📈 economics

Two-stage dynamic network game for low-carbon technology innovation and diffusion among manufacturers under government intervention

This paper employs two-stage dynamic network game models to demonstrate that government interventions—specifically reputation concerns, subsidies, and carbon taxes—combined with oligopolistic innovation and favorable small-world network topologies, are critical for overcoming barriers and accelerating low-carbon technology diffusion among manufacturers.

Original authors: Wanting Chen, Zhi-Hua Hu

Published 2026-07-20

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Wanting Chen, Zhi-Hua Hu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world of business as a giant, buzzing city where every factory is a building, and the roads connecting them are the flow of information and trade. In this city, there's a massive problem: too much smoke and pollution. To fix it, the city needs a new kind of "clean magic"—low-carbon technology. But here's the catch: inventing this magic is expensive and risky. Some buildings are huge, wealthy skyscrapers (oligopolistic manufacturers) with deep pockets, while others are small, struggling shacks (small and medium-sized manufacturers). The skyscrapers can afford to invent the magic, but the shacks can't.

Enter the City Mayor (the government). The Mayor wants everyone to be clean but can't force the shacks to invent the magic themselves. Instead, the Mayor tries to convince the skyscrapers to invent it and then share it with the shacks. This paper uses a tool called "game theory," which is like a super-advanced board game that predicts how people will act when they are trying to win but also need to cooperate. It also looks at "networks," which are just maps of who talks to whom. The big question is: How do we get the skyscrapers to invent the magic, and how do we get the shacks to use it without the whole city falling apart?

This paper sets up a two-step game to solve this puzzle. In the first step, the Mayor and the big skyscrapers play a game. The Mayor decides whether to give out cash rewards (subsidies) or just sit back and do nothing. The skyscrapers decide whether to spend their money inventing the clean tech or stick to their old, dirty ways. The paper simulates this game using a computer to see what happens. It finds that the skyscrapers only play along if the Mayor's rewards are big enough to cover the cost of invention, and if the Mayor cares enough about their reputation to actually follow through. If the Mayor is lazy or the skyscrapers think the cost is too high, the game stalls, and no new tech is invented.

Once the first step is done, the paper moves to the second step: the diffusion game. Now that the skyscrapers (hopefully) have the new tech, they need to pass it down the line to the shacks. This is where the "network" comes in. The paper imagines the factories connected in a "small-world network," which is a fancy way of saying a group where everyone has a few close friends but can also reach anyone else in the group with just a few handshakes. The computer simulates how the technology spreads through this web of connections.

The results of these simulations are quite clear. First, if the Mayor and the skyscrapers are on the same page—meaning the Mayor is supportive and the skyscrapers are eager to innovate—the cost for the small factories to learn the new tech drops, and the technology spreads like wildfire. Second, the paper tests what happens when the Mayor puts a "pollution tax" on dirty factories. The simulation shows that higher taxes make the clean tech much more attractive, speeding up the spread. Third, the paper looks at "spillovers," which is like when one factory's success accidentally helps its neighbors learn faster. The more this happens, the faster the whole city gets clean.

Interestingly, the size of the network matters, but only under certain conditions. If the taxes and spillovers are weak, adding more factories to the network doesn't help much; it might even slow things down slightly. But if the taxes are high and the spillovers are strong, a bigger network means the clean tech spreads to everyone much faster. The paper also checks out what happens if customers start demanding clean products (low-carbon preference). When customers care, the factories rush to adopt the new tech to keep their sales.

Finally, the paper looks at the shape of the network itself. It turns out that the "small-world" shape is perfect for this. If the network looks a bit more like a regular, orderly grid (where everyone has the same number of neighbors), it helps the tech spread efficiently, especially when the government and big companies are being cautious. But if the government and big companies are super active and supportive, a more random, chaotic network actually helps the tech spread even faster.

In short, the paper suggests that getting low-carbon technology to spread isn't just about one thing. It requires a team effort: the government needs to be a supportive coach, the big companies need to be willing to invest, and the connections between all the companies need to be strong enough to carry the message. Without all these pieces working together, the clean tech might get stuck in the lab, leaving the city smoky and the shacks struggling.

Technical Summary: Two-Stage Dynamic Network Game for Low-Carbon Technology Innovation and Diffusion

Problem Statement
The manufacturing sector faces significant structural hurdles in adopting and diffusing low-carbon technologies. While oligopolistic manufacturers possess the resources to lead innovation, small and medium-sized manufacturers (SMEs) encounter high technical entry barriers and cost pressures that impede adoption. Furthermore, existing literature often restricts analysis to static, single-stage interactions between governments and firms, failing to capture the dynamic, two-stage nature of policy formulation followed by industry-wide diffusion. Additionally, the role of complex network topologies—specifically how information flows and strategies evolve within heterogeneous industrial networks—remains under-explored in the context of low-carbon transitions. This study addresses the need to understand how government intervention, manufacturer heterogeneity, and network structures jointly influence the innovation and diffusion of low-carbon technologies.

Methodology
The paper employs a two-stage evolutionary game theory framework embedded within complex network models (specifically Watts-Strogatz small-world networks).

First-Stage Game (Government vs. Oligopolistic Manufacturers):
- Agents: The government and oligopolistic manufacturers.
- Strategies: The government chooses between a "supportive" strategy (providing subsidies and supervision) or an "inactive" strategy. Oligopolistic manufacturers choose between "low-carbon technological innovation" or "conventional technology maintenance."
- Dynamics: The model utilizes replicator dynamic equations to analyze the evolutionary stability of strategies. Key parameters include government reputation benefits ( $W$ ), innovation costs ( $I$ ), subsidies ( $S$ ), penalties ( $F$ ), and technology sharing profits ( $D$ ).
- Outcome: This stage determines the equilibrium probabilities of government support ( $x$ ) and manufacturer innovation ( $y$ ), which subsequently influence the cost of technology sharing for the broader market.
Second-Stage Game (Among Heterogeneous Manufacturers):
- Agents: A network of heterogeneous manufacturers (oligopolists and SMEs) modeled as nodes in a WS small-world network.
- Strategies: Each manufacturer chooses between adopting low-carbon production or maintaining conventional production.
- Dynamics: The model uses the Fermi update rule to simulate bounded rationality, where manufacturers probabilistically imitate neighbors with higher payoffs. Payoffs are influenced by carbon taxes ( $T$ ), technology spillover coefficients ( $\beta$ ), consumer low-carbon preferences ( $\lambda$ ), and the technology sharing cost coefficient ( $\sigma$ ), which is endogenously determined by the first-stage game outcomes.
- Simulation: Numerical simulations are conducted using MATLAB on networks of varying scales ( $N=50, 100, 200$ ) and rewiring probabilities ( $p$ ) to observe diffusion dynamics over time.

Key Contributions

Two-Stage Framework: The study distinguishes itself by modeling the interaction as a sequential process: a first-stage game determining policy and innovation incentives, followed by a second-stage game analyzing the resulting diffusion dynamics across a network.
Integration of Heterogeneity and Network Topology: It explicitly incorporates manufacturer heterogeneity (oligopolists vs. SMEs) and utilizes WS small-world networks to capture the "short average path length" and "high clustering coefficient" typical of real-world industrial clusters.
Endogenous Cost Mechanism: The model links the first-stage strategic equilibrium to the second-stage diffusion process by defining the technology sharing cost coefficient ( $\sigma$ ) as a function of government support and oligopolistic innovation probabilities.

Results

First-Stage Equilibrium: The optimal strategic equilibrium between the government and oligopolistic manufacturers is primarily driven by government reputation concerns, innovation investment costs, and subsidy intensity. A synergistic scenario where the government provides active support and oligopolists innovate leads to the most favorable outcomes, significantly reducing technology sharing costs.
Impact of Sharing Costs: There is a strong negative correlation between low-carbon technology sharing costs and diffusion efficiency. Lower sharing costs, achieved through government-firm synergy, accelerate diffusion.
Carbon Tax and Spillovers: Higher carbon tax rates and stronger technology spillover coefficients ( $\beta$ ) are positively correlated with diffusion. High spillovers can lead to rapid, complete diffusion, whereas low spillovers may result in stagnation.
Consumer Preferences: Consumer low-carbon preference ( $\lambda$ ) acts as a critical driver. The presence of carbon labels (simulated by $\lambda > 0$ ) effectively guides consumption behavior and motivates manufacturers to adopt low-carbon strategies.
Network Topology Effects:
- Scale: Generally, a larger network scale expands the coverage of low-carbon technologies. However, this effect is conditional; if carbon taxes or spillovers are low, network expansion may have negligible or even slightly negative effects on diffusion.
- Structure: The efficiency of diffusion depends on the network's rewiring probability ( $p$ ). When government and oligopolist strategies are conservative (low support/innovation), networks resembling regular networks (low $p$ , high clustering) facilitate better diffusion. Conversely, when policies and innovation are active (high support/innovation), networks closer to random networks (high $p$ ) accelerate diffusion speed and levels.

Significance and Claims
The paper claims that its primary significance lies in providing a systematic theoretical framework to unpack the mechanisms of low-carbon technology diffusion under government intervention. It highlights that:

Policy Synergy is Crucial: The coordination between government supportive policies and the innovation activities of oligopolistic manufacturers is essential to reduce sharing costs and break diffusion bottlenecks.
Context-Dependent Network Effects: The paper modestly claims that network topology does not have a uniform effect; rather, the "best" network structure for diffusion depends on the strategic stance of the government and leading firms.
Practical Implications: The findings suggest that for industries where products are physically indistinguishable (like food or beer), implementing carbon labeling systems and maintaining high carbon tax rates are effective levers to stimulate market demand and technological adoption.

The authors acknowledge limitations, noting that the study relies on theoretical simulations with specific update rules (Fermi) and that future research could explore more diverse network structures and heterogeneous strategy updating rules to reflect real-world competition and cooperation scenarios more accurately.

Technical Summary: Two-Stage Dynamic Network Game for Low-Carbon Technology Innovation and Diffusion

More like this