Calculating the Return on Incremental Capital Investments

Posted on Posted in Education, Investment Philosophy

“Leaving the question of price aside, the best business to own is one that over an extended period can employ large amounts of incremental capital at very high rates of return.”Warren Buffett, 1992 Shareholder Letter

I received a lot of feedback, comments and a few questions after Connor Leonard’s guest post last week. Connor’s write-up was very well articulated, and deservedly received much praise. There were a few questions which we tried to address in comments and email responses, but one comment that came up a few times was: how do you calculate the return on reinvested capital?

I think a lot of people were looking for a specific formula that they could easily calculate. I think the concept that Connor laid out is what is most important. I think his examples did a nice job of demonstrating the power of the math behind the concept. His write-up also demonstrated why Buffett said those words in the 1992 shareholder letter that I quoted at the top of the post.

So the concept here is what needs to be seared in. In terms of calculating the returns on capital, each business is different and has unique capex requirements, cash flow generating ability, reinvestment opportunities, advertising costs, R&D requirements, etc… Different business models should be analyzed individually. But the main objective is this: identify a business that has ample opportunities to reinvest capital at a high rate of return going forward. This is the so-called compounding machine. It is a rare bird, but worth searching for.

I amalgamated a few of my responses to comments and emails into some thoughts using an example or two. These are just general ways to think about the concept that Connor laid out in the previous post.

Return on “Incremental” Invested Capital

It’s simple to calculate ROIC: some use earnings (or some measure of bottom line cash flow) divided by total debt and equity. Some, like Joel Greenblatt, want to know how much tangible capital a business uses, so they define ROIC as earnings (or sometimes pretax earnings before interest payments) divided by the working capital plus net fixed assets (which is basically the same as adding the debt and equity and subtracting out goodwill and intangible assets). Usually we’ll want to subtract excess cash from the capital calculation as well, as we want to know how much capital a business actually needs to finance its operations.

But however we precisely measure ROIC, it usually only tells us the rate of return the company is generating on capital that has already been invested (sometimes many years ago). Obviously, a company that produces high returns on capital is a good business, but what we want to know is how much money the company can generate going forward on future capital investments. The first step in determining this is to look at the rate of return the company has generated on incremental investments recently.

One very rough, back-of-the-envelope way to think about the return on incremental capital investments is to look at the amount of capital the business has added over a period of time, and compare that to the amount of the incremental growth of earnings. Last year Walmart earned $14.7 billion of net income on roughly $111 billion debt and equity capital, or about a 13% return on capital. Not bad, but what do we really want to know if we were thinking about investing in Walmart?

Let’s imagine we were looking at Walmart as a possible investment 10 years ago. At that point in time, we would have wanted to make three general conclusions (leaving valuation aside for a moment):

  • How much cash Walmart would produce going forward?
  • How much of it would we see in the form of dividends or buybacks?
  • Of the portion we didn’t receive, what rate of return would the company get by keeping it and reinvesting it?

Our estimates to these questions would help us determine what Walmart’s future earning power would look like. So let’s look and see how Walmart did in the last 10 years.

In 2006, Walmart earned $11.2 billion on roughly $76 billion of capital, or around 15%.

In the subsequent 10 years, the company invested roughly $35 billion of additional debt and equity capital (Walmart’s total capital grew to $111 billion in 2016 from $76b in 2006).

Using that incremental $35 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 of about 10% on the capital that it has invested during that time.

Note: I’m simply defining total capital invested as the sum of the debt (including capitalized leases) and equity capital less the goodwill (so I’m estimating the tangible capital Walmart is operating with). I’ve previously referenced Joel Greenblatt’s book where he uses a different formula (working capital + net fixed assets), but this is just using a different route to arrive at the same basic figure for tangible capital.

There are numerous ways to calculate both the numerator and denominator in the ROIC calculation, but for now, stay out of the weeds and just focus on the concept: how much cash can be generated from a given amount of capital that is invested in the business? That’s really what any business owner would want to know before making any decision to spend money.

Reinvestment Rate

We can also look at the last 10 years and see that Walmart has reinvested roughly 23% of its earnings back in the business (the balance has been primarily used for buybacks and dividends). How do I arrive at this estimate? I simply used the amount of incremental capital that Walmart has invested over the past 10 years ($35 billion) and divide it by the total earnings that Walmart has generated during that period (about $148 billion of cumulative earnings).

An even quicker glance could simply be to look at the retained earnings on the balance sheet in 2016 ($90 billion) and compare it to the retained earnings Walmart had in 2006 ($49 billion), and arrive at a similar reinvestment rate: basically, Walmart is retaining roughly $0.25 of every $1 it earns and reinvesting it back into the business. The other $0.75 is being 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 (leaving aside capital allocation, which can increase or decrease value per share as well).

Intrinsic Value Compounding Rate = ROIC x Reinvestment Rate

There are other factors that can create higher earnings (pricing power is one big example), but this simple formula is helpful to keep in mind as a rough measure of a firm’s compounding ability.

So if Walmart can retain 25% of its capital and reinvest that capital at a 10% return, we’d expect the value of the company to grow at a rate of around 2.5% per year (10% x 25%). Stockholders will likely see higher per-share returns than that because of dividends and buybacks, but the total value of the enterprise will likely compound at roughly that rate. And over time, the change in value of the stock price tends to mirror the change in value of the enterprise plus any value added from capital allocation decisions.

Here is a table that summarizes the numbers I outlined above:


Not surprisingly, Walmart’s stock price is only about 45% higher (not including dividends) than it was 10 years ago. 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 quarter of its earnings at a 10% return.

Chipotle—A High Return on Capital Business

Let’s use the same principles to very briefly take a look at Chipotle, which is a business that has been able to reinvest its earnings at very high rates of returns over the past decade.

For Chipotle, we’ll look at the last 9 years of operating results starting at the end of 2006, which was the first year that Chipotle operated as a standalone company (McDonald’s spun it off in mid-2006).

Chipotle has great unit economics. In 2015, it cost $805,000 to build out the average restaurant which, when up and running, produces around $2.4 million in revenue and better than 25% restaurant-level margins. So an $800k initial investment produces around $600k of yearly cash flow. In other words, the average Chipotle restaurant has achieved an incredible 75% cash on cash return (this is restaurant-level ROI before corporate G&A expenses and taxes).

Over the past 9 years, Chipotle has grown from 581 restaurants to just over 2,000. They’ve invested a total of $1.25 billion to build out these restaurants which has increased its earnings by around $435 million. In other words, Chipotle saw around a 35% after-tax return on the capital it reinvested back into the business:


Chipotle was able to invest over half of its earnings at 35% returns, and using the formula described above to approximate intrinsic value growth, 57% reinvestment rate times 35% returns equals about a 20% increase in earning power.

The result is that the stock price has compounded at over 20% annually, very much in line with Chipotle’s intrinsic value.

Mature Businesses vs. Compounders as Investments

Of course, the hard part is finding these companies in advance. Also, we should remember that buying cheap and selling dear is a very good strategy, and there are plenty of opportunities in the market to do so with durable and established, but low-growth businesses. Peter Lynch called these companies stalwarts—they were big companies without a lot of growth potential, but occasionally you could buy them at a discount and sell them after a 30-50% rise (which largely comes from the valuation multiple increasing as opposed to the business value increasing).

That’s a good strategy, and there are probably more opportunities that Mr. Market offers in this particular category. But an investor should realize that these investments are not going to result in big compounding results. Lynch’s famous 10-baggers came from the rare companies that could retain their earnings and plow them back into the business at high rates of return for many years. The former is much more common and actionable, but it’s still worth hunting for stocks in the latter category in my opinion. You only need a few to make a career.

To Sum It Up

What you really want to know is this: At the end of the year, the company will have made a certain amount of money. Out of that pile of cash, you want to know:

  • How much can the company reinvest into the business?, and,
  • What the return will be on that investment?

If Walmart makes $16 billion, they might spend $4 billion on building out new stores. How much additional cash flow will come from that $4 billion investment?

Obviously, Walmart getting a 10% return on a quarter of its earnings is not nearly as attractive as Chipotle getting a 35% return on half of its earnings. The latter is going to create much more value than the former.

This is a really rough measure of how to think about return on capital, but it’s generally how I think about it.

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. There are also “investments” that don’t always get categorized as capital investments but run through the income statement—things like advertising expenses or research and development (R&D) costs. To be accurate, you’d want to know what portion of advertising is needed to maintain current earning power (akin to “maintenance capex”). The portion above that number would be similar to “growth capex”, which could be included when thinking about return on capital. R&D could be thought of the same way.

But hopefully this is a helpful example from a very general point of view on how to think about the concept. As a business owner, you want to know where you can reinvest your company’s excess cash flow and what rate of return you can get from those investments. The answers to those two main categories will in part determine how fast your business grows its earning power and increases its value.

John Huber is the portfolio manager of Saber Capital Management, LLC, an investment firm that manages separate accounts for clients. Saber employs a value investing strategy with a primary goal of patiently compounding capital for the long-term.

John also writes about investing at the blog Base Hit Investing, and can be reached at

35 thoughts on “Calculating the Return on Incremental Capital Investments

  1. Hi John, Thanks for the post. I attempted a similar calculation on IBM. This process doesn’t seem to work as cleanly as above…
    Net capital seems to have reduced and income is up some in the last 10 years.

    Goodwill Ret Earnings Debt Equity Net Capital
    12,854 52,432 13,780 28,506 29,432
    14,285 60,640 23,039 28,470 37,224
    18,226 70,353 22,689 13,465 17,928
    20,190 80,900 21,932 22,637 24,379
    25,136 92,532 21,846 23,046 19,756
    26,213 104,857 22,857 20,138 16,782
    29,247 117,641 24,088 18,860 13,701
    31,184 130,042 32,856 22,792 24,464
    30,556 137,793 35,073 11,868 16,385
    32,021 146,124 33,428 14,262 15,669

    Net Income Dividends
    9,492 -1,683
    10,418 -2,147
    12,334 -2,585
    13,425 -2,860
    14,833 -3,177
    15,855 -3,473
    16,604 -3,773
    16,483 -4,058
    12,022 -4,265
    13,190 -4,897
    12,876 -5,059
    147,532.00 (37,977.00)

    Buybacks total: (115,781.00)

    eps Shares
    6.108108108 1,554
    7.179875948 1,451
    8.924746744 1,382
    10.01118568 1,341
    11.52525253 1,287
    13.0601318 1,214
    14.37575758 1,155
    14.94378966 1,103
    11.9029703 1,010
    13.41810783 983

    1. At first glance, this appears to be a conundrum. But it is solved when you take into account two major factors – the excess cash and pension assets.

      The total cash + marketable securities was 13.7 bn in 2006, but only 8.2 bn today. We can guess that a significant amount of this is excess cash, and some of that excess was spent towards buybacks and dividends.

      The prepaid pension assets add up 20.6 bn in 2006, but only 1.7 bn today. This is due to the phase out of IBM’s pension plan post 2005.

      Also, don’t forget to take into account other intangibles than goodwill. The other intangibles were 1.7 bn in 2006, and 3.5 bn today. I think its best to exclude all intangibles for these purposes. If you want to include the value of intangibles, I would use the cumulative R+D expense (as intellectual property represents the majority of intangibles for a company like IBM) rather than the balance sheet value.

      If we look at invested capital ex-cash, ex-pensions, ex-intangibles, we get a negative figure in 2006 (-6542 mm) and a positive number (1957 mm) today. This indicates that about 8.5 billion have been invested to get the growth in earnings.

      One thing that’s interesting is that if you use the Greenblatt definition of IC (Net working capital – cash + Net PPE), this has barely budged over the 10 yrs, going from 10.6 bn in 2006 to 10.8 billion today. So by this definition of invested capital, there has barely been any additional investment to get the extra earnings. Perhaps this is what Buffett is referring to when he says that IBM has an “infinite return on capital”.

      Always a fan of these discussions.

  2. Net Income PP&E spend Acquisitions spend
    9,492 -4,362 -3,799
    10,418 -4,630 -699
    12,334 -4,171 -6,242
    13,425 -3,447 -794
    14,833 -4,185 -5,867
    15,855 -4,108 -1,797
    16,604 -4,082 -3,123
    16,483 -3,623 -2,759
    12,022 -3,740 1,701
    13,190 -3,579 -3,750
    12,876 -3,678 -6,164
    cumulative 147,532 -43,605 -33,293
    increase 3,384

    Increase in Net Income / Cumulative Acquisition & PPE Spend is only ~4.4%

    Why is IBM interesting to value investors like Buffett and Watsa?

    1. I haven’t done much work on IBM Tom, so I wouldn’t be able to help much. I’ve taken a pass on it since my general feeling is that the company is going to struggle to compete with AWS (Amazon’s cloud business that is growing fast and taking share from IBM). I have looked at the financials, and they do produce a lot of cash flow (Buffett has said he likes the cash flow and the company’s ability to buyback stock). But it’s certainly not a company that has much opportunity to reinvest that cash flow into the business (as the vast majority of their free cash flow has been used for either buybacks or acquisitions in the hope of defending its market share). I personally think the business has a lot of headwinds, but I don’t know it that well. Buffett (and others) could be right, but I’ve chosen to watch this one from the sidelines.

  3. John, thanks a lot for the follow up to Connor’s article! Nice to see the concrete examples and ever grateful for your articles.

  4. Hi John

    How do you arrived at the 147.5 Bn earnings figures. Is it sum of the retained earnings? Or is it the total FCFF?

  5. This was really a terrific adress to my question.I will keep looking at cash generating businesses that can expand their operations efficiently by earnings great returns on their incremental capital while at the same time they don’t invest in areas outside of their core operations. Thank
    you very much Mr.Huber!!

  6. thak you for this excellent post.
    I didn’t understand how you got to 671 in Goodwill/Excess Cash in 2015 ?
    goodwill is 22 and total cash is 663 , what is the calculation for Goodwill/Excess Cash ?


  7. Thanks for the great post. One comment and one question:

    One comment: since you are using net debt + shareholders’ equities – goodwill, you might want to use adj. after-tax EBIT (excluding goodwill amortization) instead of net income. My sense is this will boost ROIC somewhat but won’t change any conclusions.

    One question: now I know how fast the intrinsic value is compounding. Unfortunately, I don’t know how fast the stock (if invested today) will compound going forward. For example, Chipotle will compound intrinsic value 20% based on your calculation. If I buy stock today at $427/sh, how fast will this investment compound? CMG has a market cap of $12.5bn, or 7.0x of capital invested ($1.8bn in your calculation). Does that mean the compound rate will drop to 20%/7=~3% at current stock price level (since you are buying 7 times of invested capital)? I hardly believe that. Any thoughts?

  8. Thanks for the insightful article. I have a question on the working you proposed – lets say in 2006 X capital produced A return, and in 2016 Y capital produced B return. You are concluding that the incremental capital has produced the incremental returns i.e. Y-X capital produced B-A return. This means that the original X capital produced A return in 2016, same as 10 years before in 2006. Is this a reasonable assumption? Given that the assets which X was used for, would diminish over time (depreciation etc., maybe some even turned unproductive and were written off) isnt it reasonable that after 10 years X would produce less than A return? Which would mean that Y-X produced better than B-A return.

    Is this a reasonable assumption you are making, or does the math make it always true somehow and I am not seeing it?

    1. Hi Ramesh,

      Isn’t this already accounted for by the depreciation?

      The actual assets that existed in 2006 may lose earning power, but the above calculations assume that the depreciation accounting entity takes care of that. Then additional maintenance capex is required to overcome the effects of depreciation.

      So by 2016, the amount X actually represents all the 2006 assets, plus all the maintenance capex that has been poured into these assets for years. These maintenance capital expenditures then allow that X amount of assets to still produce A return in 2016.

      The growth capex is what takes X to Y by 2016. So by subtracting Y-X we are isolating the total amount of growth capex that was spent over those years.

      Let me know if this makes sense, or if I’m way off track.

  9. Good post. I think you have written about this topic before. Here is my question, using your method, we will always end up with very low growth projections for huge companies like JNJ, WMT, PEP etc. However, what I have observed is, if you buy these companies when they are beaten down due to company specific temporary issues (I bough JNJ when they temporary issues couple of years early and WMT late last year around 58-65) and hold it for long periods of time say 10-15 years you invariably get a return between 11-15% annualized. I use XIRR to calculate returns that includes dividend and capital appreciation. Yes this is not astronomical but if you can replicate this for a bunch of stocks, you have what I call ‘stocks which are like bonds’ reliable stream of income for retirement. What are your thoughts on this? My specific question is even if WMT has a growth projection of 3-4% (ROIIC * Reinvestment rate), it is likely to give you 12% returns if you buy it at the right time and sit tight. Why is that?

    1. Hi Pop,
      By buying WMT or JNJ when they are temporarily beaten down, you are betting on positive corrections over the long run. As Graham, says in the long run, Market is a weighing machine. Nothing wrong with that.

      But I think the point of this article was to show that betting on compounding machines over long periods of time is likely to be more rewarding than betting on mean reversion.

      Just my 2 cents.

  10. The only issue is that a long history of great return on incremental capital will attract attention and keep prices elevated and the company will usually be quite big when you find it. Conversely, a great compounder may be a spinoff from a large parent or have little operating history just at a time when its returns on incremental capital are just being discovered and high – but you may only have a short history, maybe even a year or two. I like the idea of having a mix of both, and also trying to work from qualitative factors if a company has a low or high return on capital going forward when there is not enough history or whether the value will rise or fall – none of those can be captured really in history for the faster, small compounder.

  11. John,

    I loved this post. ROIIC (Return on incremental invested capital) is a great metric to analyze a company. Your explanation is spot on and I believe the two examples were great ways to exhibit the difference between a high return on capital asset vs. one that is lesser so. I’ve dug in to Walmart in the past (as many often do) and believe that it is currently undervalued, but the opportunity ahead for the company doesn’t warrant strong consideration (and I believe your analysis supports that), albeit 2.5% intrinsic value return + ~2.75% dividend yield isn’t that shabby. In an environment of high uncertainty it wouldn’t be a bad place to invest capital. Again, I really enjoyed your post and have enjoyed many others in the past! Thanks for the great blog.

    Chris @

  12. Hi John – thank you for the post, very useful. Wouldn’t it make more sense to use EBIT or EBITDA instead of net income? since we are including capital provided by debt + equity?

  13. John

    Thanks. Two questions:

    1) How should you think about cumulative buybacks? Gross or net buybacks?

    2) What about companies where the reinvestment rate is greater than 100% — i.e., they have been issuing (common or preferred) shares or raising significant sums of debt?


  14. I struggle with and debate myself over and over and over about deducting goodwill from the denominator b/c it inflates returns when acquisitions are made and – arguably – rewards companies for over-paying on deals.

    Imagine a company borrows $100 to buy a company and 90% of that acquisition price goes into goodwill. Now you’ve got the incremental EBIT from the acq inflating the numerator but only $10 in incremental invested capital in the denominator.

    I own one such company that’s been investing capital in acquisitions and since goodwill is such a large component of the acqs, how I calculate the return can differ; from +30% excluding goodwill and +10% including it.

    I try not to get lost in the woods with this and so I really just focus on the tremendous cash flow they’re generating but everything is relative in this business. It’s critical when comparing investment ideas to be consistent and fair.

  15. Hi John,

    Thanks for the fantastic and insightful post.

    I was thinking if there was a way to calculate total shareholder return. It seems to me that any earnings that are not reinvested ought to go towards buybacks and dividends (assuming no major accumulation of excess cash). Then if you take (1-reinvestment rate)(base year earnings)/(base year market cap), you ought to get to an approximation of the additional return from buybacks and dividends.

    For example, if Walmart reinvests 25% of earnings, then 75% goes towards buybacks and dividends. So (.75)(11.2)/(192.6 bn) = 4.4% additional return from buybacks and dividends. So the total shareholder return ought to approximate the growth in intrinsic value, 2.4%, plus the additional return of capital through buybacks and dividends, 4.4%, for a total of 6.8%.

    I think if we negate the effects of p/e expansion, then the total shareholder return of Walmart over the period would probably be roughly equal to this number.

    Does this make sense? Curious for your thoughts.

  16. Hi John,

    Out of many excellent posts on your site, this is one of my favorites. I’ve incorporated a lot of this thinking into some of my valuations recently, and was wondering how you account for companies with large cash balances.

    While a part of these balances are surely used for regular operations, most of it is just sitting overseas waiting for a tax holiday and not doing all that much. While hard to estimate what portion of the cash is actually used for the business, I wanted to know if you reduce the total capital invested by some estimated amount?

    For example, when looking at Apple, the “value compounding rate” is actually on the low end (~7.75% 5/yr) if we consider total capital invested = equity + debt – goodwill. (Here’s a quick model:

    But if we assume only 30% of the cash/short term securities balance are for operating purposes, this number shoots up to ~32.75% over the last 5 years ( Quite a big difference! And if we just look at most recent 3 years, Apple’s compounding rate is actually negative because of the negative total capital invested. This makes sense, and is actually a good thing for shareholders since it shows Apple is able to grow profits without having to take on more capital to run the business.

    I’m just curious how you think about situations like this. Obviously Apple is an exception because of their immense cash pile, huge profitably, and relatively low capital requirements due to outsourcing of production, but would love to hear any additional thoughts you might have.

    Slow Appreciation

  17. John,
    Thanks for this excellent post!

    Had a question. This method of looking for incremental returns over incremental capital- would this work for companies that have been listed for less that 5 years? For example, a company that I am analyzing is using the cash from IPO to reduce debt and invest in growth capex. So the quality of balance sheet is going to improve and in the future the earnings from growth capex are going to show up. But if I look for incremental returns on incremental capital TODAY, it will misguide me, because the capital has bloated (because of the IPO) and the returns from the capex are yet to show up.

    Am I looking at it correctly?

  18. Can you speak a little to how you think about companies with negative reinvestment rates (i.e., buybacks/dividends > retained earnings), but which are still able to increase earnings and invest in new opportunities?

    Some of these companies can still compound capital despite the “negative” rate of investment, and further expand their moats.

  19. This might be a really dumb question, but where did the un-reinvested portion of Chipotle’s earnings go? The business pays no dividend and, to my knowledge, does not buy back stock.

  20. Hi John, Thanks for this great post.

    Question, in the 1st example on WMT you add Equity and Debt & Capital Lease obligations and in the 2nd one on Chipotle you add Equity and deferred obligations. How do you decide what to add?


Leave a Reply

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