"In recent years, both academia and industry have paid significant attention to economic policy uncertainty. Indeed, the greatest certainty of this era is its abundance of uncertainties. The primary metric currently used to measure economic policy uncertainty is the EPU index developed by BBD (2016). However, each region or country has only one EPU index at each time point, making it indistinguishable from time fixed effects in econometric analyses. Moreover, it cannot capture the heterogeneity in policy uncertainty perceptions across different firms."
— WeChat Official Account: Professor Nie Huihua
In 2020, Professors Nie Huihua, Ruan Rui, and Shen Ji published the paper Corporate Uncertainty Perception, Investment Decisions, and Financial Asset Allocation in The World Economy. This study proposed a methodology for calculating firm-level economic policy uncertainty indices and utilized the data to analyze how uncertainty perceptions affect corporate investment and financing decisions. The paper has garnered widespread academic attention since its publication.
To advance research on economic policy uncertainty and reduce unnecessary time costs, Professor Nie Huihua and his research team have made the Firm-level Economic Policy Uncertainty Perception Index publicly available for academic use. With authorization, CnOpenData has incorporated this dataset into the public data section of its official website, facilitating access for researchers.
To download this dataset or obtain additional information, please visit White Shark Online.
Database Usage Guide
Methodology
The economic policy uncertainty metric in this study is extracted from annual reports of listed companies using text mining methods.
With advancements in computational technologies, incorporating unstructured data such as text into corporate finance research has become increasingly common (Tetlock, 2007; Li, 2008; Tetlock et al., 2008; Loughran and McDonald, 2014, 2016). Following the approaches of Baker et al. (2016) and Hassan et al. (2019), this study employs a "keyword-based method" to screen specific textual content. If specific terms appear in a text segment, it is identified as expressing particular meanings. A sentence is classified as reflecting economic policy uncertainty if it contains both "policy-related terms" and "uncertainty-related terms."
The specific methodology is as follows:
- Convert PDF files of annual reports into text format using file conversion tools, and extract the "Management Discussion and Analysis" (MD&A) section using regular expressions (正则表达式). Remove all numbers, English letters, punctuation (except periods), and special symbols.
- Segment the MD&A text into sentences using Chinese periods as delimiters. Given Chinese linguistic conventions, sentences serve as the basic analytical unit. For firm i in year t, let S denote the total number of sentences in the MD&A section. Using Python's jieba module (结巴分词模块), perform word segmentation (分词) on each sentence while removing stopwords (停用词). To minimize ambiguity, a custom dictionary is defined, including full and abbreviated names of all A-share listed companies, accounting terms, uncertainty-related words, and government/policy-related terms.
- After segmentation, each sentence becomes a sequence of words. For each sentence (s), perform the following operations:
- Identify sentences containing uncertainty-related terms as uncertainty sentences.
- Classify sentences containing both policy-related terms and uncertainty-related terms as policy uncertainty sentences (P).
- Measure firm-level economic policy uncertainty (FEPU) as the ratio of uncertainty-related words (n) in policy uncertainty sentences to the total words (N) in the MD&A section.
References
Note: Any use of the Firm-level Economic Policy Uncertainty Perception Index must cite the following source:
- Nie, Huihua; Ruan, Rui; Shen, Ji. 2020. "Corporate Uncertainty Perception, Investment Decisions, and Financial Asset Allocation." The World Economy (《世界经济》), No. 6, pp. 77-98.
Data Update Frequency
Updated irregularly