When making investment decisions, I focus much more on whether a business is high quality than on its valuation. However, I don’t think this means that valuations should be ignored entirely. Sometimes even high quality businesses become overvalued. In practice I am often put off buying businesses where the valuation doesn’t seem to be justified by the growth prospects. But I’ve found this very hard to judge – am I getting it right?
Quality and growth are closely related. A business’s value is determined by the certainty, as well as the extent of long term growth in profits. Quality businesses possess characteristics that make long term growth more certain and so the shares more valuable. One way to think of it is that quality is certainty of growth.
The main thesis behind my investing strategy is that the highest quality businesses tend to be undervalued by the stock market because:
- Identifying quality is difficult. In contrast to value and momentum factors it is not possible to capture quality well solely using simple metrics.
- Valuing growth is difficult. Investors tend to be short-sighted, underappreciating the value of more certain compound growth over the long term.
I’ve written about identifying quality before, but less about valuing growth. Taking account of growth correctly when valuing stocks is extremely difficult, especially as growth profiles vary widely across businesses and at different points in their development.
Discounted cash flows
The best way to think about the value of a business is as the sum of all its future cash flows. This is the return you would actually receive if you owned the business outright rather than simply trading its shares on the stock market. While this is a straightforward idea, using discounted cash flow (DCF) modelling to work out valuations in practice can be challenging for a couple of reasons.
First, you immediately encounter the need to project cash flows far into the future, revealing the inescapable truth that equity valuation is inexorably tied to speculation about the future. And there is a lot of future to contend with. The time horizon is not bounded – all the profits the business will ever make are relevant to its current valuation.
Second, even if you are happy to have a stab at projecting how cash flows might grow forever, you also need to think about a discount rate to apply. The discount rate accounts for the fact that future cash flows are worth less than immediate cash in hand that you could put to work elsewhere. It also needs to account for risk i.e. that future cash flows from equity investments are often highly uncertain, again making them worth less than cash in hand. Accounting for the return you might get from putting your cash elsewhere is possible (you could buy government bonds for little risk). However, accounting for the risk involved in an equity investment is essentially pure speculation.
The result is that DCF models are often complicated yet yield apparently arbitrary results that are largely driven by the assumptions you put in to them (‘garbage-in, garbage-out’). Because of this, it’s easier to avoid thinking about discounted cash flows and instead rely solely on simple valuation multiples to assess whether an investment looks cheap. There’s nothing wrong with this but you need to be careful when making comparisons. It’s not obvious how to compare the valuation multiples of businesses with different growth characteristics. I think you need to understand the underlying logic of compounding and discounting and how this applies to some simple benchmarks to judge this well.
Long term growth
In theory it’s straightforward to calculate the valuation of an investment with constant long term growth, based on its discounted future profits. By long term here, I mean very long term – literally forever, but in practice say at least 30 years. The value of an infinite series of cash flows growing at a constant rate is given by CF/(d–g), where CF is the value of the initial cash flow, d is the discount rate and g is the growth rate. This means the valuation multiple would be 1/(d–g).
This simple formula shows that valuation should be driven by the difference between the discount rate and the long term growth rate. Shares which expect low growth and are very risky, i.e. demand a high discount rate, should trade on low multiples (and vice versa).
So what is the right discount rate to use? This depends on the risk inherent in the investment. If the investment was almost certain to grow profits at a certain rate for a very long time it might make sense to use a discount rate close to long term government bond yields (currently around 3%). This would value the growth extremely highly. In practice, for equity investments you also need to include a risk premium that compensates you for the lack of certainty. It’s not possible to determine precisely what the discount rate should be for any particular investment. However, you can compare relative valuations for an implicit degree of risk by calibrating the discount rate to a benchmark.
A discount rate of around 12.5% for the ‘average’ equity investment seems to yield fairly sensible looking valuations. It would imply that a share expecting stable profits with zero growth should be valued at eight times profits. Growth of 2.5%, more or less in line with inflation, would imply a valuation of ten times profits. Growth rates of around 6%, in line with long term average equity returns, would result in valuations of around 15 times profits, in line with long term average equity valuations.
For higher growth rates, the implied valuations start to increase exponentially. Growth of 10% would imply a valuation of 40 times profits and growth of 12% would imply a valuation of 200 times profits. Above 12.5% growth the implied valuation would be infinite. However, this assumes that the returns across the associated investments are equally certain. Higher growth would tend to imply less certainty in practice, particularly over the long term. This in turn implies a higher discount rate. I think expecting growth rates of much more than 10% to persist over the long term with much degree of certainty seems unrealistic.
This is essentially just a thought experiment. It doesn’t give you a very practical way to assess the valuation of individual companies with much precision. In fact it suggests that this may not really be possible due to the inherent uncertainty involved. It also provides some colour to other intuitions.
First, it highlights that certainty is a key driver of valuation. You can see this from the equity risk premium over bonds. Long term government bond yields currently imply a risk-free discount rate of around 3%, while long term average equity valuations and returns imply a discount rate of 12.5%. Part of this may be because bonds are currently relatively overvalued, though it also suggests there is a very substantial premium on certainty. While all equity investments are inherently uncertain, some are clearly more uncertain than others. The equity risk premium implies that equity investments that are more bond-like in the certainty of returns they deliver should be valued much more highly than less certain investments, perhaps to a greater extent than I think many investors assume.
Second, it shows that where you have a reasonably high degree of certainty, small differences in expected long term growth imply much larger differences in valuation than you might expect. You don’t necessarily need a very high growth rate to justify a high valuation. You could alternatively have more modest growth that is fairly certain to go on for a very long time.
To me these points suggest that high quality businesses with fairly modest but certain long term growth are often likely to be undervalued. This should become apparent if you hold on to them for a very long time – I’ve discussed this previously in the context of the Nifty Fifty. But it’s hard to say by how much they are undervalued. The inherent lack of precision in assessing the certainty of long term growth rates makes it hard to judge whether very high valuation multiples (e.g. above 40 times profits) are justified for specific high quality investments or not.
Medium term growth
Many investments aren’t amenable to directly estimating a long term growth rate. Typically you would expect initially rapid growth to then converge gradually over time to a lower rate. It’s far from obvious what long term rate this might average out at. This means it’s often worth trying to project growth over a medium-term period of about 5-10 years.
In the past I’ve often used the Graham rule of thumb to estimate valuation based on a medium term growth rate, as discussed here. I like that this approach is simple and gives sensible looking results but am less comfortable that I don’t know exactly how it was derived. To get a better understanding of the underlying mechanics I’ve had a go myself.
I’ve worked out the valuations implied by different average growth rates over a 5 year and a 10 year period using a highly simplified DCF model. To do this I’ve made the following assumptions:
- A discount rate of 5%, approximating a ‘risk free’ rate. This effectively assumes that the medium term growth is near-certain. I’ve chosen this for clarity – in practice it means it’s important to be conservative with the growth rate projected.
- That the shares should be valued at 20 times earnings at the end of the period (regardless of the growth rate during the period). I’ve chosen 20 as this seems a reasonable conservative baseline for long term valuation of high quality shares, based on the discussion about valuing long term growth above.
- That the growth rate is constant over the medium term. In reality in most cases the growth would be initially rapid and then slow down over time. As growth that happens earlier should be worth more, this builds in a bit of conservatism to the model.
The model adds up all the discounted cash flows to give a net present value (NPV). I’ve then used this to compare relative valuations by normalising the valuation of a share with 2.5% growth to ten times profits. The table below shows the implied multiples you get for different growth rates over a 5 year period and a 10 year period.
The results seem fairly sensible and I think provide useful benchmarks for thinking about valuation. I find the most practical way to use this is to reverse engineer the kind of growth rate you would need to justify a certain valuation multiple. This is particularly useful for judging whether more exceptional highly-rated high-growth stocks might be worth paying for, or whether the valuation demands too much. For example, a multiple of around 40 times earnings would suggest that you would need to be confident that profits are going to grow by around 20% a year for the next 10 years (or 30% for 5 years). This would be asking a lot! The assumptions I’ve put into the model means that these benchmarks are only useful for valuing high quality businesses. You need to have a relatively high degree of certainty both that the business will achieve at least this level of growth over the medium term, and that it will continue to grow at a decent rate over the long term. You also want to have a ‘margin of safety’ ideally.
Stable defensive vs exciting growth
DCF modelling pushes you to think separately about the value driven by medium term and long term growth. I find this quite a helpful discipline. The part which is more important depends on the growth profile of the business in question. In practice you need to adjust your focus depending on the nature of the business.
For stable defensives – mature businesses with long track records, stable growth rates, proven high and stable profitability and very strong market positions with pricing power (e.g. Diageo, L’Oreal, Nike etc.), long term growth is more relevant. It makes sense to focus on the certainty of this growth. Higher valuation multiples can be justified by the longevity and certainty of growth rather than necessarily by particularly rapid growth rates. The relevant question is whether any key long term risks have been overlooked e.g. from changing consumer preferences, disruptive competition or government intervention.
For exciting growth companies – smaller, emerging businesses with shorter track records and rapid or accelerating growth, a medium term perspective is more relevant. As for stable defensives, certainty is key. However you need to be more careful in interpreting valuations when growth rates are not stable. It can be hard to judge and the market often gets it wrong. Shares of high quality growth companies with strong competitive advantages in growing markets are often substantially undervalued, as investors tend not to look far enough ahead in considering the implications of this growth on valuation (or they simply understimate how long the growth for a high quality business will persist). The implied fair valuation multiples of these investments and their subsequent returns can sometimes be far above what the market deems them to be worth, especially if you identify them at an early stage. However, you need to be careful not to get caught out if the growth slows down faster than expected. Short-sightedness also means that the market can be slow to respond to decelerating growth, sometimes valuing relatively more established growth businesses on multiples that implicitly extrapolate unsustainable growth rates too far into the future.