Solid read - I would add a couple of factors (as discussed in my December '25 e-book):
- impact of inflation vs. assets at historical cost; pricing power can significantly distort assumptions on future investment needs
- utilization rate vs. capacity (reflecting all assets and liabilities on the balance sheet); some companies will have leftover capacity post-COVID and will see their ROIIC increase if you're adjusting the numbers on a like-for-like basis
- not really a big fan of the year-over-year ROIICs because it can be distorted by timing of working capital management (could be payment timing during the holidays)
- not always easy to break down the profit growth coming from new investments or just optimizing the base business
Thank you for the comment and for taking the time to spell these points out.
On inflation/historical cost and pricing power, I’m fully with you. In this piece we deliberately stuck to a “clean accounting” approach and didn’t go down the route of replacement-cost adjustments or a full real-terms analysis. The idea was mainly to show how the framework ROIC/ROIIC can be used with a concrete example (not to build a perfect economic model). But over time, historical cost + strong pricing power can clearly make future investment needs look lighter than they really are.
Same story for utilization vs installed capacity: in some industries you can see ROIIC climb for years just because you’re filling assets that were already in place. In the Meta example we largely treated that as background noise, simply because we didn’t have the operational detail to adjust for it properly (as is often the case, which is exactly why investing is a probabilistic game).
For YoY ROIIC, Ithe signal can get messy, as we highlighted (working capital, cash-flow timing, and all the other moving parts in invested capital). That’s one of the reasons we also focus on cumulative ROIIC over 2020/2024 instead of reading too much into a single year.
You’re also right to highlight the split between profit growth driven by new vs. old investment. From the outside, it’s very hard to do this "cleanly". I struggle to think of many companies where management gives enough disclosure to do it with sufficient precision. In the article, we lived with that and used Meta mainly as a way to put the framework to work, rather than as a full deep dive on the company.
The appeal of the framework, at least for me, is that even without insider data, and with a few reasonable shortcuts, it already helps frame the rough order of magnitude of value creation and gives a structured way to think about capital allocation. Of course, it’s not a scalpel, but it does force us to ask the right questions (especially when the business model is shifting or a new CapEx cycle is starting [like Meta].)
Really appreciate you pushing the conversation in this direction.
100% - the bottomline is that we shouldn't be tripping over a 10 vs. 13% ROIIC after adjusting for certain items; such ROIICs are rarely attractive. Requires a high reinvestment rate over the whole cycle, and everything's cyclical; so it requires other drivers to generate a satisfying return on our investment like margin expansion (some companies' executives overestimate their ability to drive profit margins for a long time). Conversely, if you find something with 30%, it could be 25% or maybe 40% on an adjusted basis because you didn't factor in the upside from un-utilized capacity. That's not marginally above the cost of capital ;-)
Always learning, every day. Think we can all agree that ROIIC/ROIC isn't about the best precision but understanding the direction the situation is moving, and how the managers are implementing it (hopefully in a splendid manner).
ROIC assesses total investments (old + new investments) whereas ROIIC new investment only which could be small in scale compared to the old investment.
Please explain why in detail.
What if ROIIC, (1+Gnp)/(1+Gic) & ROIC are assessed at the same time?
They’re both just measurement tools. Neither is “better” in absolute terms. It really depends on the objective and the context.
As we wrote:
“ROIIC fixes the ‘average vs incremental’ issue [of ROIC], but at the cost of worsening the ‘accounting vs economic’ issue.”
Looking at the two together, within a single framework, is exactly what we’re trying to do: reduce the averaging problem you get with ROIC alone, while also limiting the impact of accounting or operational changes that aren’t economically relevant for the analysis.
ROIIC also isn’t limited to “just the new investments”. You’re free to choose an older/longer time window. In our case we used 2020-2024, but we could just as well have chosen 2012-2018. What matters is to look at ROIIC over time blocks that make economic sense for the business and, where possible, make the necessary adjustments to reduce the noise in the data you’re measuring.
To finish, you’re suggesting adding the NOPAT vs IC growth rate into the framework. I haven’t studied it enough to have a strong view, but my feeling is that the information it brings is quite overlapping with the ROIC/ROIIC combination. I may be wrong. In any case, like any model, it’s a trade-off between simplicity, usefulness and accuracy. Depending on what you want to prioritise, including it (just like using ROIC or ROIIC on their own) can be relevant or not. It really depends on the context and how you intend to use it.
Solid read - I would add a couple of factors (as discussed in my December '25 e-book):
- impact of inflation vs. assets at historical cost; pricing power can significantly distort assumptions on future investment needs
- utilization rate vs. capacity (reflecting all assets and liabilities on the balance sheet); some companies will have leftover capacity post-COVID and will see their ROIIC increase if you're adjusting the numbers on a like-for-like basis
- not really a big fan of the year-over-year ROIICs because it can be distorted by timing of working capital management (could be payment timing during the holidays)
- not always easy to break down the profit growth coming from new investments or just optimizing the base business
It remains a very nuanced subject :))
Thank you for the comment and for taking the time to spell these points out.
On inflation/historical cost and pricing power, I’m fully with you. In this piece we deliberately stuck to a “clean accounting” approach and didn’t go down the route of replacement-cost adjustments or a full real-terms analysis. The idea was mainly to show how the framework ROIC/ROIIC can be used with a concrete example (not to build a perfect economic model). But over time, historical cost + strong pricing power can clearly make future investment needs look lighter than they really are.
Same story for utilization vs installed capacity: in some industries you can see ROIIC climb for years just because you’re filling assets that were already in place. In the Meta example we largely treated that as background noise, simply because we didn’t have the operational detail to adjust for it properly (as is often the case, which is exactly why investing is a probabilistic game).
For YoY ROIIC, Ithe signal can get messy, as we highlighted (working capital, cash-flow timing, and all the other moving parts in invested capital). That’s one of the reasons we also focus on cumulative ROIIC over 2020/2024 instead of reading too much into a single year.
You’re also right to highlight the split between profit growth driven by new vs. old investment. From the outside, it’s very hard to do this "cleanly". I struggle to think of many companies where management gives enough disclosure to do it with sufficient precision. In the article, we lived with that and used Meta mainly as a way to put the framework to work, rather than as a full deep dive on the company.
The appeal of the framework, at least for me, is that even without insider data, and with a few reasonable shortcuts, it already helps frame the rough order of magnitude of value creation and gives a structured way to think about capital allocation. Of course, it’s not a scalpel, but it does force us to ask the right questions (especially when the business model is shifting or a new CapEx cycle is starting [like Meta].)
Really appreciate you pushing the conversation in this direction.
100% - the bottomline is that we shouldn't be tripping over a 10 vs. 13% ROIIC after adjusting for certain items; such ROIICs are rarely attractive. Requires a high reinvestment rate over the whole cycle, and everything's cyclical; so it requires other drivers to generate a satisfying return on our investment like margin expansion (some companies' executives overestimate their ability to drive profit margins for a long time). Conversely, if you find something with 30%, it could be 25% or maybe 40% on an adjusted basis because you didn't factor in the upside from un-utilized capacity. That's not marginally above the cost of capital ;-)
Appreciate the discussion; very solid points.
I'd like to quote Munger "I have nothing to add" although in reality it's more like "I have more to learn".
Always learning, every day. Think we can all agree that ROIIC/ROIC isn't about the best precision but understanding the direction the situation is moving, and how the managers are implementing it (hopefully in a splendid manner).
Can ROIIC replace ROIC?
Is ROIIC superior than ROIC?
ROIC assesses total investments (old + new investments) whereas ROIIC new investment only which could be small in scale compared to the old investment.
Please explain why in detail.
What if ROIIC, (1+Gnp)/(1+Gic) & ROIC are assessed at the same time?
https://gemini.google.com/share/686374f17eb4
They’re both just measurement tools. Neither is “better” in absolute terms. It really depends on the objective and the context.
As we wrote:
“ROIIC fixes the ‘average vs incremental’ issue [of ROIC], but at the cost of worsening the ‘accounting vs economic’ issue.”
Looking at the two together, within a single framework, is exactly what we’re trying to do: reduce the averaging problem you get with ROIC alone, while also limiting the impact of accounting or operational changes that aren’t economically relevant for the analysis.
ROIIC also isn’t limited to “just the new investments”. You’re free to choose an older/longer time window. In our case we used 2020-2024, but we could just as well have chosen 2012-2018. What matters is to look at ROIIC over time blocks that make economic sense for the business and, where possible, make the necessary adjustments to reduce the noise in the data you’re measuring.
To finish, you’re suggesting adding the NOPAT vs IC growth rate into the framework. I haven’t studied it enough to have a strong view, but my feeling is that the information it brings is quite overlapping with the ROIC/ROIIC combination. I may be wrong. In any case, like any model, it’s a trade-off between simplicity, usefulness and accuracy. Depending on what you want to prioritise, including it (just like using ROIC or ROIIC on their own) can be relevant or not. It really depends on the context and how you intend to use it.
Hey ATC,
Perhaps simplistically, this is how I think about ROIC vs ROIIC.
I am a marathon runner, and ROIC is my total marathon time (eg how long it took me to run my 42.2km).
But, what about my splits? Did I start strong and fade? Did I speed up as I went? How does KM 32-42 look? That's where ROIIC has value.
To the growth rate comment I'd have to think more about it, it's a good question and adds another layer of analysis - thanks for that.
Which is exactly what we argued in the piece: ROIC and ROIIC only really come into their own when you look at them together rather than in isolation.
"ROIC is a good reflection of a company’s economic quality over its entire life.
ROIIC measures the quality of recent capital-allocation decisions and gives a glimpse of what the company is in the process of becoming.
Looking at the two side by side helps you judge whether its value-creation trajectory is improving, deteriorating, or broadly stable over time."