The choice between building technology in-house and buying it from outside is often framed as a simple trade-off. This study, based on Chinese enterprise tax data from 2016 to 2020, shows that for marine enterprises, the reality is more strategic. Both independent R&D and external equipment procurement significantly improve total factor productivity, but their effectiveness is not universal. The research identifies a clear threshold effect: when a company's internal business diversification or external energy consumption crosses certain levels, the productivity returns from these investments shift downward. This is not an argument for one path over the other. It is a warning that context determines outcome.
What stands out is the finding on strategic differentiation. High-tech enterprises achieve the best results by pursuing a dual path, combining their own research with purchased equipment. Non-high-tech firms, however, face higher costs for internal innovation and are better served by focusing solely on external procurement. This is a practical, unsentimental conclusion. It moves the conversation away from romantic notions of self-reliance and toward a more calculated view of capability and resource constraints. For a sector like the marine economy, where capital is heavy and margins can be thin, this distinction matters. It also connects to broader questions about how we measure efficiency in coastal systems, as seen in Measuring Efficiency: Untangling Natural Resource Productivity in Coastal China, where resource productivity varies significantly across regions.
The policy implications are direct. Subsidies and support programs designed to boost marine innovation cannot be one-size-fits-all. A blanket push for R&D will not help a non-high-tech enterprise that lacks the internal capacity to absorb those investments productively. Conversely, restricting equipment procurement support could undermine the very firms that rely on it as their rational, dominant strategy. The study's use of a multiply robust causal identification model adds weight to these conclusions, but the underlying logic is clear: effective support must be calibrated to the technological baseline of the enterprise. This aligns with the broader evidence that digital infrastructure, such as Data Integration Drives Low-Carbon Fisheries in China, can reshape productivity outcomes, but only when deployed in alignment with existing capabilities.
Our take is that this research offers a useful corrective to the prevailing enthusiasm for indigenous innovation at all costs. It does not dismiss the value of R&D; it simply argues that for some enterprises, the cost structure makes external procurement the more efficient route. That is not a failure of ambition. It is a rational response to resource boundaries. The finding that these effects exhibit a non-linear, regime-switching character suggests that firms should periodically reassess their position relative to these thresholds, since a change in scale or energy intensity could alter the calculus. For policymakers, the practical question is whether current funding mechanisms can accommodate this level of differentiation. The specific detail to watch is whether future support policies begin to explicitly incorporate such threshold effects, or whether they continue to treat all marine enterprises as equally positioned to benefit from R&D incentives. The answer will determine whether these productivity gains are concentrated among the capable few or distributed more broadly across the sector.