Question

You work at an investment firm. Your colleague has an investment thesis that he presents at your firm's investment committee meeting. He has built a sophisticated model, based on a number of financial statement data items (e.g., accruals, interest payment coverage, various cash flow metrics, etc.) which predicts which firms will announce, within the next three months, that they are reducing (cutting) their dividends. He performs appropriate statistical tests and finds that his model works well-it successfully predicts these announcements significantly better than chance. He also performed an event study of firms that have announced dividend cuts in the past, and confirmed his expectation that, on average, a firm's common stock underperforms significantly around the moment such announcements are made. Clearly, your firm wants to "be out of (not own) any stock when an announcement about a dividend cut is announced. He proposes that his results support the use of his model as a "screen" for your firm's long-only products. That is, any stocks that have a "dividend cut" score above a certain threshold (indicating a relatively high probability of dividend cut) will not be allowed in the portfolio, since he expects them to underperform. a) Your colleague has made a serious (but common) mistake in his analysis. What is it? Don't speculate on mistakes he might have made based on issues that haven't been mentioned (e.g., anything to do with risk costs, data errors, etc.). Describe the mistake that you know he made based just on the description given in the question b) What additional analysis could your colleague perform in order to test his idea correctly? Note that there is more than one correct answer

Answer 1

Answer :-

( a ) :-

• The principal botch is testing mistake. He has just considered firms that have reported profit cut already for his preparation informational index.
• This makes an inclination towards his speculation thus high connection between's profit cut and different factors.
• So the informational index ought to be substantial that incorporates different firms with various conditions for a progressively irregular examining.
• This can likewise prompt over fitting the model.

( b ) :-

• Extra examination can be taking many years of information of firms from various enterprises with an extensive number of factors.
• The informational index must be extremely expansive to evacuate any opportunity of under-scaling or over-scaling.
• This will give a bin of firms that are irregular and give progressively sensible examination.
• Prescient examination models like strategic relapse can be utilized to investigate this informational collection.
• There can be factors that have high relationship among's them and must be evacuated for more precision.
• The models constructed must be tried on some new organization or information of organizations that was excluded in the preparation information and check the adequacy of the model that he has made.

Note :-

if you have any doubt about the answers please ask me in comment box . but please don't give thumb down .

THANK YOU.