''Endogenous Regime Identification in China: A Markov-Switching Analysis of Innovation and Productivity''
Description
The analysis uses annual time‑series data for China covering 1985–2023. The start year reflects the earliest availability of patent application data in the World Bank Development Indicators, while the endpoint corresponds to the most recent update of Penn World Table 11.0, which extends to 2023 (Feenstra et al., 2015). Variables are drawn from PWT 11.0, the World Bank, and CNIPA patent statistics . Growth rates are calculated as log differences × 100 (Solow, 1957), while ratios are taken directly from source data. Patent applications are defined as the sum of resident and non‑resident filings at CNIPA. For 1985–2021, data is from World Bank series (IP.PAT.RESD, IP.PAT.NRES); for 2022–2023, CNIPA statistics extend the series. A cross‑check for 2010–2021 confirms exact consistency between World Bank and CNIPA data. It is important to note due to space limitation this study does not include descriptive statistics .
Files
Steps to reproduce
The dataset employs annual macroeconomic and innovation indicators from Penn World Table version 11.0 and the World Bank Development Indicators, supplemented with patent statistics. Patent application data are compiled as the sum of resident and non‑resident filings. For 1985–2021, the series is drawn from World Bank indicators (IP.PAT.RESD, IP.PAT.NRES). Missing values for 2022–2023 are filled using official statistics from the China National Intellectual Property Administration (CNIPA). Consistency is ensured through cross‑checks: patent data for 2010–2021 were compared across World Bank and CNIPA sources, confirming alignment. This harmonization yields a continuous series of resident plus non‑resident patent applications, suitable for empirical analysis of innovation dynamics