Scale matters …
As many ambitious banks continue to expand, we wondered: What is scale worth today? We looked at the relationship between cost efficiency and assets at more than 3,000 banks around the world. We found a significant relationship between banks’ cost-to-asset (C/A) ratio and their market share. Approximately tripling the market share (that is, increasing the natural logarithm of the market share by one) results in a 50-basis-point (bps) decrease in C/A.
We also tested other metrics such as risk cost, revenue margin, and leverage, which did not yield such a statistical relationship. In the meantime, better cost efficiency also means better return on equity and more substantial value creation for the banks at scale.
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at:
McKinsey_Website_Accessibility@mckinsey.com ... but scale effect differs by sector and geography ...
In general, larger banks are more cost efficient. So far, so predictable. But the research also found that scale effects vary considerably by sectors and increase the efficiency of technology-enabled businesses such as payments or cash equities. Also, they are strongest in digitally advanced markets. Consider Australia and Denmark, two places where banking is rapidly moving online. The lowest quintile of banks in these countries has a C/A ratio above 350 bps, while the top banks are considerably lower, at about 100 bps. An even bigger difference shows up in Russia, for other reasons: despite the central bank’s cleanup program, the Russian banking system is still fragmented with more than 500 banks. Many in the lowest quintile are much less efficient.
In the two big emerging markets of China and India, cost efficiency is associated with scale but to a very different extent. In China, dominated by many corporate banks holding large balance sheets, the top quintile’s C/A (92 bps) is half that of the lowest quintile (184 bps). Yet in India, while some scale effect is visible, larger banks have a C/A higher than 150 bps in all quintiles. Big banks in India operate on a higher cost base, for instance, by maintaining larger physical networks to lend in undeveloped rural areas.
Scale is less visible in the United States. Just 70 bps separate the bottom and top quintiles. That’s partly explained by the large off-balance-sheet business of top US banks; all their costs are reported, but their asset base appears smaller than it actually is. But scale effects could be expanding. US banks are on a path of digitization and might soon achieve results like top banks in Australia and Denmark. For instance, many large US banks are launching a digital bank under a different brand name—a practice already visible in Europe and Asia for a decade.
… and scale does not determine fate
These are important differences. But they do not determine a bank’s destiny. A regression analysis on the factors affecting the C/A finds that just 8 percent is explained by a bank’s size. Business mix, outsourcing, regional platform opportunities, and efficiencies in the branch network and head office are other important drivers of cost efficiency. Interestingly, digitization can also have adverse effects: while it helps with sales through digital channels and with branding on social media, next-generation technologies such as cloud computing make the heavy fixed costs of digital redundant for smaller banks.
The takeaway? Scale still matters, more so in certain geographies, and should be a primary consideration when banks look to grow. Smaller banks should consciously offset competitors’ scale benefits in three ways: by choosing a niche area where scale is less relevant (such as people-driven, local businesses in retail and corporate banking), choosing a service model that can benefit from cooperation with others on industry utilities, and partnering with fintechs to accelerate their innovation.
The Panorama team wishes to thank Amit Garg, Clara Aldea Gil de Gomez, Rushabh Kapashi, Dmitry Kholodov, Arielle Pensler, Luisa Quetti, and Gabriela Skouloudi for their contributions to this article.