Client Background
Client: A leading business school worldwide
Industry Type: R&D
Services: Research & Innovation
Organization Size: 10000+
Project Objective
Do regression modeling on the data provided, cross-country determinants of Key Audit Matters (KAMs) and its usefulness to Investors and Debt Market Participants
Project Description
USEFULNESS TO EQUITY MARKETS
- Examine whether the number and content of KAMs varies with country-level determinants.
- Explore whether the usefulness of KAMs to investors varies with country level variables such as type of law, enforcement etc.
- Examine whether adoption of the expanded auditor’s report associated with change in audit quality? Examine whether the content in the auditor’s report improves audit quality. Does this vary across countries?
- Is the adoption of the expanded auditor’s report associated with a change in audit fees?
- Explore whether the content of auditor report moderates the usefulness of KAMs to investors (also check by country-level variables)
- Can the number and content of KAMs be used to predict restatements (2017 onwards)?
In order to do the analysis and hypothesis testing, create a mapping to divide the audits into sub category and category according to the sub category and category provided in the question document. Clean the data before proceeding and calculate variables ABRET, ABVOL, CAR and CAAR according to the description provided.
Our Solution
Created a mapping for key audit matters to label the sub category and category of the audit for further analysis and merging with other datasets on the basis of the unique keys to create a final dataset we can use to calculate and do the hypothesis testing.
Calculation of variable ABRET and ABVOL is proceeded by firstly arranging the data by unique key and then the date of the data to get the sorted data. Cleaning is done on the data by removing the repetitive entries from the dataset and then selected the data around the date for which the variable is to be calculated. Similarly, calculated ABVOL in which extracted the data around the annual report filing date and mean value for 40 days interval that ends 21 days before earning announcement dates.
Couldn’t proceed because dataset provided by the client was incomplete in order to calculate ABRET.
Language/techniques used
R language to create mapping for the key audit matters and save data set for question 1.
Python pandas library to deal with dates and extract data around annual report filing date.
Skills used
Data mapping, data cleaning, data manipulation, debugging
Databases used
Key audit matter
GDP rule law
Audit fee
Trading data
Earning date
Report filing date
What are the technical Challenges Faced during Project Execution
Dataset provided by the client was too big and made my system slow when the data is loaded in the environment. Too many datasets and variables made it bit difficult to understand and time taking.
How the Technical Challenges were Solved
Calculated the number of unique identifiers in the large dataset and sorted those. Then selected the data for 1 unique identifier and sorted dates for it and append it to the dataframe and saved group of such unique identifiers to reduce the size of the dataset and performed the calculations in loop.
To tackle the difficulty of understanding the data I made a document tracking all the columns or variables present in the data.
