The rapid evolution of drone technology has positioned the drone industry as a pivotal engine of the low-altitude economy and a representative of new-quality productive forces. However, a structural tension persists between the accelerating pace of technological innovation and the lagging development of safety governance mechanisms. Under the national strategic imperative of “ensuring both development and safety”, understanding how the design and combination of development-oriented and safety-oriented policies can shape a comprehensive industrial advantage becomes critical. This study adopts a configurational perspective, employing fuzzy-set Qualitative Comparative Analysis (fsQCA) on data from 30 provinces in China to investigate how the Development–Safety policy mix drives the formation of the drone technology industry’s comprehensive advantage. We define comprehensive advantage as an integrated competitive position encompassing innovation advantage, market advantage, safety advantage, and ecological advantage, which aligns with the dual requirements of high-quality development and high-level safety. Our findings reveal three distinct policy configurations for high-level comprehensive advantage and two patterns for non-high-level outcomes. We further unpack complementary, substitution, multiplier, and bottleneck effects among policy instruments and construct a configurational model that illustrates the growth mechanism of drone technology industry advantage.
1. Theoretical Background
1.1 Comprehensive Advantage of the Drone Technology Industry
The comprehensive advantage theory extends conventional competitive advantage frameworks by emphasizing the synergistic interplay between multiple resource- and capability-based advantages. Applied to the drone technology industry, comprehensive advantage is characterized by four interrelated dimensions:
- Innovation Advantage: reflects the strength of R&D investment, patent output, and talent concentration in drone technology.
- Market Advantage: captures the scale of drone technology enterprises, market capitalization, and revenue from drone-related products.
- Safety Advantage: embodies the levels of technology security (e.g., self-reliance in core drone technology components), operational safety (e.g., airspace regulation and flight certification), and public safety (e.g., emergency response systems).
- Ecological Advantage: indicates the health of the innovation ecosystem (e.g., enterprise growth, academic events) and the governance ecosystem (e.g., multi-department coordination, industry associations).
These dimensions interact nonlinearly. For instance, a strong innovation advantage without adequate safety governance may lead to risks that undermine market trust. Conversely, excessive safety regulation without innovation stimulation can stifle growth. Therefore, comprehensive advantage requires deliberate policy orchestration that balances development (innovation and market expansion) with safety (technology, operational, and public security).
1.2 Development–Safety Policy Mix for Drone Technology
Drawing on policy mix theory, we conceptualize the Development–Safety policy mix as a multi-level construct comprising three components:
- Policy Strategy: The national vision of “ensuring both development and safety” provides overarching principles such as innovation-driven development, safety-guaranteed development, and mutual reinforcement between the two.
- Policy Instrument Mix: Using LDA topic modeling on 126 policy documents (2010–2023) from 30 provinces, we identify six types of policy instruments classified along the development–safety axis. Table 1 summarizes these instruments with their typical measures.
- Implementation Context: Provincial variations in resource endowments, industrial bases, and institutional capacity necessitate tailored combinations of instruments.
Table 1: Classification of Development–Safety Policy Instruments in Drone Technology
| Policy Dimension | Instrument Type | Top Keywords (from LDA) | Typical Measures |
|---|---|---|---|
| Development-Oriented | Innovation-Driven | innovation, R&D, talent, science, funding, collaboration, project, platform, subsidy, special fund | R&D subsidies, tax incentives, talent cultivation, industry-university-research collaboration, open competition mechanism |
| Application Expansion | commercialization, qualification, operation, market, procurement, product, scenario, pilot | commercialization funds, qualification certification, operational subsidies, government procurement, pilot demonstration projects | |
| Ecology Optimization | investment, finance, industrial park, service, apron, landing pad, charging station, exhibition, event, culture | investment guidance, industrial park construction, service platforms, infrastructure development, event support | |
| Safety-Oriented | Science & Technology Security | patent, standard, independent innovation, intellectual property, bottleneck, sanction, controllable, supply chain reinforcement | patent navigation, standard setting, IP services, countermeasures against entity lists, supply chain strengthening actions |
| Operational Safety | market access, business environment, competition, network security, data, privacy, laws, regulations, supply chain | negative list management, business environment optimization, cybersecurity regulation, data governance, supplier audits | |
| Public Safety | safety awareness, risk, prevention, hazard, precaution, emergency, life, rescue | safety education, risk monitoring, hazard inspection, emergency plans & drills, rescue team building |
1.3 Configurational Effects of the Policy Mix
Rather than assuming linear additive effects, we adopt a configurational perspective positing that the six policy instruments interact in complex ways to produce combinatorial effects on the comprehensive advantage of drone technology. Four types of effects are particularly relevant:
- Complementary Effect: Two or more instruments mutually reinforce each other, generating synergy greater than the sum of individual effects.
- Substitution Effect: Certain instruments can replace one another depending on contextual conditions, leading to equifinal configurations.
- Multiplier Effect (Lever Policy): A particular instrument acts as a lever that amplifies the impact of other instruments, creating a multiplicative boost.
- Bottleneck Effect (Shortboard Policy): The absence or weakness of a critical instrument constrains the overall outcome, functioning as a limiting factor.
These effects are embedded in the implementation context, meaning that the same policy combination may yield different results across provinces.
2. Research Design
2.1 Method: fsQCA
We employ fuzzy-set Qualitative Comparative Analysis (fsQCA) for three reasons:
- fsQCA is designed to identify configurations of conditions (i.e., policy instruments) that are sufficient or necessary for an outcome, capturing complex causal patterns that traditional regression cannot.
- The fuzzy-set calibration allows fine-grained measurement of policy intensity and comprehensive advantage, preserving both ordinal and interval information.
- fsQCA supports equifinality, enabling the discovery of multiple distinct pathways to high comprehensive advantage, which aligns with the diversity of provincial contexts.
2.2 Sample and Data
The sample consists of 30 provincial-level administrative regions in China (excluding Tibet, Hong Kong, Macau, and Taiwan due to data unavailability). These provinces represent varying stages of drone technology industry development and policy environments. Policy documents from 2010 to 2023 were collected from the Peking University Law Database (pkulaw.com) and supplemented by provincial government portals, yielding 131 valid policy documents. Quantitative indicators for comprehensive advantage were obtained from the 2025 China Civil Aviation Industry Yearbook, the National Intellectual Property Administration, the National Standard Information Platform, Qichacha, Wind Database, and official government websites.
2.3 Variable Measurement and Calibration
Policy Instrument Intensity: For each policy document, we extracted specific measures and coded them into one of the six instrument types. Each measure was scored on three dimensions (form of issuance, intervention type, policy goal) using a 5-point scale (Table 2). The three scores were multiplied to obtain the total intensity of a single measure in a given year. The maximum score for each instrument type per year was retained to avoid double-counting. Finally, the total intensity for each instrument type in each province was computed by summing the annual scores over the period 2010–2023, adjusted for policy renewals and abolitions.
Table 2: Scoring Criteria for Policy Intensity Dimensions
| Score | Form of Issuance | Intervention Type | Policy Goal |
|---|---|---|---|
| 5 | Law, Regulation | Mandatory (e.g., flight control regulations) | Very clear with detailed action plan |
| 4 | Rule, Ordinance | Priority support (e.g., pilot demonstration cities) | Very clear with brief action plan |
| 3 | Decision, Outline, Plan, Stipulation | Resource input (e.g., industry development fund) | Relatively clear with action plan but insufficient elaboration |
| 2 | Program, Opinion, Scheme, Method | Directional guidance (e.g., technology roadmap) | Relatively clear without action plan |
| 1 | Notice, Announcement, Detailed Rule | Encouragement (e.g., industry self-discipline initiative) | Vague without action plan |
Comprehensive Advantage: We constructed a multi-dimensional index using 22 indicators across four dimensions (innovation, market, safety, ecological). After min-max normalization, entropy weighting was applied to determine dimension weights: innovation 0.342, market 0.287, safety 0.204, ecological 0.167. The weighted sum formula is:
$$CA_i = 0.342 \times I_i + 0.287 \times M_i + 0.204 \times S_i + 0.167 \times E_i$$
where \(I_i, M_i, S_i, E_i\) are the normalized scores of innovation, market, safety, and ecological advantages for province \(i\). The resulting comprehensive advantage scores for the 30 provinces range from 0.212 (Qinghai) to 0.927 (Guangdong).
Calibration: We used the direct calibration method with the 75th percentile, median, and 25th percentile as anchors for full membership, crossover point, and full non-membership. Because the sample size is 30, we averaged the values at positions 7–8, 15–16, and 22–23 to avoid boundary ambiguity. Table 3 shows the calibration anchors for all variables.
Table 3: Calibration Anchors for All Variables
| Variable | Full Membership | Crossover Point | Full Non-Membership |
|---|---|---|---|
| Innovation-Driven Policy | 269 | 115 | 62 |
| Application Expansion Policy | 196 | 99.5 | 49.5 |
| Ecology Optimization Policy | 223 | 106 | 56.5 |
| Science & Technology Security Policy | 184.5 | 77 | 43.5 |
| Operational Safety Policy | 291 | 162 | 65 |
| Public Safety Policy | 163.5 | 86.5 | 49 |
| Comprehensive Advantage | 0.674 | 0.479 | 0.382 |
3. Results
3.1 Necessary Condition Analysis
Before examining configurations, we tested whether any single policy instrument is a necessary condition for high (or non-high) comprehensive advantage. Table 4 shows that all consistency and coverage values are below the 0.90 threshold, indicating that no single instrument is necessary—confirming the configurational nature of the phenomenon.
Table 4: Necessary Condition Analysis Results
| Condition | High Comprehensive Advantage | Non-High Comprehensive Advantage | ||
|---|---|---|---|---|
| Consistency | Coverage | Consistency | Coverage | |
| Innovation-Driven | 0.832 | 0.617 | 0.629 | 0.505 |
| ~Innovation-Driven | 0.730 | 0.505 | 0.661 | 0.577 |
| Application Expansion | 0.721 | 0.535 | 0.636 | 0.516 |
| ~Application Expansion | 0.693 | 0.516 | 0.784 | 0.659 |
| Ecology Optimization | 0.688 | 0.616 | 0.538 | 0.578 |
| ~Ecology Optimization | 0.595 | 0.578 | 0.777 | 0.697 |
| Science & Technology Security | 0.731 | 0.656 | 0.645 | 0.501 |
| ~Science & Technology Security | 0.608 | 0.501 | 0.698 | 0.503 |
| Operational Safety | 0.833 | 0.698 | 0.702 | 0.654 |
| ~Operational Safety | 0.754 | 0.654 | 0.776 | 0.546 |
| Public Safety | 0.856 | 0.396 | 0.552 | 0.365 |
| ~Public Safety | 0.704 | 0.365 | 0.723 | 0.604 |
Note: “~” denotes the negation (absence) of the condition.
3.2 Configurational Analysis for High-Level Comprehensive Advantage
Using a frequency threshold of 1, consistency threshold of 0.80, and PRI threshold of 0.70, we derived the intermediate solution and combined it with the parsimonious solution to distinguish core and peripheral conditions. Table 5 presents three configurations (C1–C3) that are sufficient for high comprehensive advantage in drone technology.
Table 5: Policy Configurations for High-Level Comprehensive Advantage
| Condition | C1 | C2 | C3 |
|---|---|---|---|
| Innovation-Driven | ● | ● | ● |
| Application Expansion | ● | ||
| Ecology Optimization | ● | ⊗ | ● |
| Science & Technology Security | ● | ● | |
| Operational Safety | ● | ● | ● |
| Public Safety | ● | ● | |
| Consistency | 0.912 | 0.875 | 0.863 |
| Raw Coverage | 0.301 | 0.243 | 0.210 |
| Unique Coverage | 0.115 | 0.084 | 0.078 |
| Overall Consistency | 0.883 | ||
| Overall Coverage | 0.691 | ||
| Named Pattern | Technology-Foundation & Safety-Core | Scenario-Expansion & Operation-Safeguard | Ecology-Synergy & Multidimensional Governance |
Note: ● = core condition present; ● = peripheral condition present; ⊗ = peripheral condition absent; blank = condition irrelevant.
C1: Technology-Foundation & Safety-Core Pattern. This configuration is characterized by strong innovation-driven and science & technology security policies as core conditions, complemented by ecology optimization and operational safety policies. It is exemplified by Shanghai, Hubei, and Shaanxi. These provinces leverage their rich research bases and industrial heritage to build a solid technological foundation while ensuring that critical drone technology components remain under domestic control. The result is a high innovation advantage and safety advantage.
C2: Scenario-Expansion & Operation-Safeguard Pattern. Here, innovation-driven, application expansion, and operational safety policies are core conditions, while public safety policy is peripheral and ecology optimization policy is absent (negative peripheral). Typical provinces are Zhejiang and Sichuan. They excel in market advantage by aggressively promoting drone technology application in logistics, emergency response, and agriculture, while simultaneously strengthening airspace management and flight certification to ensure safe operations. The absence of ecology optimization policy suggests that these provinces prioritize application-led growth over ecosystem infrastructure at this stage.
C3: Ecology-Synergy & Multidimensional Governance Pattern. This configuration combines innovation-driven, ecology optimization, science & technology security, and public safety policies as core conditions, with operational safety as peripheral. Guangdong and Jiangsu are representative. They have established world-class drone technology ecosystems through industrial parks, investment funds, and standards organizations. At the same time, they deploy a comprehensive governance framework covering technology security (core chips), operational security (airworthiness certification), and public security (real-name registration and anti-black-flight measures). This integrated approach yields balanced strengths across all four dimensions of comprehensive advantage.
3.3 Configurational Analysis for Non-High-Level Comprehensive Advantage
We also analyzed configurations leading to the absence of high comprehensive advantage. Table 6 shows two distinct patterns (NC1 and NC2a/NC2b).
Table 6: Policy Configurations for Non-High-Level Comprehensive Advantage
| Condition | NC1 | NC2a | NC2b |
|---|---|---|---|
| Innovation-Driven | ⊗ | ⊗ | |
| Application Expansion | ● | ||
| Ecology Optimization | ⊗ | ⊗ | ⊗ |
| Science & Technology Security | ⊗ | ⊗ | ⊗ |
| Operational Safety | ⊗ | ⊗ | |
| Public Safety | ⊗ | ⊗ | |
| Consistency | 0.918 | 0.844 | 0.906 |
| Raw Coverage | 0.285 | 0.216 | 0.129 |
| Unique Coverage | 0.101 | 0.091 | 0.062 |
| Overall Consistency | 0.875 | ||
| Overall Coverage | 0.558 | ||
| Named Pattern | Innovation Ecology Imbalance | Safety Governance Deficiency | |
NC1: Innovation Ecology Imbalance. This pattern features the presence of application expansion policy but the absence of innovation-driven, ecology optimization, and science & technology security policies. Provinces like Fujian, Gansu, and Ningxia exhibit this profile. They attempt to stimulate market demand through subsidies and pilot projects, but without sufficient investment in R&D, infrastructure, and technology security, the drone technology industry lacks the foundational innovation ecosystem to sustain long-term growth. Consequently, even if market advantage appears moderate, the overall comprehensive advantage remains low.
NC2a and NC2b: Safety Governance Deficiency. These two configurations are characterized by the comprehensive absence of all safety-oriented policies (science & technology security, operational safety, and public safety). The lack of safety governance creates vulnerabilities such as unauthorized flights, airspace conflicts, and public incidents, which erode trust and hinder industrial development. Provinces such as Heilongjiang, Guizhou, and Xinjiang fall into this category. They have invested in drone technology applications (e.g., agricultural spraying) but neglected the corresponding regulatory and security frameworks, leading to low safety advantage and, consequently, low comprehensive advantage.
3.4 Robustness Check
We performed two robustness tests. First, we shifted the crossover calibration anchor to the 45th and 55th percentiles; consistency and coverage values changed slightly but no configuration was altered. Second, we raised the consistency threshold from 0.80 to 0.90; the number of configurations remained the same, with four of the five configurations being clear subsets of the original solutions. These tests confirm the stability of our findings.
4. Discussion: Policy Effects and Configurational Model
4.1 Policy Complementary Effect
Innovation-driven policy and operational safety policy appear as core conditions in all three high-level configurations (C1, C2, C3). This pattern reveals a strong complementary effect between the two. Innovation-driven policy, through mechanisms such as R&D subsidies, talent programs, and collaborative platforms, builds resource-based dominant advantages in drone technology. Operational safety policy, through airspace management, certification systems, and network security regulations, establishes institution-based dominant advantages. When both are present, they interact symbiotically: innovation provides the technological solutions for safer operations (e.g., anti-jamming navigation), while safety regulations create a stable environment that encourages further innovation investment. This dual-wheel logic is the core engine of comprehensive advantage growth for the drone technology industry.
4.2 Policy Substitution Effect
The comparison between C1/C3 and C2 highlights a substitution effect. In C1 and C3, the combination of ecology optimization and science & technology security policies is crucial, whereas in C2, these are replaced by application expansion policy. The substitution is contingent on the implementation context. In provinces with abundant research resources and established industrial bases (e.g., Hubei), enhancing the innovation ecosystem and technology security is more effective. In provinces with vibrant private capital and diverse application scenarios (e.g., Zhejiang), deploying application expansion policies that accelerate market uptake is a better fit. Thus, the same overall outcome of high comprehensive advantage can be achieved through different policy recipes, depending on local resource endowments and institutional conditions.
4.3 Multiplier and Bottleneck Effects
Innovation-driven policy emerges as a lever policy: its presence in all high-level configurations and its core status indicate that it amplifies the effectiveness of other instruments. This multiplier effect can be expressed as:
$$\text{Comprehensive Advantage} \propto f(\text{Innovation-Driven}) \times g(\text{Other Policies})$$
Conversely, science & technology security policy acts as a shortboard or bottleneck policy. In non-high configurations (NC1, NC2a, NC2b), the absence of this policy is consistently present as a core missing condition. Without securing the supply chain of critical drone technology components (e.g., chips, sensors) and protecting intellectual property, even strong innovation or application efforts are vulnerable to external disruptions. The bottleneck effect means that the weakest policy in the mix determines the upper limit of comprehensive advantage. This finding underscores the strategic importance of “ensuring both development and safety” – specifically, that technology security is not merely a background condition but a direct enabler of sustained advantage in the drone technology industry.
4.4 Proposed Configurational Model
Based on the empirical results, we construct a configurational model depicting how the Development–Safety policy mix shapes the comprehensive advantage of the drone technology industry (Figure 1 conceptual). The model integrates three configurational pathways, each characterized by a distinct combination of policy instruments, implementation contexts, and emergent advantages. The model formalizes the following causal logic:
1. The policy mix simultaneously cultivates resource-based dominant advantages (through innovation-driven and ecology optimization policies) and institution-based dominant advantages (through operational and public safety policies).
2. These two types of dominant advantages are synthesized into dual-wheel core capabilities: innovation development capability and safety governance capability.
3. The interaction between capabilities and contextual factors (regional resource endowments, industrial base, market maturity) determines the specific profile of comprehensive advantage (innovation-dominant, market-dominant, or balanced).
4. Lever and bottleneck policies moderate the overall effect: innovation-driven policy multiplies the impact of other instruments, while the absence of science & technology security policy creates an insurmountable ceiling.
The generic equation for the configurational model can be expressed as:
$$CA = \alpha \cdot (D \otimes S) \cdot \beta(X) + \gamma(L) – \delta(B) + \epsilon$$
where \(D\) = development-oriented policy bundle, \(S\) = safety-oriented policy bundle, \(\otimes\) denotes configurational integration, \(X\) = contextual factors, \(L\) = lever policy (innovation-driven), \(B\) = bottleneck policy (science & technology security), and \(\alpha, \beta, \gamma, \delta\) are positive parameters, \(\epsilon\) is an error term.
5. Conclusions and Implications
5.1 Theoretical Contributions
This study makes several theoretical advances. First, it extends the policy mix literature by explicitly incorporating the development-safety duality into policy instrument classification and configurational analysis. Prior work often treated development and safety as separate domains; our integration reveals their interdependence in shaping industrial outcomes for the drone technology industry. Second, we advance comprehensive advantage theory by embedding safety and ecological dimensions into the framework and demonstrating how policy-driven “dual-wheel” capabilities (innovation and safety) co-evolve. Third, the identification of complementary, substitution, multiplier, and bottleneck effects provides a nuanced vocabulary for describing policy interaction mechanisms, moving beyond simple additive or multiplicative assumptions.
5.2 Practical Implications for the Drone Technology Industry
Central government: The strategic principle of “ensuring both development and safety” must be operationalized through a coherent policy framework that explicitly links innovation incentives with safety mandates. A national-level “15th Five-Year Plan” for the drone technology industry should establish cross-ministerial coordination mechanisms (e.g., among science & technology, industry, civil aviation, public security) to avoid policy fragmentation. Regular policy synergy audits can ensure that development-oriented and safety-oriented instruments remain balanced over time.
Provincial governments: Local policymakers should diagnose their own industrial context and select the configuration that best fits their resource profile. Regions with strong research bases but technology bottlenecks should adopt the Technology-Foundation & Safety-Core pattern (C1), investing heavily in innovation-driven and science & technology security policies. Regions with active markets and diverse applications but underdeveloped infrastructure should consider the Scenario-Expansion & Operation-Safeguard pattern (C2), prioritizing application pilots and operational safety frameworks. Regions with mature clusters and emerging governance challenges should implement the Ecology-Synergy & Multidimensional Governance pattern (C3). In all cases, ensuring the presence of innovation-driven policy (lever) and science & technology security policy (bottleneck) is critical: the former should be strengthened to generate multiplier effects, while the latter must never be neglected to avoid ceiling effects.
Enterprises: Drone technology companies should actively monitor policy trends and leverage available incentives. For example, R&D tax credits and talent subsidies can be used to accelerate drone technology breakthroughs, while safety-related grants can support the development of encrypted communication systems and anti-jamming navigation. By aligning corporate R&D strategies with both development and safety policy priorities, firms can turn policy support into competitive advantages and contribute to the healthy and orderly growth of the entire drone technology industry ecosystem.
5.3 Limitations and Future Research
Several limitations of this study point to directions for future research. First, our analysis is confined to the provincial level due to data availability; future studies could zoom into city-level or firm-level analyses using survey or interview data to uncover micro mechanisms of policy perception and implementation. Second, the policy quantification relies on manual coding and LDA topic modeling; future work could adopt more advanced natural language processing techniques (e.g., BERT-based embedding) to improve accuracy and reduce subjectivity. Third, our cross-sectional design captures static configurations; longitudinal data and dynamic QCA methods could reveal how policy mixes evolve over time and co-evolve with the comprehensive advantage of the drone technology industry. Finally, the generalizability of our configurational model to other emerging industries (e.g., autonomous vehicles, smart energy) warrants further investigation.