Firm-Level Uncertainty Instruments
Download Firm-Level Uncertainty Instruments Data:
1) Download Data updated until 2023
2) Original paper data and code (data ending in 2019, posted on the Harvard Dataverse)
In The Finance Uncertainty Multiplier, published in the Journal of Political Economy in 2024, Alfaro, Bloom, and Lin (ABL) develop firm-level instruments to address endogeneity concerns in measuring the causal impact of uncertainty shocks. Their identification strategy exploits cross-industry non-directional exposure to 9 different sources of aggregate uncertainty to identify exogenous variation in firm-level volatility.
The firm-level uncertainty instruments include the second moment exposures to 7 bilateral exchange rates with the US Dollar considered as "Major Currencies" by the US Fed Board (Euro, Canadian Dollar, Japanese Yen, British Pound, Swiss Franc, Australian Dollar, and Swedish Krona), crude-oil prices, and Economic Policy Uncertainty. The dataset contains all 9 second-moment instruments and their respective first-moment controls from 1992. The firm sample is for publicly listed US firms with data available in Compustat and CRSP. ABL's dataset provides different firm identifiers, including Compustat Gvkey, CRSP Permno and Permco, CUSIP, and TICKER codes. Codes for 2-digit Standard Industrial Classifications are also included. The instruments can also be used as industry-level instruments for uncertainty that can be applied to, for example, industry, private firm, or establishment datasets. The set of instruments are constructed using the realized volatilities of the 9 different aggregate sources of uncertainty.
ABL's dataset includes 2 firm-level proxies for uncertainty: 1) realized stock return volatility from CRSP and 2) forward-365-day option-implied volatility from OptionMetrics, where annual firm realized volatility is estimated as the annualized 12-month standard deviation of daily CRSP returns and implied volatility is a forward 365-day implied volatility measure built from of a mix of put and call at-the-money options. For both the realized and option-implied volatilities ABL provide: 1) annual growth rates from one year to the next (i.e., shocks to firm-level volatilities as described in the baseline measures in the paper) and 2) levels. The latter is provided for researchers wanting the level of firm volatility instead of the shock. The instruments can be used for either the shock or level of firm volatility measures.
The data also includes firm-level measures of first-moment effects: 1) 12-month compounded stock return and 2) Tobin's Q. These firm-level variables can serve as additional controls to further tease out the effects of 2nd-moment firm volatility shocks from correlated 1st moment effects.
UPDATE November 2024: We have now extended the data until year 2023 (previously ending in 2019 per the sample appearing in the published paper). We expect to provide annual updates to this data in November, at which point all the required data from the previous year has been made available (and revised for accuracy in some cases) by the raw data source providers. Due to oil prices dropping to negative values for the first time in history in 2020 (see, for example, the news article from the BBC here), and its non-trivial consequences for the computation of measures involving changes in oil prices and their volatility (e.g., mathematically undefined numbers), the oil volatility information in calendar year 2020 is not generated when constructing the oil IV and its 1st moment. Given the nature of the oil IV, which as described in the paper is a "volatility shock" involving the growth rate in volatility from one year to the next, this event affects the years 2020 and 2021. Therefore, in the extended dataset, the oil IV and its 1st moment take missing values in those 2 years.
You can find the full code and data replicating all tables and figures of the original paper in the Harvard Dataverse: https://doi.org/10.7910/DVN/0IRS7Z. The published paper is available in Open Access, no paywall, in the JPE publishers' website: https://www.journals.uchicago.edu/doi/10.1086/726230. Using the extended data through 2023, we have recreated all main empirical Tables (2 to 5) and Figure 2 in the updated results available here.
The dataset should be cited as follows: Alfaro, Iván, Nicholas Bloom, and Xiaoji Lin, The Finance Uncertainty Multiplier, Journal of Political Economy, Volume 132, Number 2, February 2024. In case of questions, please contact Iván Alfaro.