Final1

docx

School

Nova Scotia Community College *

*We aren’t endorsed by this school

Course

3430

Subject

Statistics

Date

Jan 9, 2024

Type

docx

Pages

Uploaded by DrCrown5209

Final Examination Exam Windows : Dec 2, 2:35 pm to Dec 4, 12:59 pm Due Dec 4, 12:59 pm (submissions after 12:59 pm will NOT be accepted ) Name: Student ID: Exam Overview In this final exam, you are given an academic paper and asked to replicate some of the paper’s results. The paper’s main research questions and conclusions are summarized for you. You are also provided with a brief review of the tables to replicate. Note that, for the ease of the replication process, you are asked to deviate from some of the empirical choices presented in the paper. In Part I , you will answer a set of questions related to the empirical setups (data, variables, and models). The answers to those questions will help guide your data collection process in Part II. (20 points)  In Part II , you will download data tables with needed variables from WRDS. You will also need to clean, organize, and join data tables for statistical tests in Part III. (20 points)  In Part III , you will replicate the required tables by properly filtering observations, correctly constructing variables, and conducting appropriate statistical tests. You will also interpret your replicated results. (40 points)  In Part IV , you will check whether you follow all general instructions of the exam. o Your data files, coding files, and logs are stored in appropriate folders. o You have at most one SAS coding file and one Stata coding file. o Your outputs supporting your answers are replicable. In other words, your coding file(s) should run from the first to the last line without incurring errors . o You use data libraries with SAS and relative paths with Stata. o You do not copy answers or codes from your fellow classmates. o You save your answers in a .pdf file and zip your answer folders in a .zip file. (20 points) 1

Exam Preparation (1) Download Final.zip from Blackboard “ 作业区→ Final”. (2) Unzip the file and put the entire folder in any place you like on your laptop. (3) File locations: a. You should store data files that you downloaded from WRDS inside the “Input_Data” folder. You are required to properly rename downloaded data files. b. If you created datafiles as the intermediate steps of your coding process, you should save those created data files to the “Output_Data” folder. c. You should save your SAS codes (.sas files) and Stata codes (.do files) to the “Final” folder. d. You should save the log(s) from running your coding file(s). (4) At the end of the exam, you will need to zip and upload the entire “Final” folder along with this word document. Replicability of Outputs It is crucial to ensure all your outputs supporting your answers to the final exam are replicable . Note that 20 points , i.e., one-fifth of the total points, are assigned to the replicability of your outputs. (1) The replicability is mainly graded by reviewing your log files. (2) Your coding files may also be checked if viewed necessary by the grader. You are required to use data libraries with SAS and relative paths with Stata to facilitate the replicability of your codes. Save Log Files To save the log for Stata, write log using Final.log, replace at the beginning of your Stata do file and write log close at the end of your Stata do file. Upon finishing running your codes from the first to the last line, the log file will be automatically saved to your current directory. To save the log for SAS, right-click on your “Log” window, select “File”, and then select “Save As”. 2

Using Data Libraries with SAS and Relative Paths with Stata (1) In your SAS codes, you are required to use data libraries. a. You should assign a data library for the provided datafiles: libname Input "Whole path of Input_Data folder on your laptop" ; b. You should assign another data library for your created datafiles: libname Output " Whole path of Output_Data folder on your laptop" ; c. You are required to use the data libraries to load and save data files in your SAS codes. (2) In your Stata codes, you are required to use relative paths loading and saving data files. a. At the beginning of each .do files, you should change your current directory to the “Final” folder: cd " whole path of Final folder on your laptop " b. You can use the following codes to save a Stata datafile: save ".\Data\ datafile_name ", replace Discussion of the Exam You are allowed to orally discuss the exam with your classmates, but you are NOT allowed to copy your classmates’ codes and answers . Note that direct copy will be evident from the codes you submit – none of you should share the same coding structure exactly. Copying from multiple sources will also be obvious as it will cause inconsistency in your codes and answers. 3

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

The Association between the Magnitude of Quarterly Earnings Forecast Errors and Risk-Adjusted Stock Returns Robert L. Hagerman, Mark E. Zmijewski, and Pravin Shah Journal of Accounting Research Vol. 22, No. 2 (Autumn, 1984), pp. 526-540 Literature Background Ball and Brown (1968) first conduct an empirical evaluation of the usefulness of accounting income numbers by examining the stock prices reaction to earnings announcements. They argue that if accounting income numbers are informative to investors, investors should adjust their investment decisions as those numbers become available. Consistently, Ball and Brown observe stock price changes following the announcements of accounting income numbers. In the setup of the empirical tests, Ball and Brown highlight the notions of “unexpected earnings” and “stock price reaction related to individual firms”. A firm’s earnings are determined by both economic-wide factors and firm-specific operations. The former is viewed as public knowledge that investors can acquire from various sources way before the firm’s earnings announcement date. Contrarily, investors should not be able to learn the conditions of firm- specific operations until the firm announces its earnings. Therefore, when assessing the “new” information conveyed in earnings numbers, Ball and Brown subtract the “expected” portion of earnings from the total earnings numbers to obtain “unexpected earnings”. “Unexpected earnings” is also called “earnings forecast errors” in the sense that “expected” implies “forecastable”. Similar to earnings numbers, stock prices are also affected by economic-wide factors. Therefore, the market-driven stock price changes should be isolated from the stock price changes in response to the new information in earnings numbers. This “stock price reaction related to individual firms” is also called “unsystematic returns”, “abnormal returns”, and “excess returns”. Subsequent studies follow Ball and Brown (1968)’s empirical design in evaluating the informativeness of various accounting numbers, including earnings. 4

Hagerman, Zmijewski, and Shah (1984) Since Ball and Brown (1968), numerous have examined the association between unexpected earnings and abnormal stock returns surrounding earnings announcement dates. Among those studies, Hagerman, Zmijewski, and Shah (1984) focus on the quarterly earnings numbers and compare the information content of quarterly earnings to annual earnings numbers. Research Question of Interest for Replication The primary research question of Hagerman, Zmijewski, and Shah (1984) is whether a firm’s unsystematic returns are associated with its quarterly earnings forecast errors. The empirical results to this question are presented in Table 5 and Table 8-11 , and the related summary statistics are in Table 1-4 and Table 7 . You are required to replicate Table 3 and Table 8 (with some modifications in variable selection and construction). 5

Part I: Understanding Empirical Setup (20 points) Q1: Read Section 2.2 of Hagerman, Zmijewski, and Shah (1984) and explain how to construct quarterly earnings forecast errors. You need to consider the empirical definition of earnings. Should raw earnings or earnings per share be used? Should extraordinary items be included in the earnings calculation? State and explain your answers. You should also provide the variable name and data sources for the raw data you need to construct quarterly earnings forecast errors. (4 points) Hint: whenever working with quarterly values, you should be careful with lags – should a quarterly value be compared against a value from the last quarter or a value from the same quarter of the previous year? Q2: Read Section 2.2 of Hagerman, Zmijewski, and Shah (1984) and explain how to construct cumulative stock returns around the announcement of earnings numbers. For your own ease, you can use market-adjusted returns (as in Lab6) instead of risk-adjusted returns . You should provide the variable name and data sources for the raw data you need to construct this variable. (4 points) Hint: whenever working with a stock return window, you should carefully consider the difference between a calendar day and a trading day. 10 EXTRA POINTS are given if you can use size-adjusted returns based on market cap deciles (Hint: obtain from “CRSP Portfolio Statistics and Assignment: Cap. Deciles” and “CRSP Stock File Indexes - Daily Index Built on Market Capitalization”. Think carefully about how to re-arrange and merge the data tables.) . Q3: Read Section 2.2 of Hagerman, Zmijewski, and Shah (1984) and write down the regression equation to estimate. You should use the variable names you assigned in Q1 and Q2. (4 points) 6

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Q4: Read Section 2.1 of Hagerman, Zmijewski, and Shah (1984) and list all the sample- filtering criteria. (1) What is the sample period? (1 point) (2) What is the requirement for stock listing? (1 point) (3) What is the requirement for firms’ fiscal year-end? (1 point) (4) Can the earnings announcement date be missing? Why? (1 point) (5) What are the data items needed to ensure the availability of constructed quarterly earnings forecast errors? (1 point) (6) What are the data items needed to ensure the availability of constructed cumulative stock returns? (1 point) (7) Think about the sequence you will follow when applying the above 6 filtering steps. List your filtering order. (2 points) 7

Part II: Collecting Data (20 points) Log in WRDS ( https://wrds-web.wharton.upenn.edu/ ) with our class account (username: danzen; password:). You are required to download all necessary variables you need to replicate Table 3 and Table 8 in Part III. Below is a list of common WRDS datasets and their descriptions.  CRSP/Compustat Merged Database - Fundamentals Quarterly provides a joint table with both financial information collected from corporate’s quarterly financial statements and the corporate security information. o Each firm is uniquely identified with GVKEY, and each security is identified with LPERMNO. You should pick the identifier based on the entity of interest. o A firm can issue multiple securities, e.g., a firm can have Class A shares and Class B shares at the same time. Different securities are flagged by different Security-level Identifiers (LIID) for each firm. If you wish to select only one security representing each firm, the simplest method is to restrict LIID=”01” (note that LIID is a string variable). Each combination of GVKEY and LIID corresponds to a unique LPERMNO. o LPERMNO is the same as PERMNO in other CRSP data tables. o Fiscal year (FYEARQ) is assigned based on the current year-end month (DATADATE). If the current fiscal year-end month falls in January through May, the fiscal year is the current calendar year minus one year. Otherwise, the fiscal year is the current calendar year. o Fiscal year (FQTR) numbered 1, 2, 3, 4 correspond to the first, second, third, and fourth quarter for which the firm report financial information, respectively. o When limited to domestic firms that report standardized consolidated financial statements in industrial format and U.S. dollars (both active and inactive firms included), each observation should be uniquely identified by GVKEY and fiscal year-quarter (FYEARQ and FQTR). o Earnings announcement date is stored in “RDQ -- Report Date of Quarterly Earnings”.  CRSP Daily Stock provides stock prices and returns for each security on a daily basis. o Each security is uniquely identified with PERMNO. o If the closing price is not available, the price field (PRC) number has a negative sign to indicate that it is a bid/ask average and not an actual closing price. o The return of a security on a particular day is stored in “Holding Period Return”. o The value-weighted return of the entire security market on a particular day is stored in “Value-Weighted Return (includes distributions)”, for which CRSP has adjusted distributions such as dividends and stock splits. o The stock exchange where the security is listed is stored in “Exchange Code”. You should find the corresponding code for each stock exchange in the variable definition. Note that NYSE stands for “New York Stock Exchange”, and ASE and AMEX both stand for “American Stock Exchange”. 8

Q1: Obtain related Data tables from WRDS. Take screenshots of your web queries and paste them here. Also, answer the following questions for each data table you have downloaded (your answers should be consistent with your screenshots). (10 points) (1) How many data tables you need to download? What are the source datasets on WRDS? (2) What is the appropriate date range? (3) What are the appropriate company codes? (4) List the query variables you select. Q3: Examine the raw data tables that you downloaded from WRDS. When necessary, reorganize the data tables to ensure that the primary keys not only identify each observation but also correspond to the entities of interests. You should save all raw data tables to the “Output_Data” folders and save any altered data tables to the “Input_Data” folder. Paste the screenshots of your codes and outputs here. (10 points) 9

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Part III: Replicating and Interpreting Results (40 points) Q1: Construct the final sample by appropriately filtering observations. Paste the screenshots of your codes and outputs. (1) Apply the filtering criteria in the sequence you state in Q4 of Part I. (6 points) (2) Test and remove influential observations . (4 points) (3) Deal with extreme values – remove the extreme 1% of each dependent and independent variable you will include in the regression estimates. (4 points) Q2: Replicate Table 3. Test whether your sample reasonably assembles the sample presented in Table 3. Paste the screenshots of your codes and outputs. Note that you are required to explain the reasons for choosing a particular test statistic. You are also required to interpret the test results. (10 points) Hint: refer to Session 7’s slides to pick the appropriate tests. 10

Q3: Replicate Table 8. (1) Paste the screenshots of your codes and outputs from replicating Table 8. You are NOT required to give the exact table format as in Hagerman, Zmijewski, and Shah (1984). Also, note that you should only have one dependent variable. (3 points) (2) Interpret your regression coefficients in an economic sense. (5 points) (3) Draw a scatter plot of the dependent and the independent variables with a linear fitting line. Appropriately format the plot. Paste the screenshots of your codes and outputs. ( 2 points ) (4) Verify the assumptions of linearity, normality, and homoscedasticity. Paste the screenshots of your codes and outputs. You are also required to interpret the test results. ( 6 points ) 11

Part IV: Wrap-up and Submit (20 points) Check your answers □ You have put in your name and student id on the first page of this word document. □ You have pasted all relevant screenshots of your codes, outputs, and data explorer windows to support your answers to this exam. □ None of your answers are copied from your fellow classmates. Check your codes □ You use data libraries with SAS and relative paths with Stata. □ Your coding file(s) can run from the first to the last line without incurring errors. □ You have complete log file(s) showing the integrity of your codes. □ None of your codes are copied from your fellow classmates. Check the file locations □ Datafiles you downloaded from WRDS are properly renamed and stored in the “Input_Data” folder. □ Datafiles you created as intermediate steps are properly named and stored in the “Output_Data” folder. □ You have at most one SAS coding file and one Stata coding file stored in the “Final” folder. □ You have one log file corresponding to each of your coding file, stored in the “Log” folder. Check your submission □ You have saved this word document in a .pdf file named “Final_ your student id ” . □ You have zipped your answer folders in a .zip file named “Final_ your student id ” . □ You have uploaded the PDF file and the .zip file to Blackboard “ 作业区→ Final”. This exam may be re-used, and thus the exam solutions will NOT be posted on blackboard. If you wish to review the exam, you need to schedule an in-person appointment with the instructor. Note that passing any class materials, including this exam, violates the academic integrity. 12

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Final1

Related Documents