Geoff Gannon November 10, 2020

Why Appraising a Stock Based on “Relative Valuation” vs. Peers Isn’t Enough to Guarantee You’re Getting a Bargain

Someone asked me a question about relative valuations:

“There are a couple of instances when we can value the stock based on comparables. However, for me, I feel that there is something in the middle that I can’t figure out. There is a big assumption that the stock with the cheaper valuation will go up to get close to the market. Is there an embedded assumption such as, if the stock being compared are similar in growth/roe, or basically we just are hoping for mean to reversion?

I found two similar companies in the same sector in the different country…P/B ratios are (1.5-2.0) and (2.0-3.0). P/E ratio range for the past 3 years are (9-15) and (40-60). ROE 14% vs 7% (yes, the first company has a better ROE and yet is cheaper). Both are growing at similar rate.

If I said that stock A is a good buy, because it is much cheaper in relation to B, do I miss anything in the middle? Since the investment return is largely reliant on the stock A being rerated higher.

How do you approach the relative valuation then?”

So, yes, I do use relative valuation. For example, I’ve mentioned before I own Vertu Motors (VTU). It is a U.K. car dealer. I believe the core economics of car dealer groups in the U.K. and U.S. are similar. I wasn’t able to find much evidence of differences in the way the businesses work, how they make money, how much they make, etc. There are legal differences. But, I couldn’t find any evidence these legal differences translate into actual differences in the economics, because the legal stuff didn’t seem to be the primary reason for the competitive advantages of certain incumbents in each market. A lot of the other economics seemed the same. Like, the scale economics seemed the same. The sales mix of what each dealer is selling and how much sales of other products (financing, warranties, etc.) they are doing and stuff like that seemed to work the same in one country vs. the other. A lot of the unleveraged returns on capital for similarly situated companies in each country seemed the same. It didn’t seem obvious to me that there was a material advantage to owning a U.S. based car dealer vs. a U.K. based car dealer. However, the U.S. based car dealer stocks traded at much higher relative valuations (higher P/B ratios, EV/EBITDA, ratios, etc.). Why?

One, the U.S. car dealers sometimes had greater scale. So, you would have to adjust for that. The market had – historically – seemed to reward companies sort of after the fact for scale. So, if you had improving margins, returns on capital, etc. now because you’d begun scaling up earlier – the market seemed to reward the companies when the actual earnings per share growth and such came in. They would then give them higher multiples. They wouldn’t necessarily give higher multiples to smaller dealer groups that said they’d scale up, even though we know that the operating margins on what was essentially the exact same group would be much higher once the group owned more dealerships. So, some of the appearance of higher relative valuations of U.S. car dealers vs. U.K. car dealers could just be the U.S. car dealers were sometimes already bigger. But, this would be solved over time as the U.K. car dealer groups seemed to be – and certainly said they wanted to be – consolidating much the way the U.S. car dealer groups had. So, that wouldn’t explain the difference.

Two, capital allocation was different. That’s more of a possible permanent problem. And it could easily explain why dealer groups (as stocks) in one country should be worth more than those in another. For example, if you have a country where the companies in an industry pay 100% of earnings out in dividends, don’t borrow, etc. and yet they are in a fundamentally good and stable business (like car dealers) – then, the value of that stock (because it’s a minority investment where you only get the returns the incumbent management can provide for you through their capital allocation decisions) is less than it would be if capital allocation was different. If U.S. car dealers focused on continually borrowing money, not issuing shares, not paying dividends, buying other dealers, and buying back their stock – they would have higher compound returns starting from the same price-to-book ratio as the U.K. dealers. This is a factor I did have to consider. And I had to admit that none of the U.K. car dealers was run from a capital allocation perspective exactly the way I’d want them to be. For example, some of them paid meaningful dividends. I don’t think there is any reason for a car dealer to pay dividends. The industry is highly fragmented and has increasing returns to scale. There is an advantage to having a lower cost of capital, to having bigger relationships with everyone from lenders to manufacturers to insurers to technology stuff. There are countless choices of companies to acquire within your industry. The industry is not highly differentiated in terms of how difficult acquisitions would be to integrate. Car dealers shouldn’t pay dividends. Piling up cash might make sense when nothing good can be done with it. When acquisitions can be done and funded through debt – that makes sense. At other times, buybacks make sense. Generally, if you can do acquisitions on a cash basis at a price that is in line with how the market is valuing your dealerships – you always should. You will scale up and not destroy value. When the situation is upside down – the public market appraises your stock more cheaply than dealerships you could acquire, you should buyback your own stock. Dealer groups are so small versus the total market they are in that, honestly, I am not in favor of any of them paying dividends.

So, that is the best explanation of a potentially justifiable reason for car dealers in one place trading at a discount to car dealers in another. Over time, a totally different philosophy toward capital allocation would lead to dramatically different compound returns at two car dealer stocks that started with the same P/B ratio, economics, etc. If one paid out dividends, issued stock etc. and the other borrowed as much as possible while acquiring when dealers could be bought at around the same multiple as their own shares traded for and then switched to buying back stock when their multiple was far lower than the multiple acquisitions could be done at – the second dealer group would, all else being equal, be a much better long-term investment.

There are many industries where extreme differences between the P/B, P/E, EV/EBITDA etc. multiples are fully justified. Two industries where I feel huge relative valuation differences are appropriate – yet most other investors disagree with me – are in insurance and banking. I don’t believe that two randomly chosen banks or two randomly chosen insurers should trade at similar P/B or even P/E ratios. If the business the companies are in is fundamentally the same – then, this would make sense except to the extent management at one is better than the other. But, many banks and insurers that get lumped in together are not realistic peers. A bank with an efficiency ratio that is 3 times the rate of another bank is not a true peer. An insurer writing specialized risks in a single state is totally different than an insurer writing general risks across the U.S. However, many investors consider finance companies to be essentially commodity companies (which is somewhat true, I mean the service they are providing is not very differentiated) and thus companies that should trade at similar multiples. I disagree with that in the sense that you wouldn’t value two commodity producers – one that has a $20/barrel cost of oil production and one that has a $60/barrel cost of oil production – as being worth similar multiples of book value, earnings, etc. – or of being worth similar values vs. reserves. The lower cost producer should be worth more per “x” amount of reserves.

How can you make this useful though? If one bank should be better than another – do you need to know how much better?

I don’t think so.

I think that, for the investor, there’s a pretty easy cheat here.

The hack is that you can ignore actual quantification if you can prove two things at the same time:

Company A SHOULD trade at a higher multiple than Company B (Business A > Business B)


Company A actually trades at a lower multiple than Company B (Stock A’s Price < Stock B’s Price)

I think it’s too hard to know that one bank should trade at a 40% premium to another bank in terms of P/B ratios or something. But, if you find a group of 5 banks (as I always did in the Singular Diligence peer section – we always chose 5 peers and then looked at the min, max, median, etc. multiples of those 5 peers) and you think the bank you’re looking at is better than all the peers and then also cheaper than all the peers – then, you have a good investment.

Sometimes, you can find such easy situations.

Often, you can’t have it so easy.

Vertu Motors is an example of where it wasn’t so easy. I thought the stock was clearly cheaper than peers. And I thought that fundamentally it was – from a business perspective, putting capital allocation aside – extremely similar to many dealers that traded at higher prices.

So, I thought car dealers weren’t really expensive. And Vertu was as good in rough terms as many of them. And it was trading quite a bit lower.

The problem, as you pointed out, is that value investors always assume the cheap company’s multiple will rise to meet the expensive peer’s multiple instead of vice versa.

So, couldn’t all car dealers get as cheap as Vertu?

I viewed that as unrealistic. The actual cash returns in the business were high. Cars are very easily saleable. From what I could learn, the land under most dealerships isn’t really the “highest and best use” for the land. So, the price-to-free-cash-flow multiples here were pretty low and price-to-book ratios were extremely low considering how solid and sellable the book was. If you look at car dealers, their tangible book value is basically a ton of used cars, new cars, buildings, and land. When they trade at deep discounts to tangible book value – they are selling below liquidation value. Something that is profitable virtually every year and usually – when using some debt, as almost all car dealers do use – when leveraged, earning average to above average returns on equity…when something like that sells for liquidation value, it’s too cheap.

So, did I buy Vertu based on relative valuation or liquidation?

Honestly, liquidation. I looked at car dealers in the U.S. and U.K. and asked how many earn returns so poor they should be worth less than tangible book of around 1. I found that really that doesn’t happen. I’m not saying buying at a price-to-tangible-book of 1 will always beat the market or something. They could make capital allocation mistakes. They could do dumb things. But, fundamentally, dealerships are assets where the value of land and cars as held on their books is not an unrealistic gauge of what someone should be willing to pay for them. So, if I thought you should pay about 1x tangible book for a generic car dealer (some groups have a better mix of assets, some worse – but
if you could buy the whole industry, you’d do well at 1x tangible book), then as long as I paid a discount to tangible book I’d be fine. Vertu stock often sold at like 2/3rds of tangible book or so. I really relied more on a discount to what I thought was a low but reasonable appraisal of generic assets in the industry – my belief that car dealers shouldn’t trade below tangible book value – to make a decision to buy Vertu at a discount to what I thought those assets were worth.

A similar example is a write-up I did of a cement plant. The company owns other stuff besides the plant. That’s harder to value. But, like a car dealer – I don’t think a cement plant is hard to value. I looked at it this way. Is the cement plant’s book value higher or lower than its replacement value? And: is the acquisition value of a cement plant higher or lower than replacement value?
These two questions skirt the issue of quantification.
I could try to precisely estimate replacement value. I could try to precisely estimate acquisition value.

I did look at multiples of prior deals and stuff like that.

But, I don’t think it’s important to do that. Because you can’t be very certain of that.

Numbers are hard to be sure of. Inequalities aren’t. Here’s a really easy inequality:

Tangible book value < Replacement value < Acquisition value

The logic here is that it’s not hard to prove that in today’s dollars you’d have to pay considerably more than the fully depreciated value of a plant built progressively over decades and put on the books in old dollars. Far, far more real dollars went into the plant than is shown on the books. So, replacement value is above tangible book. There’s no doubt of that. Then, you have the relationship between acquisition value and replacement value. It just does not make sense to build a new, modern cement plant within the same market as an existing one. The only time that would make sense is if the existing one was already incredibly profitable on an ROE basis and also 100% opposed to selling out to you. The two plants would each earn worse returns than just one. It always makes sense to try to offer some premium to acquire an existing plant than to try to add new capacity directly by building a new plant.

I did look at numbers of prior deals, ask about offers made for the plant, look at EV/ton for similar deals of similar plants across the U.S., etc.

But, mostly it was not an analysis based on relative valuation and careful quantification. Instead, it was based on more broad stroke micro-economic thinking and general logic. It’s difficult to quantify exactly what a plant is worth under future circumstances. I mean, as of today, it’s not hard. There were people willing to guess what a plant with “x” million tons per year production capacity located in such and such a state and with access to whatever lime at whatever price and so on was worth. But, these are really not long-term estimates. They are based on current prices, macro environment, aggressiveness of bidders, etc. The appraisal will bounce around a lot – just like on real estate. So, it’d be easy to get comparable figures. And if you look at something like Value Investors Club – that’s all they need. If you can buy this plant at $100/ton of annual capacity and all the recent deals in the industry have been done at $300/ton of annual capacity – then, buy this stock. That’s it.

If you make enough bets like that in enough different industries – it’ll work. Like, if you combine as many insurers in your portfolio with as many cement plants – this approach will work. But, if you bet all on one asset, you’ll sometimes have problems. Like a relative valuation of natural gas producers 7 years ago would just result in you buying stuff that was cheap vs. peers but that still declined by a huge amount. They were all really overvalued versus the future.

So, I like to focus on valuation methods that aren’t just relative valuation as of now. Instead, they are more about establishing a logically sound floor you don’t think the stock can be worth less than. Like, I figured in both the case of car dealers and cement plants I could prove that if you can buy an average or above average asset in those industries – you’d be getting a bargain as long as you were paying less than tangible book. In other industries, I’m not sure a discount to tangible book is enough. But, in car dealerships and cement plants – as long as it’s not an especially bad asset you’re buying – getting in at a discount to tangible book is all the valuation work you need to do. Like, I honestly don’t spend time debating whether Vertu is worth 100% of tangible book or 125%. I don’t feel like you need to do that kind of appraisal. I know I’m not going to buy the thing at 110% of tangible book. But, I know I will buy it at 55% of tangible book. So, why does it matter what sort of premium above book it’s really worth? I don’t think it does. It just matters whether book value keeps compounding, whether book value stays solid, and then whenever it’s way below book – I can buy. All other times – I can’t.
This fits with my general belief that actual appraisal of stocks is unimportant in investing. You don’t need to actually quantify the value of anything you buy. As an investor, you just need to know the merchandise is good enough and cheap enough. How good and how cheap shouldn’t be important once it surpasses what you need for your margin of safety.

Like, I talked about Cinemark (CNK) recently when it was trading at $8/share. I said it might be worth $45 a share if COVID had never happened. I could be wrong, and it’d be worth $30 a share without COVID. But, the stock was at less than $10/share. So, it really doesn’t matter. You don’t want to spend a lot of time worrying about whether something is trading at 60 cents on the dollar, 30 cents on the dollar, or 15 cents on the dollar. If you can buy something at between 15 cents on the dollar and 60 cents on the dollar – the questions to ask are not how cheap it is but how good, how safe, and most importantly how sure are you of it as an asset. You never need to establish cheapness below a 40% discount. If you’re in the right asset, a 40% discount is always enough. What you need to establish is whether you should even be in car dealers and cement plants and banks and movie theaters and so on – or not.

So, I always do use relative valuation. But, it’s mostly just to see if there are cheaper stocks out there. And especially to check if there are better businesses out there that are just as cheap. I mean, you never want to buy the #3 best movie theater chain if it turns out the #1 chain in terms of business quality is selling for the same low EV/EBITDA multiple. You always want to buy the best peer you can when they’re offered at the same price.