This paper examines the co-evolution of technological diversification and international collaboration, and how they affect the intensity of innovation in a country. A two-step analysis is applied on a global panel dataset consisting of patents and macroeconomic data for 54 countries, covering a period of 40 years. First, the co-evolution patterns and characteristics of diversification and collaboration are explored. Then, a series of econometric techniques are employed in an attempt to explain the observed patterns. This step involves conducting the Toda–Yamamoto and Dolado–Lutkepohl (TYDL) Granger causality test to analyze the directions of the causal effects of technological diversification and international collaboration on innovation. Such version of the Granger causality test is valid and consistent regardless of whether a series is stationary at level, first order or second order difference; and non-cointegrated or cointegrated of any arbitrary order. In addition, reduced form vector autoregression (VAR) models are estimated to determine the scale of the impacts of both diversification and collaboration on a country's innovation performance. Our empirical results show that there is a bidirectional causality between technological diversification and innovation. This result is robust across different time periods and groups of countries. Furthermore, international collaboration is found to positively influence the intensity of innovation in a country while technological diversification has a negative effect.