Set Problem #1
Econ 381, Prof. Evans Due Monday, Sep. 15, 3 p.m.
NOTE: You are encouraged to work on this problem set in a group of up to four members. When nished, each group should turn in one copy of the problem set to the class inbox in 130 FOB. Each completed problem set should list the names of the group members who worked on the assignment. As noted in the syllabus, no late assignments will be graded.
1. Hands dirty with data (3 points). In this problem, you will retrieve and manipulate macroeconomic data and verify some relationships that inform macroeconomic theory. (a) Go the Federal Reserve Bank of St. Louis Federal Reserve Economic Data (FRED) website at http://research.stlouisfed.org/fred2/ and download the following three data series into an Excel le.1 You can then do the subsequent analysis using either Excel or Stata2 : Real Gross Domestic Product, 3 decimal (GDPC96), percent change from year ago, deviation from trend. Make sure you set the units of account to percent change from year ago on the download page. Otherwise, you will simply have to manually calculate the percent change from a year ago from the quarterly levels. You transform this series to deviation of GDP growth from trend in the following way. If you look at the GDP percentage change series, youll notice that the average percentage change (growth rate) is above zero. So calculate the average growth rate for the whole series. Then the deviation from trend equals the growth rate minus the average growth rate. This measure shows how the economy is growing relative to its average or natural rate. Total Nonfarm Payroll Employment (PAYEMS), year over year percentage change, deviation from trend. Make sure you set the units of account to percent change from year ago on the download page. Follow the same procedure from above in creating the series for deviation from trend in employment growth. Because the GDP series is quarterly and the employment series is monthly, youll need to transform the employment series to quarterly. Just
Once you nd the desired series at the FRED website, click on the series ID (e.g., GDPC96) which should take you to the series data page. At the top of the page, choose the option, Download Data. Then choose the units you want the data in, choose Excel format, and then click Download Data. 2 Stata is available in both the economics computer lab in 136 FOB and in the SWKT computer lab on the rst oor (basement). See the Appendix for instructions on how to do the data analyses below. Econ 381, PS#1 Page 1 of 5
1
set each quarters employment equal to the average of the rst three months employment growth rate. Appendices A.1.1 and A.2.3 give instructions on how to transform a monthly series into a quarterly series. Consumer Price Index for All Urban Consumers (CPILFESL), all items less food and energy, year over year percentage change, deviation from trend. Make sure you set the units of account to percent change from year ago on the download page. Follow the same procedure from above in creating the series for deviation from trend in employment growth. Also, follow the same procedure as with the PAYEMS series to transform CPILFESL from a monthly series to a quarterly series. (b) Create a chart that plots both Real GDP (% chg. from year ago, deviation from trend) and total nonfarm payroll employment (% chg. from year ago, deviation from trend) from 1948Q1 to 2008Q2. Calculate the correlation of the two series over the period 1948Q1 to 2008Q2.3 How do you explain the two series being related in this way? (c) Create a chart that plots both Real GDP (% chg. from year ago, deviation from trend) and consumer price index, all urban consumers, less food and energy (% chg. from year ago, deviation from trend) from 1958Q1 to 2008Q2. Calculate the correlation of the two series over the period 1958Q1 to 2008Q2. How do you explain the two series being related in this way? 2. Chapter 2, Problems and Applications (2 points): #6 3. Chapter 2 related employment question (1 point): Suppose an economy is made up of 100 people who are in the following mutually exclusive categories shown in Table 1: (a) What is the percent of the workforce not in the labor force? (b) What is the labor force participation rate? (c) What is the unemployment rate? 4. Chapter 3, Problems and Applications (4 points): #3,#6,#9,#10
In Excel, you calculate the correlation between two series by using the =CORREL(series1,series2) command. In Stata, you use the correlate var1 var2 command. Econ 381, PS#1 Page 2 of 5
3
Table 1: Employment categories for fake economy
Number 5 5 5 5 10 50 10 5 5 Category under 16 years old disabled have been unemployed for a long time and have decided to stop looking for a job stay-at-home spouses retired currently working have jobs that are currently o for the summer had to leave jobs temporarily because of hurricane Ike and unemployed looking for work
Econ 381, PS#1
Page 3 of 5
APPENDIX A.1
A.1.1
Excel Instructions
Transforming series from monthly to quarterly in Excel
The Total Nonfarm Payroll Employment (PAYEMS) series comes at a monthly frequency, but we want to compare it to the GDP series that comes at a quarterly frequency. The solution is to make each quarter the average of the corresponding three months data. However, it is more dicult to do this in Excel in a formula rather than having to manually type the average command in each row. A solution is to type the following formula into Excel. Assume that the monthly data for June 1947 (1947-06-01) is in cell A117. You can place the formula for second quarter 1947 in cell E117 by inputting the following formula: = IN DEX $A$1 : $K$971, 3 ROW ($B117) 236, COLU M N ($B117) + ... IN DEX $A$1 : $K$971, 3 ROW ($B117) 235, COLU M N ($B117) + ... IN DEX $A$1 : $K$971, 3 ROW ($B117) 234, COLU M N ($B117) /3
You can then copy this formula down the column and it will augment of row number of the three cells begin averaged by a multiple of three. The key is that the minus terms in the formula (e.g., -236, -235,-234) must be changed to t the given series. This new quarterly series can then be matched up to the quarterly GDP series.
A.2
A.2.1
Stata Instructions
Opening Stata in SWKT computer lab
When in the Kimball Tower Student Computing Center (101 SWKT) on the rst oor (basement) of the SWKT, you open Stata in the following way. Log onto the computer by entering your NetID and password. Go to the Start button in the lower left-hand corner. Click on All Programs. Then select Class specic, and click on Stata 10. The Economics Department computer lab (136 FOB) is slightly easier, although signicantly smaller, than the SWKT computer lab. You simply log into one of the computers using your NetID and password, then click on Stata 9 from the All programs section of the Start menu. There is not any dierence that you will notice between Stata 9 in the Econ computer lab and Stata 10 in the SWKT computer lab
A.2.2
Importing data into Stata
One easy way to import data into Stata is to save it as an Excel le and then just cut and paste the data into Statas data editor. The Stata data editor is accessed by
Econ 381, PS#1 Page 4 of 5
either clicking on the spreadsheet looking icon or by going to the menu and clicking on Data and data editor. Make sure that data series are in numerical format (numbers should be black, text should be red). One easy way to make sure that your Excel numbers get pasted as numbers and not text is to avoid pasting any #N/A values from Excel. Otherwise you will have to use a command in Stata like destring. Once the variables have been input into Stata, you can change their names. Suppose the variables that you have imported into Stata are named var1 and var2. You can change the name of var1 to newname1 by typing in the command line rename var1 newname1. You can save your data by going to the le menu and selecting save as. You will probably want to save to and work from a ash drive.
A.2.3
Transforming series from monthly to quarterly in Stata
Once you have the monthly data series in Stata, you can transform it to a quarterly series in the following way. First, rename the date variable. Suppose your date variable is called var1 and it is a string variable str10 (use the describe command to conrm this). Rename var1 to something like datetext by typing rename var1 datetext. Now convert datetext to a numerical date variable called datevar by typing gen datevar = date(datetext,YMD, 2008). This creates a numeric variable that Stata recognizes as the date given in the text. Now just format that variable to display as something more readable by typing format datevar %td. Next, create a numeric year variable by typing gen year = year(datevar), a numeric quarter variable by typing, gen quarter = quarter(datevar), and a numeric month variable by typing gen month = month(datevar). These variables now tell you which year, month, and quarter each months data is in. Lastly, you just need to calculate the quarterly average and then collapse the data. First, you must sort the data by typing sort year quarter month. Suppose your employment growth variable is named empgrowth. Create the quarterly average by typing by year quarter: egen qempgrowth = mean(empgrowth). Now you just drop the second and third month observations in each quarter because they are duplicates by typing, by year quarter: drop if n = 1. Now you have your quarterly employment growth rate series, which is the average of the growth rates of the three months in that quarter. In Stata, there are always many ways to accomplish the same task, and some are more direct than others. I dont claim that my way is the easiest way. But this is the way that I know how to do it.
Econ 381, PS#1
Page 5 of 5