Measuring the impact of the Russian attack on the firm default probability through a non-linear model
Editors note: This column is part of the Vox debate on the economic consequences of war.
Events such as the Covid-19 epidemic or Russia’s war against Ukraine and its impact on the stability of corporate debt and disruption of the supply chain illustrate the importance of monitoring appropriate financial stability. Stress testing tools should capture the specific effects of external shocks in different economic sectors to ensure that banks are adequately supplied with the most affected sectors. This column introduces a approach that combines macroeconomic shocks and sectoral disparities with countries that create different possibilities for defaults for companies across different sectors.
Although stress tests are an important tool of financial stability, predicting corporate default rates represents a major modeling challenge. The main reason is its non-linear relationship with the macroeconomy, especially in the deep recession. Structural credit risk models are only responsible for the average level of the default rate at the overall portfolio level and ignore the original firm-specific data. This granular data, however, is important when predicting the path of the default rate over time, as the recession shock affects the economic sector differently. In addition, different sectors may respond differently to external shocks and ignoring such cross-sectional dispersions can lead to biased predictions.
To provide a broader view on the possible nature of the default rate, this column introduces a panel quantile regression model, which is applied to granular firm-level data, allowing non-linearity of the sectoral default probability compared to a set of regressors. On the tail of the distribution, with additional push at the sectoral level. The approach is aimed at macroeconomic conditions, which are commonly used in stress testing, at an estimated default rate at the country and sector level. The concern that forms the basis of our new application is that the presence of sectoral differences and non-linearity is often overlooked.
The model has been developed using long-term series of country and sector-level data, as well as macroeconomic variables, to significantly improve the predictive accuracy of the default point-in-time probability. Similar to Kalemli-Özcan et al. (2015), the corporate balance sheet data of five million companies from the Orbis global database was used to create representative sample countries, covering all eurozone countries from 2001 to late 2020, with macroeconomic variables derived from OECD. The default variable is firmly defined as a binary indicator that uses balance-sheet data following Gourinchas et al. (2020), where flow default rates are used at the NACE-2 sector level (for more details, see Connaught et al. 2022). The default rates are withdrawn in contrast to several control variables using a fixed effect quantile regression model introduced by Rios-Avila (2020) for diversification in the macroeconomic push of specific industries and potential non-linear sensitivities. Following this approach Adrian et al. (2019) and Cathcart et al. (2020) reflects strong fundamentals on the choice of variables and their effects for default risk. Macroeconomic variables include housing prices, unemployment, and interest rates, and – unlike other specifications and model choices – are supplemented by sectoral gross value-added shocks. Thus, induced by changes in the country-level scenario and sectoral variation, the model allows to predict the sectoral probability of default for any amount on the selected view horizon.
Assess the impact of the Russian invasion on the corporate potential of the default
Disruptions in the energy, food and commodities markets can significantly damage a firm’s creditworthiness (Lafrogne-Joussier et al. 2022) and bank viability, although eurozone banks’ overall direct exposure to the energy and commodities business is low (SSM 2022). For the purpose of vulnerability analysis, two possible tail situations were examined, which were anchored in the March 2022 ECB Staff Macro Economic Projection (ECB 2022: Box 6). Adverse and deadly adverse conditions are designed to capture the direct and indirect effects of war on the corporate potential of default at the sectoral level.
Figure 1 shows the probability of default estimates for unfavorable and severely unfavorable situations, grouped by the corporate, vulnerable and low-risk sectors at 75 percent. Intercortile ranges represent the possibility of default distribution at the country and sector level. Sectors are identified as weak when they are affected by macroeconomic shocks (Atinasi et al. 2022, Bachmann et al. 2022, Winkler and Wister 2022). The data collected combine data from the OECD input-output table and NACE-2 level estimates of total value addition at the country level produced by Oxford Economics. Sectors that are either too open to trade with Russia or feel a negative impact on their total value addition on the three-year horizon are classified as weak (such as mining and quarrying, oil and gas extraction, and the automotive and parts sector). The model estimates show different risk levels across the sector, including the upward shift in the likelihood of default for the weaker sector in both adverse conditions. Consistent with the average difference between the probability of weak and low-protected sectors defaulting on the three-year horizon, the weaker sectors contribute to the increased potential of the euro area with a default of about 20 basis points.1 As a result, Other things being equalThe bank’s general equity tier 1 (CET1) declines by about 30 basis points,2 Above the three-year horizon of the scene. Although moderate in size for the episodes considered and situations, this differential shock sensitivity across the sector reflects the model’s ability to capture the variability of firms in the approximate probabilities of default in a given situation.
Figure 1 Sectoral potential of the projected euro area of the default forecasting firms
Formula: ECB and the authors’ own calculations.
Note:: x-axis shows the projection horizon in years. The y-axis is the percentage point that indicates the probability of defaulting to the 75th percentile. The bars cover the interquartile range; The yellow dot is the middle. Blue bars and orange bars combine the default country and sectoral probabilities that are more or less (less-protected vs. weaker) affected by the additional total value-added shock. The underlying macro-shocks, with the exception of the gross value-added element, are the same across all firms in a given country. The most risky sectors are mining and quarrying (NACE 07-09), oil and gas extraction (NACE 06), and motor vehicles and parts (NACE 29).
Looking at the euro area potential for default estimates from 2005 to 2024, the lower quintiles seem to be moving closer together, while the upper quintiles are characterized by a greater sensitivity to macroeconomic shocks (Figure 2). This holds for perceptions observed during global crises, with the possibility of default between 2015 and the end of 2020 and slight fluctuations between the latest estimates. For example, the Quantile model proves to capture the same dynamics with different amplitudes / intensities across the Quantile and detects pockets of weakness under tail-event shock. Estimates of the Ukraine war appear to show a similar increase in default risk in the case of the global crisis and the European sovereign debt crisis.
Figure 2 Euro area probability of firm default estimate by quintile
Formula: ECB and the authors’ own calculations.
Note:: x-axis is the year, the vertical black line marks the beginning of the projection horizon between 2022 and 2025. The y-axis shows the probability of default as a percentage point for adverse conditions. The colored lines represent the gross value-added weighted average of the default euro area potential of non-financial corporations in the 25th, 50th, 75th, and 90th quintiles that is not parallel and parallel to real estate.
Consider financial stability
Banks and regulators are constantly searching for reliable measures by default. Models relying on data collected at the country level may pass a set of basic criteria for economic and financial well-being but may imply a response to unnecessarily benign risk parameters in potentially adverse conditions. So it creates a risk for a bank to be significantly less-provisioned depending on its sectoral concentration.
It is known that a number of factors contribute to the likelihood of default, including deterioration of the macroeconomic environment, high short- and long-term rates, or a decline in housing prices (e.g., for loans secured by real estate). Although the scenario is the same for all sectors within a country, sectoral differences and over-sensitivity to push in terms of their respective gross value-added causes a strong downturn in debt stability for the most affected sectors. As credit risk increases for banks centralized in debt in companies exposed to external shocks such as strength and legacy Covid-19 vulnerabilities, the inclusion of sectoral level estimates of the likelihood of default has become an important component for a strong assessment of credit risk. Stress testing. Thus, the approach presented in this column combines macroeconomic shocks with country and sectoral disparities that create different possibilities for defaults for companies across different sectors. The sectoral decay of corporate prospects by default not only allows for the design of more appropriate scenarios that include targeted shocks such as the impact of Russian aggression on various product sectors, but also supports pocket identification of weaknesses in the bank’s balance sheet. More granular layer.
Adrian, T., N. Boarchenko and D. Giannon (2019), “Vulnerable Growth”, American Economic Review 109 (4): 1263–89.
Attinasi, MG, R Gerinovics, V Gunnella, M Mancini and L Metelli (2022), “Global Supply Chains Rattled by Winds of War”, VoexEU.org, 9 June.
Bachmann, R, D Baqaee, C Bayer, M Kuhn, A Löschel, B Moll, A Peichl, K Pittel and M Schularick (2022), “What if Germany separates from Russian power?”, VoxEU.org, 25 March.
Cathcart, L. A. Dufour, L. Rossi and S. Varoto (2020), “Differential Impact of Leverage on Default Risk of Small and Large Companies”, Corporate Finance Journal 60: 101541.
European Central Bank (2022), “Review of Financial Stability”, Box 6, “Assessing the Resilience of the Eurozone Banking Sector in the Light of the Russia-Ukraine War”, May.
Gourinchas, PO, S Kalemli-Özcan, V Penciakova and N Sander (2020), “COVID-19 and sme failure”, NBER Working Paper 27877.
Kalemli-Özcan, S, B Sorensen, C Villegas-sanchez, V Volosovych and S Yesiltas (2015), “How to Create Nationally Representative Solid Level Data from Orbis Global Database: New Data and Overall Impact”, NBER Working Paper 21558.
Konietschke, P, J Metzler and AP Marques (2022), “Credit risk country and sector level specificity. A Quantile Approach to Modeling Sectoral Probability by Default ”, ECB Working Papers, forthcoming.
Lafrogne-Joussier, R, A Levchenko, J Martin and I Mejean (2022), “Outside the Macro: The Strong-Level Impact of Russian Power Cutting”, VoxEU.org, 19 April.
Rios-Avila, F (2020), “Recent Recent Impact Functions (RIFs) in Stata: RIF regression and RIF decomposition”, Stata Journal 20 (1): 51-94.
Single Supervisor Mechanism (2022), “Supervision Newsletter”, 18 May.
Winkler, D and L Wuester (2022), “The Impact of Russia’s Ukraine Invasion for the Intervaluation Chain”, VoxEU.org, 11 May.
1 For the low-risk and high-risk sectors on the three-year horizon, 3.3% and 3.5% in adverse conditions, respectively; 4% and 4.2% in severe adverse conditions.
2 The total starting point Euro Area Banking System CET1 ratio was about 15% (end of 2021). These estimates include the expected losses of performing and non-performing loans (Phase 3 loans). It is assumed that the other effects and risks remain the same (ceteris paribus).