Impact of risk metrics on EMD: Data Files Guide / Matt Hamilton / Saunvit Pandya
Identifying Emerging Markets Bond Mispricing with Economic and Social Development Metrics
Guide to Data Files
- File: AshtaMaster Data[proprietary]
- The Master Data File is an excel workbook that contains the dependent and independent variables used in our analysis.
- The File is structured in three segments: Data Master, Dependent Variables and Independent Variables
- The Data Master Tab is in Purple and it is an aggregation of all of the metrics used in our analysis. The data is sorted by country, year, and category. A variety of Index Match excel functions pull the data from each individual worksheet. Data from this tab is the foundation of the regressions.
- Dependent variable data is to the right of the green dependent variable page separator tab. The individual tabs are color coded cream. There are three tabs that have different credit spreads from different data sources. In some instances the data has been normalized and each data sheet has the specific source and data normalization method mentioned.
- Independent variable data is to the right of the green independent variable page separator tab. Each independent variable data tab is color coded light green, with the exception of the CIA data tabs that are color coded light blue. The light green tabs are structured by category and have the specific source and normalization method mentioned. The light blue tabs are structured by year with multiple categories within each tab.
- File: AshtaRegressions
- Ashta regression analyses are split into multiple files. Each file pairs a dependent variable dataset (EMBI or CEMBI) with an independent variable dataset (WB, CIA, or ICRG). The pairing is reflected in the file’s name.
- Each file typically consists of four types of tabs: An overall model summary (usually titled “Models” or “Model Summary”) briefly noting what assumptions were used in each model, a data tab (either titled “Final Data” or the name of the dependent variable), model tabs, and a Model Summaries tab that features a comparison of all models (which is one of the automated outputs of the RegressIT program).
- To be able to replicate the results in the file, one must download and install the RegressIT program (formerly FSBForecast) from the Fuqua Software site. This program is the standard multivariate regression program mandated to Fuqua MBAs in the Forecasting elective course. Without the program installed, the Excel files will contain only the static version of the output. With the program installed, the user will be able to run additional regressions based on the model data.
- The program operates similarly to FSBStats. Named ranges are created from variables in the data tab. The program allows multivariate regressions to be run through a point and click interface. The interface, while simple in design and conception, is exceedingly powerful: it allows for time series components (lead and lagged), intercept and no intercept regressions, and operations upon the independent and dependent variables (including log-log transformations).
- The actual running of the regressions is fairly routine and standardized. However, the key is understanding the output and choosing parameters for consequent modeling. Our thought process on how we did this is not documented in the workbook itself but rather in the regression analysis portion of our writeup.
- File: Ashta Trading Strategy Data [proprietary]
- Ashta Trading Strategy Data is an excel workbook that contains the relevant pricing for the various 10Y Sovereign CDS (Source: Bloomberg) as well as the actual EMBI and modeled sovereign spreads over time. Modeled spreads were calculated from regression output coefficients and applied to normalized data inputs.
Page 1 of 3