Christopher S. Sutherland, "Forward Guidance and Expectation Formation: A Narrative Approach", Journal of Applied Econometrics, Vol. 38, No. 2, 2023, pp. 222-241. --------------------- SOFTWARE REQUIREMENTS --------------------- Stata (code was run with StataMP 17) Some of the tables rely upon the following two nonstandard Stata commands, so researchers seeking to replicate the tables in this paper should first install those commands using the following commands: ssc install xtabond2 ssc install coefplot All files are zipped in cs-files.zip. ------------------------- OVERVIEW OF DATA AND CODE ------------------------- 1. Consensus Economics data, Excel files (proprietary) 2. Central Bank Communication Panel Data, Excel file (authour-generated) 3. Firm Mnemonics, Excel file 4. Data Import, Cleaning, and Consolidation .do file 5. Empirical analysis .do file -------------------------------------- DETAILS OF NOT PUBLICLY AVAILABLE DATA -------------------------------------- 1. Consensus Economics data These data require a license from Consensus Economics. The Consensus Economics variables are described in the column headers of each month's regional forecast Excel file. More details about the variables are given in a guidebook that is provided by Consensus Economics. You can also find information about each variable in the monthly PDF summaries of the data that accompany each monthly, regional Excel file. Note, however, that this paper does not use all of those variables. It only uses a subset. All of those variables are described in the paper in detail. The Consensus Economics forecasts files are organized into many subfolders. The organization of those subfolders is based upon the manner in which Consensus Economics provides the data. Someone who has a license to Consensus Economics data would probably have the data organized in the very same manner. The precise details of how the files are organized can be found in the Stata .do file called "Data Import, Cleaning, and Consolidation.do". For replication, one would simply need to change the file pathway names to match how the files are stored on their own drives. --------------------------- DETAILS OF CONSTRUCTED DATA --------------------------- 2. Central Bank Communication Panel Data.xlsx This Excel file contains all of the central bank data collected by the author. The forward guidance, quantitative easing, effective lower bound, policy rate, central bank projections, media article links, and associated metadata information is all contained in the file. Each variable is described in a column header that sits directly above the row of variable names that are directly imported into and used in all empirical analysis. There are comments and hyperlinks contained all throughout the file to help researchers to understand how the data were constructed and to provide additional information on variables if needed. The use of cell shading is also employed to denote blocks of data that are related to each other to make the data easier to read. Both the paper and its accompanying online appendix provide detailed descriptions of all of the relevant data in this file. There are 5152 rows and the columns used directly in the empirical analysis go from column A to column BK. 3. All Mnemonics.xlsx This Excel file provides the information needed to track individual forecasters over time. Each variable is described in a self descriptive column header that also serves as the variable name that is directly imported into the data import and cleaning file described immediately below. This file tracks firms through mergers and acquisitions across time by country. There are two separate tabs: the CF tab includes the European, Scandinavian, and North American countries; the AP tab includes Australia and New Zealand. This file allowed for the consolidation of many individual forecaster time series that were seemingly separate into fewer, harmonized individual forecaster time series. This decreased the total number of firms in the sample data and increased the number of observations per firm. Essentially, this file is merged with the Consensus Economics forecasts data in the .do file below and then not directly used again. ------------------- DESCRIPTION OF CODE ------------------- 4. Data Import, Cleaning, and Consolidation.do This Stata .do file imports, cleans, and consolidates all of the data. First, the Consensus Economics data are imported and cleaned directly from the proprietary Consensus Economics Excel files referenced above. One additional file that researchers will need for replication is the All Mnemonics.xlsx file described immediately above. Second, the Consensus Economics data are consolidated into one .dta file. Next, the Consensus Economics data are merged with the Central Bank Communication Panel Data.xlsx. The panel data set was designed specifically so that it could be merged appropriately with the Consensus Economics data. Some file pathways correspond to the data appendix folder: those are files that do not include proprietary Consensus Economics data. Other file pathways correspond to the authour's personal drive: those are files that do include proprietary Consensus Economics data and therefore cannot be disseminated publicly. However, a researcher interested in replication could amend the latter pathways for replication once they have acquired the Consensus Economics data. Finally, many variables needed for empirical analysis are created and some minor data cleaning is performed, which is described by notes contained in the .do file. The file includes comments throughout to help researchers understand the purpose of each block of code. 5. Empirical analysis.do The Empirical analysis.do file provides all of the code for the tables from the main paper (first), tables from the online appendix (second), and figures from both the paper and online appendix (third). Each table and figure has a description as well as table notes to help researchers follow the various steps of the analysis.