Trademark filings and patent application count time series are structurally near-identical and cointegrated: Implications for studies in innovation

TL;DR

This study demonstrates that trademark filings and patent applications are structurally similar, cointegrated, and exhibit synchronized long-term dynamics, implying a close relationship that supports their joint use in innovation research.

Contribution

It provides empirical evidence of the structural and cointegration relationship between trademarks and patents over 40 years, extending the understanding of their interconnectedness in innovation studies.

Findings

Trademark and patent time series are structurally similar.

They are cointegrated, indicating a long-term equilibrium.

Structural breaks occur simultaneously, suggesting common external influences.

Abstract

Through time series analysis, this paper empirically explores, confirms and extends the trademark/patent inter-relationship as proposed in the normative intellectual-property (IP)-oriented Innovation Agenda view of the science and technology (S&T) firm. Beyond simple correlation, it is shown that trademark-filing (Trademarks) and patent-application counts (Patents) have similar (if not, identical) structural attributes (including similar distribution characteristics and seasonal variation, cross-wavelet synchronicity/coherency (short-term cross-periodicity) and structural breaks) and are cointegrated (integration order of 1) over a period of approximately 40 years (given the monthly observations). The existence of cointegration strongly suggests a "long-run" equilibrium between the two indices; that is, there is (are) exogenous force(s) restraining the two indices from diverging from one another. Structural breakpoints in the chrono-dynamics of the indices supports the existence of potentially similar exogeneous forces(s), as the break dates are simultaneous/near-simultaneous (Trademarks: 1987, 1993, 1999, 2005, 2011; Patents: 1988, 1994, 2000, and 2011). A discussion of potential triggers (affecting both time series) causing these breaks, and the concept of equilibrium in the context of these proxy measures are presented. The cointegration order and structural co-movements resemble other macro-economic variables, stoking the opportunity of using econometrics approaches to further analyze these data. As a corollary, this work further supports the inclusion of trademark analysis in innovation studies. Lastly, the data and corresponding analysis tools (R program) are presented as Supplementary Materials for reproducibility and convenience to conduct future work for interested readers.