Expert Dialog

Question #17: Degrees of freedom and Fama-French Three Factor Model

I was wondering as to how the formula for the standard deviation, the cross-sectional test (CAAR) and the Patell Z (CAAR) need to be adjusted if instead of the market model I would use a model with three factors and one constant?

Answer by Dr. Simon Müller:
The degree of freedom of a linear model is the number of observations (n) minus the number of independent variables (k) plus 1 (n - (k + 1)).
Question #16: short term event study

I am conducting a clinical event study on the IPO of Visa Inc and I was wondering how to adjust for thin trading in my alpha and beta estimates in the market model for calculating AR in the short term. I didn't understand the Scholes and Williams paper very well.

Answer by Dr. Simon Müller:
Dear Soheil, Our tool provides the Scholes/Wiliams model and it is free. Best, Simo
Question #15: short term event study

I am conducting a clinical event study on the IPO of Visa Inc and I was wondering how to adjust for thin trading in my alpha and beta estimates in the market model for calculating AR in the short term. I didn't understand the Scholes and Williams paper very well.

Answer by Dr. Simon Müller:
Our tool provides the Scholes/Wiliams model and it is free.
Question #14: Aggregation across time and firms in Event Study

Dear Sir,
I intend to use event study methodology for the purpose of my research for accessing the impact of quarterly financial results and stock price behavior for a sample of firms .
Kindly help me to understand as to how the aggregation across time and firms is done for the sample.
Data for 30 stocks and 600 quarters.
In case of clarification on the query please let me know.

Regds,
Nayan

Answer by Dr. Simon Müller:
Dear Nayan,please visit out test statistic page. On the top of the page you will find the aggregation formulas.Best,Simon
Question #13: Training workshop online or onsite

Hello sir/madam

I am writing to find out if you will offer any workshop so that I can have the opportunity to be trained using this important method for research purposes. If not, what is the best way to grasph this technically thoroughly?

Thank you very much for your help.

Best
Deli Yang

Answer by Dr. Simon Müller:
Dear Mrs. Yang,we do not provide online or onside tutorial at the moment. I may recommend you the excellent book: "Event Studies for Financial Research" from Kliger and Gurevich.Best,Simon
Question #12: CAR t-test

Dear Simon,

I'm a MSc student working on my final thesis. Could you please explain me why the denominator of the t-statistic for CAR is not divided by the square root of N?

Thank you in advance

Kind regards

Paolo

Answer by Dr. Simon Müller:
Hi Paolo,as we have in the nominator a sum and not a mean. Just try to calculate the expectation and variance of CAR_i, then you will see that CAR_i ~ N(0, L_2 sigma_i^2). As we estimate sigma_i^2 in the estimation window you get that CAR_i is t-distributed with M - 2 degrees of freedom (for the market model).Best,Simon
Question #11: Finding p-values

Hi!

How do you calculate the p-values? We use excel and look the t-statistic up in a normal distribution with a mean of 0 and a variance of 1 + cumulative = 0. Is that the right way to do so? And can we assume that all the test statistics should be looked up in a normal distribution and not a t-distribution?

Answer by Dr. Simon Müller:
Hi,on you test statistic page (http://www.eventstudytools.com/significance-tests) the brackets indicates which distribution to use for p value calculation, e.g. for the Patell test we have [4] Patell or Standardized Residual Test (Abbr.: Patell Z), so you use here z distribution (normal distribution with mean 0 and sd 1), a T indicates T distribution with degree of freedoms is depending on which model you use. Best, Simon
Question #10: IPO CAR

I'm trying to calculate Cumulative Abnormal Return for IPO firms after listing for 3 year trading period, by calculating monthly abnormal return and cumulating for the duration.

I went through the guides on CAR, but all of them cover event studies which include an estimation window. But as estimation windows is not available for IPO firms, I am confused how to calculate CAR. I have attached the formulas which I am considering to use.

Can you please help me to calculate CAR?

Answer by Dr. Simon Müller:
You may use the market adjusted model in your case. As our API uses an estimation window for calculating sigma it is not possible to solve your problem with our API. Best, Simon
Question #9: GRANK - T and GRANK - Z estimation

Hi, Firstly, I'm very impressed with the kind of event study calculation summary you have provided in your webpage. I have a question regarding the GRANK -T and GRANK - Z non parametric estimations. Are the GSAR ranked in ascending i.e. lowest rank for the lowest GSAR or vice versa descending order i.e. Highest rank for the lowest GSAR. I have checked other academic sources including Kolari and Pynnonen, (2011) but they don't seem to state it specifically. If I consider ranking in ascending order the T statistics are +ve for negative CAAR values whereas when ranked in decending order they are negative for -ve CAAR values. Looking forward to hear from you. Thanks for your time.

Regards,
John Pereira

Answer by Dr. Simon Müller:
Hello John Pereira,thank you for your question. This point is not really specified in the literatur of test statistics for event studies. In your question, you already answered it yourself. For consistency reasons (from a paractical view), it would be plausible to have a positive t statistic if the GSAR of the cumulative event day belongs to the top 50% in this sequence, vice versa. Theoretically, it does not matter if you order them in the ascending or descending way, the absolute t-value will be in both cases the same, just the sign changes (as you already noted).Best,Simon 
Question #8: Can we conduct event study using SPSS

Dear Sir,
I am doing my Ph.D research, where I require to conduct event study (here event study is to be conducted by combining on multiple events)
Events are identified in the data series (both upper tails and lower tails)
Can it be possible to conduct this type of event study by using SPSS?
Kindly help me in this regard.

Answer by Dr. Simon Müller:
Hello,It should be possible with some programming to conduct this type of event studies in SPSS. Best

Pages