The cost efficiency improvement of Norwegian banks can be explained by automation and digitalisation

Andersen, Henrik Research Report The cost efficiency improvement of Norwegian banks can be explained by automation and digitalisation Staff Memo, No. 9/2020 Provided in Cooperation with: Norges Bank, Oslo Suggested Citation: Andersen, Henrik (2020) : The cost efficiency improvement of Norwegian banks can be explained by automation and digitalisation, Staff Memo, No. 9/2020, ISBN 978-82-8379-176-1, Norges Bank, Oslo, https://hdl.handle.net/11250/2690371 This Version is available at: https://hdl.handle.net/10419/246149 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/deed.no STAFF MEMO The cost efficiency improvement of Norwegian banks can be explained by automation and digitalisation NO. 9 | 2020 HENRIK ANDERSEN 2 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION The papers in the Staff Memo series present reports and documentation written by staff members and other authors affiliated with Norges Bank, the central bank of Norway. The views and conclusions expressed in these papers do not necessarily represent those of Norges Bank. © 2020 Norges Bank This paper may be quoted or referenced provided the author and Norges Bank are acknowledged as the source. ISSN 1504-2596 (online) ISBN 978-82-8379-176-1 (online) 3 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION The cost efficiency improvement of Norwegian banks can be explained by automation and digitalisation Henrik Andersen1 Operating costs in the Norwegian banking sector have been reduced considerably in recent decades, both as a share of income and assets. This has increased banks’ resilience to increased losses and reduced the risk of crises. In this article, I analyse how costs have been reduced and the main drivers of the cost efficiency improvement. The results suggest that automation and the digitalisation of banks’ operations have played a key role in improving cost efficiency. Key words: banks, costs, digitalisation, regulation, business cycle 1. Introduction Banks provide a number of services that are crucial to economic activity.2 As access to these services is often disrupted during banking crises, the cost to society of such crises is high. Cutting costs boosts banks’ resilience to increased losses and reduces the risk of crises. Historically, most banking crises have been caused by losses on lending and financial instruments. The first line of defence against such losses is banks’ profits. In isolation, cost cuts increase profits krone for krone. Banks therefore reduce costs to improve profitability and thus their first line of defence. For example, Andersson et al (2018) show that the euro area banks demonstrating the highest improvement in profitability after the financial crisis have reduced their cost-to-income ratios. Banks’ cost cuts can also improve monetary policy transmission and contribute to higher growth. According to Jonas and King (2008), cost-efficient banks adjust their loan volumes to the policy rate more than less cost-efficient banks. Cost cuts can also enable banks to offer cheaper services. Lower bank lending rates can for example push up both business investment and consumption (see Andersen and Walle (2015) and Andersen et al. (2016)). Norwegian banks3 are cost-efficient compared with banks in other countries. Operating costs have been reduced considerably over the past three decades, both as a share of income (cost-to-income ratio) and assets (cost-to-assets 1 The views and conclusions in this publication are the author’s own and do not necessarily reflect, and must not be reported as those of Norges Banks. I thank Håkon Astrup (DNB Markets), Eleonora Granziera, Torbjørn Hægeland, Tom Høiberg (Finance Norway), Einar Nordbø, Knut Sandal, Norman Spencer, Ylva Søvik, Bent Vale, Sindre Weme and Terje Åmås for useful comments and input, as well as Kaja Dørum Haug and John Henrik Mulelid for their kind assistance with background information and charts. 2 Banks provide loans, accept deposits, execute payments and help customers with managing risk. 3 All banks and mortgage companies in Norway unless otherwise. 4 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION ratio) (Chart 1). As a result, Norwegian banks have the lowest average costto-income ratio of all the EEA countries (Chart 2). After the 2008 financial crisis, the same key ratio has increased for euro area banks. According to Andersson et al, euro area banks’ cost-to-income ratios have risen because of substantial growth in wage and personnel expenses. Andersson et al. also refer to euro area banks’ weak income developments and large stocks of nonperforming loans. 4 In addition, Huljak et al (2020) show that productivity growth in the euro area banking sector has fallen. Chart 1 Operating costs in the Norwegian banking sector as a share of operating income and assets.1 1987 – 2019 1) See Appendix 1 for more information on the data series. Source: Norges Bank Chart 2 Average cost-to-income ratio for Norwegian1 and European2 banks for the period 2019 Q3 – 2020 Q2. Percent 1) All banks except branches of foreign banks in Norway. 2) 147 European banks. Sources: European Banking Authority (EBA) and Norges Bank 4 Banks may need to devote substantial resources to non-performing loans (see eg Fell et al (2017). 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 10 20 30 40 50 60 70 80 90 100 1987 1991 1995 1999 2003 2007 2011 2015 2019 Financial crises Share of income Share of assets (r.h.s.) 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Germany France Belgium Denmark Malta Cyprus Italy Luxembourg Ireland Austria Slovenia Finland Hungary Netherlands Slovakia Iceland UK Portugal Latvia Poland Spain Sweden Romania Czech Rep. Bulgaria Estonia Croatia Greece Lithuania Norway EU/EEA 5 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Developments during the Covid-19 pandemic indicate that cost efficiency improvements have increased Norwegian banks’ resilience to higher losses. Despite increased credit losses, Norwegian banks have posted profits and largely maintained credit supply. By comparison, the largest German, Belgian and Italian banks as a whole posted losses in 2020 Q1, even though their credit losses were lower than those of Norwegian banks. In this Staff Memo, I analyse how Norwegian banks have reduced their costto-assets ratios and the main drivers of cost efficiency improvements. Section 2 describes the dataset. Section 3 decomposes banks’ costs and assess how Norwegian banks have reduced their cost-to-assets ratios. Section 4 discusses possible drivers of cost developments. Section 5 models developments in the cost-to-assets ratio, and Section 6 concludes. 2. Data I use several data sources to analyse how Norwegian banks have reduced their cost-to-assets ratios. Data from the ORBOF 5 bank statistics show developments in banks’ costs and assets. The ORBOF bank statistics provide a rough allocation of costs back to 1987 and a more detailed allocation back to 1998. Developments in banks’ labour costs are analysed using the ORBOF bank statistics on banks’ labour costs and number of employees as well as data from Statistics Norway on hours worked per employee and wage per hour worked. In addition, I complement the ORBOF bank statistics with statistics from Finance Norway on the number of bank offices in Norway to assess developments in other costs. I analyse the drivers of the improvements in Norwegian banks’ cost efficiency with data from a number of different sources. The effects of automation and digitalisation are analysed using Norges Bank data on the number of ATMs, payment terminals and various kinds of payment transaction in Norway. In addition, I use data from Statistics Norway and Finance Norway on Norwegians’ use of the internet, smart phones and online banking services. Other potential drivers of the cost developments are analysed using data from the ORBOF bank statistics, Finanstilsynet (Financial Supervisory Authority of Norway) and the Norwegian State Administration Database. 3. Decomposition of costs In this section, I decompose developments in banks’ costs to provide a better basis for assessing how Norwegian banks have reduced their cost-to-asset ratios. The decomposition shows that nearly half of the decline in the cost-toassets ratio is due to reductions in wage and personnel expenses by banks relative to assets, which inter alia reflect a fall in the number of employees. The decline in this ratio has been restrained by the average wage of bank employees, which has risen faster that the average wage in Norway. Other 5 Banks’ and financial undertakings’ financial reporting to the Norwegian authorities (ORBOF). 6 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION operating expenses have also fallen relative to assets, among other reasons, as a result of the reduction in the number of bank offices. Higher costs for IT and external services have dampened the decline in the cost-to-assets ratio. A large part of the improvements in cost efficiency can be explained by developments in wage and personnel expenses. Wage and personnel expenses and costs for external services fell from 1.4 percent of assets in 1987 to 0.4 percent in 2019 (Chart 3). Developments in these costs explain nearly half of the decline in the total cost-to-assets ratio since 1987. 6 The share attributed to other operating costs also fell. Chart 3 Norwegian banks’ operating costs as share of assets. 1987 – 2019 Source: Norges Bank A more detailed allocation of banks’ operating costs shows that the cost composition has been fairly stable over the past two decades (Chart 4). Wages and personnel expenses have generally accounted for around half of total operating costs. On the other hand, IT costs have accounted for an ever greater share. Since 2014, IT costs, which include costs for licences, software and other ICT equipment and costs for external IT services, have contributed to increasing the total cost-to-assets ratio. In this period, IT costs have risen by two-thirds, while other costs have only risen by 3 percent. Costs for external services have also accounted for an ever greater share of operating costs, and in many of the years after the financial crisis, these costs have contributed to an increase in the cost-to-assets ratio. 7 This may be because banks have outsourced larger parts of their operations, for example payroll administration, accounting, canteen operation and janitorial and cleaning services. 6 Measured as a share of assets, wages, personnel expenses and costs for external services overall fell by 1 percentage point between 1987 to 2019, that is, developments in these costs explain nearly half of the 2.2 percentage point decline in the total cost-to-assets ratio. 7 During these years, costs for external services increased more relative to assets than other costs. 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 1987 1991 1995 1999 2003 2007 2011 2015 2019 Financial crises Wage, personnel expenses and external services Changes in the value of non-financial assets Other operating costs 7 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Chart 4 Norwegian banks’ operating costs decomposed into subgroups. Percent of total operating costs. 1998 – 20191 1) In 2018, the statistics were restructured and a number of the financial statement items were changed. Changes between 2017 and 2018 must therefore be interpreted with caution. Source: Norges Bank Chart 5 Number of employees and number of bank offices and branches.11984 – 20192 1) All banks in Norway. 2) Statistics for the number of bank offices and branches in 2019 are not available. The number is therefore approximated using information from bank websites. Sources: Bankplassregisteret, Finance Norge, Statistics Norway and Norges Bank Norwegian banks have reduced their cost-to-assets ratios by downsizing (Chart 5). The number of employees has fallen by over 14 percent since 1990, even though banks’ assets have increased more than twelve-fold. Finance and insurance are among the industries in Norway that have replaced labour with capital in production the most (see Hagelund et al (2017)), and productivity growth in the industry has been high since the start of the 1990s (Chart 6). This downsizing has, in isolation, reduced banks’ wage and personnel expenses. The downsizing may have also contributed to reducing 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 1998 2002 2006 2010 2014 2018 IT costs Sales, advertising and representation costs Costs for renting office premises Other operating costs Changes in the value of non-financial assets Other external services Wage and personnel expenses 0 400 800 1,200 1,600 2,000 2,400 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 1984 1988 1992 1996 2000 2004 2008 2012 2016 Employees in the banking sector Employees in finance and insurance Bank offices and branches (r.h.s.) 8 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION other personnel expenses and other costs, eg costs associated with office premises and office equipment. On the other hand, some of the downsizing may be due to outsourcing, which has increased the costs for external services. In addition, the average wage of bank employees has risen faster than the average wage in Norway, which has contributed to keeping labour costs elevated. In 1987, the average hourly wage for employees in finance and insurance was 16 percent higher than the average wage in Norway. In 2019, this gap was 40 percent. This may be because low-wage employees were replaced by more highly paid new hires, for example, because digitalisation and more burdensome regulation have changed banks’ skill needs (see Sections 4.1 and 4.2). 8 In addition, the wages of other bank employees may also have risen more than the average wage in Norway. Chart 6 Gross product per hour worked. Change from previous year. Constant prices. Five-year moving average. Percent. 1984 – 2019 Source: Statistics Norway Chart 7 Rent per square metre for prime office space in Oslo. Banks’1 average rental cost per office. In NOK. 1987 – 20192 1) All banks and mortgage companies in Norway. 2) 1998-2018 for average annual rent per office. Sources: CBRE and Norges Bank 8 The number of hours worked per employee in finance and insurance have risen only marginally in the analysis period. -10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10 1984 1988 1992 1996 2000 2004 2008 2012 2016 Finance and insurance All industries together 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 1987 1991 1995 1999 2003 2007 2011 2015 2019 Average annual rental cost per office Annual rent (r.h.s.) 15 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION provide an economic net gain, if the direct cost to banks is less than the gain from curbing market failures (see Borchgrevink et al (2013)). A number of studies indicate that it is costly for banks to comply with regulation (compliance costs). Older studies and investigations find that compliance costs account for between 6 and 14 percent of banks’ total costs (see Thornton (1993) and Elliehausen (1998)). Results from more recent studies and surveys suggest that compliance costs represent an ever larger share (see Cyree (2016), Deloitte (2017), Depman (2016), Hogan and Burns (2019) and Hui et al (2016)). Hogan and Burns point out that regulatory changes have increased costs in the form of data processing, consultants, lawyers and auditors. Several factors suggest that the regulatory burden has risen for Norwegian banks. The capital adequacy rules have increased in scope and complexity. These rules were modified already in the 1990s 13 , but the biggest change came in 2007, when the 30-page Basel I framework was replaced by the 347page Basel II framework (see Basel Committee (1988) and Basel Committee (2006)). The increase in the number of pages alone implies that banks needed to devote considerably more resources to complying with Basel II than with Basel I. In addition, Basel II permitted banks to calculate their own capital requirements for credit risk using internal models (IRB approach). 14 This has led the largest Norwegian banks to spend more on development, validation and follow-up of risk models. The financial crisis uncovered a number of weaknesses of Basel II, and in December 2010, the Basel Committee proposed an even more extensive set of rules totalling 616 pages (Basel III) (see Basel Committee (2010), Haldane (2012) and Lund and Nordal (2017)). Among other measures, Basel III introduced quantitative liquidity requirements and a number of new capital buffer requirements for banks. These rules were phased in from summer 2013 in Norway. 15 Norwegian banking regulation has probably also become more burdensome in other areas. The Norwegian authorities introduced guidelines and requirements for prudent lending standards for residential mortgages from 2010 and consumer credit from 2017. In autumn 2018, the Ministry of Finance introduced a new Money Laundering Act, which according to Finance Norway, entailed increased efforts by financial institutions (see Finance Norway (2018b)). This is consistent with Depman’s findings that anti-money laundering work, consumer protection and credit standards are the primary drivers of costs associated with regulatory compliance. In recent year, banks have also 13 A simple unweighted capital ratio was replaced by a risk-weighted capital ratio (Basel I) in 1991. The riskweighted requirement was calculated using standardised, fixed weights, and the asset classes were relatively broad (see Basel Committee (1988) and Haldane (2012)). In 1996, banks were permitted to calculate capital requirements for market risk using internal models (see Basel Committee (1996)). This made the rules considerably more complex. 14 Basel II also introduced capital requirements for operational risk. 15 Looking ahead, the capital adequacy rules may become somewhat less complex. The Basel Committee has recommended that the authorities rescind the right to use internal models to calculate capital requirements for operational risk and for calculating capital requirements for credit risk in certain areas (see Basel Committee (2017)).The recommendations are to be phased in between 2023 and 2028. 16 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION devoted resources on adapting to the revised Payment Services Directive (PSD2), including by developing systems that give third-party providers access to account information. Finanstilsynet is responsible for banks’ regulatory compliance. The number of employees of this supervisory authority can therefore be a reliable proxy for how burdensome the regulations are. Since 1987, the number of employees of Finanstilsynet has risen sharply (Chart 13). At the same time, the number of banks has fallen. The number of supervisory FTEs per bank has thus risen from 0.4 in 1987 to 2.1 in 2019. This may indicate that regulating Norwegian banks has become significantly more burdensome. On the other hand, banks’ average size has increased in the period (see Section 4.3). This may explain some of the increase in the number of supervisory FTEs per bank. Chart 13 Number of Finanstilsynet employees. In number of persons and as a share of the number of banks. 1987 – 2019 Sources: Finanstilsynet, Norwegian State Administration Database and Norges Bank 4.3. Economies of scale An explanation for Norwegian banks’ efficiency gains may be that they have become bigger. When banks get bigger, they can exploit economies of scale, ie the costs per produced service falls when the number of produced services rises. There may be a number of sources of economies of scale in banking. Banks may have expertise, systems and office space that can handle more customers without incurring substantial additional costs. 16 Thus, organic growth, acquisitions and mergers can reduce the cost-to-asset ratio. Bank mergers are therefore often motivated by economies of scale (see Schmitz and Tirpák (2017)). For example, the two Spanish banks CaixaBank and 16 According to Roades (1998), bank mergers result in cost saving in the form of downsizing and improved IT system and infrastructure utilisation, with downsizing representing in many cases more than half of the saving. 0.0 0.4 0.8 1.2 1.6 2.0 2.4 0 50 100 150 200 250 300 350 1987 1991 1995 1999 2003 2007 2011 2015 2019 Number of employees (l.h.s.) Number of employees as a share of number of banks 17 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Bankia expect an annual cost reduction of at least EUR 770 million if they merge (see CaixaBank (2020)). A number of studies document economies of scale in the banking sector (see eg Beccalli et al (2015), Berger and Mester (1997), Demirgüç-Kunt and Huizinga (2011), Humphrey and Vale (2004) and Dijkstra (2013)). However, results from several other studies suggests that economies of scale diminish or disappear when banks reach a certain size (see Andreeva et al (2019)), Berger and Mester, Demirgüç-Kunt and Huizinga, Feng and Serletis (2009) and Huljak et al (2020)). Huljak et al explain the diminishing economies of scale by noting that large banks often employ more sophisticated business models and are more difficult to manage. Demirgüç-Kunt and Huizinga (2011) point out that banks that become large relative to domestic GDP may run out of profitable business opportunities. Moreover, in Norway, DNB, which is Norway’s largest bank, with assets of over NOK 3 trillion, and the large foreign banks have higher wage and personnel expenses per employee than the small and medium-sized banks that are members of the SpareBank 1 Alliance and the Eika Group (Chart 14). On the other hand, the SpareBank 1 Alliance and Eika Group banks have more employees per NOK billion under management than the large banks (Chart 14). Chart 14 Wage and personnel expenses in millions of NOK as a share of the number of employees. Number of employees as a share of total assets in billions of NOK. 2019 Source: Norges Bank Banks in Norway have likely been exploiting economies of scale since the 1980s. Humphrey and Vale’s analyses indicate that Norwegian banks exploited economies of scale in the period 1987-1998. Mergers and acquisitions reduced the number of banks, especially in the 1980s and beginning of the 1990s (Chart 15). This has led to higher concentration in the 0 1 2 3 4 5 0 1 2 3 4 5 DNB and foreign banks SpareBank 1alliansen Eika-gruppen Other banks Wage and personnel expenses per employee Number of employees per NOK billion 18 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Norwegian banking sector. 17 In addition, organic growth has boosted banks’ average size especially in the 2000s (Chart 15). Chart 15 Number of banks in Norway. Banks’ average total assets as a share of mainland GDP.1 1984 – 2019 1) Market value at current prices. Sources: Norges Bank and Statistics Norway Both organic growth and cooperation have likely resulted in economies of scale for the small and medium-sized banks in Norway. Small and mediumsized banks have grown more overall than the largest banks. 18 In addition, many of the small and medium-sized banks have become members of the SpareBank Alliance and the Eika Group. The aim of these alliances is to exploit economies of scale by sharing expertise and collaborating on banking services, branding, payments and IT infrastructure. Nevertheless, the cost-toasset ratios of small banks is high compared with large banks (Chart 16). 17 According to Ulltveit-Moe et al (2013), the Norwegian banking sector was already highly concentrated in 2013, in comparison to both other countries and other sectors. Ulltveit-Moe et al explain the high concentration by citing economies of scale, among other factors. 18 At the end of 2019, seven Norwegian banks had assets of more than NOK 100 billion. Since 2000, these seven banks have grown less than the Norwegian banking sector as a whole. 0.0 0.4 0.8 1.2 1.6 2.0 0 50 100 150 200 250 300 1984 1988 1992 1996 2000 2004 2008 2012 2016 Number of banks Average size 19 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Chart 16 Banks’1cost-to-assets ratios (vertical scale). Banks ranked from smallest (1) to largest (116) by assets. Percent. 2019 1) Parent bank data for all banks in Norway, excluding branches of foreign banks and four banks with a cost-to-assets ratio above 5 percent. Source: Norges Bank Cost developments at Norwegian banks show no indication that economies of scale have diminished or disappeared as banks have become larger. Since the end of the 1980s, the largest banks have reduced their cost-to-assets ratios more than small banks. DNB has reduced its cost-to-assets ratio more than the SpareBank 1 Alliance and Eika Group banks (Chart 17). At the same time, the SpareBank 1 Alliance, which is dominated by four savings banks with assets over NOK 100 billion, has reduced its cost-to-assets ratio more than the Eika Group. Chart 17 Banks’ cost-to-assets ratios. Percent. 1987 – 2019 Source: Norges Bank Both digitalisation and more burdensome regulation may have increased the advantages of being large (see Sections 4.1 and 4.2). Digitalisation requires substantial investment, and in many cases, IT systems may only be profitable if used on a large scale (see Amel et al (2004)). In the DNB Markets survey, 0 1 2 3 4 5 0 1 2 3 4 5 020 40 60 80 100 120 SpareBank 1 banks Eika banks Other banks 0 1 2 3 4 0 1 2 3 4 1987 1991 1995 1999 2003 2007 2011 2015 2019 DNB SpareBank 1 Alliance Eika Group 20 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION 37 percent of the banks responded that economies of scale associated with digitalisation are the primary driver of consolidation, while 61 percent of the banks responded that the primary driver of consolidation are economies of scale associated with regulation and reporting. 19 In addition, use of the IRB approach, which lowers banks’ capital requirements, is limited to large and medium-sized banks. 20 Looking ahead, mergers and acquisitions may further increase the average size of Norwegian banks. On average, the banks in the DNB Markets survey expect that the number of banks will be reduced by nearly a third in the coming decade. However, the ownership structure of Norwegian savings banks may limit consolidation, because mergers require the approval of shareholders, employees, customers and politically appointed representatives (see Bøhren (2014)). 21 4.4. Competition Some of the fall in banks’ cost-to-assets ratios may reflect increased competition. Increased competition can force banks to cut costs, because margins and profitability come under pressure. A number of studies find correlations between competition and banks’ margins and costs. Lian (2017) and Joaquim et al (2019) both show that increased competition lowers banks’ lending margins, while Carbó et al (2009) and Nguyen and Nghiem (2017) document that lower interest margins and increased competition, respectively, often coincide with greater cost efficiency. The literature measures competition using several indicators. There is often low competition in sectors dominated by firms with substantial market power. The literature therefore often uses measures of market share and concentration as indicators of competition, for example the Herfindahl Index or the largest firms’ market share (see Carbó et al). The Herfindahl Index is calculated as the sum of the squares of the market shares of all firms in a sector. High values indicate high market concentration and thus low competition, whereas low values indicate high competition. The literature also uses interest rate margin and return on total capital as indicators of competition, because increased competition can weaken margins and profitability. In Norway, Ulltveit-Moe et al (2013) pointed out three factors that contribute to competition in the Norwegian banking sector. First, Norwegian banks compete with foreign banks with branches and subsidiaries in Norway. Second, 19 The results from the survey suggest that the small banks have the greatest need to invest in digitalisation. 64 percent of the small banks in the survey expect that they will invest more in digitalisation, while only 13 percent of the large banks expect an increase. 20 According to Finanstilsynet, banks with less than NOK 30 billion in corporate exposures cannot expect permission to use the IRB approach (see Finanstilsynet (2018)). 21 The board of trustees/supervisory board, which is a savings bank’s highest decision-making body, shall comprise customers, employees, public sector representatives and owners, if any. 21 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION digitalisation has increased competition from institutions that offer banking services without physical presence. Third, Norwegian banks compete with the bond market to finance firms. Ulltveit-Moe et al concluded that competition from foreign banks has increased, while non-financial enterprises’ bond debt had risen less than their bank debt. The competition indicators from the literature do not provide an unambiguous conclusion on how competition has evolved in the Norwegian banking sector. Foreign banks’ market shares and interest margins indicate increased competition (Charts 18 and 19). However, the fall in the interest margin may just as likely be a result of cost cutting as of changes in the competitive environment. For example, the return on total capital has remained at broadly the same level after the banking crisis, which indicates stable competition. Moreover, the Herfindahl Index signals that competition has declined (Chart 20). Tradable debt securities as a share of total debt (Chart 18) and the market shares of the largest banks (Chart 20) suggest that competition has remained relatively stable. Chart 18 Foreign banks’ market share.1 Tradable debt securities as a share of total domestic debt (C2) owed by the general public. Percent. 1987 – 2019 1) Market share is calculated by each bank’s total exposure in Norway. Sources: Statistics Norway and Norges Bank 0 2 4 6 8 10 0 5 10 15 20 25 30 35 40 1987 1991 1995 1999 2003 2007 2011 2015 2019 Foreign banks’ market share Share of tradable debt securities (r.h.s.) 22 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Chart 19 Interest margin and after-tax return on total capital for the Norwegian banking sector.1 Percent. 1987 – 2019 1) Return on total capital is calculated for all banks and mortgage companies in Norway. Sources: Statistics Norway and Norges Bank Chart 20 Herfindahl Index for all banks in Norway and market share of the five largest banks in Norway.1 1987 – 2019 1) For more information on the data series, see Appendix 1. Source: Norges Bank Competition in the Norwegian banking sector may increase ahead. The Herfindahl Index indicates lower competition in the Norwegian banking sector than in the Swedish and Danish banking sectors (Chart 21), and the majority of the 50 largest banks in Norway fear increased competition in the coming years (see DNB Markets (2020)). 22 Digitalisation may increase competition by making information more easily available, reducing the importance of physical presence and making it easier to switch banks. In addition, PSD2 has 22 In the DNB Markets’ survey, 71 percent of banks responded that their greatest concerns for the next three years were increased competition and pressure on margins. -3 -2 -1 0 1 2 3 0 1 2 3 4 5 6 1987 1991 1995 1999 2003 2007 2011 2015 2019 Interest margin banks Interest margin banks and mortgage companies Return on total capital (r.h.s.) 0 10 20 30 40 50 60 70 80 0 300 600 900 1,200 1,500 1,800 2,100 1987 1991 1995 1999 2003 2007 2011 2015 2019 Herfindahl Index (l.h.s.) Market share of the five largest banks 23 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION promoted new services that make it easier to compare the terms of banking services. For example, a number of banks’ online banking portals now enable customers to check account information in other banks. This may contribute to increasing competition. Chart 21 Herfindahl Index for selected Nordic countries.1 2019 1) Calculated for credit institutions in Denmark, Sweden and Finland. Sources: ECB and Norges Bank 4.5. Changes in the importance of business areas Banks’ cost-to-assets ratios may be affected by changes in business areas’ relative size. Banks’ most important business areas are lending activities, payments and transactions, insurance, estate agency and trading in financial instruments and currencies. Cost-to-assets ratios often vary across business areas. Banks may choose to focus more on a business area with a high costto-assets ratio, among other reasons, because the business area’s return on equity is high. An example of such a business area may be non-life insurance. In recent years, the cost-to-asset ratios and return on equity of large Nordic non-life insurance companies 23 were higher than those of Norwegian banks. Banks’ income does not indicate any substantial changes in banks’ focus on different business areas. Over the past three decades, net interest income, which is primarily generated by lending activity, has accounted for around three-fourths of banks’ total operating income (Chart 22). This suggests that the importance of lending activity is broadly the same today as at the end of the 1980s. At the same time, commission income has edged somewhat down over the past decade (Chart 22). This may indicate a slight reduction in the importance of business areas such as insurance, estate agency, payments and transactions, but the decline in this share may also be a result of 23 Gjensidige Forsikring, Top Danmark, Tryg and Sampo Oyj. 0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000 Norway Sweden Denmark Finland 24 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION increased competition and pressure on margins. Other income, such as gains and losses on financial instruments, has risen a little. Chart 22 Norwegian banks’ income items as a share or total operating income. Percent. 1987 – 2019 Source: Norges Bank Cost-to-assets ratios may also be affected by changes in the relative size in lending to various sectors and groups of borrowers. Reported data for banks’ mortgage companies suggest that the cost-to-assets ratio for residential mortgages loans is lower than for other loans. This may be because residential mortgages are a simpler and more homogeneous product than corporate loans, which makes it easier to automate the credit process for residential mortgages. Chart 23 shows that residential mortgages have accounted for an ever increasing share of banks’ total lending. This may explain some of the decline in banks’ cost-to-asset ratios. Banks have also automated the credit process for other retail market loans, including consumer credit. However, since 1987, such lending has become less important for banks (Chart 23). Chart 23 Norwegian banks’ residential mortgage and other retail market loans as a share of gross lending. Percent. 1987 – 2019 Source: Norges Bank -20 0 20 40 60 80 100 -20 0 20 40 60 80 100 1987 1991 1995 1999 2003 2007 2011 2015 2019 Net interest income Net commission income Other income 0 10 20 30 40 50 60 0 10 20 30 40 50 60 1987 1991 1995 1999 2003 2007 2011 2015 2019 Residential mortgage loans Other retail market loans 31 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Tabell 1 Forecast error (RMSFE) and deviation between actual and predicted growth in the cost-to-assets ratio. Percentage points. 2010 – 2019 Source: Norges Bank Norges Bank estimates Norwegian banks’ costs in an annual stress test of the banks. The preferred model may be useful in this work. With actual values for the explanatory variables, the model estimates that the cost-to-assets ratio will increase from 0.75 percent in 2019 to 0.76 percent in 2020. According to the model, more burdensome regulation will pull up the cost-to-assets ratio, while automation and digitalisation of banking services will pull it down. With GDP growth from the stress test in Financial Stability Report 2019 and simple assumptions for the other explanatory variables, the cost-to-assets ratio increases to 0.86 percent in 2023. 33 Negative GDP growth in 2020 and 2021 increases the cost-to-assets ratio in the two succeeding years, while automation and digitalisation pull in the opposite direction. 6. Conclusion Norwegian banks have reduced operating costs substantially in recent decades, both as a share of income and assets. As a result, Norwegian banks’ average cost-to-income ratio is the lowest in the entire EEA area. These cost-efficiency improvements have made Norwegian banks more resilient to higher losses and reduced the risk of costly crises. The developments during the Covid-19 pandemic may be an example of that. Despite higher credit losses, Norwegian banks have turned a profit and have largely maintained credit supply. By comparison, the largest German, Belgian and Italian banks as a whole posted losses in 2020 Q1, even though their credit losses were lower than Norwegian banks’. Our dataset shows that Norwegian banks have reduced all large cost items relative to assets. Nearly half of the decline in cost-to-assets ratios is due to banks’ reduction of wage and personnel expenses relative to assets, which 33 I project the indicators for burdensome regulation and automation of payment services using average growth for the period 2018-2019. The indicator for internet usage is projected using population projections from Statistics Norway (Chart 10). Preferred model AR(1) model Random walk 2010 1.8 -1.3 2.6 2011 5.0 -8.4 -7.0 2012 -1.9 -4.8 3.4 2013 -7.3 -14.0 -9.1 2014 -0.1 -1.9 11.8 2015 -1.1 2.1 4.3 2016 2.4 -0.3 -2.2 2017 -5.9 -11.0 -10.7 2018 4.1 -1.7 9.0 2019 -7.2 -10.9 -9.0 RMSFE 4.4 7.4 12.1 32 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION primarily reflects a reduction in the number of employees. The decline in the cost-to-assets ratio has been dampened by the average wage of bank employees, which has risen faster than the average wage in Norway. Other operating costs have also fallen relative to assets, among other reasons, as a result of the reduction in the number of bank offices. Increased costs for external services and IT has restrained the decline in cost-to-assets ratios. To ascertain the primary drivers of improvements in Norwegian banks’ cost efficiency, I have estimated a model for developments in the cost-to-assets ratio using a number of indicators from theory and literature. The model explains cost developments well. According to the model, automation and digitalisation have reduced the cost-to-assets ratio, while more extensive and complex regulation have contributed to keeping the cost-to-income ratio elevated. In addition, the results suggest that lower economic activity lowers the cost-to-income ratio and vice versa. The model estimates indicate that automation and digitalisation of banking services have been decisive for Norwegian banks’ cost-efficiency improvements. According to the model, the automation of payment services was the primary driver of the cost-efficiency improvements at the end of the 1980s and much of the 1990s. The model estimates also suggest that the transition to online banking services has been crucial for the cost-efficiency improvements in the past two decades. Internet usage in Norway has grown considerably since the end of the 1990s, and in 2019, the share of online banking users in Norway was the highest in Europe. Online and mobile banking, payment apps and other web-based services have made bank customers more self-sufficient and reduced the need for bank personnel and bank offices. Moreover, digitalisation has also enabled banks to automate other aspects of banking, including customer contact and processing of loan applications. In addition, digitalisation may have increased competition and economies of scale in the banking sector, which may have further contributed to cost-efficiency improvements. According to the model, automation and digitalisation may contribute to further reductions in cost-to-asset ratios ahead. A number of studies indicate that the banking sector will continue to invest substantially in digitalisation. In addition, the Covid-19 pandemic may further speed up digitalisation, because increasing digital customer contact is an effective containment measure. 33 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION References Amel, D., C. Barnes, F. Panetta and C. Salleo (2004): “Consolidation and efficiency in the financial sector: A review of the international evidence”, Journal of Banking and Finance, vol. 28, issue 10, pp 2493-2519. Andersen, H. and M.A. Walle (2015): “What explains developments in business investment?”, Staff Memo 2/2015, Norges Bank. Andersen, H., E. Husabø and M.A. Walle (2016): “What influences household demand for goods and services?”, Staff Memo 4/2016, Norges Bank. Andersson, M., C. Kok, H. Mirza, C. Móré and J. Mosthaf (2018): “How can euro area banks reach sustainable profitability in the future?”, ECB Financial Stability Review, November 2018. Andreeva, D., M. Grodzicki, C. Móré and A. Reghezza (2019): “Euro area bank profitability: where can consolidation help?”, ECB Financial Stability Review, November 2019. Arts Council Norway (2014): “Oslonett 1993-94”, December 2014 (Norwegian only). Basel Committee (1988): “International convergence of capital measurement and capital standards”, Basel Committee on Banking Supervision, July 1988. Basel Committee (1996): “Overview of the amendment to the capital accord to incorporate market risks”, Basel Committee on Banking Supervision, January 1996. Basel Committee (2006): “International convergence of capital measurement and capital standards – A revised framework. Comprehensive version”, Basel Committee on Banking Supervision, July 2006. Basel Committee (2010): “Basel III: A global regulatory framework for more resilient banks and banking systems”, Basel Committee on Banking Supervision, December 2010. Basel Committee (2017): “Basel III: Finalising post-crisis reforms”, Basel Committee on Banking Supervision, December 2017. Beccalli, E. (2007): “Does IT investment improve bank performance? Evidence from Europe”, Journal of Banking & Finance, no 31 (7), pp 2205–2230. Beccalli, E., M. Anolli and G. Borello (2015): “Are European banks too big? Evidence on economies of scale”, Journal of Banking and Finance, no 58, pp 232-246. Berger, A.N. and L.J. Mester (1997): “Inside the black box: What explains differences in the efficiencies of financial institutions?”, Journal of Banking and Finance, no 21, pp 895-947. Berger, A.N. (2003): “The economic effects of technological progress: Evidence from the banking industry”, Journal of Money, Credit and Banking, Vol. 35, no 2, pp 141–176. 34 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Bessen, J.E. and C. Righi (2019): “Information Technology and Firm Employment”, Boston Univ. School of Law, Law and Economics Research Paper, no 19-6. Borchgrevink, H., Y. Søvik and B. Vale (2013): “Why regulate banks?”, Staff Memo 16/2013, Norges Bank. Bøhren, Ø (2014): “Økonomiske særtrekk ved stiftelser” [Financial characteristics of foundations], Nordisk Tidsskrift for Selskapsrett, pp 1–14 (Norwegian only). CaixaBank (2020): “CaixaBank and Bankia Boards of Directors approve merger plan to create Spain’s leading bank”, Press release, September 2020. Carbó, S., D. Humphrey, J. Maudos and P. Molyneux (2009): “Cross-country comparisons of competition and pricing power in European banking”, Journal of International Money and Finance, vol. 28, no 1, pp 115-134. Cyree, K.B (2016): “The effects of regulatory compliance for small banks around crisis-based regulation”, The Journal of Financial Research, no 3, pp 215-245. Deloitte (2017): “Regulatory productivity: Is there an answer?”, RegTech position paper, Financial Services. Demirgüç-Kunt, A. and H. Huizinga (2011): “Do We Need Big Banks? Evidence on Performance, Strategy and Market Discipline”, World Bank Policy Research Working Paper, no 5576. Depman, J. (2016): “2016 Regional and Community Banking Industry Outlook Survey”, KPMG. Dijkstra, M. (2013): “Economies of Scale and Scope in the European Banking Sector 2002-2011”, Amsterdam Center for Law & Economics Working Paper, no 2013-11. DNB (2019a): “DNB Presentation 4Q18”, February 2019. DNB (2019b): Annual report 2018, March 2019. DNB (2020): Annual report 2019, March 2020. DNB Markets (2020): “2020 Norwegian Bank Survey: Overall position of strength”, August 2020. EBA (2019): “Risk Assessment Questionnaire – Summary of the Results. Autumn 2019, European Banking Authority. Elliehausen, G.E. (1998): “The cost of banking regulation: a review of the evidence”, Staff Studies 171, Board of Governors of the Federal Reserve System (US). 35 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Fell, J., M. Grodzicki, D. Krušec, R. Martin and E. O’Brien (2017): “Overcoming non-performing loan market failures with transaction platforms”, ECB Financial Stability Review, November 2017. Feng, G. and A. Serletis (2009): “Efficiency and productivity of the US banking industry, 1998–2005: Evidence from the Fourier cost function satisfying global regularity conditions”, Journal of Applied Econometrics, no 24, pp 105–138. Finance Norway (2013): “Vi har banken i lomma” [We have the bank in our pockets], March 2013 (Norwegian only). Finance Norway (2015): “Mobilbanken – den nye hverdagsbanken?” [Mobile bank – the new everyday bank?], March 2015 (Norwegian only). Finance Norway (2018a): “Finansnæringens arbeid mot kriminalitet – Trusler og sårbarheter“ [The financial industry’s efforts to combat crime – Threats and vulnerabilities], January 2018 (Norwegian only). Finance Norway (2018b): “Ny hvitvaskingslov: Større ansvar tildeles finansnæringen” [New Money Laundering Act: Greater responsibility for the financial industry], October 2018 (Norwegian only). Finance Norway (2020): “Forbrukerog finanstrender 2020“ [Consumer and finance trends 2020], May 2020 (Norwegian only). Financial Industry Authorisation schemes (2020): “Forbrukerog finanstrender 2020” [Consumer and finance trends 2020], May 2020 (Norwegian only). Finanstilsynet (2018): “Krav til banker som søker om IRB” [Requirements for banks applying for IRB], November 2018 (Norwegian only). Greer S., G. Lodge, J. Mazzini and E. Yanagawa (2019): “Global Tech Spending Forecast: Banking Edition, 2019”, Celent Report, March 2019. Hagelund, K., E.W. Nordbø and L. Sauvik (2017): “Lønnsandelen” [The labour cost share], Aktuell kommentar, 9/2017, Norges Bank (Norwegian only). Haldane, A. (2012): “The dog and the frisbee”, Speech at the Federal Reserve Bank of Kansas City’s 366th economic policy symposium, “The changing policy landscape”, Jackson Hole, Wyoming, August 2012. Handelsbanken (2020): “Handelsbanken gathering its forces at branches, accelerating the pace of digital development and cutting costs”, Press release, September 2020. Hogan, T.L. and S. Burns (2019): “Has Dodd–Frank affected bank expenses?”, Journal of Regulatory Economics, no 55, pp 214–236. Hui, V., R. Myers and K. Seymour (2016): “Regulatory Burden Financial Impact Study”, Credit Union National Association, February 2016. Huljak, I., K. Mikkonen, C. Móré and C. Perales (2018): “Digitalisation and its impact on banks’ costs and profitability”, ECB Financial Stability Review, November 2018. Box A. 36 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Huljak, I., R. Martin and D. Moccero (2020): “Cost-Efficiency and Productivity of euro area banks”, SUERF Policy Note, no 135. Humphrey, D.B. and B. Vale (2004): “Scale economies, bank mergers, and electronic payments: A spline function approach”, Journal of Banking Finance, no 28 (7), pp 1671 - 1696. Joaquim, G., B. Van Doornik and J.R. Ornelas (2019): “Bank Competition, Cost of Credit and Economic Activity: evidence from Brazil”, Banco Central Do Brasil Working Paper, no 508. Jonas, M.R. and S.K. King (2008): “Bank efficiency and the effectiveness of monetary policy”, Contemporary Economic Policy, Vol. 26, no 4, pp 579-589. KPMG (2016): “Rise of the Robots”, April 2016. Kjærnsrød, S. (2001): “Et skråblikk på Oslonetts webtjenester i perioden 1993 til 1997” [A sideways glance at Oslonett’s web services in the period 1993 to 1997], University Center for Information Technology (USIT), December 2001 (Norwegian only). Lian, Y. (2017): “Bank competition and the cost of bank loans”, Review of Quantitative Finance and Accounting, no 51, pp 253–282. Lindquist, K.-G. (2002): “The effect of new technology in payment services on banks' intermediation”, Working Paper 2/2002, Norges Bank. Lund, A. and K.B. Nordal (2017): “Endringer i bankreguleringen etter finanskrisen i 2008” [Changes in banking regulation following the financial crisis in 2008], Aktuell kommentar 5/2017, Norges Bank (Norwegian only). Nguyen, T. and S.H. Nghiem (2017): “The effects of competition on efficiency: the Vietnamese banking industry experience”, The Singapore Economic Review, vol. 63, no 1. Norges Bank (2014): “Costs in the Norwegian payment system”, Norges Bank Papers 5/2014, Norges Bank. Norges Bank (2020): Retail payment services 2019, Norges Bank Papers 1/2020, Norges Bank. Rhoades, S.A. (1998): “The efficiency effects of bank mergers: An overview of case studies of nine mergers”, Journal of Banking and Finance, no 22, pp 273–291. Skule, S. and T. Grytli (1997): “Teknologisk utvikling og samfunnsendring - Eksempler fra oljehistorien og bankhistorien” [Technological developments and social changes: Examples from the histories of oil and banking], Fafo Report no 217 (Norwegian only). SpareBank 1 Østlandet (2016): “20 år siden nettbank-kupp” [Twenty years since online banking coup], September 2016 (Norwegian only). 37 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Schmitz, M. and M. Tirpák (2017): “Cross-border banking in the euro area since the crisis: what is driving the great retrenchment?”, ECB Financial Stability Review, Special Feature C, November 2017. SpareBank 1 Nord-Norge (2020): “Stenger 16 bankkontorer” [Closing 16 bank offices], Oslo Børs company announcement, September 2020. Thornton, G. (1993): “Regulatory Burden - The Cost to Community Banks”, Study prepared for the Independent Bankers Association of America, January 1993. Ulltveit-Moe, K.H, B. Vale, M.H. Grindaker and E. Skancke (2013): “Competitiveness and regulation of Norwegian banks”, Staff Memo 18/2013. Norges Bank. 38 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Appendix 1 – Data series Banks’ operating costs Total operating costs for all banks and mortgage companies in Norway. 1987 – 2019. Annual data. In NOK Banks’ operating income Total operating income for all banks and mortgage companies in Norway. 1987 – 2019. Annual data. In NOK Banks’ total assets Aggregate total assets of all banks and mortgage companies in Norway. Average total assets for the year in question. Average total assets for 1987 is estimated using data for the end of January and November 1987. 1987 – 2019. Annual data. In NOK Banks’ return on total capital Total after-tax profit as a percentage of average total assets for all banks and mortgage companies in Norway. Percent. 1987 – 2019 Banks’ size Average total assets of all banks and mortgage companies in Norway as a share of GDP mainland Norway measured as market value at current prices. 1987 – 2019 Number of banks Number of banks in Norway. 1982 – 2019 Herfindahl Index for the Norwegian banking sector Herfindahl Index for all banks in Norway (parent bank data). The index is calculated by summing the squares of banks’ market shares in percent. Market shares are measured by total assets. The index generates values between 0 and 10000. 1987 – 2019 Market share of the five largest banks Market share of the five largest banks in Norway measured by total assets. Percent. 1987 – 2019 Foreign banks’ market share Foreign banks’ market share measured by each bank’s total exposures in Norway. Percent. 1987 – 2019 Number of ATMs Number of ATMs in Norway. 1982 – 2019 Number of payment terminals Number of payment terminals owned by banks and others in Norway. Observations for the period 19911993 are estimated using the number of bank-owned payment terminals owing to insufficient data. 1987 – 2019 39 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Share of electronic payments Number of electronic payment transactions as a share of total debit and credit transfer (giro), payment card (goods purchases) and cheque transactions in Norway. Calculated share does not include payment card transactions prior to 1991. Data for electronic giros prior to 2002 do not include miscellaneous credit transfers, including standing payment orders. 1984 – 2019 Share of persons who have used the internet Share of a representative sample of the Norwegian population that has used the internet on an average day. 1997 – 2019. Estimates for 1987-1996. The internet was not commercially available in Norway prior to 1993. I therefore assume that the share was 0 up to and including 1993 and that it increased linearly until 1997. Projections for the period 2020-2030 are based on Statistics Norway’s population projections (main alternative) and assumptions that the current population will maintain its internet usage and internet usage of new residents will be at the same level as today’s younger population cohorts Share of persons who use online banking Reported share of respondents to grocery market survey who use online banking services. The survey is conducted by Kantar TNS in collaboration with Finance Norway. 2000 – 2018. The share is approximated by interpolating for the years 2001, 2002, 2004, 2006 and 2007 owing to insufficient observations. From 2019, the share of online banking and mobile banking is reported together Share of persons using mobile banking Reported share of respondents to grocery market survey who use online banking services. 2010 – 2018 Number of Finanstilsynet employees Number of Finanstilsynet employees. 1987 – 2019 GDP GDP for mainland Norway. Rebased volume. 1971 – 2019 Financial crises Financial crises in Norway. 1987 – 2019 Share of residential mortgages Total residential mortgage lending of all banks and mortgage companies in Norway as a share of gross lending. 1987 – 2019 40 NORGES BANK STAFF MEMO NO. 9 | 2020 THE COST EFFICIENCY IMPROVEMENT OF NORWEGIAN BANKS CAN BE EXPLAINED BY AUTOMATION AND DIGITALISATION Share of retail market loans Total retail market lending of all banks and mortgage companies in Norway as a share of gross lending. 1987 – 2019 Share of net interest income Net interest income of all banks and mortgage companies in Norway as a share of total operating income. 1987 – 2019 Share of net commission incomer Net commission income of all banks and mortgage companies in Norway as a share of total operating income. 1987 – 2019 Appendix 2 – Documentation of estimations Table 1 Preferred model Dependent Variable: D(LOG(COSTS)-LOG(ASSETS)) Method: Least Squares Date: 09/01/20 Time: 09:33 Sample (adjusted): 1989 2019 Included observations: 31 after adjustments Variable Coefficient Std. Error t-Statistic Prob. C 0.019711 0.023589 0.835583 0.4110 D(LOG(ELECTRONIC_PAYMENT(-1))) -0.405302 0.100481 -4.033620 0.0004 D((INTERNET(-2))) -0.010640 0.003462 -3.073715 0.0049 D(LOG(FSA_EMPLOY2(-1))) 0.507363 0.200310 2.532895 0.0177 D(LOG(GDP(-2))) -1.251506 0.589337 -2.123584 0.0434 R-squared 0.548377 Mean dependent var -0.045209 Adjusted R-squared 0.478897 S.D. dependent var 0.073678 S.E. of regression 0.053186 Akaike info criterion -2.883337 Sum squared resid 0.073549 Schwarz criterion -2.652049 Log likelihood 49.69173 Hannan-Quinn criter. -2.807943 F-statistic 7.892539 Durbin-Watson stat 2.013578 Prob(F-statistic) 0.000264 Table 2 Stationarity test Null Hypothesis: RESIDUAL has a unit root Exogenous: Constant Lag Length: 0 (Automatic - based on SIC, maxlag=7) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -5.542469 0.0001 Test critical values: 1% level -3.670170 5% level -2.963972 10% level -2.621007 *MacKinnon (1996) one-sided p-values.