We contribute to the literature on the assessment of innovation systems by relating the amount of inputs available to the system and its performance through the concept of returns to scale (increasing, constant or decreasing). We study to what extent the size or scale of innovation systems relates to their performance, which is estimated through frontier Data Envelopment Analysis-TOPSIS methods, which overcome several limitations of the standard DEA approach.

Using the same data provided by the European Innovation Scoreboard (EIS) for years 2010, 2013 and 2016, our results indicate that countries with a high innovation scale tend to overinvest in innovation inputs. This results into scale inefficiencies stemming from decreasing returns, leading to lower productivity levels. Thanks to DEA-TOPSIS we identify the best and worst performing innovation systems. This provides helpful information by setting suitable reference benchmarks for policy analysis and decision-making.

Our results question the current allocation of resources and call for a reconsideration of how innovation policies are designed in many European countries. We conclude that for the EIS to become a useful instrument for the definition of innovation policies, it should consider the nature of returns to scale. This would allow policymakers to identify problems and limitations related to the size of their respective innovation systems, and hence, design holistic innovation policies to act upon them.

我们通过规模报酬（增加、不变或减少）的概念将系统可用的投入量与其绩效联系起来，从而为有关创新系统评估的文献做出贡献。我们研究创新系统的规模或规模在多大程度上与其绩效相关，这是通过前沿数据包络分析-TOPSIS 方法估计的，该方法克服了标准 DEA 方法的几个局限性。

使用欧洲创新记分牌 ( EIS)在 2010、2013 和 2016 年提供的相同数据，我们的结果表明，创新规模高的国家倾向于对创新投入进行过度投资。这会导致收益减少导致规模效率低下，从而导致生产力水平降低。多亏了 DEA-TOPSIS，我们才能确定最佳和最差的创新系统。这通过为政策分析和决策制定合适的参考基准来提供有用的信息。

我们的结果质疑当前的资源分配，并呼吁重新考虑许多欧洲国家的创新政策是如何设计的。我们的结论是，要使EIS成为定义创新政策的有用工具，它应该考虑规模收益的性质。这将使政策制定者能够确定与其各自创新系统规模相关的问题和限制，从而设计整体创新政策来对其采取行动。