds in the Bush is Worth One in the Hand”, Banque de France Working Paper No. 953, 16 July.
Claessens, S and L F Klapper (2002), “Bankruptcy Around the World: Explanations of its Relative Use”, World Bank Policy Research Working Paper 2865, July.
Davis, S J and J Haltiwanger (1998), “Measuring Gross Worker and Job Flows”, in Labor Statistics Measurement Issues, University of Chicago Press.
Simmons, G, J Suarez and P McGhee (2016), “The importance of better EU insolvency regimes”, VoxEU.org, 29 April.
