You are here

Question #21: VOLUME EVENT STUDY

I would like to use EventStudyTools for volume eventstudy.

The market data required "mean of log percentage of trading volume of index".

Could you advise me how to derived the value from raw data "volume of index".

thank you

Answer by Dr. Simon Müller:
There is no easy way to get the volume of an index. You have to get the volumes of all companies from that index and then merge it into one variable.
Question #20: Should I use Market Model( OLS ) or Market Model ( Garch)?

Dear Professor,

I conduct ARCH effect tests on the firms in my samples, and find that around half of the firms has significant ARCH effect, while the other half has no ARCH effect.

I wonder which model I should use. The Market Model or the GARCH Model?

Best wishes,
Chelsea

Answer by Dr. Simon Müller:
Hi. I would generally advice to use the GARCH model, as this model can handle volatility changes.
Question #19: Multiple Event Windows

How to prepare the request sheet when looking at more than 1 event window? The sample request sheet is only calculating 1 event window.

Eric Romero

Answer by Dr. Markus Schimmer:
You treat this second event window as an individual observation - meaning, you add another line to the request sheet and calculate the other event window. 
Question #18: Alfa Beta in aARC

In have one question what are alfa and beta in Analysis Report. I know what they mean if I use Market Model

Equation:
http://image.slidesharecdn.com/l2flashcardsportfoliomanagement-ss18-1312...

BUT

I don't know what Alfa and Beta are in FamaFrench 3 Factor Model

Equation
http://3.bp.blogspot.com/-dqWugq0Wotw/Ut8PnxHP-6I/AAAAAAAAAAA/d6_cqU81vt...

Could you help me and explain this?

John

Answer by Dr. Simon Müller:
Hi. You may write the model as: R_it = alpha + beta * R_im + s_i * SMB_t + h_i + epsilon_it. Then it should be clear what alpha and beta is. Simon
Question #17: AR CAR signifiance tests

Hi, I'am Eray. I'am phd student.

My thesis field is price manipulation. Manipulation period is more than one day. So, event date is (5, 20. 45, 100, 250 etc). For example, a company's manipulation period is start 03.08.2000 and end 14.12.2000. (03.08.2000-14.12.2000) As seen 95 day manipulation period.

I want to calculate each company for pre-manipulation, manipulation and post-manipulation period cumulative abnormal return. My estimation period, before 03.08.2000 100 trading day.

Estimation period pre-man manip period post-manip
----------------------------///-------------------///--------------------////----------------------
100 trading day -30 day 95 day + 30 gün

What can I do? I want to buy ESM. Can I calculate?

Answer by Dr. Simon Müller:
ESM? You can do this with our tool for free (just one citation in you thesis). Please contact me per email (info@muon-stat.com).
Question #16: EVENT STUDY

Hi, I have a question about using free event study which is available on the website. the problem is that when I use a credit rating base index as the market price and I have 5 broad categories for the market price in each day.My problem is that I don't know how to list the five market price for each date in the Market data file and how to clarify it clarify my firm dataset and request file? I cannot understand what is the variable group in request dataset.
I already have the file for abnormal return and I just want to use this file to calculate t statistic, is there any simple way for that?
could you please help me to solve this issue?

Answer by Dr. Simon Müller:
Both request data files are in the long format as described on http://www.eventstudytools.com/instructions-axc. So firms are identified by the first column, dates by the second, and on the third you put the closing price. For the market data you have in the first column the market index indentifier, on the second the date and on the third the closing price. You can calculate test statistics with our free web tool.
Question #15: Application Error

Hi

I try to run your AR calculator with your sample data. It doesn't work: I got an Application error message ???

Philippe

Answer by Dr. Simon Müller:
Possibly your data is wrong formatted.
Question #14: long term event study analysis

I want to measure long term performance (after 1 year, 3 years and 5 years) of IPO for 400 companies between 2000 to 2010. I want to use BHAR to measure the performance. but I don't know how prepare input data to be used in STATA. I have data in table format in excel as follows:
1. daily market return between 2000 to 2010
2. daily company return between 2000 to 2010
3. IPO listing for each company.

the questions is:
1. how can I prepare my data as you suggested using Comma-separate-value because I have up to 400 companies to go one by one?
2. how can I start my analysis?

Answer by Dr. Simon Müller:
You may use the reshape procedure in STATA. You find an example ont the idre UCLA site: http://www.ats.ucla.edu/stat/stata/modules/reshapel.htm
Question #13: Event Study on CDS

Hi everyone,

I want to analyze abnormal Returns of daily CDS Prices over the past 10 years, especially at certain Events. The CDS are still collected via Bloomberg (different maturities, Senior and Sub CDS, for nearly 100 European Banks). For the estimated return I'd take an iTraxx CDS index.

Would you consider to analyze analogous to your stock return Event study examples? I doubt, that the market model will fit?

Maybe I can upload the dataset for analyzing?

Many thanks in advance,
Sascha

Answer by Dr. Simon Müller:
Feel free to analyze your data with our tools. They are developed for that. If you get into trouble, please contact us per email (info@muon-stat.com).
Question #12: Competitor Portfolio

I am examining an individual event. According to your recommendations a parametric t-test (Nr. 1 on the website) is applicable. Besides the effects this event has on the stock returns of the company responsible for the event, I'd also like to examine the effects on the firm's competitors. Treating individual competitor stocks as independet is almost surely inappropriate, since I would expect a fair amount of correlation among competitor retruns. To adjust for this possibility, I create a portfolio composed of all relevant competitors and calculated the abnormal returns of this portfolio.
1.) Can I apply the same test statistic to test for significant AR of the competitor portfolio as I do for the individual focal firm?
2.) Is there a way to avoid portfolio creation that still yields statistically meaningfull/ unbiased results? Which test would be applicable in that case?

Answer by Dr. Markus Schimmer:
Since you do a "single event"-event study, you can use the grouping variable available in our research app to group all competitors and keep the other firm separate. The tests statistics will consider the grouping as described in the test statistics section. 

Pages