Does digitalisation increase bank profitability?

Does digitalisation increase bank profitability?

September 2019

Giulia Bassani, Sara Emiliani, Lorenzo Prosperi

Digitalisation can be an important and permanent cost-saving strategy for banks, particularly in countries with a dense branch network

 

Bank profitability in Italy and in the euro area has improved from the post-crisis lows. However, for many banks, earnings are still below what is required by investors and the recent slowdown in economic growth could threaten the recovery of banks’ profits. Why is this a problem? Why does banks’ profitability matter overall, not only for their shareholders? The reason is that persistently low profitability can limit banks’ ability to generate capital and can make raising capital very costly. Therefore, it becomes harder to build up buffers against unexpected shocks and this can bind banks’ capability to provide financial services to households and businesses.

Weak profitability is one of the key challenges facing the euro area banking sector. Our analysis explores some of its drivers, including digitalisation.

Reducing costs and improving efficiency are necessary steps. Possible strategies to achieve them range from consolidation via mergers and acquisitions, downsizing, as well as branch closures and stuff reduction. But another, not necessarily alternative, way would be to adopt new, cost-saving technologies aimed at digitalising financial intermediation services, in particular those with low added value.

Digitalisation can be an important and permanent cost-saving strategy for banks, particularly in countries with a dense branch network, although depending on structural factors such as labour laws, population density and the level of digitalisation in society. Greater automation and more efficient processes allow cost savings and relieve employees of routine work, allowing them, if appropriately qualified, to take on more demanding tasks. An efficient and modern IT infrastructure is one of the prerequisites to achieve such efficiencies and must, at the same time, be safeguarded against its heightened vulnerability to cyber attacks. In addition, digital ‘leaders’ may also benefit from additional revenues via market share gains. This will require considerable investment by banks in the short to medium term, with cost savings typically likely to materialise in the medium to long term. At the same time, a higher reliance on digitalised forms of financial services may also threaten the stability of revenues, as it becomes easier for customers to shop around and compare banks’ products and prices. This trade off makes it particularly interesting to have a closer look at the overall impact of digitalisation on bank profitability.

Some individual banks have managed to invest large amounts of money in digitalisation in recent years notwithstanding the poor profitability of the euro area banking system as a whole. Figures 1-2 show tech expenses for a sample of European and Italian banks. In terms of IT expenses, Italian banks seem to lag behind: in 2018, IT costs were just 0.12% of total assets in Italy against 0.17% on average for their European peers. In fact, euro area banks (excluding Italian banks) have invested in IT even during the financial crisis or soon after (the ratio increased already in 2013), while it has been at the same level for Italian banks, thus widening the gap with European peers.

 
Fig. 1 EMU (excl. Italy), IT costs - % values Fig. 2 Italy, IT costs - % values
Source: Prometeia calculation on SNL, ABI and financial statement data
 

What can explain this gap in IT investments? And can investments in IT really boost profitability? To answer these questions, we estimate a panel data model with country fixed effects on a sample of 124 European banks from 2006 to 2018 and test if additional investments in IT improve profitability. We also investigate the determinants of the choice of investing in Information & Technology.

 
Impact of IT expenses on bank’s profitability

Results from the regressions run show that an increase in IT expenses can improve banks’ profitability, up to 1.1 percentage points (for the sample including all banks) and 0.8 percentage points for the Italian banks (Fig.3). The positive and statistically significant coefficient of credit risk in the EMU sample can be explained by the fact that the dependent variable (ROA) is computed as pre-impairment profits to total assets and therefore it does not incorporate the impact on profitability of loan loss provisions. Higher credit risk may result from riskier assets, such as riskier loans, which entail higher pre-impairment returns. Also higher personnel expenses determine greater return on assets. This result may be explained by the fact that staff expenses in banks that invest more in IT might already reflect the compensation of more skilled and specialised employees. Finally, concerning total assets, the empirical results show that the size of a bank does not necessarily affect its profitability.

Fig. 3 Results of the regression on bank profitability (ROA)
Source: Prometeia calculation on SNL, ABI and financial statement data
 
Determinants of the choice of investing more in digitalisation

Balance sheets’ weaknesses affect the ability to invest in IT. The results of the second regression (Fig. 4) show that higher credit risk entails lower investments in digitalisation. Moreover, higher cost of funding has a negative effect on the decision of investing in IT. On the contrary, higher profitability the previous year is particularly relevant to the choice of allocating more resources to IT investments. Also labour costs to total assets contributes positively, since the variable may reflect the severance payments following recent reorganizations, aimed precisely at enhancing the digital transformation of the bank, or it could already incorporate the higher cost of labour of a more specialised workforce, which typically goes hand-in-hand with more IT investments. Moreover, banks with a more solid capital position (equity to total assets) are more capable of investing in digitalisation. Finally, the loan-to-deposit ratio – that can be interpreted as a business model variable – does not have any significant impact.

Fig. 4 Results of the regression on IT expenses
Source: Prometeia calculation on SNL, ABI and financial statement data
 
The contribution of some determinants for the main European countries

We also tried to understand whether the relevance of some particular variables in driving higher investments in IT might vary across countries.  Figure 5-7 show results for Italy, Spain, Germany and France. Italian and Spanish banks appear to rely on previous year’s profitability when deciding whether to invest more in technology or not. The share of loan loss provisions over total assets is a great disincentive for Italian banks, especially if compared to the other countries. On the contrary, Italian banks’ IT investments are the least affected by their cost of funding, while investments by French banks appear to suffer significantly from higher costs of funding.

 
 

In conclusion, our analysis shows that banks’ digital transformation can lead to profitability gains, but it needs to be accompanied by structural changes and balance sheet adjustments that create an environment conducive to digitalisation. Even if there is no “one-size-fits-all” strategy for banks to return to sustainable profitability, investments in information technology may play an important role. Supporting the digital transformation will be crucial to further adapt banks’ business models to cope with the rapid pace of technological innovation in financial services, increasing competition from Big Tech and fintech companies as well as the need to strengthen their resilience to cyber threats.

 
[1] To do so, we computed the contribution of each variable by multiplying the coefficients of regression (2) by the correspondent observation. We then computed the mean of these new vectors, per year and per country. Figures 5-7 show results for the 2017, being the most recent year with a high number of observations.
Read All