Subsidies can be a race to the bottom when network effects are weak or products are close substitutes; but with strong network externalities and differentiated products, competing industrial policy can create mutual gains—especially when it spurs product innovation rather than process improvements.
Industrial policy has returned to the centre of economic governance, particularly in the high-tech sectors where positive network externalities in demand make market dominance self-reinforcing. This paper studies the welfare effects of an industrial policy targeting a sector with network externalities in a two-country model with strategic trade and R&D investment. We show how the welfare consequences of this policy are determined by the interaction between the strength of the externality, the type of R&D, and the degree of product differentiation between the home and the imported goods. When externalities are weak or the goods are close substitutes, the business-stealing effect produces a race to the bottom that dissipates more surplus than it creates. Under sufficiently strong externalities and weak substitutability or complementarity of the goods, industrial policy competition can make both countries simultaneously better off compared to the laissez-faire outcome because of the mutual business-enhancement effect. The case is stronger for the product innovation than for the process innovation, as the former directly affects the demand and triggers a stronger network effects than the latter which operates indirectly through the supply. Thus, the network externalities create an opportunity for a win-win industrial policies, but its realisation depends on the market structure and the nature of innovation.
Summary
Main Finding
Industrial policy (R&D subsidies combined with import-side taxes) in sectors with network externalities can produce two qualitatively different international outcomes. When network externalities are weak or goods are close substitutes, non-cooperative policy competition generates a “race to the bottom” that destroys more surplus than it creates. But when network externalities are strong and goods are sufficiently differentiated (or complementary), subsidies (especially to demand-enhancing/product R&D) can trigger mutual business-enhancement and yield a win‑win: both countries can be better off relative to laissez‑faire. The win‑win is much more likely under product (quality/demand-side) R&D than under process (cost‑reducing) R&D.
Key Points
- Model setup
- Two countries, two firms (one per country); home country is the sole consumer market in the model (importer), foreign is exporter.
- Goods are differentiated and exhibit positive network externalities in demand. Substitutability/complementarity is captured by parameter m ∈ [−1,1].
- Firms choose R&D and Cournot quantities; governments simultaneously choose R&D subsidies for domestic firm and (home) import sales tax on the foreign firm. Game solved by backward induction.
- Demand specification
- Representative consumer utility quasi-linear; inverse demands: p_i = a + r_i − (1 − b_i) q_i − m q_j.
- ai = a + b_i Q_i captures network externalities (b_i strength); r_i is product quality from product R&D.
- R&D types and channels
- Process R&D: cost‑reducing (shifts supply), affects network externality only indirectly via changed quantities.
- Product R&D: quality‑enhancing (shifts demand directly), directly amplifies network effects.
- Strategic policy results
- Both governments optimally subsidise their firms’ R&D (consistent with observed subsidy escalation).
- Sign and welfare impact of the home import tax depend jointly on (i) network externality strength b, (ii) degree of product differentiation/substitutability m, and (iii) whether R&D is product or process type.
- Weak externalities or strong substitutability → business‑stealing dominates: policy competition reduces global welfare (race to the bottom).
- Strong externalities + weak substitutability or complementarity → mutual business‑enhancement: home may optimally subsidise (not tax) foreign sales under product R&D, enlarging the network and creating aggregate welfare gains (win‑win).
- Product R&D more likely to deliver win‑win because it directly increases willingness‑to‑pay and hence network value; process R&D operates indirectly and is less effective at internalising demand externalities.
- Interpretation and generality
- The network externality formulation can be read either as consumer-side adoption feedback or as derived (downstream) demand / agglomeration effects for intermediate inputs.
- The qualitative results illuminate high‑tech policy arenas (semiconductors, EV charging, AI platforms) where adoption and ecosystem size matter.
Data & Methods
- This is a theoretical, analytical paper (no empirical data).
- Methods:
- Two-stage simultaneous‑move game: Stage 1 governments choose subsidies and import tax; Stage 2 firms choose R&D and Cournot outputs.
- Demand with additive network externality and product‑quality term; closed‑form inverse demand used to derive firm best responses.
- Two R&D specifications analysed separately: process (cost-reduction) and product (quality improvement).
- Analytical equilibrium derivations for firm and government choices where tractable; supplemented by numerical exercises to map welfare regions across parameter space (network strength b and substitutability m).
- Welfare measure:
- Aggregate welfare accounts for consumer surplus, firm profits, and government fiscal transfers/costs associated with subsidies and taxes (the usual strategic‑trade welfare aggregation).
Implications for AI Economics
- Relevance: AI platforms and many AI-enabled markets exhibit strong network/externality features (user base, developer ecosystem, model/data complementarities). This paper provides a formal framework to think about subsidies, trade/tax-like barriers, and cross‑border policy toward AI goods/services.
- Policy design takeaways
- Evaluate the nature of innovation: policies that primarily foster product/quality/demand‑side improvements (interoperability, open standards, ecosystem incentives) are more likely to generate positive cross‑border externalities than pure cost‑reduction support.
- Gauge network strength and differentiation: when network externalities are strong and products are differentiated (or complementary across borders), enabling larger international ecosystems (even via subsidising foreign provision or removing trade frictions) can create global gains—i.e., Pareto‑improving cooperation is feasible.
- Avoid automatic tit‑for‑tat subsidy escalation when externalities are weak or products are close substitutes: such subsidy races may be welfare‑destroying.
- Trade measures (taxes, export controls, procurement barriers) can have non‑intuitive demand‑side effects; a tax on a foreign platform may reduce the foreign network’s value for domestic users, harming consumer welfare beyond simple profit‑shifting.
- Coordination and strategy
- International coordination or explicitly ecosystem‑focused policies (promoting compatibility, data‑sharing standards, cross‑border R&D collaboration) can help internalise positive externalities and avoid destructive competition.
- For AI policy: prioritize interventions that expand ecosystem adoption and compatibility (product‑R&D analogues: standards, model‑sharing, developer incentives) rather than only firm‑level cost or capacity subsidies.
- Caveats for application
- Model abstracts from multi‑market/platform multi‑sidedness, heterogenous firms, and dynamic strategic entry; real AI markets involve complementarities (data, compute, developers), multi‑sided pricing, and possible global consumer bases — quantitative outcomes will depend on richer features.
- The paper’s consumer‑only-in‑home assumption simplifies incidence and cross‑border demand; empirical diagnosis of network strength and substitutability is necessary before policy prescriptions are implemented.
Overall, the paper highlights that network externalities change the conventional strategic‑trade intuition: under the right conditions (strong demand externalities + demand‑side innovation + product differentiation), industrial policy can shift from a zero‑sum rent contest into a mutually beneficial ecosystem strategy—important guidance for AI industrial policy design.
Assessment
Claims (5)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| The welfare consequences of an industrial policy targeting a sector with network externalities are determined by the interaction between the strength of the externality, the type of R&D, and the degree of product differentiation between the home and the imported goods. Consumer Welfare | mixed | high | aggregate welfare (welfare consequences of policy) |
0.12
|
| When externalities are weak or the goods are close substitutes, the business-stealing effect produces a race to the bottom that dissipates more surplus than the industrial policy creates. Consumer Welfare | negative | high | aggregate welfare (net surplus created/dissipated by policy) |
0.12
|
| Under sufficiently strong network externalities and weak substitutability (or weak complementarity) of the goods, industrial policy competition can make both countries simultaneously better off compared to the laissez-faire outcome because of a mutual business-enhancement effect. Consumer Welfare | positive | high | aggregate welfare (comparison to laissez-faire) |
0.12
|
| The case for mutually beneficial industrial policy is stronger for product innovation than for process innovation, because product innovation directly affects demand and triggers stronger network effects while process innovation operates indirectly through supply. Consumer Welfare | positive | high | magnitude/likelihood of welfare gains from industrial policy (product vs process R&D) |
0.12
|
| Network externalities create an opportunity for win-win industrial policies, but the realisation of such mutually beneficial outcomes depends on market structure (product differentiation/substitutability) and the nature of innovation (product vs process). Consumer Welfare | mixed | high | possibility/conditions for mutual welfare improvement from industrial policy |
0.12
|