Supervisory policy stimulus: evidence from the euro area dividend recommendation

© European Central Bank, 2023

Postal address 60640 Frankfurt am Main, Germany

Telephone +49 69 1344 0

Website www.ecb.europa.eu

All rights reserved. Any reproduction, publication and reprint in the form of a different publication, whether printed or produced electronically, in whole or in part, is permitted only with the explicit written authorisation of the ECB or the authors.

This paper can be downloaded without charge from www.ecb.europa.eu, from the Social Science Research Network electronic libraryor from RePEc: Research Papers in Economics. Information on all of the papers published in the ECB Working Paper Series can be found on the ECB’s website.

PDF ISBN 978-92-899-5992-6 ISSN 1725-2806 doi:10.2866/62005 QB-AR-23-033-EN-N

[1] On 27 March 2020, the ECB adopted the recommendation that at least until 1 October 2020 no significant institutions should pay out dividends, and no irrevocable commitment to pay out dividends should be undertaken by the credit institutions for the financial years 2019 and 2020. This recommendation was addressed to significant institutions directly supervised by the ECB. For further information see the ECB press release from 27 March 2020 at: https://www.bankingsupervision.europa.eu/press/pr/date/2020/html/ ssm.pr200327~d4d8f81a53.en.html. This was followed by an EBA statement on 31 March urging banks to follow prudent dividend and other distribution policies, including variable remuneration”. Many NCAs subsequently issued their own regulatory announcements in a similar vein.

[2] In a technical report for euro area significant institutions, Dautovi´c et al. (2021) show that the treated banks increased their loan loss provisions by around 5.5% relative to the control group, thereby strengthening their relative capacity to absorb future losses.

[3]Ampudia et al. (2023) also provide evidence on the impact of the recommendation on selected aggregates (bank lending, bank valuations, dividend expectations).

[4] See Table 1 in Abreu and Gulamhussen (2013) for a summary of the empirical evidence on dividend payout policies and stock prices.

[5] The following criteria are applied to determine if a bank is a significant institution and hence should be directly supervised by the ECB and not by the national competent authorities (NCAs) : 1) Size: the total value of its assets exceeds €30 billion; 2) Economic importance for the specific country or the EU economy as a whole; 3) Cross-border activities: the total value of its assets exceeds €5 billion and the ratio of its cross-border assets/liabilities in more than one other participating eurozone member state to its total assets/liabilities is above 20%; 4) Direct public financial assistance: the bank has requested or received funding from the European Stability Mechanism or the European Financial Stability Facility; 5) A supervised bank can also be considered significant if it is one of the three most significant banks established in a particular country. For the full definition of eurozone significant institutions see: https://www.bankingsupervision. europa.eu/banking/list/criteria/html/index.en.html.

[6] This calculation is performed holding the average regulatory capital ratio, and risk-weights, of euro area banks fixed at the end of 2019. For ECB banking supervision data, see the publicly available statistics at: https://www.bankingsupervision.europa.eu/banking/statistics/html/index.en.html.

[7] It is worth noting that from a welfare perspective, restricting dividend distributions to generate additional lending is superior to easing capital requirements. The latter measure indeed does not increase banks’ loss absorption capacity, Imbierowicz et al. (2018). Therefore, from a policy perspective restricting dividends can be more effective in supporting the real economy, particularly when combined with a regulatory capital release. Moreover, in a modelling framework, Mun~oz (2021) and Fischer and Kessler (2022) show that prudential policies on dividends can be superior to conventional macroprudential policies in smoothing the financial cycle and providing additional welfare gains.

[8] See the ECB press release of 28 July 2020 at: https://www.bankingsupervision.europa.eu/press/ pr/date/2020/html/ssm.pr200728_1~42a74a0b86.en.html

[9] See Copeland et al. (2021); Demiralp et al. (2021); Ryan and Whelan (2021) for a similar approach on asset purchases. The amount of deposits held by commercial banks at the central bank also serves as a measure of the costs that a negative interest rate policy imposes on each financial institution, Heider et al. (2019), Bubeck et al. (2020).

[10] For an overview of ECB asset purchase programs before and during the Covid-19 pandemic see: https:

//www.ecb.europa.eu/mopo/implement/app/html/index.en.html

[11] Evidence of the effects of ECB recommendation for the case of Spain is also provided by Mart´inezMiera and Vegas (2021). The authors analysed credit registry data and found larger effects on credit supply, ranging from 11.9% to 14.5%. However, these large point estimates may be the result of not controlling for the simultaneity of unprecedented monetary and fiscal policies which are all statistically and economically significant in our specifications.

[12] In our dataset, this variation of compliance implies that around 60% of observations are linked to banks that deviated from their initial distribution plans and are thus assigned to the treated group of observations.

[13]AnaCredit is the analytical credit register of the Eurosystem and additional documentation can be found here: https://www.ecb.europa.eu/stats/money_credit_banking/anacredit/html/index.en.html

[14] Covid-19 guaranteed loans have been identified by using registry information - e.g. LEIs and Register of Institutions and Affiliates Data (RIAD) codes - of the promotional lenders charged with this task in each country (for example, ICO in Spain, KFW in Germany, BPI in France and SACE/Fondo di Garanzia in Italy). In addition to the registry information of the guarantor, the starting date of the public guarantee scheme has also been used as an identifying device.

[15] In a robustness check in Section 5.1, we provide variation of our baseline specification by replacing borrower-time fixed effects with industry-location-size-time fixed effects and including thus also single bank relationship firms in the estimation sample. In a further robustness check in Section 5.2 we remove banks that did not plan to distribute dividends already prior to the recommendation and keep only banks with ex-ante positive dividend plans.

[16] The MDA is a limit on dividend distribution that decreases linearly when a bank’s capital level falls below the regulatory minimum capital requirement. In other words, the distance from the MDA explains the amount of capital a bank has to expand its assets.

[17] See the SME definition of the EU Commission at this link https://ec.europa.eu/growth/smes/ sme-definition_en. In accordance with this definition, we use the following dummy variables to classify enterprise size: Micro is a dummy variable that is equal to 1 for enterprises that employ fewer than 10 employees and whose annual turnover and/or annual balance sheet total does not exceed EUR 2 million, and 0 otherwise. Small is a dummy variable that takes the value 1 for enterprises that employ fewer than 50 employees and have an annual turnover and/or annual balance sheet total that does not exceed EUR 10 million, and 0 otherwise. Medium is a dummy variable that takes the value of 1 for enterprises that employ less than 250 but more than 50 employees, have an annual turnover not exceeding EUR 50 million and/or an annual balance sheet total not exceeding EUR 43 million, and 0 otherwise.

[18] NACE Rev. 2 classification is based on a hierarchical structure, which consists of first level sections (alphabetical code), second level divisions (2-digit numerical code), third level groups (3-digit numerical code), and fourth level classes (4-digit numerical code). For more information, refer to EU Commission NACE classification.

[19] Table A2 in the Annex reports robustness of baseline results by using alternative clustering of standard errors.

[20] This dummy variable may not account for new loans granted to riskier enterprises that did not have pre-existing banking relationships prior to the pandemic. Nevertheless, such loans represent a negligible fraction of the sample.

[21] In a bank-level setting, Dautovi´c et al. (2021) show that loss absorption and capital conservation drive the reallocation of non-distributed dividends.

[22] The March ECB press releaseasks banks not to pay dividends until at least October 2020.

[23] The July ECB press releaseextends recommendation not to pay dividends until January 2021.

[24] Strictly speaking, those banks are effectively complaint with the ECB recommendation, but they had already committed some of the funds and received approval just few days prior to the ECB announcement at their shareholders’ meetings, which paved the way for their inevitable disbursement.