The announcement by MAS to accept digital banking licenses in Singapore brought back memories of my visits to the bank in the late 1990s-early 2000s. The local banks had just introduced Internet banking. Functions were limited, and adoption was slow. Most people relied on human assistance at the banks’ branches to fulfill their banking needs.
A trip to the bank was an arduous affair for the family. My mother would plan meticulously for the visits. Arriving in the morning meant that you had to stand in line with the elders who tended to ask numerous questions. Plus, you had to wait for the staff to “set up their system.” Arriving near or during lunchtime was a strict no-no, as the office workers would be running their banking errands at this time. Arriving in the early afternoon would be risky as the post-lunch food coma could make the staff less productive. According to my mother, the optimal time to visit would be slightly before the closing hour. The risk of having to work over-time would translate to the staff having the highest sense of urgency to process customers’ requests.
Further, she would double and triple-check that the papers were in place. These documents included a whole range of miscellaneous records, such as photocopies of every family member’s passport details, phone bills, bank passbooks, key registration titles, and certifications, etc.…just in case the staff asked for them. Every successful visit would see my mother wear a proud grin for the rest of the day.
Fast forward to today, one can safely execute most banking requests online, and I do not recall when I last set foot in a bank. In the age of everything digital, it only seemed a matter of time before banking would go fully digital too.
Several countries have already seen digital banks operating in their territories. For instance, China’s WeBank (owned by Tencent) obtained a private bank license in 2014 [1]. UK’s Starling Bank received its banking license in 2016 [2]. Other Asian states, including Hong Kong, Malaysia, and Taiwan, have also introduced digital banking licenses in the past year. Figure 1 shows 50 digital banks that were ranked by the Financial IT magazine in 2017 [3]. Unsurprisingly, Tencent’s WeBank and Alibaba’s MYBank clinched the top two spots, respectively.
In Singapore, the MAS has received 21 applications as of 7 January 2020 [5]. Various consortiums have sprung up, from telcos partnering ride-share companies to gaming companies joining forces with supermarkets. Each must demonstrate how they plan to address unmet and underserved customer needs. Increased convenience, more personalized services, greater financial inclusion, and better interest rates are some benefits that consumers can expect.
A topic that has been on my mind is the cost structure of digital banks and the implications on profitability. If there are so many interested applicants, then surely this has to be a profitable venture, right?
In this post, I adopt a data-driven approach to compare the financial metrics of some banks. My goal is to determine whether digital banks are indeed more profitable than traditional banks.
Digital Banks Vs. Traditional Banks
One of the most contrasting differences between conventional and digital banks is that the latter do not operate physical branches and tellers. Therefore, digital banks save a large chunk of the operating expenses that traditional banks incur, such as staff and property costs. How do the overall operating costs between the two types of banks differ? What do the differences in operating models mean for the banks’ profitability?
As an engineer by training, I could not resist the urge to dissect the financial statements of the banks to compare their profitability metrics. My approach is summarized as follows.
(i) Identify suitable digital banks and traditional banks for the study,
(ii) Obtain their respective financial statements, and
(iii) Compare and contrast the banks’ profitability metrics over some time
Identifying suitable banks
This proved to be a much harder task than I had expected. While there are various digital banks in the world, it is challenging to access the financial statements of these banks. Many are still start-ups (e.g., Germany’s N26) or private companies (e.g., Japan’s Jibun Bank) that do not need to disclose their financial statements. For the publicly listed companies (e.g., Tencent and its Webank), many tend to provide consolidated numbers for the parent company. Admittedly, the availability of financial statements guided my choice of banks. I was lucky to find the financials of Sony Bank, a digital bank in Japan. I will use the findings from Sony Bank to represent digital banks in this post henceforth.
A 2014 study by a group of economists showed a negative relationship between a bank’s size and its operating cost [6]. While the research was conducted on US banks, I have assumed that the negative correlation also applies to Japanese banks. Hence, to control for this size variable, I have chosen two banks of similar size (measured by total assets) as Sony Bank. They are Yamagata Bank and Tsukuba Bank, two regional banks in Japan.
A summary of the selected banks is as follows.
· Sony Bank (“Sony”): A digital bank that is a subsidiary of Sony Financial and part of the Japanese electronics giant Sony Corporation. It began operations in 2001 and serves consumers mainly in Japan. It was set up against a backdrop of financial deregulation initiatives that Japan had embarked on since the late 1990s. Japan is one of the first movers globally to welcome digital banks.
· Yamagata Bank (“Yamagata”): A Japanese regional bank headquartered in Yamagata City, Yamagata prefecture. The prefecture has a population of approximately 1.09 million, and its primary industries are farming (agricultural products like rice) and manufacturing (traditional crafts and electronics). The bank has 80 branches across the country, of which 69 are in Yamagata prefecture.
· Tsukuba Bank (“Tsukuba”): A Japanese regional bank headquartered in Tsukuba City, Ibaraki Prefecture. It serves consumers predominantly in Japan, with a vast majority of its branches (134 out of 148) located in Ibaraki prefecture. The prefecture’s primary industries include agriculture and energy production. Tsukuba Bank was formed in 2010 with the merger of Kanto Tsukuba Bank and Ibaraki Bank.
Obtaining financial statements
For this exercise, I considered a 10-year period from FY09 to FY18. I will put the links to the financial statements in the references section below.
Reviewing the metrics
I constructed a simple valuation model to compare the profitability metrics of the banks from FY09 to FY18. I referenced the consolidated figures for this exercise. Table 2 below shows the numbers for FY18.
Results
A couple of interesting observations stood out to me.
(1) Sony had a higher average cost-to-income ratio (CIR) than Yamagata and Tsukuba.
The CIR, calculated by dividing the operating expense by the operating income, measures how efficiently the bank is run. Generally, the lower the CIR, the more profitable the bank will be.
Operating expenses consist of various items, such as interest expenses, fees and commissions, and general and administrative (G&A) expenses. In particular, G&A expenses include salaries, rent and utilities, marketing, and legal fees.
Since digital banks do not operate physical branches, they incur less G&A expenses, and I had expected their CIR to be lower. However, the data from FY09 to FY18 did not support this hypothesis (Figure 2). Sony recorded an average CIR of 84.6%, higher than that of both Yamagata (69.3%) and Tsukuba (80.8%) (Figure 3).
(2) However, Sony incurred a lower proportion of G&A expenses. The G&A expenses were, on average, 0.71x (as a proportion of operating expense) and 0.81x (as a proportion of operating income) relative to the expenses of the traditional banks.
Taking the analysis a step further, I decided to break down the operating costs. I was particularly keen to study the trends in G&A expenses, as they are typically the most significant cost item. I considered two metrics:
(a) G&A expenses as a proportion of total operating expense, and
(b) G&A expenses as a proportion of total operating income
The data showed that Sony incurred lower G&A expenses for both (a) and (b) compared to the other two banks (Figure 4, 5), and this trend was pretty consistent from FY09 to FY18.
While G&A expenses typically make up the majority of a bank’s total operating costs, it is no surprise that the proportion spent on G&A is lower for digital banks as they incur less property and staff costs. The data from FY09 to FY18 allowed us to quantify this difference — Sony’s G&A expenses were, on average, 0.71x (as a proportion of total operating expense) and 0.81x (as a proportion of total operating income) compared to that of Yamagata and Tsukuba (Figure 6). A remarkable difference indeed.
(3) Sony’s deposit base included foreign currency deposits that generated a higher yield. Total interest expense on deposits was, on average, 4.0x relative to conventional banks. As a result, Sony had a higher average CIR.
Why then did Sony have a higher CIR than Yamagata and Tsukuba?
I believe the answer lies in the second largest cost item — the interest expense.
The interest expense is the cost of borrowing money, such as through deposits, bonds, convertible debt, or other instruments. The amount of each borrowing and the corresponding interest expense depend on the bank’s funding structure.
Figure 7 shows each bank’s average cost item from FY09 to FY18. Sony had an average interest expense of 9,021 million yen, which made up approximately 29% of its total operating cost. This spending was 3.1x and 3.5x relative to the interest expense of Yamagata and Tsukuba, respectively.
The bulk of Sony’s interest expense was for bank deposits, which was, on average, 4.0x relative to that of Yamagata and Tsukuba (Figure 8). All three banks carried similar amounts of deposits as of 31 March 2019 (Sony: 2,358 billion yen; Yamagata: 2,213 billion yen; Tsukuba: 2,246 billion yen). However, the composition of the deposit base was not homogeneous — Yamagata and Tsukuba had mostly local currency deposits (LCD), while Sony also carried foreign currency deposits (FCD). FCD yield a premium for exchange rate risk. Hence the interest on FCD was significantly higher at 1.18% compared to 0.09% for LCD in FY18. Overall, although FCD only made up 18% of the total deposits in FY18, close to 73% of total interest expense was spent on FCD.
(I could not find information on Yamagata and Tsukuba’s yield for LCD, but I suspect that it would be similar to Sony’s yield. One may also reason that Sony’s yield could be higher to win over consumers from incumbent banks.)
For Sony, the yield on FCD followed an upward trend from 0.99% in FY19 to 1.18% in FY18, and interest expense on FCD increased from 3,059 million yen to 4,891 million yen accordingly (Figure 9). On the other hand, interest expense on LCD saw a reduction. This reduction outweighed the increase in interest expense on FCD and led to an overall decrease in interest expense on bank deposits from 8,535 million yen (FY09) to 6,757 million yen (FY18).
In summary, although Sony generated G&A savings, its reliance on FCD brought higher interest costs that led to a higher average CIR compared to Yamagata and Tsukuba. However, total interest expense decreased from FY09 to FY18, driven by a reduction in interest expense on LCD. As a result, the CIR also observed a downward trend (Figure 2).
(4) Sony had a return on equity (ROE) that was, on average, 1.26x to 1.40x relative to the ROE of traditional banks.
ROE is a critical profitability ratio that measures how effectively the company is generating profits from the money that equity investors have put in.
At one glance, we can see that Sony’s ROE was generally higher, and it trended upwards from 1.96% in FY09 to 7.10% in FY18 (Figure 10) — an impressive 15.4% CAGR!
The average ROE from FY09 to FY18 of Sony was 4.43%, and this was 1.26x and 1.40x relative to the average ROE of Yamagata and Tsukuba, respectively (Figure 11). The data showed that Sony, being a digital bank, was more profitable than the traditional banks.
Concluding remarks
The study focused on the cost factors, and we see that Sony’s online operating model and financing structure differentiated it from the cost structure of Yamagata and Tsukuba. Overall, Sony proved to be a more profitable bank with higher ROE.
Most digital banks we see today are start-ups that are still bleeding cash. Like most start-ups, they spend aggressively on customer acquisition to chase “growth” rather than “profitability” at this stage. Will these digital banks ever become profitable? I sense that they could, and their profitability levels could even exceed that of traditional banks, as shown by my study above of the Japanese banks. However, the challenge for these digital banks would be to demonstrate profitability first before they run out of VC money.
Lastly, please note that this post is strictly for my academic interests. I recognize that there are limitations to my study, such as reviewing a small sample size of banks. However, faced with the difficulty of obtaining suitable financial statements, this was the best that I could do. If you have suggestions to improve the study or also have an interest in this topic, please feel free to reach out.
References
[1] Gavin Gunning, S&P Global Ratings. 17 July 2019. The Future of Banking: Virtual Banks Chase the Dream in Asia-Pacific. https://www.spglobal.com/en/research-insights/articles/the-future-of-banking-virtual-banks-chase-the-dream-in-asia-pacific
[2] Starling Bank. 14 July 2016. Starling Bank receives Banking License and announces Faster Payments. https://www.starlingbank.com/news/starling-bank-receives-banking-licence/
[3] Financial IT. 2017. Top 50 Digital Only Banks Ranking 2017, Special Sibos & Money 20/20 Issue. https://financialit.net/sites/default/files/top50_digital_banks.pdf
[4] Fintech News Hong Kong. 13 October 2017. Top 50 Digital Only Banks Ranking 2017. https://fintechnews.hk/2649/various/top-50-digital-banks-ranking-2017/
[5] Monetary Authority of Singapore. 7 January 2020. MAS Receives 21 Applications for Digital Bank Licenses. https://www.mas.gov.sg/news/media-releases/2020/mas-receives-21-applications-for-digital-bank-licences
[6] Anna Kovner, James Vickery, and Lily Zhou. Liberty Street Economics. 25 March 2014. Do Big Banks Have Lower Operating Costs? https://libertystreeteconomics.newyorkfed.org/2014/03/do-big-banks-have-lower-operating-costs.html
Financial statements of banks
· Sony Bank: https://www.sonyfh.co.jp/en/financial_info/annualreport/
· Yamagata Bank: http://www.yamagatabank.co.jp/english/
· Tsukuba Bank: https://www.tsukubabank.co.jp/ir/disclosure/annual.html
