BIST market power firm value dataset.
Description
This dataset contains the firm-level panel used in the study "Market Power, Pricing Power and Firm Value: Evidence from Borsa Istanbul under High Inflation". It is an unbalanced firm-year panel of 460 non-financial firms listed on Borsa Istanbul, covering fiscal years 2014–2025 (4,338 firm-year observations). Market-based variables are available for 426 firms (3,573 firm-years). Data sources: Annual consolidated financial statements and year-end market data (last trading day of December) were retrieved from İş Yatırım Menkul Değerler A.Ş. (standard industrial template). Banks, insurance, leasing and factoring companies, brokerage houses, asset management companies, real estate and other investment trusts, holding companies and sports clubs are excluded. Firm-years with zero or negative sales or total assets are also excluded. Variables: The file provides financial statement items and market data, together with constructed variables: Pricing power: gross margin and the Lerner index. Within-industry position and concentration: market share, Herfindahl–Hirschman index. Firm value: Tobin's q, market-to-book and their logarithms. Profitability risk: ROA volatility, the standard deviation of ROA over the four-year window ending in year t. Controls: size, leverage, ROA, tangibility, sales growth. Indicators: high-inflation period (2022–2025) and IAS 29 reporting (2023 onwards). The workbook contains five sheets: README: dataset description and usage notes. Data: 42 variables. Variables: definitions and source item codes. Descriptive_Statistics: summary statistics. Sector_Year: sector-year concentration and margins. Notes for users: All values are reported as raw (non-winsorized) figures. In the study, ratio variables are winsorized at the 1st and 99th percentiles before estimation. From fiscal year 2023, firms report under IAS 29 (Financial Reporting in Hyperinflationary Economies), so monetary amounts from 2023 onwards are restated and not directly comparable with earlier years. Ratio variables and year fixed effects are recommended. Monetary values are in Turkish lira. Market share and HHI are computed from listed sample firms only and measure within-exchange concentration. Sales growth is not computed for 2023 because of the accounting transition. The dataset can be used to replicate the study's results and for further research on product market competition, firm value and inflation accounting in emerging markets.
Files
Steps to reproduce
Data collection: Annual consolidated financial statements (standard industrial template) for all firms listed on Borsa Istanbul were retrieved from İş Yatırım Menkul Değerler A.Ş. (www.isyatirim.com.tr) for fiscal years 2014–2025. Year-end closing prices, market capitalization and USD/TRY rates were retrieved from İş Yatırım historical stock data for the last trading day of December of each year. Sample construction: Firms in financial and related sectors were excluded: banks, insurance, leasing and factoring companies, brokerage houses, asset management companies, real estate and other investment trusts, holding companies and sports clubs. Firm-years with zero or negative sales or total assets were removed. Expense items, reported as negative values in the source, were converted to positive values. Variable construction: Variables were computed as defined in the "Variables" sheet: Gross margin = (sales − cost of sales) / sales. Lerner index = (sales − cost of sales − selling, marketing and general administrative expenses) / sales. Market share = firm sales / total sales of sample firms in the same sector-year; HHI = sum of squared market shares. Tobin's q = (market capitalization + total liabilities) / total assets. ROA volatility = standard deviation of ROA over the four-year window ending in year t (minimum three observations). Sales growth is set to missing for 2023 because of the IAS 29 transition. Winsorization: Ratio variables were winsorized at the 1st and 99th percentiles before estimation. Values in the Data sheet are raw and non-winsorized. Estimation: Panel regressions were estimated in Python with the linearmodels package (PanelOLS). The dependent variable is ln Tobin's q, regressed on one-year lagged gross margin, Lerner index, market share and HHI. Controls are size, leverage, tangibility, sales growth and ROA. All models include firm and year fixed effects, with standard errors clustered by firm. Robustness tests use two-way (firm and year) clustering. Inflation analysis: Market power measures were interacted with the high-inflation indicator (2022–2025). Models were also estimated separately for 2014–2021 and 2022–2025, and separately before and after the IAS 29 transition (2023). Robustness: Models were re-estimated with ln market-to-book as the dependent variable, with only firms listed since 2014, and with continuing firms only. Following these steps with the Data sheet reproduces the tables reported in the article.
Institutions
- Adnan Menderes UniversityAydın, Aydin