Investment Philosophy

Importance of ROIC: “Reinvestment” vs “Legacy” Moats

I’ve talked a lot about the importance of the concept of return on invested capital (ROIC), and how it is a key driver of value in a business. Feel free to go back and read some of those posts here. In this particular post, the discussion is continued. This post is something new for BHI: it’s a guest post written by my good friend Connor Leonard (see his brief bio at the end of the post).

Connor and I live in the same area (Raleigh, North Carolina), and we get together on a regular basis to discuss businesses we follow as well as investment strategy. He and I think very similarly when it comes to investing in high quality businesses that can create significant value over the long-run. Connor is a smart investor, and I appreciate his willingness to be a sounding board for me at times.

This post contains his own thoughts (unedited by me) regarding the importance of a company being able to retain and reinvest its cash flow at high rates of return. I think he articulates the concept very well.

Here is Connor’s post:

The Reinvestment Moat

Outstanding companies are often described as having a “moat”, a term popularized by Warren Buffett where a durable competitive advantage enables a business to earn high returns on capital for many years[1]. These businesses are rare and form a small group, however I bifurcate the group further into what I classify as “Legacy Moats” and “Reinvestment Moats”. I find that most businesses with a durable competitive advantage belong in the Legacy Moat bucket, meaning the companies earn strong returns on capital but do not have compelling opportunities to deploy incremental capital at similar rates.

There is an even more elite category of quality businesses that I classify as having a Reinvestment Moat. These businesses have all of the advantages of a Legacy Moat, but also have opportunities to deploy incremental capital at high rates. Businesses with long runways of high-return investment opportunities can compound capital for long stretches, and a portfolio of these exceptional businesses is likely to produce years of strong returns. It will take some work and a lot of discipline to filter down to the true compounding machines, however I will outline what factors to look for and how many of the “bargains” hide in plain sight.

The “Legacy Moat”:

Businesses with a Legacy Moat possess a solid competitive position that results in healthy profits and strong returns on invested capital. In exceptional cases, a company with a Legacy Moat employs no tangible capital and can modestly grow without requiring additional capital. However because there are no reinvestment opportunities offering those same high returns, whatever cash the business generates needs to be deployed elsewhere or shipped back to the owners.

Think of a self-storage facility in a rural town with a high occupancy rate and little competition. This location may be generating $200,000 of annual free cash flow, a solid yield on the $1,000,000 of capital used to build the facility. As long as they run a tight operation, and a competing storage facility doesn’t open across the street, the owner can be reasonably assured that the earnings power will persist or modestly grow over time.

But what does the owner do with the $200,000 that the operation generates each year? The town can’t really support another location, and nearby towns are already serving the storage demand adequately. So maybe the owner invests it in another private business, or puts it towards savings, or maybe buys a lake house. But wherever that capital goes, it likely won’t be at the same 20% return earned on the initial facility.

This same dilemma applies to many larger businesses such as Hershey’s, Coca-Cola, McDonald’s or Proctor & Gamble. These four companies on average distributed 82.4% of their 2015 net income out to shareholders as dividends. For these companies that decision makes sense, they do not have enough attractive reinvestment opportunities to justify retaining the capital.

Even though these Legacy Moat businesses demonstrate high returns on invested capital (ROIC), if you purchase their stock today and own it for ten years it is unlikely you as an investor will achieve exceptional returns. This is because their high ROIC reflects returns on prior invested capital rather than incremental invested capital. In other words, a 20% reported ROIC today is not worth as much to an investor if there are no more 20% ROIC opportunities available to direct the profits.

Equity ownership in these businesses ends up resembling a high-yield bond with a coupon that should increase over time. There is absolutely nothing wrong with this, businesses like Proctor & Gamble and Hershey’s provide a steady yield and are excellent at preserving capital but not necessarily for creating wealth. If you are looking to compound your capital at unusually high rates, the focus needs to shift to identifying businesses that also possess a “Reinvestment Moat”.

The “Reinvestment Moat”:

There is a second group of companies that have all the benefits of a Legacy Moat, but also have opportunities to deploy incremental capital at high rates because they have a Reinvestment Moat. These companies have their current profits protected by a Legacy Moat, so the core earnings power should be maintained. But instead of shipping the earnings back to the owner at the end of each year, the vast majority of the capital will be retained and deployed into opportunities that stand a high likelihood of producing high returns.

Think of Wal-Mart in 1972. There were 51 locations open at the time and the overall business generated a 52% pre-tax return on net tangible assets. Clearly their early stores were working, they dominated small towns with a differentiated format and a fanatical devotion to low prices. Within the 51 towns, I would bet that each store had a moat and Sam Walton could be reasonably assured the earnings power would hold steady or grow over time. Mr. Walton also had a pretty simple job when it came to deploying the cash those stores generated each year. The clear path was to reinvest the earnings right back into opening more Wal-Marts for as long as possible. Today there are over 11,000 Wal-Mart locations throughout the world and both sales and net income are up over 5,000x from 1972 levels.

How To Identify a Reinvestment Moat:

When searching for a Reinvestment Moat, I’m essentially looking for a business that defies capitalism. Isolated profits in a small market is one thing, but continuing to achieve high returns on incremental dollars for years should in theory not be attainable. As a business gets bigger, and the profits become more meaningful, it will attract more and more competition and returns should eventually compress. Instead I’m looking for a business that actually becomes stronger as it gets bigger. In my opinion there are two models that lend itself to this kind of positive reinforcement cycle over time: companies with low cost production or scale advantages and companies with a two-sided network effect.

Low Cost / Scale:

Going back to the early Wal-Mart example, the stores were so big compared to traditional five and dimes that Mr. Walton could sell each item at a lower margin than competitors and still operate profitably due to the large volume of shoppers. The more people that shopped at a given Wal-Mart, or the more Wal-Marts that were built, only furthered this cost advantage and widened the moat. The lower prices enticed more shoppers, and the cycle continues to reinforce itself. So by the time there were 1,000 Wal-Marts in existence, the moat was significantly wider than when there were 51. Other businesses such as Costco, GEICO, and Amazon have followed a similar playbook, creating a “flywheel” that accelerates as the business grows[2].

Example: Sketch of the Amazon Flywheel

Connor-Amazon Flywheel

Two-Sided Network:

Creating a two-sided network such as an auction or marketplace business requires both buyers and sellers, and each group is only going to show up if they believe the other side will also be present. Once the network is established however, it actually becomes stronger as more participants from either side engage. This is because the network is stronger for the “n + 1,000th” participant compared to the “n” participant directly as a function of adding 1,000 participants to the market[3].  Another way of describing this: as more buyers show up it will attract more sellers, and that in turn will attract more buyers. Once this positive cycle is in place, it becomes nearly impossible to convince either buyer or seller to leave and join a new platform. Businesses such as Copart, eBay, and Airbnb have built up strong two-sided networks over time.

Andreesen Horowitz’s example of Airbnb’s two-sided network:

Judging the “Runway” to Reinvest:

Many investors focus purely on growth rates, driving up the valuation of a company growing at high rates even if the growth does not carry positive economics. The key to Reinvestment Moats is not the specific growth rate forecasted for next year, but instead having conviction that there is a very long runway and the competitive advantages that produce those high returns will remain or strengthen over time. Instead of focusing on next quarter or next year, the key is to step back and envision if this company can be 5x or 10x today’s size in a decade or two? My guess is for 99% of businesses you will find that it is almost impossible to have that kind of conviction. That’s fine, be patient and focus your energy on identifying the 1%.

Admittedly this is the most difficult step of the process, with many variables and uncertainties. Each situation will be different, but below are items I look for as positive indicators of a long runway and also red flags that the runway is concluding:

Positive Indicators:

If a two-sided network is consistently increasing key metrics like users or gross transactions but is still a small percentage of the overall market:

  • Focus on companies with a high “flow through” margin on an incremental user or transaction, which will help the company expand margins as the network grows.

If a company has a structural advantage that leads to a lower cost model than competitors:

  • This could be a differentiated business model such as selling direct rather than through agents. Or it could be an advantage developed over time such as technology that results in greater automation. The structural advantage has to be difficult for larger incumbents to duplicate.

If a multi-unit retailer currently has less than 100 locations but foresees an end market of over 1,000:

  • Focus on companies with a consistent, profitable, and replicable model. The company should primarily be “stamping out” the same prototype over and over while producing consistent unit economics.

Red Flags:

If a company that claims to have a long runway begins shifting into new or different markets:

  • If the future is so bright, then why deviate from the plan? The management may already know the runway is limited and is making a pivot.

If management’s definition of the total addressable market (TAM) is suspect:

  • Some management teams like to throw out a massive TAM number in a slide deck for investors to focus on. Check the underlying source of that number, if their definition is overly broad they may be trying to mislead investors.

If the recent vintages of growth investments are producing lower returns:

  • If the most recent stores opened are producing lower sales and margins, but cost just as much, the runway is showing some cracks. Many multi-unit businesses begin to show lower unit returns once outside of core markets.

Why Value Investors Often “Miss” the Reinvestment Moats:

I believe identifying these businesses with Reinvestment Moats is possible with some work, but many value investors struggle with identifying a “reasonable” price. My theory is that these Reinvestment Moats tend to “hide in plain sight” because most investors underappreciate the impact of compounding.

When assessing a quality business, value investors will often point to a P/E ratio over 20x or the EV/EBITDA multiple of 10x+ to show that Ben Graham would surely shake his head in disgust over such a purchase[4]. However let’s consider two investments and determine which will yield better results over a ten year horizon. The first business, Reinvestment Corp., has the ability to deploy all retained earnings at a high rate because of the strong Reinvestment Moat it possesses. Of course investors acknowledge this likelihood, meaning the entry price is fairly high at 20x earnings, leading most bargain hunters to pass. On the other hand, Undervalued Corp. has all of Graham and Doddsville in a buzz because it’s a steady business with a nice dividend selling for a bargain of only 10x earnings! Assume that over time both companies will be valued in-line with the market at 15x:

Connor-Reinvestment vs Legacy 1
*Assumes all earnings not reinvested are distributed as dividends                                                                                            
**Pre-tax IRR, factoring in tax rates will only further the advantage of Reinvestment Corp.               

This example illustrates a concept Charlie Munger outlined in “The Art of Stockpicking”:

“If the business earns 6% on capital over 40 years and you hold it for that 40 years, you’re not going to make much different than a 6% return—even if you originally buy it at a huge discount. Conversely, if a business earns 18% on capital over 20 or 30 years, even if you pay an expensive looking price, you’ll end up with a fine result.”

To create an even more extreme example, if you find a business that you believe is capable of earning strong returns over a decade, look at how much you can “overpay” and still earn a return equal to a typical business:

Connor-Reinvestment vs Legacy 2
*Assumes all earnings not reinvested are distributed as dividends       
                                                                                     **Pre-tax IRR           

Time is certainly the friend of a great business. But does this mean that only businesses with a Reinvestment Moat should be considered for a long-term investment?

Earning High Returns Investing in Legacy Moats:

A solid Legacy Moat paired with the right management team and strategy can be a wealth creator for shareholders over many years. In order to accomplish this, the playbook has to change into one more focused on capital allocation, specifically a systematic focus on acquisitions and managing the capital structure. In a sense the management team’s capital allocation prowess must become the Reinvestment Moat.

While the research on the negative consequences of M&A for corporations is extensive, I think there are a select group of management teams that can actually reinvest the company’s capital better than individual shareholders could do on their own. They tend to operate solely within their circle of competence, which is typically the sector where the underlying business resides. With deep industry knowledge, access to deal flow, and the ability to achieve operational synergies, these companies can operate like a private equity fund with permanent capital (and without the fee structure). Notable examples include TransDigm Group, Danaher Corporation, and Constellation Software.

Typically the management team consists of at least one “Operator” and a sole “Allocator”. The Operator is tightly managing the existing businesses to maintain their competitive positions. The Allocator functions more as an investor than a CEO, seeking out opportunities to deploy capital at high rates while also optimizing the capital structure. For the Allocator the capital structure is another means of creating shareholder value, and it is common to see special dividends, strategic use of leverage, and lumpy share buybacks that only occur when the stock is undervalued. William Thorndike’s book “The Outsiders” does a fantastic job of detailing these unique management teams with a talent for capital allocation.

I’ve found that the best way to find these companies is by reading annual shareholder letters and picking up on certain qualitative patterns.  First, a thoughtful and informative annual letter is key because it shows the managers view the shareholders more as business partners and co-owners rather than a pesky group they have to deal with each quarter. While I prefer to do further research, typically the letter is so insightful to a potential owner that they could make an informed investment decision simply by reading it each year. The letters typically contain terms like “intrinsic value”, “return on capital employed”, and “free cash flow per share” rather than simply discussing sales growth. These businesses tend to view frugality as a source of pride, with the home office setting the tone that each dollar is valuable because it ultimately belongs to the shareholder. If you happen to come across one of these companies, and you think the management team has a number of years remaining with plenty of attractive M&A targets, my advice is to buy the shares and let them take care of the compounding for you.


Most of the companies that are identified as having a “moat” tend to be Legacy Moats that produce consistent, protected earnings and strong returns on prior invested capital. These are perfectly good businesses and can produce nice returns for investors in a comfortable fashion.

However if you aiming to compound capital at high rates, I believe you should spend time focusing on businesses possessing a Reinvestment Moat with a very long runway. These businesses exhibit strong economics today, but more importantly possess a long runway of opportunities to deploy capital at high incremental rates. If these are hard to come by, the next best alternative is a business with the combination of a Legacy Moat and an exceptionally strong capital allocator. It will take some work and a lot of discipline to filter down to the true compounding machines, however a portfolio of these exceptional businesses acquired sensibly is likely to produce years of strong returns.

[1] If you are new to the concept of “moats”, this video of Buffett speaking to MBA students at the University of Florida does as far better job of describing the concept than I can:  

[2] The concept of a flywheel is popularized by Jim Collins, you can read more about it here:

[3] This description of the strength of a network business is taken from venture capitalist Bill Gurley

[4] According to the postscript to the revised edition of the Intelligent Investor, Ben Graham made more money off his stake in GEICO (a true Reinvestment Moat with lost-cost advantages) than he did from every other investment in his partnership over twenty years combined. At the time of the GEICO purchase, Graham allocated about 25% of the partnership’s funds towards the investment.

John’s Comment: Thanks again to Connor for putting this guest post together. I think it is a good extension of some of the investment concepts we’ve talked about here before. For related posts on this topic, please review the ROIC label as well as a recent post I did summarizing the talk that Connor referenced where Buffett does a particularly great job summarizing some of his investment tenets, including the concepts discussed in this post.

Connor Leonard is the Public Securities Manager at Investors Management Corporation (IMC) where he runs a concentrated portfolio utilizing a value investing philosophy. IMC is a privately-held holding company based in Raleigh, NC and modeled after Berkshire Hathaway. IMC looks to partner with exceptional management teams and is focused on being a long-term owner of a family of companies. 

47 thoughts on “Importance of ROIC: “Reinvestment” vs “Legacy” Moats

  1. Very nice post! I liked how you explained it graphically with the Reinvestment Corp. Thanks Connor and John!

    Thinking about the Runway left and total addressable market for a company is something a lot of people forget. Often they just projekt growth rates from the past into the future.

    I think Tom Gayner from Markel once said something like “If you can just focus on one thing about a company it should be their capital allocation.”
    Could be that it was in that video (I recommend it for your road trips):

    “If a company that claims to have a long runway begins shifting into new or different markets” could in some cases still make sense of course.



  2. Really enjoyed this post. However, I think it’s worth giving a little more consideration to the downside risk when paying-up for high-return businesses. If you use the exact same starting assumptions as above but then assume Reinvestment Corp delivers an 18% ROIIC (rather than 25%); the resulting IRR would be inferior to a purchase of Undervalued Corp over a ten-year horizon (all else being equal).

    To me, therein lies the danger with paying-up for businesses. Even if you correctly identify a meaningful moat and thereby anticipate that Reinvestment Corp’s 10 year ROIIC will be very impressive (18% would still be a very strong return); you run the risk of overpaying because the high initial-multiple leaves little margin for error in your ROIC estimate…How confident can you really be that the ROIIC for any given business will be exactly 25% rather than 15% or 18% over a ten year time-horizon? If you’re even moderately wrong in the ROIIC you assume, you could end up generating a weak IRR if you’ve paid a very high starting multiple.

    I absolutely agree that paying-up for high ROIC businesses is justifiable in theory; but implementing that strategy in reality is incredibly difficult because it all-too-often requires the kind of perfect foresight that most of us simply don’t possess. In the case of “boring” Undervalued Corp, you’re only reliant on the business continuing to reinvest at close to cost of capital in the long-run (which most ordinary businesses in decent industries should be able to do) to generate a nice IRR for you as investor. The outperformance may not be as eye-grabbing; but the margin of safety in such cases is significantly higher.

    1. I’ll respond in my own words, and we’ll see if Connor wants to address any of these comments (if so, I’ll post them here in the comments section). So I’m not speaking for him, but my own comments (without checking your math or getting into the numbers), from just a conceptual point of view, I’d say the point is that paying up for great businesses often makes a lot of sense.

      In other words, sometimes a company with a high P/E ratio can actually be extremely cheap. Obviously, you have to analyze the value of the business and determine that the value is in fact significantly greater than the price, but the point that I took away from this post is that it’s important to understand the math and the power behind a company that has the ability to reinvest its earnings at high rates of return.

      One other consideration is taxes, as Connor mentioned. Buying cheap stocks at 10 P/E and selling them each year will reduce the after tax returns to a greater degree than the tax liabilities associated with the long-term compounder.

      Thanks for reading, and thanks for the comment.

    1. So you can pay 36x and be wrong on ROIIC estimate by 36% and still break-even relatively against the typical companay. That’s a pretty good margin of safety compared to the probability of typical company’s business deteriorating.

  3. Well Ben Graham made not more money from his investment in Geico compared to other investments but made more money from the Geico investment compared to the sum of the money he made in other investments in his lifetime !!!
    How does somebody calculate the amount of capital that was invested from retained earnings for a given year (change in PPE) back in the business ???
    —Amazing article,just like the terrific job that John does .–

  4. Great post, thanks Connor. It is very difficult to identify businesses that will compound at very high rates of return for long periods ahead of time. Yet, if you run a concentrated portfolio and only have one of two of these companies during your career, you will likely do very well, especially if the other things you own do okay. This was Graham’s experience with Geico. It’s not like in order to be successful you have to identify an entire portfolio of these types of stocks. Just one or two will do. Second to finding them in the first place, the most difficult aspect of owning such stocks is hanging onto them. I’ve often wondered what the results would be long-term if someone vowed at the inception of their investing career to never sell a stock once purchased, over say a 30 year period. Would the flexibility given up under such strict rules be more than compensated for by one or two major winners? I guess such an approach would only be possible if you had additional funds for investment consistently available over time like in a 401-k or something.

  5. This is a great post. The mathematical example of how a high-return business with reinvestment opportunities can actually be inexpensive at 20x or even 30x earnings is particularly revealing. One question on that example though. It assumes that current earnings power of $100 cannot grow at all without at least some reinvestment. (Mathematically, the example suggests that earnings power would still be $100 in year 10 if the reinvestment rate was 0%. The growth to $163 is completely due to reinvestment.) But what if “Undervalued Corp” is a “legacy moat” business that can grow earnings without additional capital reinvestment? (Or for that matter, what if “Reinvestment Corp” can grow its earnings without reinvestment?) There aren’t too many businesses that can do that, but they certainly do exist. I wonder if you think that the analysis above would be more complete if it allowed for that possibility. Or do you think that it is impractical to expect earnings growth without at least some level of reinvestment?

    I’m not trying to nit-pick. The message about the power of businesses that can compound value over long periods of time remains the same either way and is clearly the most important takeaway, and this does a great job of illustrating that. But I’m interested in understanding whether you think that the point about growth that’s independent of reinvestment makes sense? It seems to me that it could play a meaningful role in analyzing some companies.

    1. That’s a good point Chris. I think the math that Connor laid out was designed to demonstrate the power of a business that can reinvest at high rates (and thus justify a higher current earnings multiple), but certainly there are some businesses that can grow without capital. Some businesses can grow through consistent pricing power (See’s Candies is one well-known example). Moody’s is another example of a business that has experienced growth with little to no capital requirements. These are rare birds, but they do exist. With both See’s and Moody’s, the businesses needed virtually no new capital, and so the cash flow was sent straight to Omaha for Buffett to deploy elsewhere or in Moody’s case, they used the cash flow to buy back their own stock.

      Those are great businesses because it doesn’t take any investment capital to raise prices, but the analysis has to center on how much pricing power is available. This a different discussion, but a business only has a finite amount of pricing power in real terms before the product or service is fully valued in the eyes of the customer. But as Buffett says, the best business is one that can soak up all of its cash flow that it generates and redeploy it back into the business at high rates of return. Imagine if See’s had the same pricing power, but in addition it had the ability to reinvest all of its cash flow into new locations nationwide that achieved similar returns on capital. The result would have been an even greater business (although Buffett did fairly well reinvesting See’s cash flow at decent rates himself) 🙂

  6. The ideas presented here are very similar to the 1996 paper “Increasing Returns” which predicted a shift towards situations where a single winner is able to dominate entire industries by virtue of increasing returns to scale.

    I think a major issue with the investment methodology presented here is that in the early stage there are dozens if not hundreds of competitors vying for control of the same market. By the time the dust has settled, the winner is already clear and every other company is worth nothing.

    Uber is probably the most visible current business which benefits from network effects, but its valuation is already stratospheric and they haven’t even gone public.

  7. Useful article, is there a particular metric that might help in identifying these all elusive reinvestment moat businesses?

  8. In your analysis, looking at the second example (36x multiple for Reinvestment Corp and 10x for Typical Corp), if you modify just one line item – change 50% of Earnings Reinvested for Typical Corp to 100% of Earnings Reinvested, in other words, both Companies are reinvesting all their earnings, but one does it at 25% and the other at 10%, you will see that in 10 years both companies would generate the same IRR (around 10%) and would trade at the same multiple on original investment (about 2.6x). What this tells us is that when you paid 36X for your “Reinvestment Corp”, even though it reinvested all its earnings at a much higher rate (2.5x) than the “Typical Corp”, and even though “Typical Corp” compounded at a much lower rate, only 50% of its earnings – the returns were exactly the same because you initially paid too high of a price for your Reinvestment Corp. In reality, its much more difficult to find companies that compound at 25% for 10 years – free capital markets dictate that excess returns will ultimately be competed away. And that is why you may be far better off buying things cheaply – you have a much greater margin of safety on your 10% ROIC (in that its not likely to erode due to competition as much as the 25% ROIC)… If you modify your analysis of the “Reinvestment Corp” to change ROIC from 25% to say 15% mid-life (in 5 years) – your returns will be worse than if you just bought a cheap “Typical Corp”. Bottom line is that when you find these “reinvestment Corps” in real life – you’ve to be pretty sure that ROIC will stay at 25% for the 10 years, and that is a much more difficult thing to predict than just buying cheaply…..

    1. Well I don’t think Connor is necessarily advising paying 36x earnings for companies, but I think he’s saying that even at that earnings multiple, a business that can compound at high rates is often fairly (or sometimes even undervalued) at that level. Google was hugely undervalued at 80 times earnings in 2005 for example. Obviously, that is a cherry-picked example and not necessarily something that most of us could replicate. But the point I took is that identifying a great business that can compound earning power is very valuable–much more valuable than a business that produces relatively low returns on capital.

      Also, while a business that does 10% ROIC is probably only worth around 10 times earnings or so, it’s unlikely that the earnings multiple of the business that just did 25% ROIC for 10 years would ever trade as low as 10x. So I think the math is fairly conservative in the example.

      But I would caution on getting too tied up in the specific math of the spreadsheet. Just focus on the broader concepts that he laid out in the post. That’s my 2 cents anyhow.

      1. Good point John. The notion that you can pay high multiples on some of these compounders and still come out ahead is really important and hard to grasp. Its a concept that took a long time for me to get comfortable with, and I still struggle at times. You mentioned Google. Personally, I knew Google well enough at the time to know it would be a compounder. Just did not feel comfortable paying 70-80x earnings multiple only to learn 5-6 years later. I saw myself having the same bias again, this time with Amazon.

        Great article! Have read it twice already and need to keep coming back to it to make it a habit. This is the point Buffett makes often with his quote, “Much better to make a great business at a good business than a good business at a cheap price”

      2. John – you entirely missed the point of my comment. In the example you use, its is ASSUMED that the company can reinvest predictably and reliably at 25% for the foreseeable future. Historical reinvestment rates are not indicators of future reinvestment rates. In fact, capitalism usually takes care of that – if a company in the past earned a return on capital far in excess of its cost of capital, competition will likely depress that spread. Most comments to this article hail the revelation that high returns on capital should be rewarded with higher multiple (if that’s not obvious to an investor – they really should put pencils down on stock picking an buy an ETF). But most comments entirely miss the point that its incredibly difficult to predict which businesses will have their moat eroded in 5 years and as a result, see ROICs collapse to cost of capital. And that is the reason why most people pay high P./Es for what seems at the time like high ROICs, end up losing money.

        1. Sam, I agree that it is hard to find great businesses like this. I don’t think anyone is suggesting it’s easy. The suggestion is that if you find one (obviously you have to be accurate in your analysis) that it is worth paying up for. If you’re wrong in your analysis and the moat gets narrowed (and return on capital shrinks), then you’ll end up with a mediocre result. The idea is to find businesses that have durability and a runway to reinvest earnings. Obviously, they are out there, but I completely agree that it isn’t easy (if it was, we’d all be doing it) 🙂

    2. Good point! That said, I would add a couple of points: a) 36x is really used to drive home a point, b) Typical Corp likely cannot afford the luxury of reinvesting 100% of its retained earnings at 10% return. They would have to pay a dividend or the multiples might shrink..

  9. How do you guys calculate the amount of money that was reinvested back into the business from the retained earnings???

    1. I would like to hear Connor’s and/or John’s thoughts on this. Do you use a concrete formula and hard numbers to try and calculate a precise number or do you look at a company’s history and find more general trends/numbers and then think abstractly to factor in competitive threats, moats etc?

      i.e Knowing what Buffet wrote in the 1993 Chairman’s Letter, “it is better to be approximately right or precisely wrong”, do you follow a less hard and fast approach and if so, what are your collective thoughts on this?



  10. Thanks for all the feedback. Sorry I’m late to the discussion here, as John knows I try to limit computer time to just a few instances a week. Fortunately John has summed up almost all of my responses eloquently so I really have “nothing to add” as Munger would say. I agree these companies are exceptionally hard to find, but all it takes is one or two over a career if you have a concentrated portfolio. For example a large part of Ted Weschler’s success is tied to his decade-long holdings in Davita and WR Grace which were both highly concentrated investments. I currently own 7 stocks, I’m of course hoping for 7/7 but realistically I only need to be totally right on a couple as long as the others perform ok. As far as the return calculations, John is correct that they are intended to be more illustrative in nature. I certainly advocate a margin of safety, personally I want an investment to be a “no brainer” rather than barely hitting my IRR hurdle if everything goes perfect. What I was hoping to show is the power of compounding and that if you bring a differentiated time horizon to your valuation work you can sometimes find companies that are “cheap” even if their current multiple of earnings would not show up on a value screen. – Connor

  11. That was amazing. Wow. This just woke me up. It really shows why picking low P/E stocks would not really get us any favorable returns over time. What’s really important is how well companies are managing/reinvesting its returns. Of course it is easier for any companies to just distribute cash through dividends or buy back stocks but if any companies can reinvest cash flows to generate more more cash flows then that’s way better stocks to buy. Thanks for sharing!


    1. One quick and dirty way is to look at the amount of capital the business has added over a period of time, and compare that to the amount of incremental growth of earnings. Last year Walmart earned $14.7 billion of net income on roughly $125 billion debt and equity capital, or just under 12% return on capital. Not bad, but what we really want to know if we are going to buy Walmart is a) how much of their earnings will they retain and reinvest in the business going forward? and b) what will the return on that reinvested capital be?

      10 years ago in fiscal 2006, Walmart earned $11.2 billion on roughly $83 billion of capital, or around 13.5%. But in the subsequent 10 years, they invested roughly $42 billion of additional debt and equity capital ($125b invested in 2016 and $83b invested in 2006), and using that incremental $42 billion they were able to grow earnings by about $3.5 billion (earnings grew from $11.2 billion in 2006 to around $14.7 billion in 2016). So in the past 10 years, Walmart has seen a rather mediocre return on the capital that it has invested during that time (roughly 8%).

      We can also look at the last 10 years and see that Walmart has retained roughly 35% of its earnings to reinvest back in the business (the balance has been primarily used for buybacks and dividends). As I’ve mentioned before, a company will see its intrinsic value will compound at a rate that roughly equals the product of its ROIC and its reinvestment rate. So if Walmart can retain 35% of its capital and reinvest that capital at an 8% return, we’d expect a modest growth of intrinsic value of around 3% per year. Stockholders will see total returns higher than that because of dividends, but the value of the enterprise will likely compound at roughly that rate. And we can see that over the previous 10 years, Walmart’s stock has grown around 45% not including dividends. So unless you are banking on an increase in P/E ratios, you’re unlikely to achieve a great result buying a business that can only invest a third of its earnings at 8% returns.

      This is a really rough measure, and this back of the envelope method works okay with a large, mature company like Walmart. But what you really want to know is what will the business retain going forward and what will the return be on the capital it retains and reinvests? Of course, there are different ways to measure returns (you might use operating income, net income, free cash flow, etc…) and there are many ways to measure the capital that is employed. But hopefully this is a helpful example from a general point of view.

      1. That’s a really useful explanation, thanks John. Are there any valuation/accounting books that have been useful for you for this way of viewing the value of an enterprise?

  12. Terrific article. I think success can be found in both the cheap stocks approach and the Munger approach. However, measured over a long (10yr+) holding period, the quality business approach is the clear winner. But that is an unfair comparison. Schloss and others advocate a max 3 yr holding period for deep value/bargain stocks. So in comparing these two approaches, assuming identical turnover is somewhat meaningless.

    Additionally, while over a long holding period one’s returns on an investment in a business with a high return on incremental invented capital will approach that rate of return, somewhat irrespective of purchase price, one can also achieve a terrific return in the legacy moat arena by purchasing those businesses when they are selling close to book value and management is committed to returning capital to shareholders. Those opportunities are rare but they do occur with some regularity among microcaps. This may seem like a totally obvious comment but it’s worth reminding people that there are many ways to skin a cat.

  13. Hello John,

    I had a doubt on the workings posted above for undervalued corporation. While I figured out that the earnings at the end of 10 years is 163$, I am unable to figure out how the dividends come up to 629$. I have done workings and find that the dividend should be about $71. I am unable to attach the workings here. However the way I have approached it as taking 100$ retained earnings at year 0, increasing earnings by 10% i.e 10$, dividing it equally among dividends and retained earnings, this exercise is done for 10 years and dividends work out 71$. Please let me know what I am missing. Also wanted to know how the multiple at original investment is arrived at.

  14. I appreciate this writting. Great job!
    I would like to underline an idea concerning Wal-Mart and Amazon. They both have an enormous bargaining power over their suppliers which allows to the former to have a negative working capital and to the latter to have a negative cash conversion cycle. The negative CCC is, of course, a huge cash-generating advantage of Amazon,as to the Wal-Marts the situation differs a little: the cash conversion cycle is positive, but the working capital is negative. Moreover, they are structurely in this position.
    That is, the both are being financed by their suppliers. Although every single supplier is paid in time, all the cohort of suppliers represents a resource of capital that may persist through a plethora of operating cycles. It may sound as a bad idea to finance an investment cycle by court-term resource, but that’s their natural advantage.
    So my idea is that it needs to be reflected, by somehow, in computation of return on (incremental) invested capital.

  15. This was really interesting and helpful.

    I’m wondering about using the “excess” capital in your example for buybacks rather than dividends. It seems like that would generate a better IRR. My rough math is ~15% (based on 10% compounding from initial earnings yield (assuming your repurchase multiple stays at 10x; even if you linearly ramp it up to 15x repurchase price, I don’t think that has a huge impact on the IRR) + 5% compounding from incremental earnings + some funny/confusing cross-compounding. One additional advantage to buybacks is that you don’t need to deal with the “redeployment” of your dividends.

    Am I missing something there?

  16. Although on a conceptual level I find return on incremental invested capital sound and worthwhile, the challenge is in finding the companies that will retain their dominance and returns for very long time. We are in a an era where technological or business model disruption is causing moats to shorten. Predicting beyond 10 years has become impossible. One good example is challenge posed by Patanjali who is an Indian herbal based FMCG products like soap, shampoo, toothpaste to global FMCG MNCs like Colgate, Lever, P&G etc operating in India. In just 3-4 years the global MNCs find their margins shrunk and sales dwindling in India. What worked in 20th century for Buffett and others may not work for us in 21st century going forward if you intend to find companies that will thrive and retain their MOATS for next 20-30 years.

  17. Pingback: My Blog
  18. As with any argument, Connor’s would benefit from examining the strongest counterarguments.

    Go back to 1972 and see if there were any companies with the ‘reinvestment moat’ profile of Walmart. If there weren’t, what were the top 5 candidates? What happened to each of them? Among them will be your strongest counterarguments.

Leave a Reply

Your email address will not be published. Required fields are marked *