How to analyse a bank? Fear not the acronyms

As a follow up to my last post, number of readers have asked me to recommend my favourite publication for explaining the basics of bank analysis. Thinking about this question, I realised that most of these fall into one of two categories – either too technical, or too general. So I decided to take matters into my own hands and here is the first instalment, of what depending on the reader feedback, may become a continuous series. 

It is often said banks are black boxes but that's not really true. Anything designed by people rather than being a given (e.g. fundamentals laws of physics) has to have underlying logic to it and so it is with banks. Perhaps the best place to start is asking how are banks different from corporates? First of all, it is much tougher to follow the cash flow cycle in a bank than a corporate. Let's take a traditional company that manufactures physical products. It sells products to the customers and the customers pay right away (cash) or later (accounts receivable that are later converted to cash), which are then used to pay employee wages, buy raw materials, make investments etc. As for ascertaining the debt sustainability of a corporate the market generally thinks in terms of Net Debt to earnings before interest, tax, depreciation and amortisation (EBITDA), debt servicing cost to operating cash flow or similar ratios. For banks none of these metrics are relevant. Why is that? Simply put a bank generally doesn't actually reduce its debt burden but simply rolls it over. In other words when a bank bond comes up for redemption a bank generally issues a new one to replace it. Why? Because a borrowing and lending is in fact banks’ core business as opposed to just being part of capital structure as it is for corporates. 

The reason it is quite difficult to analyse a bank is that ultimately the quality of banks assets is the first derivative of the type of borrowers which it lends to. In other words if a bank has financially strong clients that so are likely its assets, and vice versa if its clients are weak, so are likely the loans to those clients (ie much more likely to become non-performing). But the information on bank clients is generally not available in the public disclosures, beyond the current degree of non-performing loans and write-offs. So whereas if you're analysing a phone company you can make certain assumptions about them as a function of a market their serving (ie how many people will get a phone etc), this is much more difficult to do for a bank. 

One of the key reasons for banks to exist is maturity transformation - the idea that if you pool short term deposits you have the ability to make long term loans which are of course a necessity for investment and in order to stimulate economic growth. The reason it is important to keep in mind is that banks face asset and liability maturity mismatch by design - it is not a bug of the banking model rather a feature. And of course maturity transformation is one of the ways that banks make money. 

Asset Liability Mismatch Example

So how do banks make money? The three key line items on the income statement are Net Interest Income, Fee Income and Others (for most banks it will be the ALM book / trading income). Of course things can get considerably more complicated for a large global institutions with Corporate and Investment Banking and other business lines but I'm starting with the basics. 

Net Interest Income is the difference between the interest income you get from clients and the interest expense you have to pay to fund those loans / positions. It is worth remembering that there is almost always a trade-off between risk and returns. For example, while mortgage loans usually considered low risk as they are secured (of course depends on Loan to Value ratio etc) they also to earn lower margins for the bank. Credit cards are on the other side of the spectrum with high interest margins but also higher cost of risk (ie cost of loan writeoffs). 

Net Interest Margin chart

Source: European Banking Authority

The Fee Income would include things like loan origination fees, asset management fees etc. Others (of which largest component tends to be treasury ALM income) could really be anything and vary quite considerably depending on a bank. 

As for Expenses line item, the key ratio to watch here is cost/income ratio or sometimes also referred as efficiency ratio which is just what is says - cost as a % of income. For an average retail bank this shouldn't be more than 50% but the truth is it very much depends on the business model. There is usually a trade off in that more capital intensive business (e.g. lending) will have lower cost / income ratio but higher capital consumption and on the other hand business such as private banking or asset management will have higher cost/income ratios but lower capital consumption. There is no such thing as better or worse business model it's all about where you come out with ROE at the end of the day. Well unless that is a bank is in a capital intensive, high cost/income business model (and there is no shortage of candidates in the European banking universe). 

Cost Income Ratio chart

Source: European Banking Authority

As for Capital, in most jurisdictions capital is measured as a % of risk weighted assets (RWA) not gross assets. The principle with RWA is that banks should hold more capital against riskier assets and vice versa. So for example whereas a mortgage loan might have risk weighting of say 20%, a credit card loan might have risk weighting of 150%. In other words let's say as a bank you're targeting a Core Equity Tier 1 (CET1) ratio of 15%. You have a 100 mortgage and 100 credit card portfolio. Assuming a RW of 20% for mortgages you need to set (100*20%*15%) = 3 of capital for mortgages. For credit cards this would be (100*150%*15%) = 22.5 of capital. So you see that the credit card margins have to be at least 7.5 times higher in order for you to make a similar ROE assuming the same cost structure. This is why banks spend so much effort with trying to lower RWAs. 

CET1 ratio chart

Source: European Banking Authority

Another point related to capital - while some of this is logical (e.g. the fact that riskier assets should have more capital held against their potential losses), some of it is not - for instance why is it that 15% CET1 is regarded as a strong level whereas 10% is on the cusp of being considered weak? No logic that's just market standard and this can vary quite considerably across jurisdictions. 

Sample Bank Capital Stack

Why are there so many layers of capital - e.g. CET1, AT1, Tier 2, TLAC/MREL etc? While the rules for what makes each instrument eligible can get quite complicated, the principle here is again logical and pretty straightforward. Basically all the non-equity capital instruments (e.g. AT1, Tier 2 etc) are a cheaper form of capital so that banks can: (1) lend more as if the cost of their capital stack is cheaper, that presumably means so is their hurdle rate for lending and (2) lower the overall cost of their capital stack (form issuer's point of view) which boosts return on equity (ROE). To put it another way of course a bank can chose to simply use only CET1 capital in its capital stack and this would be perfectly fine - but it would also be a totally uneconomical decision as equity is the most expensive type of capital and AT1 and Tier 2 are considerably cheaper (particularly after tax).

Perhaps this is a good point to stop and take a break. As I noted in the intro, ideally I’d like to expand on this guide based on incoming questions, so please don’t be shy. Looking forward to hearing any and every feedback.