There is usually at least one metric that gives valuation-based support for an investment, even if this is contradicted by other indicators of relative or absolute value. You may have heard comments such as “… but it looks cheap on EV/EBITDA” to help justify a particular investment recommendation.
We examine why different multiples can give conflicting indications of relative value. For example, food-on-the-go stock Greggs trades at a 35% discount to rival Dominos Pizza, based on EV/EBITDA, but at a 24% premium using a price earnings ratio. Which multiple (if either) gives the correct relative valuation?
Valuation multiples are simplistic, but at the same time complex. They are simplistic because they explain value using a single variable for a single period. They are complex because multiples are affected by many different value drivers, any one of which could explain differences between stocks. Identifying and allowing for these differences in value drivers when assessing whether one stock is cheaper than another is complex and challenging.1Our ‘Target Multiple Calculator’ is one way to go about it, but applying and interpreting this is still difficult.
You may argue that some sort of discounted cash flow approach2There are several ways to structure a DCF valuation. In our article ‘DCF valuation: Financial leverage and the debt tax shield’, we describe (and reconcile) three approaches. is superior to simplistic yet complex multiples, but DCF presents its own challenges. Even if your main valuation approach is DCF, we recommend that multiples are used as a reality check – it is often easier to identify aggressive DCF model inputs by examining implied current and forward priced multiples, than it is by examining the DCF input variables themselves. One way or another, most investors use valuation multiples to describe value differences and as an input to stock selection.
How should you deal with inconsistent messages from different multiples?
Arguably, the two most common valuation multiples used are EV/EBITDA and the price earnings ratio (P/E). Often these multiples will give consistent messages, and stocks look either cheap or expensive using both approaches; however, this is not always the case. The question is what should you do when multiples contradict – which multiple (if either) is most reliable as a description of value?
A good example of conflicting valuation multiples can be seen using data for Greggs and Dominos Pizza. We have used these UK FTSE250 companies several times in our training events given their easily understood business models, yet interesting (and different) accounting complexities. Using Factset consensus forecast data for FY 2025, Greggs trades at a 35% discount compared with Dominos based on EV/EBITDA; however, the opposite occurs when using price earnings ratios, with Greggs then appearing to be 24% more expensive.
Relative valuation of Greggs and Dominos
Factset consensus forecast data for FY 2025. Data current on 22nd November 2024
In our view, neither of these relative value descriptions are particularly useful. The problem is that EV/EBITDA is incomplete, and price earnings ratios are not comparable. EV/EBITDA misses out important expenses that should affect value. If these expenses differ between stocks, then they should trade on different EV/EBITDA multiples. In the case of price earnings ratios, the multiples are affected by other differences in stock characteristics. Because the significance of these differences varies, and their influence is hard to quantify, price earnings ratios are complex metrics that are less comparable and difficult to interpret.
Valuation multiples – completeness and comparability
The Footnotes Analyst
We think investors should focus more on EV/NOPAT.
Why EV/EBITDA is incomplete
It is the ‘TDA’ portion of EBITDA that is the problem.
Taxation is clearly an important expense and one that should affect value. Companies that operate in low tax jurisdictions, or which obtain tax incentives that reduce their effective tax rate, should (all other things being equal) have a higher value and trade on a higher multiple of pre-tax profit. If more pre-tax profit is available to investors because less tax is paid, the pre-tax profit must be relatively more valuable.
For enterprise value multiples it is the tax related to operating profit that matters. The overall tax charge also reflects the tax effects of investment income and net financing expenses. There is no disaggregation of taxation along these lines in financial statements; however, investors need to estimate this analysis to be able to correctly calculate the post-tax enterprise value multiple EV/NOPAT. Nevertheless, using the overall effective tax rate is close enough in many situations.
In the case of Greggs and Dominos we can easily observe the overall forecast effective tax rate. Not surprisingly, given they largely operate in the same tax jurisdiction, the forecast rates for 2025 are similar – 25.9% for Greggs and 24.9% for Dominos. However, if you compare across jurisdictions, or are analysing multinational companies, tax differences may well be material.
Differences in depreciation and amortisation (D&A) are a more common problem, even with companies operating in the same jurisdiction and same industry.
Don’t ignore D&A just because it is supposedly non-cash
Although depreciation and amortisation are regarded as non-cash expenses, they are very real costs and should not be ignored. In fact, they are also cash flows, except that the cash effect – capital expenditure – arises in a different accounting period, typically the period in which the asset is purchased. Ignoring depreciation and amortisation and the related capital expenditure means ignoring a key driver of value. If more EBITDA is required to fund ongoing (maintenance) capital expenditure, less is available for capital providers, and enterprise value is inevitably lower. Consequently, fixed asset heavy companies tend to (correctly) trade on lower EV/EBITDA multiples.3We also discussed the merits of EBITDA in our article ‘EBITDA-AL: More letters but no more insight’.
In the case of Greggs and Dominos, D&A is a key driver of the difference we observe in EV/EBITDA. Greggs has a much higher D&A expense. Although this is largely a result of its business model (more owned assets and less franchising) rather than a sign of any inefficiency, the difference still matters in terms of analysing valuation multiples.
Fixed asset expense: Greggs versus Dominos
Factset consensus forecast data for FY 2025. Data current on 22nd November 2024
Post D&A profit metrics (such as NOPAT) should be preferred. However, be aware that what really drives value differences is future capital expenditure, and specifically maintenance capex, not backward-looking allocations of past sunk costs. Depreciation and amortisation will usually be a good proxy for future (maintenance) fixed asset purchases, but not always – a good example of this is the amortisation of intangibles arising in a business combination.
Inconsistent recognition of intangible assets may require an adjustment to NOPAT
Due to the inconsistent recognition of purchased and internally generated intangibles, many intangibles that are capitalised because of a business combination will not be replaced by new capitalised expenditure when the acquired assets are consumed. This results in a component of amortisation having no equivalent future capital expenditure. Instead, the cost of replacing these assets is a regular operating expense. We call these ‘replacement-expensed’ intangibles.
In effect, this creates an element of double counting, with profit and loss impacted by both the amortisation charge and the ongoing non-capitalised expense. Many companies, therefore, add back this component of amortisation in their non-GAAP metrics – a practice we think is valid.
Accordingly, we think that NOPAT should be stated before deducting the amortisation of replacement-expensed intangibles arising from a business combination. In other words, the D&A deducted to calculate NOPAT should be a ‘replacement cost’ measure. It should only reflect the depreciation and amortisation of assets where the replacement expenditure will actually be capitalised. Replacement cost depreciation and amortisation may also be adjusted for fixed assets price changes, although this is rare outside inflationary economies.4For more about the treatment of intangible assets capitalised as a result of a business combination, see our article ‘Should you ignore intangible amortisation?’.
In the case of Greggs and Dominos, acquired intangibles are relatively small and we do not think any adjustment is required when calculating NOPAT.
In summary, we think a post-tax operating profit (after potentially adjusting the D&A) is a better and more complete basis for an enterprise value multiple than EBITDA. Using EV/NOPAT, Greggs and Dominos valuations appear much closer – based on Factset consensus forecast data for 2025 we estimate multiples of 19.0x and 18.4x respectively.
Why price earnings ratios may not be comparable
Operating enterprise value multiples focus on the market’s implied valuation of operating activities. The value of the equity shareholders’ claim is after adding to the operating EV the value of non-operating assets, such as investments, and deducting non-common share claims, such as debt and non-controlling interests. These other components of equity value are not ignored in EV multiples; they are simply included in the estimate of EV itself.
In equity multiples, such as the price earnings ratio, non-operating assets and non-common share claims are incorporated through the related income and expense being included in earnings or through a separate dilution calculation. However, the effect of this is that the price earnings ratio is impacted by the scale and implied multiple for these other components. The result is that equity multiples are less easily compared than EV multiples – assuming a realistic value of these other assets and claims is included in the calculation of enterprise value, and the numerator and denominator of the multiple are consistent.5For more about common mistakes in enterprise value calculations, see our article ‘Enterprise value: Calculation and mis-calculation’.
Investments and non-core assets
The price earnings ratio is, in effect, the weighted average6It is not a normal weighted average but instead a weighted harmonic mean, i.e. the reciprocal of the weighted average of the reciprocals of the multiples. of the EV/NOPAT multiple and the ratio of the value of the investments compared with the post-tax income generated from those investments. The extent of the difference between these multiples and the relative size of the investments determines the price earnings ratio.
This is perhaps best illustrated by an example: Assume a company has an EV/NOPAT multiple of 19x, has no investments or leverage, and equity is purely ordinary shares. This means the price earnings ratio will also be 19.0x. Now assume another company has operations with a similar value, but in addition owns minority equity investments. The price earnings ratio of this company will depend on the size of these investments and on the multiple of investment value to its post-tax contribution to net income.
Price earnings ratios are further complicated by the different accounting treatment of equity investments
Unfortunately, the earnings contribution of equity investments depends on whether these investments are classified as an associate and, if not, the IFRS (not US GAAP) accounting policy choice. In each case the earnings contribution differs. Here is an illustration of how this can affect the price earnings ratio:
Impact of equity investments on the price earnings ratio
The Footnotes Analyst
We assume that the investment, if classified as an associate, has a P/E ratio of 10x, that the investment post-tax cost of equity is 7% and dividend yield is 3%. Multiples are based on forecast data and, in the case of FVPL accounting, this forecast includes a price change that reflects the cost of equity.
This table demonstrates how difficult it can be to interpret price earnings ratios when a company owns minority equity investments. In each of the four scenarios in our example the value of the stock is the same, and you would think each should trade on the same multiple. Only EV/NOPAT shows this – none of the price earnings ratios are comparable.
For Greggs and Dominos, the effect of investments is not material and does not explain the differences we observe in price earnings ratios. In their case it is leverage that matters.
Financial leverage
The effect of financial leverage on the price earnings ratio is like the effect of investments, except reversed, given their opposite effect on equity value. Fortunately, there is no (significant) difference in accounting to complicate things in this case. What matters is the debt multiple relative to the EV/NOPAT multiple and the amount of leverage. The debt multiple is the reciprocal of the post-tax cost of debt. Considering current developed market interest rates, the debt multiple (we estimate generally over 30x) will be higher than EV/NOPAT for all but the most highly rated companies. This means that higher leverage results in a lower price earnings ratio, which is exactly what applies to Greggs and Dominos.
Dominos has much higher financial leverage than Greggs – debt as a percentage of EV (according to Factset) is about 28% for Dominos compared with just 6% for Greggs. As we would expect, this results in a significantly lower price earnings ratio. It is difficult to derive the relevant numbers from Factset consensus forecasts considering the way data is summarised, but here are our estimates of the drivers of the effect of financial leverage on the price earnings ratios of these companies.
EV/NOPAT versus price earnings ratios for Greggs and Dominos
Factset consensus forecast data for FY 2025 and The Footnotes Analyst estimates. Data current on 22nd November 2024
The main economic explanation for the lower multiple is likely to be the higher cost of equity due to the higher leverage.
Other equity claims
The other claims that are part of enterprise value, and which create a deduction in calculating the earnings attributable to common shareholders used in the price earnings ratio, are also relevant. This includes non-controlling interests, debt-like securities classified as equity7See our article ‘Analysing complex capital structures – perpetual bonds’., employee stock options and the equity component of convertible bonds. The impact of these on price earnings ratios can be determined in a manner similar to what we describe above. In most cases the NCI effect will not be significant (assuming NCI has been correctly dealt with in the EV multiples); however, if material, stock-options can be more troubling.
We do not consider this aspect further in this article because the effect on Greggs and Dominos does not seem to be material. However, the diluting effects of options and convertibles has been the subject of other Footnotes Analyst articles – see ‘The diluted EPS calculation is 50 years out of date’ and ‘Dot-com bubble accounting still going strong’.
An adjustment factor approach
We have often used a factor approach in trying to explain the difference between EV/EBITDA and price earnings ratios. The differences described above can be presented as a series of factors that, when combined, explain the reasons for differences in multiples.
Adjustment factors to reconcile EV/EBITDA to price earnings ratios
The Footnotes Analyst
Here is this approach applied to Greggs and Dominos.
Reconciling EV/EBITDA and price earnings ratios
Factset consensus forecast data for FY 2025 and The Footnotes Analyst estimates. Data current on 22nd November 2024
If you would like to see how we calculated these adjustment factors, you can download the model we used here:
Please enter your email address to receive an excel version of this model
This analysis highlights the key reasons for the difference between EV/EBITDA and how these factors differ between companies. Alternatively, you could simply use EV/NOPAT in the first place.
Finally – don’t forget the real estate effect
We think EV/NOPAT is the best way to summarise relative value differences due to the multiple not being distorted by depreciation and tax differences or by how the entity is financed. However, this multiple does not allow for differences in the nature of the underlying operating business.
Clearly, the quality of the business as represented by the key value drivers of growth, returns and risk will affect how the stock is valued and, consequently, the EV/NOPAT multiple. Outside a DCF analysis, it is difficult to quantify the effect of these value driver differences and therefore judgement is inevitably required. However, there is one difference between underlying operating businesses that can easily be factored into your analysis through a disaggregation of the EV/NOPAT multiple – the real estate effect.
Companies that use real estate8The same applies to other major ‘strategic’ assets such as airplanes in the case of airlines. as a significant business input can access this asset in different ways. Some companies may own the real estate they use in operations, whereas others may lease properties and use different length leases. Each approach exposes the company to real estate to a different extent, which should affect how that company is valued, and the EV/NOPAT multiple.
Use Opco-Propco analysis when companies have different exposure to real estate
Different exposure to real estate is not a reason to use EV/EBITDA – that just ignores the issue. To obtain more comparability of the multiples of similar operating businesses, but businesses which have different exposures to real estate, it is better to still use EV/NOPAT, but separate the real estate activity from the rest of operations – so-called Opco-Propco analysis.
We do not provide a full explanation of this analysis here. To see the mechanics of this disaggregation, and how it can help in ensuring that valuation multiples (and other key metrics such as return on capital) are comparable, see our previous article ‘Real-estate and equity valuation – Opco-Propco analysis’.
For Greggs and Dominos, the real estate difference appears to be significant. Dominos seems to have a lower real estate exposure due to the company owning fewer properties and using more short-term leasing, and also more franchising. Applying Opco-Propco analysis to further disaggregate the EV/NOPAT multiples of these companies would seem particularly useful. Although they appear to have similar EV/NOPAT the market may not attribute similar value to each business.
Insights for investors
- EV/EBITDA is incomplete. Depreciation and amortisation may be classified as ‘non-cash’, but it represents a very real cost. Differences in both D&A and taxation matter in equity valuation.
- Amortisation of acquired intangibles that are ‘replacement-expensed’ is different. We think stating NOPAT before this amortisation (after allowing for the tax effect) is preferable.
- Price earnings ratios are complex. They not only reflect differences in underlying operating business value but are also affected by differences in investments and capital structure.
- Higher leverage usually (but not always) results in a lower price earnings ratio, but this does not necessarily indicate a better value stock.
- Prefer the EV/NOPAT multiple for relative value comparisons. However, be sure that your enterprise value metric is complete, correctly reflects fair values and is consistent with NOPAT.
- If exposure to real estate differs, consider additional Opco-Propco based analysis of valuation multiples and other analytical metrics.