Defensive Value Investing: A 20-week course
“First, let’s analyze why the financial crisis occurred. If I could sum up the catastrophe in one word, it would be ‘leverage.’ “Steve Eisman (reimagined as Mark Baum in The Big Short)
Over the last few weeks we’ve looked at a collection of metrics designed to identify “above average” companies. These included:
- Growth Rate, which measures per share growth across revenues, capital employed and dividends over the last ten years
- Growth Quality, which counts revenue, earnings and dividend increases over the last ten years
- Profitability, which measures return on sales and return on lease-adjusted capital employed, and
- Debt Ratio, which compares total borrowings to ten-year average earnings
These metrics work for the majority of companies, but not for banks. So this week I want to focus on what makes banks different to other businesses and, more importantly, what bank-specific adjustments need to be made so they can be included in our investable universe.
How banks are different
Most companies are either trading businesses or service businesses.
Very simplistically, a trading business buys in raw materials, does something to add value to those raw materials (perhaps turning wood into chairs or stacking tins of beans on a convenient supermarket shelf) and then sells the end product to customers, hopefully for a profit.
Trading businesses require capital assets such as offices, factories, tools and machinery to add value to their raw materials. These are assets on the balance sheet which are balanced off against liabilities owed mostly to shareholders (shareholder equity), debtholders (borrowings) and leasors (lease liabilities).
Service businesses are similar to trading businesses, except their raw materials are the people or technologies which provide the service. In many cases, people-based service businesses don’t require much in the way of capital assets as their main asset is people, and people are effectively rented on a rolling short-term basis and so don’t show up on the balance sheet.
Banks are different to trading and service businesses because their raw material isn’t wood, steel or people, it’s money. They borrow money from those that have it and lend it out to those that need it, which means that banks are one of the economy’s major links between savers and borrowers. Because of this business model, banks have very different income statements to other businesses.
For example, a bank’s main income is interest payments on loans made to customers, while its main expense is interest on deposit accounts and borrowings from other banks or institutional lenders. If a bank receives £100 million in interest and pays out £90 million in interest, its net interest income is £10 million. These interest-related numbers are what you’ll see at the top of a bank’s income statement instead of revenues.
As well as having different income statements, banks also have different balance sheets. This is because banks fund their core lending operations almost entirely from debt (such as customer deposits or institutional borrowings) rather than equity (such as retained earnings or rights issue proceeds). The result is that banks have huge debts that would be unthinkable in almost any other business.
This doesn’t mean that banks are necessarily high risk or uninvestable, but it does mean we need to use bank-specific metrics to analyse their accounts.
Updating Growth Rate for banks
When I begin to analyse a company the first thing I look at is its Growth Rate, so let’s start there.
The first problem with Growth Rate is that it includes revenue per share growth. Banks don’t report revenue, so clearly something has to replace revenues. But what?
Replacing revenue growth with total asset growth
The Growth Rate metric includes revenue because it’s the total amount of money coming into a business from its customers. This is important because it’s the original source of cash from which all future earnings and dividends flow. If revenues per share aren’t growing then any earnings and dividend increases will not be sustainable in the long run.
For banks, the primary measure of income is usually interest income, or net interest income. However, interest income is affected by interest rates, so a bank can see its interest income increase simply because it’s lending to higher risk borrowers who have to pay higher interest on their loans. This is not sustainable income growth as a bank cannot go on lending to ever-riskier borrowers.
I think a more fundamental source of bank income is the total amount of loans made. Growing income by lending more money to more people is far more sustainable than lending to ever-riskier borrowers, so for me total loans outstanding is the original source from which future bank earnings and dividends will flow.
Outstanding loans are recorded on the balance sheet as an asset (under various names), and it’s usually the main asset by some margin. Given that the rest of a bank’s assets exist to support its lending business, I think a reasonable simplification is to focus on the growth of a bank’s total assets.
So our first tweak for banks will be to use total assets per share in the Growth Rate metric instead of revenues.
Replacing capital employed growth with net asset growth
The Growth Rate metric also measures lease-adjusted capital employed growth, which is the sum of equity capital (shareholder equity or net assets), debt capital (operational borrowings) and leased capital (lease liabilities). Banks make use of all these forms of capital, so in theory we don’t need to make any changes.
However, banks don’t typically use much in the way of operational borrowings, i.e. debt which is used to purchase property or equipment instead of being lent to customers. What operational borrowings they do use are often buried in the notes at the back of the accounts. And when it comes to banks, there are far bigger things to worry about on their balance sheet than operational borrowings (we’ll cover what those things are shortly). This means it can sometimes take quite a bit of effort to dig up figures which ultimately don’t make much difference to the overall picture.
So in the case of banks, instead of measuring lease-adjusted capital employed growth I measure leased-adjusted net asset growth, which is basically the same but ignores operational borrowings.
The last item measured by Growth Rate is dividend growth, and dividends are no different for banks so there’s nothing to change.
In summary then, Growth Rate for banks will measure:
- Assets per share growth
- Lease-adjusted net asset per share growth
- Dividend per share
Calculating Growth Rate for banks
Steps to calculate Growth Rate for banks
Note: The caret symbol (^) is used to identify exponents, so in the calculation above, “^1/7” should be read as “raised to the power 1/7”. The seven is there because two three-year periods at the start and end of a ten-year period are effectively seven years apart.
Updating Growth Quality for banks
Growth Quality measures the consistency of growth across three important factors: revenues per share, earnings per share and dividends per share.
Respectively these represent the total income from customers, the net income after all expenses and the amount paid out to shareholders.
Banks don’t report revenue, so Growth Rate uses assets per share instead. This applies to Growth Quality as well, giving the following bank-specific calculation:
Steps to calculate Growth Quality for banks
Updating Profitability for banks
My Profitability metric is based on average returns on sales (ROS) and average net returns on lease-adjusted capital employed (net ROLACE). Here’s a quick refresher on how to calculate these ratios:
Return on sales is the ratio of net earnings to revenues:
ROS = net earnings / revenues * 100%
Net return on lease-adjusted capital employed is the ratio of net earnings to the sum of shareholder capital, debt capital and leased capital:
Net ROLACE = net earnings / (net assets + borrowings + lease liabilities) * 100%
To calculate Profitability, take the ten-year average of ROS and net ROLACE and calculate their combined average, so:
Profitability = (10yr avg. ROS + 10yr avg. net ROLACE) / 2
To make this ratio work for banks, we need to replace ROS and net ROLACE with their banking equivalents.
In place of return on sales we’ll use return on assets (ROA), and in place of net return on lease-adjusted capital employed we’ll use net return on lease-adjusted net assets (net ROLANA).
Steps to calculate Profitability for banks
Banking rules of thumb
To keep things simple, I mostly use the same rules of thumb for banks as I do for other companies. However, one difference does appear with return on assets, which replaces return on sales.
For return on sales, I have a rule of thumb minimum of 5%, but banks almost never produce a return on assets of more than 5%. This isn’t necessarily a weakness, but it does reflect the relatively commoditised nature of banking (money is money, regardless of which bank you borrow it from, so it’s hard for banks to differentiate enough to earn high rates of return on such a commoditised asset).
Return on assets for most UK banks has historically been below or close to 1%, so I have set my return on asset rule of thumb at 1%, minimum. Here’s the initial set of banking rules of thumb:
Defensive value rules of thumb
Additional ratios for banks
So far we’ve updated some of the core metrics so they work for banks. In addition, I use some bank-specific ratios which help me assess the strength and profitability of a bank’s balance sheet, both of which are important factors given their highly leverage nature.
The Net Interest Return ratio
In the previous section we looked at return on assets as part of a bank’s Profitability score. The problem with return on assets is that returns can be skewed upwards if a bank has significant non-lending activities such as fund or wealth management.
It would be useful to isolate the performance of the bank’s lending business, and one way to do that is to look at the ratio of net interest to tangible assets.
Since tangible assets are mostly made up of loans to customers, this ratio tells us how much income the bank is generating from its lending activities, net of the cost of those loans to the bank.
It’s a simple ratio which isolates the profitability of a bank’s lending activities, and I use the following rule of thumb:
Defensive value rule of thumb
Most banks don’t exceed that 2% hurdle rate, which means that only the most profitable banks will ever end up in a defensive value portfolio.
The Common Equity Tier 1 Ratio
Imagine a bank with £100m of customer deposits held across thousands of current accounts. Those are the bank’s liabilities (otherwise known as its sources of funds). The bank has loaned £99m of those deposits out to small businesses and homebuyers, leaving £1m in its vault as a daily cash float. Together the loans and cash float make up the bank’s assets (otherwise known as its uses of funds).
In that simplified example everything works fine when all loans are repaid on time, but in the real world some loans are not repaid on time and some are not repaid at all. If £5m of the bank’s loans are not repaid (i.e. are defaulted on) the bank would have £100m in liabilities (those deposit accounts) but only £95m in assets (£94m in loans and £1m in cash). The bank would no longer have enough money to pay back its depositors and so it would be technically insolvent; shareholders would almost certainly be asked to put up additional capital through a rights issue to make the company solvent again.
To avoid this situation, banks use additional sources of funds which act as a buffer to protect customer deposits. This may involve some form of long-term unsecured debt, but the primary buffer is shareholder equity.
Imagine the same bank as before, but this time instead of having £100m of deposits the bank has £90m. It now also has £10m of equity capital in the from of cash, giving the bank the same £100m of funds as before. If £99m is lent out and if the same £5m of loans is not repaid, it has the same default rate as before, but this time the bank’s remaining assets (£94m of loans that will be repaid and £1m of cash in the vault) are still worth more than its customers’ deposits of £90m. In this scenario, current account holders can still get all their money back and so a rights issue would not be required to raise additional funds.
The difference with the second example is that shareholders have absorbed the losses from defaulted loans rather than customer deposits. More specifically, shareholder equity has absorbed the losses by declining in value from £10m before those loans were defaulted on to £5m afterwards.
This sounds bad for shareholders, but if the bank is well run then it should be possible to rebuild the equity during good times (i.e. when loan default rates are low) in order to prepare for increasing defaults during the next economic downturn.
The important point is that by having a sufficient buffer of shareholder equity, a well run bank should be able to absorb loan defaults without having to suspend its dividend or raise additional equity capital through a rights issue.
To measure the ratio between a bank’s equity buffer and the amount of loans it has made, I use a ratio called the Common Equity Tier 1 Ratio (CET1 ratio). Common equity tier 1 is basically tangible shareholder equity with a few adjustments which we don’t need to worry about. It is, according to the latest banking regulations, the “highest quality component of a bank’s capital”.
So looking back at that previous example, with its £ 10m of shareholder equity and £ 100m of loans and cash (ignoring the risk weighting of the loans for the sake of simplicity), the CET1 ratio for the bank would be £ 10m divided by £ 100m, which is 10%.
One thing I should point out is that while the CET1 ratio is found in the annual and interim results, you will struggle to find it in older results because it’s fairly new. However, its predecessor, the core equity tier 1 ratio, is very similar and for our purposes the two are interchangeable. If I refer to CET1 then take that to mean either the common or core equity tier 1 ratio, depending on what was in use at the time.
Table 5.1 shows the CET1 ratio for several leading banks, before the financial crisis and after.
|Bank||2007 (pre-crisis)||2013 (post-crisis)|
|HBOS||7.4%||Taken over by Lloyds in 2009|
Table 5.1: Common equity tier 1 ratios for several leading banks before and after the financial crisis
Before the financial crisis, all of the surviving banks in Table 5.1 had lower CET1 ratios than they did a few years after the crisis. It’s interesting to see that HBOS, with the lowest pre-crisis ratio, was taken over, while Standard Chartered, with the highest pre-crisis ratio, was barely hurt by the crisis at all (although there were also many other factors, not least of which was the fact that Standard Chartered’s business operates primarily in markets that were not hurt so immediately by the crisis).
Of course, there’s much more to calculating the risks faced by a large bank than simply looking at its CET1 ratio. However, I think the ratio has a lot of merit to it. It’s usually easy to find the number in the annual reports, it’s used extensively by the banks themselves and it does appear to have a reasonable correlation to how well each bank withstood the stresses of the financial crisis.
As with many of the financial ratios and metrics in this book, I prefer to look at the average CET1 ratio over a number of years.
Calculating the average CET1 ratio
To calculate the average CET1 ratio you’ll need:
- Common equity tier 1 ratio (CET1 ratio) – For the most recent five years. This isn’t part of the balance sheet but it should be found fairly prominently towards the beginning of a bank’s annual results. Older results may refer to the similar core equity tier 1 ratio instead.
There is only one step to calculating this metric:
Steps for calculating the average CET1 ratio
What is a reasonable minimum CET1 ratio?
The latest banking regulations state that a bank must have a CET1 ratio of at least 4.5% at all times. On top of this, an additional buffer of 2.5% (taking the total to 7%) should be built up during good times so that it may, if necessary, be drawn down in bad times. Finally, there is an additional buffer for systemically important banks of up to 2.5%, taking the maximum requirement at any time to 9.5%.
That seems like a fairly reasonable minimum to me. As Table 5.1 shows, before the financial crisis most banks had a CET1 ratio of less than 9.5%. After the crisis they have all moved to ratios above 9.5%.
I want any bank that I invest in to have been slightly more cautious than the maximum caution stipulated by the regulators, so initially my rule of thumb was to insist that the average of this ratio be 10% at least.
However, Standard Chartered passed (one of my previous holdings) passed this test and subsequently ran into major problems, so now by CET1 requirements are even higher .
Defensive value rule of thumb
Given that Standard Chartered has had problems, let’s calculate its average CET1 ratio to see if it passes the new stricter rule of thumb.
Standard Chartered’s common equity tier 1 ratio
Standard Chartered is a FTSE 100-listed bank which operates primarily in Asia, Africa and the Middle East. For a while it was popular with investors because it came through the financial crisis with barely a scratch on it, largely thanks to its small exposure to Western markets and a relatively strong balance sheet.
The common equity tier 1 ratio is quoted in a bank’s annual results and you can see Standard Chartered’s CET1 ratios over the five years to 2018 in Table 5.2.
Table 5.2: Standard Chartered CET1 ratio for the five years to 2018
After running into significant problems in 2015 and 2016, Standard Chartered raised its CET1 target to between 13% and 14%, which is unusually high.
Given that my preferred average CET1 minimum is 12%, Standard Chartered does make it past this hurdle, and is one of very few UK banks to do so.
The Tangible Common Equity Ratio
Although the CET1 ratio is the standard measure of balance sheet strength for banks, I prefer simple ratios which I can calculate myself, directly from the income statement, cash flow statement or balance sheet.
A good candidate for a simple balance sheet ratio for banks is the tangible common equity (TCE) ratio, which is the ratio of tangible shareholder equity to tangible assets.
The key difference between CET1 and TCE is simplicity. The equity part of the ratio is just tangible equity rather than Common Tier 1 equity, and assets are taken as tangible assets rather than risk-weighted assets.
Steps for calculating the tangible common equity ratio
As with CET1, TCE shows us how much of an equity buffer the bank has to protect non-equity sources of funds (such as deposit accounts and subordinated borrowings) in the event of large-scale loan defaults.
A sensible minimum for the TCE ratio
If we look back at the financial crisis it’s easy see what was and wasn’t a sensible TCE ratio. It’s easy because most UK banks had very significant problems during that crisis, and most of those problems were self inflicted through excessive leverage and excessively thin equity buffers.
For example, in 2008, all large UK banks had TCE ratios of less than 4% and in some cases less than 2%. In other words, some UK banks would become insolvent if just 2% of the loans they expected to be paid in full were instead defaulted on. That was a recklessly thin margin of safety as the banks soon found out.
Since then most banks have increased their equity buffers and today many have a TCE ratio of 5% or more. That’s obviously better than 2% and it meets the current regulatory standards, but it isn’t good enough for me.
I don’t have to invest in banks, so if I’m going to invest in a company which is highly leveraged by nature, then I only want to invest in those with abnormal levels of prudence and balance sheet strength.
For me that means demanding an average TCE ratio which is comfortably above the average of any large UK bank:
Defensive value rule of thumb
In recent years, only small and specialist lenders have been able to exceed that 7% hurdle rate. That’s fine by me, because in my experience niche lenders tend to make better investments than the highly commoditised high street banks.
Close Brothers tangible common equity ratio
Close Brothers is a UK-focused merchant bank listed in the FTSE 250. It’s the only bank I’ve owned over the last few years and it has an unusually strong balance sheet. This shows up as a high average TCE ratio, which you can see in Table 5.3.
|Year||Tan. Equity||Tan. Assets||TCE ratio|
Table 5.3: The tangible common equity ratio for Close Brothers
As you can see, Close Brothers has consistently had a TCE ratio far above 7%, which means by that measure at least, it has a very wide equity buffer to protect it when loan default rates surge during lean economic times.
This is in fact a key part of the company’s strategy. Close Brothers nurtures deep relationships with corporate customers who need to know that their bank will continue to lend even during deep recessions. That is precisely when many of its customers need additional emergency funds, either to replace profits which are temporarily depressed or to invest in expansion while other companies are retreating.
Conservative borrowings for banks
Earlier I said that my standard Debt Ratio didn’t work for banks, but that’s not quite true. It does work, but with banks you have to separate out the different types of borrowings:
- Operational borrowings: Used to fund the purchase of capital assets such as property, equipment and so on (also known as other borrowings, loans, overdrafts etc)
- Structural borrowings: Used to provide an additional capital buffer to protect customer deposits if loan defaults are significantly higher than expected (also known as subordinated capital, Tier 2 capital etc)
- Source of funds borrowings: Used a source of funds for lending (also known as debt securities, loan backed securities, etc)
As you can see, these different forms of borrowings can have many different names and in many cases you’ll need to dig through the notes to the accounts, rather than the balance sheet, to find their values.
Once you have a bank’s operational borrowings, you can use them to calculate the debt ratio:
Calculating the Debt Ratio for banks
The related rule of thumb is the same as for other companies. Banking is a cyclical sector, so that gives the following:
Defensive value rule of thumb
Rules of thumb for banks
Bank-specific rules of thumb
Next week: Insurers
Having covered banks this week, our next step will be to look at insurers, which in many ways are more like banks than other companies. We’ll be reusing some of the updated metrics from this week and introducing some insurer-specific measures as well.