Capitalized consumer surplus vs. market capitalization. Points above the line create more surplus for their consumers than the market prices into the equity. The bars that span the points represent the linear-demand and CES estimates at each sector's elasticity band. When a firm's bar is longer than it's distance to the parity line, the side it falls on represents the features of the demand system the company is in rather than any concrete items regarding the firm's financial performance/metrics.

So what? 355 of the 497 companies sit above the line, and the median firm generates two dollars of capitalized consumer surplus per each dollar of market value. This roughly aligns with theory, showing that in any market in real competition most of the gains from trade go to the buyer.

The bars matter too, equal if not more than the dots. Where a bar is longer than the firm's distance to the line, which side it ends up being on is a property of the assumed demand curve rather than any property of the firm itself.

Looking at the sector totals, we see some misleading information since substitution in real competition happens mostly within a sector.

Dashed line represents a parity - utilities and staples sit high because regulation and competition can be observed capping the markup, not because they are better managed firms. The ordering certainly doesn't represent anything virtuous about any of these industries. Utilities is high because a regulator already capped the markup. IT scores low because pricing power is the sector's main defensive goal. This is a good way to split up industry structure, rather than management. Healthcare needs a caveat set out under the tree map before you start quoting it...

Producer surplus as a share of total surplus, PS/(PS+CS), by sector. Under CES this equals 1/elasticity. Almost a quarter of the index earns less than its cost of capital and so captures nothing based on this definition. Thick bar spans the interquartile range and the open circle is the sector median.

Median capture is about 2% in utilities and 24% in energy while a quarter of the index sits ar 0%. These firms earn less than their cost of capital, so on an economic definition they capture no surplus (technically)

The 30 largest S&P500 firms by market capitalization, reordered by capitalized consumer surplus. Both columns are ranked within these 30, but the rank across all 497 companies is in parentheses on the right side.

UnitedHealth moves up 29 places, BofA 17, and Exxon by fifteen. Nvidia falls by 16, AMD falls 14, and Tesla 13. This seems like a paradox, but maybe not. High markups here imply a high elasticity under CES, and a high elasticity implies a very thin surplus triangle. The firms with the strongest pricing power generate the least consumer surplus per dollar of revenue. Semis fall for the same reasons that utilities rise.

The ordering certainly doesn't represent anything virtuous about any of these industries. Utilities is high because a regulator already capped the markup. IT scores low because pricing power is the sector's main defensive goal. This is a good way to split up industry structure, rather than management. Healthcare needs a caveat set out under the tree map before you start quoting it...

Top 90 S&P500 firms by capitalised consumer surplus. The area is consumer surplus and the color represents the share that the producer keeps - blue keeps less and red keeps more, diverging at the index median.

Apple and Nvidia are small and red, but McKesson, Cencora, and Cardinal Health are enormous and blue. This is a limitation of this exercise - consumer surplus scales with revenue, and pharma distributors book hundreds of billions of low-single-digit margins. This model then makes them the largest creators of consumer value in the index, and some of that is probably true given the business type. Anchoring on revenue when a distributor's revenue is usually the price passed through is the reason for this, but healthcare's sector-level 11.5x is driven by the same three names. Firm-level surplus doesn't help this picture. If one firm disappeared, it's rivals would absorb most of the demand, so these are good representations of the upper bounds for each firm rather than a complete decomposition.

Log-scaled - when the ratio is high, CES is extrapolating demand into a region with no transactions (so there's no evidence to use). Under CES, consumer surplus is revenue x (markup - 1), so this rises with pricing power.

This is a shortcut that needs no elasticity assumption at all. Under the CES the Lerner index gives ε = μ/(μ − 1). This substitutes back to CS = Revenue × (μ − 1), and markups are straight out of accounting data.

The problem is what it is and is what it says. It ranks firms by markup, so the higher a firm prices, the more social value it appears to deliver to society. CES being internally consistent doesn't make it a great representation of social value. But, with the median 1.65x multiple above the central estimate, we can see the worst case for several hundred companies. Financials and IT show the widest gaps, where CES extrapolates the hardest.