The real effects of banks nationalization – evidence from the UK

ABSTRACT

How did the nationalization of UK operating banks as a result of the 2008 banking crisis impact their client firms’ performance? We use unique firm-bank data and a propensity score matching technique and find that firms that borrowed from nationalized banks show a slight decrease in the growth of investment and innovation relative to firms that borrowed from non-nationalized banks. Interestingly, we find that firms that borrowed from nationalized banks slightly increase employment, short-term debt and cash holdings. Overall, these firms were able to maintain performance as a result of policy intervention.

KEYWORDS:

• Firm performance
• bank nationalization
• financial crisis
• United Kingdom

JEL CLASSIFICATION:

• G21
• G34
• O16
• O30

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ www.telegraph.co.uk/finance/financialcrisis/3187946/Financial-crisis-Banks-nationalized-by-Government.html

² The objective of the government intervention as presented in HM Treasury (2009), were to consolidate the banking sector, to provide greater protection for banks’ customers and ‘[to] ensure banks are more willing to lend’. Arguably, these interventions restarted the flow of credit to firms, possibly resulting in improvements in, or at least stopping the decline in firms’ performance.

³ Highly leveraged firms before the crisis are defined in more details in Section 2.

⁴ This is a total value of peak provisions when summed, not a temporal peak in the total value of the guarantees.

⁵ All these databases are provided by Bureau van Dijk and have yearly detailed information about firms and banks.

⁶ We compare our data with the distribution of firms in different industries from the Annual Business Survey, 2018 Revised Results, Office for National Statistics, UK. We use Wilcoxon matched-pairs signed-rank test for our distribution of firms and the distribution of firms from the Annual Business Survey and we cannot reject the hypothesis that both distributions are the same (for ex., for the year 2009, the Z statistics from Wilcoxon matched-pairs signed-rank test is 0.1719, significantly greater than 0.05, therefore we cannot reject that the samples were selected from populations having the same distribution).

⁷ Our identification strategy is robust to post nationalization government schemes that gave banks additional funding for lending to UK businesses and households. From 2011–12 both ‘Project Merlin’ and the subsequent 2012–15 ‘Funding for Lending’ schemes did impact lending across both other banks and nationalized banks. Our identification simply tracks whether nationalized banks performed differently, inclusive of these schemes.

⁸ The number of banks in the initial sample was high (reflecting the relationship between firms and subsidiary banks or even foreign banks). In the PSM sample that we use, the number of banks is reduced to 74 banks out of which 27 are nationalized banks.

⁹ In our dataset, there are 1.2% firms with multiple lending relationship. The results presented in paper are for firms with a single lending relationship. As a robustness check, we re-estimate the regressions including firms with multiple lending relationships (as long as the firms do not belonging simultaneously to the control and the treatment group). The results are similar with the previous results (and available upon request).

¹⁰ We use the application year of granted patents since it is closer to the actual date of innovation (Griliches 1990).

¹¹ Classifying highly leveraged firms as firms in the top third of the leverage distribution before the crisis (2006) does not change the results.