BUSI4434 - Research Methods for Management Studies Assignment

Assignment Task

Question

1. A credit card company wants to run an OLS regression to explain variations over time in the use of their cards. They have data on a number of variables for 216 weeks. The regression output is as follows:

The regression variables are defined as follows:

USE is the number of times the cards were used each week.

AD is the company’s weekly advertising expenditure in thousand £.

INT is the rate of interest charged on their cards measured in per cent (%).

EXP is the aggregate weekly consumer expenditure for the UK in thousand £.

ADC is the weekly advertising expenditure of their principal competitor in thousand £.

(a) Interpret and comment on these results. Please explain in detail and make use of all information available to you.

(b) The company believes that card use is higher in July and August due to holiday spending patterns. Explain how you might adapt the model to test for this. What results would this new model need to yield to support the company’s belief?

(c) What changes (if any) to the specification of the model would you make and why? Please make use of all information available to you.

(d) Why might multicollinearity be a problem in this model and how would you test for it? Explain how the tests you propose would help determine the presence or absence of problematic multicollinearity.

2. Dr Görtz is a labour economist who is interested in explaining why people work in the UK manufacturing industry. He intends to estimate the following equation

• Dr Görtz initially decides to estimate the equation using an OLS model. Outline what issues are likely to arise and how these would affect the interpretation of his results.
• After consulting Dr Riegler, Dr Görtz opts to estimate the above equation using a logit estimator. Having estimated a logit model, Dr Görtz obtains the output below. Interpret and comment on these results. Please explain in detail and make use of all information available to you.
• Discuss two ways of measuring the success of Logit models

3. The Chief Executive Officer (CEO) of a company wishes to understand the factors that determine CEO salaries. He asks his research analyst to survey 177 CEOs and collect data on a number of variables. The data are then used to run an OLS regression, with the results given below:

• Interpret the above results, paying careful attention to the signs and statistical significance of the coefficients.
• How well does the model fit the data? What are the limitations of R-square?

• A recent paper suggests that the relationship between CEO salary and the firm’ market value is different for CEOs with and without an MBA degree. Describe how you can amend the model above to find the answer to the question.

• A female CEO with an MBA degree has accumulated 5 years working experience as a CEO, and she was 50 in 2018. In the same year, the firm’s market value is 55,000,000£ and the firm’s profit margin is 0.1. What is her predicted wage in 2019?

• In the above model, what is the purpose of introducing the variable Ceotensq alongside the variable Ceoten ? Perform appropriate calculations to explain what information can be obtained from the estimated coefficients on these variables.

4. The World Food Programme (WFP) usually supplies potatoes to war-torn countries in South Asia. The organisation has decided to invest into the cultivation of the crop. To help them understand the relationship between potato crop output and inputs, the WFP have collected a balanced panel data from 44 different farmers over a period (1990-1997).

The WFP economist decided to run three different estimations based on three different model specifications. Table 1 presents the summary of the estimation results from the model above:

Table 1: Regression Output Summarised

• What is the value of X (total sample size) in Table 1? Please explain how you derived this from the information available to you.

• The WFP research team is debating between reporting the estimates from the pooled regression or those from one of the panel models (i.e. the fixed effects or random effects model as per the above). Which of these two model types do you find to be most appropriate to understand the relationship between potato crop output and inputs? Explain your answer, stating clearly any hypotheses you would formulate to address this debate, how these hypotheses are to be tested, and what test outcomes would support which of the two options being debated.

• Assume you determine a panel model is most appropriate. Now you are torn between using the fixed effects or the random effects model. Evaluate which of these two alternative panel model specifications you would use. What important assumption would you have to make in choosing between a fixed effects or random effects model? Support your answer using the information from Table 2.

This BUSI4434 – Management has been solved by our PHD Experts at My Uni Paper.