Digitization and the Macro-Economics of Superstars

Summary

"Digitization and the Macro-Economics of Superstars" by Anton Korinek (University of Virginia and NBER) and Ding Xuan Ng (Johns Hopkins on the manuscript's title page; listed at the Monetary Authority of Singapore on the July 2019 ECB presentation) examines how digital innovation creates "superstar" dynamics in the macroeconomy — a phenomenon in which a small number of firms or individuals capture a disproportionate share of market value because information can be copied at negligible cost yet is excludable, generating increasing returns and market power. The manuscript's abstract describes superstars "as arising from digital innovations, in which an entrepreneur pays a fixed cost to automate a fraction of the tasks in production at zero marginal cost, generating a form of increasing returns to scale"; Korinek's site summarizes it as "digital technologies intrinsically lead to higher markups and a lower labor share, favoring the creation of superstars." The paper was presented in successive versions at the 5th IMF Statistical Forum (November 2017, under the title "The Macro-Economics of Superstars") and the ECB Conference on Challenges in the Digital Age (July 2019), and circulated as an academic working paper; the full manuscript posted on Korinek's website (listed there as Dec. 2018; the PDF's title page is dated December 2019) carries no working-paper series number from the IMF or any other institution, and Korinek's CV likewise lists it without one — it appears never to have been issued as a numbered IMF Working Paper.

The work sits at the intersection of the superstar firms literature and the broader debate over whether rising corporate profits reflect economic rent extraction rather than competitive returns to capital. Its relevance to the Georgist framework is that digitization may create returns to position, scale, and network control that are structurally analogous to land rents — returns to scarcity and location rather than to marginal production.

Core Argument

The paper builds a general-equilibrium model in which superstars arise from digital innovation, defined as an innovation that "replaces a fraction of production tasks by a digital process that can be scaled at negligible cost." The mechanism turns on a specific property of information as an input: it is non-rival — it "can be copied at negligible cost" — yet excludable — it "may generate monopoly power." Together these features mean, in the deck's words, that "information technology supercharges the superstar effect." The following mechanisms are the model's own, verified against the authors' presentation text:

1. Fixed cost plus near-zero marginal cost → increasing returns. In the model an entrepreneur adopts a technology that "imposes a fixed cost ξ ≥ 0 but automates a fraction γ ∈ (0,1) of production tasks at negligible marginal cost." The fixed cost "generates increasing return" and the exclusive right to the innovation (e.g. a patent) "generates market power." A superstar's marginal cost falls to (1 − γ) times the traditional unit cost, so the most productive supplier can undercut competitors while charging a markup.
2. Market power and market concentration. Because superstars can serve the whole market at a markup, they "capture large market share, earn rents (in contrast to models of factor-biased technological change)." The paper derives explicit implications for market concentration and the superstar profit share.
3. A pricing regime that switches with the size of the cost saving. When cost savings are small (γ below a threshold set by the elasticity of substitution), the superstar is constrained by competition from traditional firms and prices at the traditional unit cost, absorbing the entire cost saving as profit; when cost savings are large, it charges the optimal monopoly markup and cuts prices to expand quantity. This yields a "labor-saving effect" of innovation in the first regime (output and employment diverge) and an "output scale effect" in the second.

The macroeconomic implications the paper derives are effects on factor prices and shares, market concentration, income distribution, and public policy. In general equilibrium, superstars earn a profit share of σ = min{γ, 1/θ} (where θ is the elasticity of substitution), with the residual split into a capital share of α(1 − σ) and a labor share of (1 − α)(1 − σ) — so a larger superstar share mechanically lowers the labor share. These results connect to the same trends documented by De Loecker, Eeckhout & Unger (2020) on rising markups and by Barkai (2020) on the rising pure-profit share; the presentation decks include a figure estimating a rising "superstar profit share in national income, 1984–2014" based on the authors' calculations from Barkai (2017) and Piketty and Saez (2017), and the manuscript's empirical-motivation section builds directly on the Barkai (2017) and De Loecker–Eeckhout (2017) profit-share and markup evidence.

Note that the model's driver is the fixed-cost/zero-marginal-cost structure of digital innovation and the resulting market power, not network effects, platform lock-in, or global scaling per se. Earlier drafts of this page attributed the mechanism to network externalities and global reach; those framings are common in the adjacent literature but are not the levers of Korinek and Ng's formal model. The full manuscript confirms this: network effects appear only as an extension (Section 4.2, "Network effects and increasing returns"), and the extension's own punchline is that "network effects provide for additional productivity gains but do not change the monopoly markup and factor shares of traditional factor owners," since "the optimal monopoly markup is driven by the demand elasticity of consumers not by the specific production technology of firms." Other extensions cover oligopolistic market-share dynamics (where letting rival firms share cost-saving digital innovations raises both consumer and producer surplus), the case of an elasticity of substitution at or below one (which "raises the dismal specter of all the gains from innovation going to the superstars"), and factor bias / skill-biased innovation; there is no global-scaling or platform-lock-in extension.

Relation to the Rent Debate

The paper's significance for the Georgist case is that it provides an IMF-platformed analysis of how digital technology generates returns that look like economic rent — income from positional advantage and scarcity rather than from productive contribution at the margin. This connects to several strands of the wiki's rent literature:

• The outcome that corporate profits increasingly reflect rents draws on markup studies and firm-level return dispersion to argue that a growing share of profit is extraction rather than reward. Korinek and Ng's digitization-superstar mechanism supplies a theoretical channel for how technology creates these rents.
• The superstar firms concept, as developed by Autor et al. (2020), offers a partly efficiency-based reading of the same concentration trend. Korinek and Ng's paper provides the digital-technology mechanism behind superstar dynamics, and it does take a position on the efficiency question: its welfare analysis finds that the decentralized equilibrium exhibits "insufficient digital innovation" and "inefficiently low quantities," because monopoly markups distort both the innovation decision and post-innovation output. In other words, the model treats superstar rents as a real monopoly distortion (with deadweight loss) rather than as an efficient competitive outcome — while also noting that once superstars reach their optimal monopoly markup, further innovation lowers prices and "expands income for all."
• The broader argument that rent-seeking extends beyond land to digital platforms is developed on this wiki in Soil to Servers: Digital Rents and in Mazzucato's mapping of modern rents.

Nuances and Caveats

• The efficiency counter-reading. As with the broader superstar-firms literature, high profits at digital firms may partly reflect genuine productivity and scale advantages rather than pure rent extraction. Crouzet & Eberly (2019) show that intangible capital (software, IP, brand) can explain a portion of measured profits, and Autor et al. (2020) argue that winner-take-most competition can be efficiency-driven. Korinek and Ng partly internalize this tension: in their model the increasing returns from digital innovation are genuine and socially valuable (the decentralized economy actually under-invests in digital innovation), yet the exclusive control of that innovation produces monopoly markups and rents that are a distortion. Their policy remedies — public investment to finance digital innovation, and offsetting monopoly markups via subsidy while charging consumers fixed-plus-variable cost — are aimed precisely at capturing the efficiency gains while neutralizing the rent. So the paper is not a pure rent-extraction story; it is a story in which the same technology generates both efficiency and rent.
• A working paper / conference paper, not a formal IMF policy document. This paper circulated as an academic working paper and was presented at IMF and ECB research events; the full manuscript (December 2019 draft) is self-posted on Korinek's website with no institutional working-paper imprint or number. As an unpublished research paper it does not represent an institutional position of the IMF or of the Monetary Authority of Singapore.
• Scope. The model is a stylized general-equilibrium theory motivated by US trends. The full manuscript's empirical-motivation section uses Compustat data on US publicly listed firms (following De Loecker and Eeckhout's markup methodology, and constructing a "superstar" sample of the top-4 firms by market share within 4-digit NAICS industries in 2014); it documents that the revenue share of variable inputs across US listed firms "has fallen to just slightly over 0.6 in 2014." International evidence on the declining labor share is cited from the literature (Karabarbounis and Neiman 2013), but the paper contains no cross-country empirics of its own and is not an empirical study of any developing economy. (The mid-1980s-to-mid-2010s labor-share figures quoted in earlier drafts of this page come from the authors' presentation decks rather than the manuscript.)

Bears On

• Outcome: Corporate profits increasingly reflect economic rents — the paper provides a theoretical mechanism (digital innovation → increasing returns and market power → superstar rent capture) that supports the claim that rising profits are partly rents. The paper frames its own contribution as explaining "rising superstar profits as [the] main driver" of the declining labor share and rising profit share, and its welfare analysis explicitly characterizes those profits as involving "significant monopoly rents and inequality" — so the assignment to this outcome is defensible from the paper's own framing.
• Concept: Superstar Firms — the paper's digitization-superstar framework complements the Autor et al. (2020) superstar-firms concept by identifying the technological driver.
• Concept: Economic Rent · Rent-Seeking
• Narrative: The Rentier Economy — digital superstar rents are part of the broader rentier-economy thesis.
• Research: Soil to Servers: Digital Rents — a Georgist analysis making the analogous argument that digital platform rents repeat the pattern of land rents.

See Also

• Korinek–Stiglitz: AI and income distribution — the macro/AI companion by the same lead author
• Corporate profits increasingly reflect economic rents
• Superstar Firms
• Economic Rent
• The Rentier Economy
• Soil to Servers: Digital Rents

Sources

1. Anton Korinek & Ding Xuan Ng, "Digitization and the Macro-Economics of Superstars," full working-paper manuscript, December 2019 (University of Virginia and NBER / Johns Hopkins). PDF via korinek.com (linked from korinek.com/research) — the full text, fetched and read directly for this page — used for, and confirms, the model structure presented in the decks: non-rival/excludable digital innovation with fixed cost and zero marginal cost; the general-equilibrium superstar profit share σ = min{γ, 1/ε} with wage and capital income w = (1−α)(1−σ)Y and RK = α(1−σ)Y (the manuscript writes the elasticity of substitution as ε where the deck uses θ); Proposition 3's welfare result that "the decentralized equilibrium exhibits (i) insufficient digital innovation and (ii) inefficiently low quantities in superstar sectors"; and the three corrective policies (publicly financing the fixed cost of innovation and making it freely available; non-linear pricing — a fixed charge plus marginal-cost pricing; or a subsidy s_i = σP_i offsetting the monopoly markup). Also the source for the Section 4 extensions and the Compustat-based empirical motivation described above.
2. Anton Korinek & Ding Xuan Ng, "The Macro-Economics of Superstars," presentation at the 5th IMF Statistical Forum, November 2017 (Johns Hopkins University and NBER). Semantic Scholar PDF · IMF-hosted deck — used to verify the model structure as presented at the IMF forum: the fixed-cost/negligible-marginal-cost digital innovation, the small-vs-large-cost-saving pricing regimes, the factor-share result σ = min{γ, 1/θ}, the monopoly-distortion welfare analysis, and the policy remedies. Quoted phrases in the Core Argument and Bears On sections not otherwise attributed to the manuscript are taken verbatim from this deck.
3. Anton Korinek & Ding Xuan Ng, "Digitization and the Macro-Economics of Superstars," presentation at the ECB Conference on Challenges in the Digital Age, July 2019 (UVA and NBER / Monetary Authority of Singapore). ECB PDF — used to confirm the current title, Ng's Monetary Authority of Singapore listing at that event, and the extension of the argument to monetary-policy implications (the manuscript's Section 5 covers macroeconomic dynamics and the Phillips curve).
4. Anton Korinek, curriculum vitae, updated August 2025. PDF — used to confirm authorship and that the paper is listed among Korinek's works without a formal IMF Working Paper number.
5. Korinek's summary line, retrieved via korinek.com/research: "Digital technologies intrinsically lead to higher markups and a lower labor share, favoring the creation of superstars." The site lists the paper as "with Ding Xuan Ng, Dec. 2018"; the linked PDF's title page is dated December 2019.

Bibliographic correction: A prior version of this page cited the paper as "IMF Working Paper WP/21/177 (2021)" with the URL wpiea2021177-print-pdf.ashx. That working-paper number and URL in fact belong to a different paper — Kazuko Shirono et al., "Is Mobile Money Part of Money? Understanding the Trends and Measurement," IMF WP/2021/177 (July 2021) — confirmed via IMF eLibrary. The Korinek–Ng superstars paper does not appear to have been issued with that number.

Verification note: The full working-paper text (December 2019 manuscript) was fetched directly from the author's website (Source 1) and checked against the presentation decks: the core model, the factor-share results, the welfare propositions, and the policy remedies are consistent across versions, and no version located carries an IMF Working Paper number. The manuscript remains an unpublished working paper; small drafting differences from the decks are noted inline where they matter (elasticity notation, the labor-share figures, and the network-effects extension).