Empirical Assignments

Econ 503 Empirical Assignments & Summaries/Presentations

General Guidelines for Empirical Assignments (assignments that do not follow these guidelines will receive half credit or no credit):

1. Print or type answers to questions in neat, easy-to-read form on one side of the page only. Staple or clip multiple pages together.

2. Include the title "Week 'X' Empirical Assignment"

3. To receive credit, each assignment is due at the beginning of class for the following week. For instance, the Week 1 Empirical Assignment is due at the beginning of class for Week 2.

4. Take care with the overall appearance and organization of results. All graphs/tables should be clearly labeled and organized in a reasonable way. Do not turn in a “data dump” – a huge amount of graphs/tables. No assignment requires that.

5. Multiple sheets must be connected by a staple or paper clip. All papers should be in good shape with no tears, folds, ...

6. You can save the output as a pdf file or cut/paste, snip/paste it to a Word file if you wish.

General Guidelines for Summaries/Presentations

1. Summarize the reading(s) in 1-2 single-spaced, 12-font pages (min-max).

2. The summary should attempt to explain the question(s) and specifics of the methods/results to undergraduate econ junior/seniors. It should not be written on merely a newspaper level. Clarity is important but so are key details.

3. Prepare a 20-minute (max) powerpoint presentation to accompany your summary. The presentation should reflect the summary, but it can include some additional info (tables/graphs/...) that may not fit into the summary.

4. Well-prepared PPT presentations emphasize key points and provide visualizations. They are not just an extended outline with a lot of words. They also include a reasonable number of slides. For 20 minutes, 20 is too many but 2 is too few.

5. Print a handout version of the PPT presentation to turn in but also be prepared to make an in-class presentation.

Week 1 Empirical Assignment (due at beginning of class for Week 2)
Basic Macro Data Manipulation and Growth
Part I

*Navigate to the St Louis Fed data website FRED

*Generate a graph of (nominal) GDP and add a line for Real GDP in 2012 dollars both using quarterly data at SAAR, 1947-2018

*Print the graph (Note: unless you have a color printer, you should reformat the graph to use different lines (solid, dash, dash-dot) for each country using the Edit Graph>Format options)

*Generate & Print a new graph for Real GDP changing the units to percent change from a year ago; then, add a line using Real GDP changing the units to the compounded annual rate of change

Part II
* Download Gretl (free statistical software)

* Use File>Databases>Gretl native>StL Fed (search for name)
* Retrieve nominal GDP (gdp, quarterly bil $ 1 decimal from 1947)
* Retrieve CPI (convert to annual by default averaging method that pops up)
* Compute GDP in 2018 constant $ (“real gdp”)
(use Add>New Variable> gdp18 = put your formula here)
*Compute percent changes in real GDP (both year to year and annualized quarterly) and compute Add>log differences of real GDP

*Generate and print summary statistics for percent changes in real GDP (annualized, year to year, and log differences

Questions:
1. What are the beginning and ending values for GDP and real GDP?

2. How do the plots of percent changes in real GDP differ by the 3 methods of computing them?

3. What are the average growth rates and annual variation in percent changes in real GDP (all 3 ways)?

Week 2 Empirical Assignment: Long-Run Growth & Living Standards Graphics

* Go to the St.Louis Fed data website FRED

* Create a graph for "Constant GDP per capita for World". To this graph (using Edit Graph) add lines for the same variable for the United States, United Kingdom, South Korea, China, and the Democratic Republic of Congo

Print the graph (or download to save)

Questions:
1. What are the most recent values of GDP per capita in these countries?

2. What is the GDP per capita in each country relative to the U.S. and to the World average?
3. What has happened to the value of World GDP per capita and U.S. GDP per capita since 1960? what is the average growth rate?

4. Describe the similarities and differences in per capita GDP in recent years across these countries

5. If the (PCEPI) price level index in the U.S. were 72 in 1995 and is 108 in 2018, what would the 2018 GDP per capita be in 2018 constant dollars? What would 2018 GDP per capita be in 1995 constant dollars?

Summary/Presentation of Empirical Macro Topic (Due in Week 4)

Assigned Topics

-- Issues in measuring & using income, wealth, & saving:
StL Fed & Richmond Fed (first half)
Wealth or Income More Persistent Across Generations (StL Fed) 1

-- Measuring consumption v. income and its relevance for health care expenditures
CAIC & health expenditures (Random Critical Analysis)
AIC and disposable income (RCA)
(background data AIC & World Bank ICP )

Week 5 Empirical Assignment: Short-Run Economic Fluctuations & Recessions

*Go to FRED. Generate a graph of the unemployment rate (unrate). Edit the graph to add real gdp (quarterly SAAR) to the graphic and change its units to percent change from a year ago. Also, add all sectors debt securites and loans (formerly total credit market liabilities) It will automatically graph as percent changes from a year ago.
*Print the graph (If you do not have a color printer, select different line styles for each line).

Questions:
1. How many recessions since 1950 (NBER-defined recessions appear as shaded areas on the graph)?

2. Summarize the length and frequency of recessions over this time frame.
3. Which of the lines match the shaded recession periods the closest?
4. How would the size of some recessions change if defined by unemployment behavior or changes in liabilities?

5. For the 1981-82 recession and 2007-09 recession, how long did it take for the unemployment rate to decline to 5.5 percent or less?

Summary/Presentation of Empirical Macro Topic (Due in Week 6)

Assigned Topics:

-- Students (Group 1): Construction of Price Indexes + estimating housing prices:
CPI v PCE CPI v PCE KC Fed VoxEu KC Fed
-- Students (Group 2): Adjusting prices for changes in medical care (and other):
NBER Murphy & Topel + Video

Instructor:

- Technology & Macro measurement
Tech & price levels Kessler WSJ Tech, services, & consumer surplus

PPT

-- Macro impacts of oil price shocks:
StL Fed Hamilton NBER

Week 7 Empirical Assignment: Money & Inflation (due in Week 8)

Part I: Money Supply & Velocity
*Open Gretl and download these variables from St Louis Fed database (cpi, money supply mzm seasonally adjusted, gdp, and real gdp)
Note: use “interpolate” to convert gdp quarterly data to monthly

*Use data from 1959 to 2017 for stats below
*Compute inflation (annual percent changes in cpi), annual percent changes in mzm, velocity (mzm/gdp), annual percent changes in velocity, real gdp (100*Nominal GDP/CPI), and and annual percent changes in real gdp
*Plot percent changes in mzm and percent changes in velocity of mzm
*Regress inflation = b0 + b1*pct change mzm + b2*pct change velocity + b3* pct change real gdp

**Regress inflation = b0 + b1*pct change mzm + b2*pct change velocity

*Regress inflation = b0 + b1*pct change mzm

*Compute 5-year averages for inflation and percent changes in mzm
*Regress inflation = b0 + b1*pct change mzm using the 5-year average data

*Print the regressions

Questions:
1. Describe the behavior of percent changes in money supply versus changes in velocity
2. What are the long run changes in velocity? What might explain these changes?
3. Describe the results of the inflation regression (R2, coefficients, …)

4. What happens to the relationship between pct change in m2 and inflation as the time unit is changed from monthly to 5 years?

Week 8 Empirical Assignment: Monetary Policy & Rates

*Open Gretl and download cpi, fed funds rate, and 3-month treasury rate secondary market, 10 year treasury rate constant maturity, and 10 year inflation indexed treasury rate
*Compute the rate of inflation (use annual percent change)
*Compute an estimate of the real 3 month rate by r3m = 3month rate - inflation

*Plot the r3m and the 10 year inflation indexed treasury rate on the same graph and print
*Plot the fed funds rate, 3 month treasury rate, and the 10-year treasury rate on the same graph and print

*Generate a correlation matrix for these three rates and print
*Estimate a VAR for the ff rate, 3 month rate, and 10 treasury year rate . Print the impulse response functions for the VAR using the variables in this order for the Cholesky ordering

*Recompute the IRF by switching the ordering for the FF and 3-month rates. Pring the IRF.

Questions:
1. Using the plots and correlations, describe the relationship between the ff rate, the 3 month rate, and the 10 year rate
2. How closely do the Fisher equation real rate and the inflation-adjusted market real rate mirror each other?

3. What do the IRFs indicate about how these rates influence each other? Does changing the ordering make a difference?

Week 9 Empirical Assignment: Issues in Estimating Policy Impact
*Go to FRED and generate a graph for real federal government current expenditure (in levels), percent change in real gdp, and and the unemployment rate for the period from 2007-2012. (you will need to change the right axis to percent). Print the graph

*Generate the same graphic except switch mzm money (in levels) for government expenditure. Print these graph.

*Open Gretl and download mzm, fed runds rate, and unemployment rate.

*Compute percent change in mzm, change in fed funds rate, and change in unemployment rate

*Estimate regression pct change mzm = b0 + b1mzm(-1) + b2mzm(-2) + b3mzm(-3)

*Save predicted and residual values of mzm pct changes

*Estimate regression change fed funds rate = b0 + b1*predicted mzm pct changes + b2*residual mzm pct changes

*Estimate regression change unemployment rate = b0 + b1*predicted mzm pct changes + b2*residual mzm pct changes with lags 1 to 6 on these variables

Questions:

1. What was the size of the increase in federal expenditure associated with the 2007-09 recession? What was the size of the increase in the money supply?

2. Describe the behavior of real GDP and unemployment during and after the recession.

3. How do the predicted and residual values of mzm impact the fed funds rate and unemployment?

4. Why might the predicted and residual values have different impacts?

Week 10 Empirical Assignment: EU Data

*In FRED, print a graph that displays the unemployment rates for the UK, France, Germany, Italy, and Spain for 1995-2017.

*Print a graph that displays unemployment rates for these countries for individuals aged 18-24 for the same time frame

*Print a graph that displays 10-year government bond rates for these countries for this time period

Questions:

1. Describe the variation of unemployment rates across these countries and over time

2. Describe the variation of youth unemployment rates across these countries and over time

3. Describe the variation of 10-year government bond rates for these countries and over time

Summary/Presentation of Empirical Macro Issue (due Week 12)

Assigned Topic:

-- Tracking GDP in Real Time:

Tracking Economy with NowCasts (StL Fed)

Chicago Fed CFNIA (Chicago Fed) Phil Fed Real-Time Research Data Center ATL Fed;

-- Using the yield curve to predict economic performance:
Yield Curve as a Predictor of Recessions (NY Fed) Yield Curve (Clev Fed)
Breaking Down Yield Curve Shifts (SF Fed)

Summary/Presentation of Empirical Macro Issue (due Week 14)

Assigned Topic:

1. Impact of Corporate Income Taxes:

Survey of Theory & Evidence (AEI)

Estimating Impact of Corporate Income Taxes (Azemar & Hubbard)

2. Assessing China's Econ Data:

China growth overstated? NBER Can we trust Chinese data? (Stl Fed)
China Transparency Challenges (Brookings)

- Estimating Real Interest Rates:
Equilibrium Real Rate (Hamilton et al) Natural Rate (Fed)

Archived Assignments
Predicting GDP, Inflation

* In FRED, generate a graph for the 10-year Treasury Rate Constant Maturity and under "customize data" ad the 3-month Treasury Rate Secondary Market; then compute the difference between the two using the Formula option and entering a-b

*To this graph, add an additional line for the 1-year Treasury Rate Constant Maturity (back to 1954) and repeat the procedure for computing the difference between it and the 3-month rate

*Print this graphic

*In Gretl, download the interest rates used in Fred and compute these difference. Add recession dummies.

*Estimate a regression for Recession = b0 + b1*Ten Year -3 month and the same using 1-Year - 3 month. In doing so, explore how many lags maximize the predictive power. You should also see if adding the squared term difference improves the results

*Print the best result in terms of highest R2

Questions:

1. Using the FRED graphs, describe the levels of the term spread variables near recessions

2. Which term spread variable did a better job of predicting recessions? How many lags were useful?