91: Going Deeper on John Huber's 3 Engines of Value

This is an excellent framework to think about investment returns. Doing the math also helps you understand what's going on beneath the surface.

You’re reading the weekly free version of Watchlist Investing. If you’re not already subscribed, click here to join 1,850 others.

Want more in-depth and focused analysis on good businesses? Check out some sample issues of Watchlist Investing Deep Dives.

For less than $17/month, you can join corporate executives, professional money managers, and students of value investing receiving 10-12 issues per year. In addition, you’ll gain access to the archives, now 18 issues and growing!

---

The 3 Engines of Value

John Huber of Saber Capital Management recently put out a new memo and interview about the source of stock market returns. The interview hasn’t been made public but I’m sure if you emailed John he’d send it to you. In the meantime, check out his memo on incremental returns on capital and others on his site. When John puts out a piece of content I jump to read it as I always learn something.

Such was the case this time. John breaks down the return of any stock into three main components:

1. Earnings growth (ROIC x retention ratio)

2. Change in price/earnings ratio

3. Change in shares outstanding

Among a few examples John cites is Copart (shameless plug: CPRT was covered here and as a Deep Dive for paid Watchlist Investing subscribers). John notes that over a 10 year period CPRT grew earnings 5.6x (19% per year), its P/E expanded from 20 to 28 or 1.4x, and it’s shares declined by 21% contributing 1.3x. Multiply 5.6 x 1.4 x 1.3 and you get a 10x return over ten years or 26% per year.

This type of analysis really appeals to me as it’s simple and intuitive. It shows you exactly where your returns are coming from. It also highlights the fact that things other than the business itself can act as tailwinds or headwinds, even though over time returns follow very closely to underlying business results.

For example, John cites a hypothetical example of a company that grows earnings 20% per year but sees its P/E fall and its share count rise because of dilution. That 20% is cut to a meager 6.4% return because two of the three engines were running in reverse. The lesson here: beware of overpaying for companies and don’t forget that exit multiples matter. Just because it’s worth paying up for something now doesn’t mean the market will continue to reward the company with the same P/E in the future. More likely it’ll revert to a market P/E.

What about the 4th engine?

What about dividends? To John’s credit he does include dividends when discussing Home Depot. I think it should probably be included as a 4th engine of value. Or perhaps included in the third engine under the heading of capital returned to shareholders.

What about value-destroying share repurchases?

This leads me to another question. This model simply states that the change in shares outstanding is a component of returns. It says nothing about valuation. I emailed John this question and he admitted that there wasn’t a neat solution to it. Purchases above intrinsic value still destroy value, but the math is such that it factors into returns.

Going deeper on the math…

I wanted to understand what was going on so I built an Excel model. My tendency is to overcomplicate things and my model soon got unweildy. So I turned pencil to paper and simplified using a one year model vs trying to model a decade in Excel. Here are some takeaways:

1. I assumed a company with capital of $1,000 earning a 15% ROIC paying out 50% of earnings as a dividend. I also assumed I paid a P/E of 10x.

2. I held the P/E constant at 10x.

3. Under these assumptions earnings growth is 15% x 50% = 7.5%.

4. The dividend return was 5%. Why? Because I paid 10x earnings for this company (the 10 P/E) that meant I paid $1,500. 50% of the $150 in earnings is $75. Divide that into $1,500 and you get 5%.

5. Another way to look at the dividend return is to take the payout of 7.5% of capital and divide by the premium to capital. In this case I paid $1,500 for a company with underlying capital of $1,000 or 1.50x. Take 7.5% divided by 1.5 and you get 5%. The math checks out.

6. Another check: Earnings grew 7.5% to $161.25. Capitalize that at our 10 P/E and you get $1,612.50. Add $75 for dividends and you get $1,687.50. Divide that by the $1,500 purchase price and you get the same 12.5%. Beautiful stuff.

How about buybacks?

Next I wanted to change dividends into buybacks but keep everything else constant. I assumed the same basics for the company: $1,000 in capital earning 15% and retaining 50%. I also assumed there were 1,000 shares outstanding.

1. Same math as above for earnings growth: $1,000 x 15% = $150 x 50% = $75 or 7.5%

2. Price/earnings stays at 10x so no contribution from that area

3. Assuming 1,000 shares outstanding that means each share earned $150 / 1,000 = $0.15. So each share is valued at $1.50 at a 10 P/E. I have the remaining 50% of earnings left or $75. Divide that $75 by $1.50 per share and you would repurchase 50 shares and have 950 shares remaining.

4. The contribution from the change in shares is 50 / 1000 or 5%, the same as dividends above.

We can pause here for a minute and make a few observations. One is that valuation matters. In the dividend example the fact that we paid a premium to underlying capital meant that the dividend component of return was reduced by the premium (in our case 1.50x, bringing 7.5% down to 5%). When it comes to buybacks the same dynamics are at play. The company is reducing shares but it’s paying a premium to underlying capital too, which means the return is cut down. If the price/earnings we pay matches what the company pays for buybacks then the capital return is the same in both cases.

Unlocking the mysteries of buybacks

This leads us to a place of understanding about capital returns and how buybacks affect shareholder returns. No matter what price is paid for buybacks there will be a return. If we paid 50x earnings in the example above we’d have paid $0.15 x 50 = $7.50 per share. That means we’d only have repurchased $75 / $7.5 = ten shares. 10 / 1000 = 1%. Our total return would have been the 7.5% earnings growth + 0% from a change in P/E (this assumes the company bought back the shares expensively but the market remained the same) + 1% from buybacks or 8.5%.

Opportunity cost

Clearly these buybacks destroyed value, right? I think the test should be the opportunity cost of that capital. If it could have been reinvested internally instead of paid out then the opportunity cost is 7.5%. If all earnings could have been retained then earnings would have grown by 15%. If 50% was all the company could retain then dividends probably would have made the most sense.

Another way to look at it is the increase in earnings per share. Under the 50x P/E buyback scenario above EPS rises from $0.15 to $0.1629 or 8.59% [Edit on 4/18/24: This is corrected from the original version which stated the increase was 8.75%]. That’s more than the 7.5% under the first scenario but in this case, the investor doesn’t get a dividend. She could have gotten another 5% through dividends. So value was destroyed by buying back shares. I think this is the sneaky way companies present buybacks to shareholders. They only present the gross value “returned” to shareholders or highlight the increase in EPS but not the alternative uses of capital. This simple math shows that capital wasn’t returned to shareholders, it was destroyed.

Clearly, management teams aren’t omniscient. If share repurchases are made at a P/E higher than the market some people will say they should have paid out the capital in dividends. To me, it’s more important that they get it roughly right. A management paying 50x earnings for buybacks is almost always going to be destroying value. But the difference between 10x or 12x or 15x is probably too small to quibble over unless there’s evidence to the contrary. In the end I think you have to make your own decision as to opportunity cost. If 10% is your discount rate then purchases of stock by a company that give you this yield are going to add value.

The Greenwald Put

There’s a way to ensure that buybacks are neutral to your returns. In his book Value Investing: From Graham to Buffett and Beyond, 2nd Ed., Greenwald makes the point that you can sell a proportionate number of your shares as the size of the company’s overall repurchase program and negate any value destruction (or accretion) through that action. If buybacks are being made at a value-destroying price at least you’re getting paid a premium for the shares you’re selling. Most people probably aren’t going to do this but it’s worth considering.

Summing up / simplifying

The beauty of looking at the engines of value is helping you understand where returns come from. Earnings growth is a simple function of returns times the amount of capital retained. You want to then think about how ROIC might change over time. You might set earnings growth to a steady state ROIC x retain ratio if that makes sense. For P/E it probably makes sense to assume reversion to the mean. Pick your period of time, whether it’s five years or a decade, and amortize that change in P/E into your return. It’s almost always going to be a headwind if you’re buying a company at a premium. Going from 20x to 10x over ten years is a 6.7% annual headwind. Then thinking of capital returns, either from dividends or share buybacks, in relation to opportunity cost of capital will help determine the attractiveness of that capital allocation decision.

Investing is simple but it’s not easy.

Stay rational! —Adam

---