Blog Page

What Makes a Business Durable?

Hey Geoff,

I noticed in one of the recent videos you mentioned that the durability of the business matters more then the returns. I believe you were answering a viewer question on return on capital for the subject matter. When you mention durability are you talking about recurring revenue, sustainability of the business going forward (like waste collection industry or railroads etc…)?

Answer: A Good Place to Start is With the Oldest Companies in the Oldest Industries

Yes. So, this is something Buffett has said before too – and it’s true. It’s the reason why I’m not a fan of “The Magic Formula”. The Magic Formula is a system that might work empirically – but, it isn’t based on sound logic. That’s different from something like the Piotroski F-Score or Ben Graham’s 2/3rds of NCAV rule. Both of those approaches are logically sound and then can be tested to see if they work empirically. I don’t think it’s a good idea to use a system that has been back tested to show good results, but that doesn’t seem logically sound. I’m unconvinced of the logic of The Magic Formula – because, it is basically buying high current return on capital stocks without asking if they have a moat. It’s not the Buffett approach. It’s actually very different from Warren Buffett’s approach. His approach is to figure out why a company has had a high return on capital in the past, has a high return in the present, and is likely to have a high return in the future. Once he knows the reason for the high return – the company’s “moat” – he can judge how durable that moat is. This is also similar – though slightly different – from the Phil Fisher approach. Phil Fisher’s approach focused more on the organization and whether it is built for the long run: is it investing enough in marketing, is it investing enough in R&D, does it have good enough people at lower levels in the organization, are the markets the company is in likely to grow for a long time to come, has the company had success releasing new products regularly to replace old products, etc. That’s very similar – though from a different angle – to the Buffett approach. Those two approaches – Buffett and Fisher – are qualitative looks into the future. We can debate how accurately a human being can judge the likely future of a company. But, the logic of trying to do that is sound. The logic of assuming that a currently high return on capital is less likely to decline than a medium or low return on capital makes little sense. There has been some research in this area and the answer is – it depends on the industry. Some industries do show high persistence of relative returns among the firms in those industries. So, for example, the leading movie studio or beverage brand or condiment maker might tend to

Someone sent Geoff this email:

Hi Geoff,

On regular days your focusing on overlooked stocks makes much sense to me and is the right place to look for good companies at a decent price.

But in this market selloff, wouldn’t it be better to buy blue chip stocks such as Nike, Disney, BrKb, etc.. if they fall to the right price (and I’m not sure they are there yet…).

Also, wouldn’t it be better to diversify? To own 10-12 stocks instead of 5-7.

Answer: It Depends On Which Stocks – Conglomerates like Berkshire and Disney Give You a Lot of Diversification, Some Other Stocks Don’t

(Note from Geoff: I’m going to split this email into two responses – this one will talk about diversification, tomorrow’s response will talk about overlooked versus big cap stocks)

I don’t know about whether it makes sense to own 10-12 stocks instead of 5-7. I’m not sure it does. It depends on which stocks you’d add to the portfolio. Remember, there is a lot of diversification in certain stocks already. So, you mentioned Disney (DIS) and Berkshire (BRK.B). Berkshire already gives you more diversification in terms of its underwriting than many insurers do. Plus, Berkshire has a big stock portfolio. Plus it owns a railroad. Plus it owns an electric utility. You could separately buy other insurers, stocks like Apple (AAPL) and Bank of America (BAC) and Wells Fargo (WFC) and Coke (KO). You could buy a publicly traded railroad. You could buy a regulated utility. If you add up what you’d put in all those stocks if they were in a 12 stock portfolio – maybe you’d put 33% of your portfolio or more into recreating what Berkshire already has. So, should you buy a bank, an insurer, a railroad, a utility, etc. – or just buy Berkshire? I don’t know the answer. If you pick the right bank, insurer, railroad, and utility – I’m sure you could beat just buying Berkshire. But, if you think you’re somehow safer putting 10% each into one bank, one insurer, one railroad, and one utility than 40% in Berkshire – I don’t think that’s true.

Disney is similar. The company is diversified across a couple big entertainment segments. It has a big allocation to parks and resorts. This includes things like cruise ships. But, it’s mainly theme parks. Disney has the most visited theme parks in several parts of several different continents . So, you could buy SeaWorld (SEAS), Six Flags (SIX), or Cedar Fair (FUN) – or you could get a similar allocation by owning 3-5 times more of Disney than you would of any one of those companies. If Disney is say 20-33% parks by value (you’d have to appraise the stock yourself to make this decision) then simply buying a 3-5 times bigger position in Disney and you’d get the same allocation to theme parks you would from buying one of those other stocks. Disney has a big movie business (it is now, by

Interpublic (IPG) is one of the big ad agency holding companies around the world. Other examples include Omnicom (also a U.S. company), WPP (a U.K. company), Publicis (a French company), and Dentsu (a Japanese company). There are countless other publicly traded advertising stocks. Some are affiliated with one of those big groups. Others aren’t. Interpublic might be the first real ad agency holding company. The name Interpublic dates back to 1961. If you read the 10-K, you’ll notice PriceWaterhouse has been auditing Interpublic’s financials since the 1950s. Before the name change in 1961, the company was called McCann-Erikson. Late in the TV series Mad Men this is the agency that buys up the company all the main characters work for. Mad Men ends with the “Hilltop ad” (I’d like to buy the world a Coke) for Coca-Cola. That’s not made up just for the show. McCann-Erikson did make that ad. The McCann-Erikson combination dates back to the 1930s. Each of those two agencies (McCann and Erikson) were founded in the first decade of the 1900s. Omnicom and other advertising agency holding companies that came later were probably based in part on Interpublic in the early 1960s. For decades now, it’s been a common strategy for publicly traded advertising companies to own different agencies and provide certain centralized functions (basic corporate functions, capital allocation, setting compensation of top executives at the agencies, managing conflicts of interest between agencies they own, etc.). To some extent the individual agencies are independent and run sort of like Berkshire Hathaway runs its subsidiaries. But, in other ways they aren’t very independent. For example, Interpublic explicitly says that corporate HQ provides guidance, advice, etc. on both certain human relations stuff and on real estate. The cost structure of ad agencies is very different from most companies. At Interpublic, close to 65% of revenues are spent on base salaries, benefits, rent and office expenses. A huge proportion of the company’s total cost structure is really just base salary and rent. These are both very fixed. This is a pure service business. It’s largely “cost plus” – though how that works is a bit complicated.

So, the actual way ad agencies bill can be pretty complicated. Interpublic gets 50-60% of its revenue from its top 100 clients. Usually, client retention rates among big accounts are incredibly high. This is not due to contracts. The industry standard is for all contracts to allow either the agency or the client to fire the other with 30-90 days’ notice. Each of the contracts are customized. So, in theory, profits could vary a lot relative to the amount of revenue booked depending on the client. But, in reality, this does not happen. The contract can be created to involve a lot of cost plus, a lot of commission, flat fees, incentives for hitting certain quantitative targets for the effectiveness of campaigns, etc. If you look at the long-term margins of all different ad agencies around the world that once took very simple …

Someone Geoff  this email ( to ask your own question: just send me an email ) :

My question is a simple one.

“What do you think is a fair price to pay for Omnicom?”

Answer: Omnicom (OMC) Could be Worth Anywhere Between 1 to 1.5 Times Sales – so, About $70 to $100 a Share – But: You Probably Want At Least a 35% Margin of Safety, So Try Not To Pay More than $45 a Share for Omnicom

This is a simple question. But, it’s tougher to answer than it appears. I see 3 big factors to consider:

1) What is the organic growth (or decay) in Omnicom’s business likely to be?

2) How much will stock buybacks drive growth in earnings per share that’s higher than organic growth?

3) How bad and how long will an ad recession be?

Because Omnicom uses so much of its free cash flow to buyback stock, there’s an element of circular logic to valuing the stock. Bizarrely, the cheaper the stock gets today – the more valuable it should be in the future. Assume Omnicom uses 50% of its EPS to buyback stock. I’ll also assume there is 1% share dilution normally before buybacks. This means that the formula that tells us how much Omnicom’s EPS will grow BEYOND its organic growth is:

(0.5 * Earnings Yield) – 1% = Inorganic EPS growth caused by stock buybacks

Let’s run this number for the following P/E ratios:

P/E = 4

P/E = 6

P/E = 8

P/E = 10

P/E = 12

P/E = 14

P/E = 16

P/E = 4 (this would be like 50% cheaper than today’s price) makes the formula…

(0.5 * 25%) – 1% = 11.5% CAGR in EPS

P/E = 6 gives us…

(0.5 * 17%) – 1% = 7.5% CAGR in EPS and so on. The “and so ons” work out as follows..

P/E 4 => 11.50%

P/E 6 => 7.50%

P/E 8 => 5.25%

P/E 10 => 4.00%

P.E 12 => 3.17%

P/E 14 => 2.57%

P/E 16 => 2.13%

Now, if we assume organic growth is equal to 0% – which may or may not be an accurate assumption – then, we can keep the fair value formula pretty simple. In theory, Omnicom should be worth its stream of dividends. EPS can be ignored as a direct thing. Instead we just assume that 50% of EPS is paid out in dividends and 50% is used to do buybacks. Again, I assume 1% is just constant share dilution in the form of stock grants to employees.

That means only two things matter. One is the buyback rate. I’ve shown that above. It ranges from a little over 2% to something like 11-12% a year depending on whether the stock trades as low as a P/E of 4 or as high as a P/E of 16. The dividend yield – if we assume a 50% dividend payout – is really easy to calculate. We just take …

Miller Industries (MLR) has a lot of things to like about it. But, the timing of buying this stock now definitely isn’t one of them. Miller is in a very cyclical, highly durable capital good industry – it produces “car wreckers” and “car carriers” – that depends heavily on business confidence and especially access to capital. It is very easy to defer the purchase of a new wrecker or carrier. And it is very hard – impossible, really – to sell wreckers or carriers without easily available credit. To give you some idea of how important credit is in this industry – Miller Industries is currently promising lenders to its distributors (these are all technically “independent distributors”) that it will buy back up to $74 million of its own wreckers and carriers if the lender repossess that collateral from the distributor. Miller makes this kind of promise all the time. In recent years, it has not had to buy back any of its equipment. But, you see the problem. It might have to do so. And, the fact that distributors are using financing that depends on the lender getting a promise from Miller (the original equipment manufacturer) gives you some idea of how important credit is in this industry. The distributors – there are 80 of them in the U.S., and Miller estimates that about 68 of them don’t actually sell any wreckers or carriers other than Miller products – rely heavily on floorplan financing.

The industry is also very cyclical. In the last economic cycle, Miller’s sales peaked at $409 million in 2006 and bottomed out at $238 million (down 42%) in 2009. Gross profit dropped by the same percentage (42%). Operating profit – however – went from $33 million in 2006 to $7 million in 2008 (down 80%). Could the same thing happen in this recession?

Yes, it could.

So, Miller’s P/E, P/S, etc. ratios are all very suspect right now. Maybe price-to-tangible book value would be a better guide to the company’s valuation. The good news is that the P/E and P/S ratios here are low. But, that’s what you’d expect with a cyclical stock that everyone now knows is at the very top of its cyclical (the recession has already started as I write this, and Miller reported earnings less than a month ago that were its best ever – so, we can call this the official peak). The P/E ratio on those peak earnings is between 7 and 8. The P/S ratio is about 0.4 times. The company has a tangible book value of $21.60. As I write this – the stock is trading at $25.89. So, the P/B ratio here is 1.2 times. That makes it pretty easy to compare this company’s long-term history with its current stock price. Miller probably doesn’t convert all its reported earnings into cash. So, if stocks generally return say 8-10% a year – and we use that as the hurdle rate you, as an investor are looking for …

Someone sent Geoff this email (to ask your own question: just send me an email):

I think that car dealers face a lot of competition, especially for used cars and service, that I could never thoroughly understand the competitive advantages. But car dealers tend to have very stable margin. So I do think that they have certain competitive advantage thanks to location, long-time presence, as well as customer relationship built up through its new car business. So, I’d like to ask you about your thoughts on the competition for car dealers in:

1. Used cars where car dealers have to compete not only with franchised dealer of another brand in the same area but also with independent used car dealers.

2. Parts and services where car dealers have to compete with independent auto service shops and chains.

Answer: I Think Big Car Dealer Groups Have Certain Advantages in Capturing Additional Profits from a Customer Relationship Based on a New Car Sale – But, the Internet May Change Things Over Time

It’s a good question. I think it’s possible that an independent seller of just used cars might be a better business – especially if it combined locations with a really, really strong internet presence.

Big car dealer groups do have to compete with service shops. And perhaps they have lost share. But, the number of dealer locations relative to the amount of population (and certainly to drivers) has been declining throughout much of the history of the industry. So, I think the business has tended to get better not worse.

What’s the big difference?

I think two things benefit dealers.

Once, they potentially have a better source of leads for their new, used, and service businesses – in the form of a new car territory they control for a brand. So, often, when someone is in the market for a new Toyota they trade in their old Toyota. The dealer can take the trade in (getting a better supply of used cars than others might) and can get the possibility that the owner of the new car will buy financing products etc. through them and – most importantly – stay with them for servicing. So, they are in a better position than some to turn a new car sale into out sources of revenue – a possible service relationship, a good price on a trade-in used car, etc. This is why I think it’d be hard to compete in new cars even if there wasn’t the dealer system that there is. New cars are largely just a good way to have a customer relationship that can make you more money form things over time.

Second – and I think for most dealers, this is the bigger issue – I think dealer groups that include new car sales, used car sales, and servicing in one operation have a lower cost of capital than competitors would. It has been very easy for new competitors in the industry to succeed in stuff

Middleby (MIDD) is a stock I was excited to write-up, because it’s rarely been cheap. I’ve seen 10-year financial type data on the company. I’ve seen it show up on screens. And now the government response to coronavirus – shutting down so many of the restaurants that Middleby supplies – looked like a once in a lifetime opportunity to buy the stock. The company also has quite a bit of debt. That can make a stock get cheap quickly in a time like this. So, I was looking forward to writing up Middleby.

I’ll spoil this write-up for you now and tell you I didn’t like what I found. The company’s investor presentation, 10-K, etc. was a disappointing read for me. And I won’t be buying Middleby stock – or even looking into it further. Why not?

Middleby is in an industry I like. The company has 3 segments. The biggest profit contributor is commercial foodservice – supplying restaurant chains like: Burger King, IHOP, Chili’s, etc. – with kitchen equipment. The company sells equipment that cooks, bakes, warms, cools, freezes, stores, dispenses, etc. It sells a very broad range of equipment. A lot of it is good equipment. A lot of the brands the company carries are well known. Plenty of them are pretty innovative. However, I think that innovation most likely happened before – not after – Middleby bought those companies. Middleby talks a lot about innovation – but, it only spends about 1-2% of its sales on research and development. That’s not a high number for a company in an industry like this. Middleby is building this stuff itself – in the U.S. and around the world (in both owned and leased facilities – and it’s making it for some pretty large scale orders. Plenty of the chains Middleby serves have thousands of locations. Gross margins in the commercial foodservice business are 40% or a bit better. Although most of this stuff is sold under just a one year warranty – some is under warranty for up to 10 years. Plenty of these products do last longer than 10 years. For a company making 40%+ gross margins on sales of key capital equipment to big business customers – Middleby doesn’t do a lot of R&D. That fact bothered me a little. It started to bother me a lot more when I looked closer at the company’s acquisitions.

Middleby has bought some leading brands. Chances are – if you’re reading this – you know more about home kitchens than commercial and industrial kitchens. You may eat at Chipotle. But, you don’t know what equipment Chipotle cooks on. You do – however – know brands you might find in home kitchens. Several years ago, Middleby bought the Viking brand and later the Aga brand (Aga is a big, premium brand in the U.K. – it’s less well known in the U.S.). In the company’s investor presentation, they show the EBITDA margins for these acquisitions at the time they were made …

Someone sent Geoff this email:

Dear Geoff,

In this video   you mention a PE of 26 for OTC and yet use an implicit growth rate of 6% which is not the inverse of 26. I would like to understand why that is.

Answer: If a Company Converts 100% or More of Its Earnings Into Free Cash Flow While Also Growing – a Normal P/E Ratio Would Deeply Undervalue the Stock

I’d assume it’s because 10% = (1/26) + 6%

In other words…

10% – 6% = 4%

And 1/25 (not 26) is 4%

OTCMarkets is unusual. What matters for a stock isn’t actually earnings – it’s truly free cash flow. So, a stock that retains 100% of its earnings and grows 6% a year while doing that still isn’t worth more than many stocks growing 0% and paying all earnings out.

Some stocks – OTCMarkets is one – have a lot of “float”. As a result, they are capable of paying out 100% (actually more than 100% while they are growing) of their earnings in dividends and buybacks while growing. They don’t actually have to retain earnings. You can see that with OTCM. The asset that has grown over time is cash. A tip-off to this fact is that ROIC is often calculated as being very high, infinite, or even negative by some websites (a negative ROIC means that the company’s cash balance is greater than what it has actually invested in the business in stuff like PP&E, receivables, inventory, etc.).

So, my point is that if OTCMarkets grows by 6% a year and you need a 10% return as your discount rate…

Well, OTCMarkets can grows at 6% a year AND pay a 4% a year dividend yield if priced at 25 times P/E.

Note that OTCM does not actually pay such a high dividend though. So, it’s somewhat debatable what the stock is worth if management keeps cash on the balance sheet rather than using all FCF to buy back stock and pay dividends.

For comparison, look at Omnicom (OMC). It actually does pay out dividends and buy back stock equal to at least 100% of reported earnings. So, the calculation there is simple. Assume OMC grows at 0%. Assume your discount rate is 10%. What is the correct price for OMC?

It’s a 10x P/E. Because OMC at 10 times EPS, OMC will pay you dividends and buy back stock equal to 1/10th of your purchase price. What this will actually do is make EPS grow – even when the company’s actual earnings don’t grow, just because share count is falling – plus you get the dividend. If the combination of the dividend and the EPS growth equals 10%, you’ve hit your hurdle rate.

So, what I said about OTCM is somewhat unique to that company and other business models like it, generally:

– Ad agencies

– Subscription services

– Some software companies

– Etc.

Usually, you have to be doing something fairly intangible (you can’t

Hamilton Beach Brands (HBB) looks like a cheap stock. As I write this, the stock is trading at $8.71 a share. The company – or, at least the continuing part of the company now that Kitchen Collection is gone – earned anywhere from $1.50 to $2.00 a share over the last 3 years. That puts the P/E ratio at something like 4-6. Other ways to look at the stock include pricing it off of its EBIT or free cash flow. Over the last 5 years, the continuing portion of the company reported EBIT of between $33 million and $41 million. The company’s enterprise value is about $186 million (just take the market cap and add about $55 million in debt). So, that gives us an EV/EBIT ratio of anywhere from 5 to 6 times. Again, not expensive. A “normal” EV/EBIT ratio given today’s tax rates would probably be about 12 times. Free cash flow has averaged just under $30 million a year over the last 3 years. That gives you an EV/FCF ratio of about 6. All of these point to the stock being pretty cheap. The company also has a plan to one day achieve revenues of between $750 million and $1 billion and EBIT margins of 9-10%. It’s a long-term plan. But, a company that hit even the bottom end of that range could be worth closer to $800 million than the less than $200 million enterprise value at which Hamilton Beach Brands now sits.

The cheapness of Hamilton Beach Brands stock is the good news here. There is some bad news. And I’ll start with the news that concerns me the most – the balance sheet. On the one hand, Hamilton Beach has a solid balance sheet. Current assets exceed total liabilities. That’s usually a great sign. The company is always solidly EBIT profitable. However, it isn’t always very solidly cash flow positive. If we look at the make-up of Hamilton Beach’s balance sheet we can see why. At present, 38% of HBB’s assets are held in the form of receivables. 36% of assets are held in the form of inventory. Less than 1% of assets are in the form of cash. The company’s undrawn portion of its credit line is only about 20% of its total balance sheet as it stands now. In fact, the amount HBB can draw on that line is less than 50% of either its receivables or its inventory. Furthermore, HBB is a seasonal company. It does things like selling its receivables off and drawing on its credit line in a normal year. Normally, HBB builds up inventory during the first half of the year. Things turn around starting in the fall. And then the cash comes flowing in as we go through the holiday season. Will that happen this year? Hamilton Beach has already promised it will. The company had – as of December 31^st, 2019 – promised to buy $210 million worth of inventory this year. That’s against …

Someone sent me this email:

…you mentioned Omnicom, and you said that per share numbers should be avoided when computing growth…as you mentioned Omnicom buys back its stock, therefore why remove this information I have about this company from the implicit hurdle rate? Suppose we have 3 companies: a historically savvy buy backer, an automatic buy backer and a company who doesn’t buyback, shouldn’t we be able to discriminate between them if they were to all have the same FCF yield?

Answer: It’s Best to Assume Future Growth in Earnings Per Share Caused by Buybacks Will Be Tied to the Price at Which Future Buybacks Are Done

The reason I said you shouldn’t use per share numbers is to avoid double counting. Here’s what I mean.

You buy Omnicom at a P/E of 8.6 and a dividend yield of 5%. You know free cash flow is normally equal to or greater than earnings per share. However, in a recession like the one we’re in now for Omnicom – it’s likely free cash flow will come in below reported earnings. This would happen if there was a decline in billings (similar to how an insurer’s “float” drops an insurer takes in less premiums this year than last year).

So, what does that P/E of 8.6 and dividend yield of 5% buy you?

What I’m worried about is investors saying: “Well, 1/8.6 = 11.6%. My earnings yield is 11.6% AND the company has grown sales PER SHARE by 5% a year over the last 10 years. So, My return is 11.6% + 5% = 16.6% for as long as I own the stock.”

Now, I don’t doubt your return if you bought Omnicom today and sold it in the future could be 17% a year or higher. However, you’ll get that return from multiple expansion. Without the P/E multiple expanding, you won’t make 17% a year in this stock.

Why not?

There are two ways of looking at this.

One, you CAN count the earnings per share growth (or sales per share growth if we’re avoiding looking at changes in EBIT margins) and count the dividend. However, if you count the EPS growth – you CAN’T count the money used on stock buybacks. So, it’s a question of EITHER credit the company’s use of stock buybacks as if it’s as good as a dividend OR count the growth in per share figures caused by stock buybacks, but don’t count the growth in EPS caused by the falling share count.

Personally, I prefer counting all of the free cash flow used on buybacks and dividends as your “hold” return in the stock. I have a good reason for doing this that has to do with timing an investment in OMC to maximize your returns.

However, it is theoretically permissible to do the calculations as:

Dividend Yield = 5%

Sales Per Share Growth = 5%

“Hold” Return = 10%

I think this undervalues the stock as of today. But, that is more consistent with