A package for conducting event studies
Project description
Event Study Toolkit
event-study-toolkit is a package designed for economists and researchers to conduct event studies on stocks and bond.
Overview
The Event Study Toolkit provides a Python framework to perform event studies. Event studies are empirical investigations that typically focus on the effects of specific events on the stock and/or bond prices of firms. To learn more about event studies, please see the excellent resource by the folks at Eventstudytools
Leveraging the Pandas functionality, this package allows one to design an event study and calculate the relevent statistics. With this toolkit, researchers can measure the impact of specific events on the stock and/or bond prices of involved firms, control for market movements, and perform various statistical tests. Moreover, this package provides access to multiple statistical tests, both parametric and nonparametric for full sample and group level tests.
Features
- Data Preparation: Easily load and preprocess your data for analysis.
- Model Estimation: Fit your data to different models, both standard and custom, in order to estimate normal returns. Standard models include Market, CAPM, and Fama-French (column names must match up)
- Prediction: Predict returns during the event period.
- Abnormal Returns: Calculate the abnormal returns, which are the actual returns minus the predicted returns.
- Cumulative Abnormal Returns: Group and sum abnormal returns to get cumulative values.
- Test Statistics: Obtain various test statistics on both a full-sample and group-level basis.
- Rank-based Tests: Implement rank-based tests, such as the Generalized Rank Test and the Wilcoxon Rank-Sum Test.
Significance Tests Available for Use
Full Sample
- T Test: parametric test
- Sign Test: nonparametric test, Arnold Richard Cowan (1992)
- Patell Test: nonparametric test, James M. Patell (1976)
- Genaralized Sign test: nonparametric test, Arnold Richard Cowan (1992)
Group Level
- T Test: parametric test
- Sign Test: nonparametric test, Arnold Richard Cowan (1992)
- Patell Test: nonparametric test, James M. Patell (1976)
- Genaralized Sign test: nonparametric test, Arnold Richard Cowan (1992)
- Generalized Rank Z test: nonparametric test, Kolari and Pynnönen (2011)
- Wilcoxon signed-rank Test: nonparametric test, Frank Wilcoxon (1945)
Usage
Below are brief descriptions of the primary functions and their purposes:
-
fitModel(modelType)
: This function fits the data to the provided model type to estimate normal returns. -
getPredictedReturns(modelType)
: Retrieves predicted returns for all groups. -
getAbnormalReturns(modelType)
: Calculates abnormal returns for all data. -
getCARS(modelType)
: Returns the Cumulative Abnormal Returns (CARs) by summing abnormal returns for specific groups, as well as, other statistics. -
getFullSampleTestStatistic(modelType)
: Returns full sample test statistics for the data. -
getGroupLevelTestStatistics(modelType, GRP)
: Provides test statistics based on groupings specified by the user. -
getGRANK(modelType, GRP)
: Implements the Generalized Rank Z Test for the specified group. -
getWilcoxon(modelType, GRP)
: Implements the Wilcoxon Rank-Sum Test for the specified group.
Getting Started
See the following link containing an example notebook: Tutorial
-
Installation: Ensure you have all the required libraries installed, pip should handle this. This toolkit heavily relies on libraries such as pandas, numpy, and scipy.
-
Data Preparation:
- Prepare your data in a suitable format. Ensure you have columns for unique identifiers, return data, and event dates.
- Load your data into the toolkit using the available methods.
-
Model Estimation & Prediction:
- Choose the model type you wish to use. This toolkit allows for standard models and also offers flexibility for custom models.
- Predict the returns during the event period using the estimated models.
-
Analysis:
- Compute abnormal and cumulative abnormal returns.
- Obtain various test statistics, both on a full-sample and group-level basis.
-
Advanced Testing: If you need more advanced testing, make use of rank-based tests available in the toolkit.
Contribution
Feel free to contribute to this project by submitting pull requests or raising issues on the project's GitHub page.
License
This project is licensed under the MIT License. Refer to the LICENSE
file for more details.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for event_study_toolkit-1.0.0.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 98f772fd9eed6a2ad6adc9aa865e1805f1aacabd22666dada053fbc3179394ba |
|
MD5 | cbbd276b2063607ee942a54839ca45fd |
|
BLAKE2b-256 | 7c40d76c27cc213bc726e2926f0a2452cd3a85ace1bacadc5f333584080f2d1c |
Hashes for event_study_toolkit-1.0.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | fc83023d4b4bcf9e03aa8be4b76f1039a2d381065246d926195bc30ada31cd0d |
|
MD5 | eba32518ce78561989a0a23909131490 |
|
BLAKE2b-256 | ee2c984b6cdb401be4ce98bc454db436a90de3cde104a11af56abe71b08fa7da |