When we talked about small caps and banks being different last month, we skipped over the part as to why banks and lending businesses are more successful as large caps than small caps. One key reason why lending businesses tend to succeed more as large caps and fail more as small caps is driven by scale-based diversification.

On the asset side of the balance sheet – the loans extended to customers – greater diversification reduces potential maximum drawdown. For example, take a lender that just specializes in equipment lending to NSW wheat farmers. You can spin the book as being diversified by postcode, equipment type, gender, etc – but the reality is this loan book is one East Coast drought away from significant losses. A larger business can diversify across multiple exposures (multiple industry sectors in a commercial loan book, plus a consumer book, plus a mortgage book, plus a geographically diverse book, etc.). This doesn’t replace the need for strong origination and risk controls, but all things being equal, the likely maximum loss position on a well-diversified loan book should be much lower than a concentrated one if the sources of diversification are uncorrelated.

This also applies to the liability side of the balance sheet. The ‘liability side’ simply means that to generate acceptable returns to equity investors from a portfolio of credit assets, you need to leverage up the credit spread. Diversification helps here as well. Having a mix of funding sources (deposits, bonds, loans, private credit, securitization, etc.) and a mix of funding parties (institutions, consumers, local, offshore, etc.) reduces risk on the funding side of the business. Larger lending businesses tend to have more diversified and less correlated sources of funding which lowers the risk here also.

Most finance business lever around 10x, which means every dollar of equity supports $10 of assets (loans) and $9 of debt. This means 2% losses at the asset level translate to 20% losses at the equity level, driving the need to control the downside risk through diversification and risk management.

Money is the ultimate commodity. It is very hard to differentiate as a product.

Related to the commodity nature of money is the difficulty of sustaining large excess returns on equity (ROE) in the lending business. The big four Australian banks are somewhat global outliers in their average ROE’s being close to 15% over the past 25+ years. In most markets, and most segments the more normal outcome are ROE’s that are closer to 10%, and exhibit a strong mean reverting tendency to that level.

ROE mean reversion tends to result from a lack of sustainable pricing power. Companies can move their ROE up temporarily by improving the terms for shareholders – such that the earnings spread between their assets and liabilities expands, however if this is at the customer’s expense (higher than competitor interest rates, poorer terms) it is likely the customer will seek an alternative. After all, money is money. Or there may be temporary economic cycle factors which improve or reduce margins and ROEs for all players, but again eventually competitive forces rebalance over time.

One of the risks we are alert to as investors is misclassification. It doesn’t happen often, but sometimes the market will completely misclassify a business. One example was Costa Group in early 2018. After a couple of good years of blueberry sales due to high prices, the market decided Costa was no longer seen as a tenant farming business trading at almost 7x book value, but rather it was a consumer food brand trading at a deserved 20x P/E multiple. Having all the farmers in Coffs Harbour (NSW) replace their banana crops with blueberries and the resulting supply increase reducing prices for everyone (including Costa) eventually disabused investors of what Costa really was, and the stock traded back to sub 2x book value. We would say book was always the appropriate valuation anchor and it was never a real branded consumer business.

Likewise, the ‘software originated’ lending boom of 2020 / 2021 (Credi, Prospa, Plenti, MoneyMe, Money3, Zip, Flexigroup/Humm, etc.) was anchored on the idea of a 4x EV / Sales multiple (per the IPO pricing in prospectuses). These businesses have none of the economics of a SaaS business, and the more appropriate multiple was 1.1 – 1.3x book value in our view which we expressed at the time. This would have resulted in market pricing 80% lower than achieved via IPO, so of course it was not in the issuers’ interest to view it like this, but as shown over the subsequent three years share price performance, it was absolutely in the investors’ interest.

Another example was Resimac (RMC). RMC wasn’t ever promoted as anything other than a very well run non-bank lender. After a few terrific years the ROE had expanded from 14% in F2017 to 38% by F2021 taking EPS from 4cps to 26cps. A number of commentators were puzzled as to why the market wouldn’t pay more than 8x EPS for a business with such impressive historic EPS growth and high returns on equity, but as a balance sheet driven lending business, the market was already paying a big premium at 2.5x book value. The strong force of ROE mean reversion has since exerted itself and RMC ROE returned to 8.3% in F2024 and EPS fell back to 9cps. Since May 2021, RMC delivered total returns to shareholders of -48% as the price to book declined from 2.5x to 0.8x and Westpac (as an example) delivered +57% as a larger, more diversified and arguably less mispriced alternative (price to book was 1.3x in mid 2021 vs 1.5x today).

Clearly our fund performance would have been much better had we owned Zip Co. ZIP is one of our largest relative detractors over the past 12 months. However as quality investors, we struggle to see the ability for what is ultimately a lending business to sustain returns high enough to warrant the >5x book value multiple it is priced at today.

What about a year ago? Well credit to the ZIP management team, they have done an amazing job of moving the business to a profitable, cash flow positive position and even having a chance of generating mid-teen returns on capital (consensus estimates ROE of around 12.5% in F2026 and 14.5% in F2027). They also seem to understand that a real way of adding value through the cycle is to raise capital when the market is paying a high multiple on book for their shares – as they did earlier this year at more than 4x book value. If there is one criticism of major banks (and there are many from which to choose) it would be their historic tendency to buy back stock when times are good and price to book multiples are high, only to then issue stock at much lower prices when Black Swans appear. Buying high and selling low works about as well for long-term bank shareholder returns as it does in all other forms of investing.

One of the reasons we didn’t buy ZIP a year ago is the same reason we are not buying Humm (HUM) or Solvar (SVR) or Pepper Money (PPM) today at less than 1x book value. They may be cheap. They may have good fortune ahead the market cannot see, and this may cause the PBV to increase again to 1x or 2x or even more. But unlike the business we do seek to own, we struggle to see how in the long run being small in this type of business isn’t a disadvantage and a risk to long-term capital compounding.