INAF U8145, Spring 2024 Problem Set 1: Development Indicators and Cross-Country Growth

1. Comparing different measures of GDP/capita

1.1. Get the data. From the course webpage on Courseworks, go to the Assignments section, and from the

folder for Problem Set 1 (Files > Problem Sets > ps1), download the Stata dataset ps1data_1.dta. On the Citrix system, you should create a folder for the class in your home folder (call it, say, u8145 in your Documents

folder), and then a folder for the problem set (call it ps1) in the u8145 folder. Stata gets confused by spaces in the pathnames of folders; to avoid the confusion, make sure that none of the folders in your folder structure  has a space in it.

Note: If you have specific questions on Stata, they can often be answered by searching through the help files using the Help drop-down menu. The complete Stata documentation is available in pdf form; in the Stata

interactive window, go to Help > PDF Documentation. The User’s Guide is the place to start in the Stata

documentation. Several tutorials for learning Stata are available at

http://www.stata.com/links/resources1.html. You may also want to refer to the Stata Guide (from another

class) that is posted in the Problem Set 1 folder. Some handy Stata “cheat sheets” are available at http://geocenter.github.io/StataTraining/portfolio/01_resource/.

The dataset contains information from the World Bank’s World Development Indicators (WDI) dataset. The  data in ps1data_1.dta are for the year 2000 for 157 countries for which reasonably complete data are available. The variables are the following:

1.2. Get the starter Stata program and run it. From Files > Problem Sets > ps1 in Courseworks, download the program ps1starterprogram.do. Open a Stata interactive window, and open the do-file editor. Open

ps1starterprogram.do in the do-file editor. You will have to change the directory in the “cd

\\sipaxafsc\users\ev2124\Documents\u8145\ps1\” statement to correspond to the directory where you put  the dataset in 1.1 above (e.g. put your UNI in place of my UNI, ev2124). The easiest way to run the program is to click on the “do current file” button. The program should create a log file (ps1.log), 2 Stata graph files

(graph1.gph, graph2.gph), and 2 pdf files (graph1.pdf, graph2.pdf). Make sure that you can see the graphs:

graph1 should contain a simple scatterplot; graph2 a scatterplot with a regression line (i.e. the line minimizing the sum of the squares of the vertical distances between the points and the line.) Also, once the program has  run, you can open the Stata data browser to see the resulting dataset. (You should have a dataset with 106

observations and 7 variables: country, cty_cod1, year, rgdp_cap, lrgdp_cap, gr6098, and cont.)

1.3. Create two new variables, gdp_usd and gdp_id, equal to GDP/capita in current U.S. dollars using the

exchange rate conversion and GDP/capita in current international dollars using the PPP conversion. (Hint: to convert from local currency to US dollars or international dollars, you divide by the exchange rate or PPP

exchange rate, respectively.)

1.4. Make a scatterplot with GDP/capita (exchange rate conversion) on the x-axis and GDP per capita (PPP conversion) on the y-axis. Save a pdf of your graph.

1.5. Make a scatterplot with log(GDP/capita) (exchange rate conversion) on the x-axis and log(GDP per capita  (PPP conversion)) on the y-axis. Save a pdf of it. Which graph is easier to interpret in your opinion, the one in 1.4 or the one in 1.5? Why do you say so?

1.6. Modify the graph command for 1.5 in two ways:

(a) Have it draw a line indicating where the points would be if GDP/capita (exchange rate conversion) were

exactly equal to GDP/capita (PPP conversion.) This is what we would ordinarily refer to as a 45-degree line, if  the scales on the x and y-axes were symmetric. (Hint: You can just include the x-axis variable as an additional y- axis variable, i.e. “graph twoway scatter lgdp_id lgdp_usd lgdp_usd, msymbol(oh) connect(i l) …” where

lgdp_id and lgdp_usd refer to log GDP/capita with PPP conversion and XR conversion respectively.

(b) Modify the graph command so that it uses the variable cty_cod1 (country code) as the plotting symbol. (Hint: put “msymbol(i i) mlabel(cty_cod1) mlabsize(tiny) mlabposition(0)” in the graph command.)

Save a pdf of the new graph.

1.7. List five countries for which GDP/capita (PPP conversion) is particularly large relative to GDP/capita  (exchange rate conversion). List five countries for which GDP/capita (PPP conversion) is particularly small relative to GDP/capita (XR conversion).

1.8. Based on your graph from 1.6, what generalization can you make about the relationship between how rich a country is and the relative magnitudes of GDP/capita (PPP conversion) and GDP/capita (XR conversion)?

What is an economic reason for this pattern?

2. Comparing GDP/capita with other development indicators

2.1. Infant mortality is a good indicator of the extent to which basic needs are being met. Using the same data

as above, make a scatterplot with log(GDP/capita) (PPP conversion) on the x-axis and infant mortality (under 5 yrs.) per 1,000 live births on the y-axis.

2.2. Add a regression line (of infant mortality on log(GDP/capita), PPP conversion) to the data. Save a pdf. Compare your graph to Ray figure 2.9.

2.3. Make a graph similar to the one for part 2.2 using the adult literacy rate in place of infant mortality on the y-axis. Save a pdf.

2.4. Write a short paragraph (i.e. 2-3 sentences) using your graphs from part 2.2 and 2.3 as evidence in support   of the assertion by Robert Lucas in the quote at the beginning of Chapter 2 of Ray (see also Ray’s discussion on p. 9) that average per capita income is a reasonable way to measure economic development. Next write a short   paragraph using your graphs as evidence against Lucas’ argument. Finally, write a sentence or two expressing

your own view about whether GDP/capita is a reasonable measure of development.

3. Comparing the Human Development Index with GDP/capita

3.1. Get the data. In this part of the problem set, you will be using the data from the UNDP Human

Development Report, which your friendly neighborhood development economics professor has arranged in

convenient form in ps1data_2.dta, on Courseworks in the Files > Problem Sets > ps1 folder. For more

information on the dataset, go tohttp://hdr.undp.org/en/statistics/. The data are from the year 2008, for 167 countries with complete data. The variables are the following:

country:              Country name

cty_cod2:           Country code

life_exp:            Life expectancy at birth, total (years)

exp_sch_child:   Expected years of schooling, for current children

sch_adult:          Mean years of schooling, for current adults

gnipc:                 Gross National Income per capita (PPP conversion)

gdppc:                Gross Domestic Product per capita (PPP conversion)

3.1. Create a new variable, hdi, representing the Human Development Index (HDI) for each of the 167

countries. Refer to the notes from lecture and to the technical notes to the Human Development Report 2016 here:http://hdr.undp.org/sites/default/files/hdr2016_technical_notes.pdf. Use the goalposts reported in the Technical Notes for the 2016 report. (Note that your HDI will not match exactly the HDI published in the

report, since the data you are using is from 2008, not 2016. Note also that you should first re-define any

variables that are above the maximum goalpost or below the minimum one. The Stata syntax for value above the maximum is “replace varname = max_goalpost if varname>max_goalpost & varname~=.”)

3.2. Calculate the rank of each country according to the HDI in descending order of “human development” – i.e. the country with the highest HDI is 1, the second-highest is 2, etc. Calculate the rank of each country

according to GDP/capita (PPP conversion) – again, richest country is 1, second richest is 2, etc. The command is “egen hdi_rank = rank(-hdi)”; the negative sign ensures that the rank will be in descending order. (Note: you  should use GDP/capita, not GNI/capita, to calculate the GDP/capita rank.)

3.3. List five countries with HDI ranks that are particularly poor (i.e. corresponding to low levels of human

development) relative to their GDP/capita ranks, and list their values of the HDI. List five countries with HDI ranks that are particularly good (i.e. correponding to high levels of human development) relative to their

GDP/capita ranks, and list their values of the HDI. List five countries that have poor ranks for both HDI and GDP/capita, and list their values of the HDI.

3.4. What generalizations can you draw about these three groups: countries with poor HDI relative to

GDP/capita, countries with good HDI relative to GDP/capita, and countries with both poor HDI and low

GDP/capita? (Note: no single right answer here. Hint: a useful exercise is to calculate the difference in HDI

rank vs. GDP/capita rank for each country, subtracting one from the other. Which countries tend to have large positive differences? Which countries tend to have large negative differences?)

4. O Convergence, Where Art Thou?

4.1. Get the data. In this part of the problem set, you will be using the data from the Penn World Table, version

10.0, which is in ps1data_3.dta on Courseworks in the Files > Problem Sets > ps1 folder. For more information on this dataset, go tohttps://www.rug.nl/ggdc/productivity/pwt/?lang=en. This file contains real GDP/capita (PPP  conversion) in constant 2017 international dollars over the period 1960-2019 for 109 countries for which complete  data are available.

4.2. Read through the code in ps1starterprogram.do under the “*** Question 4.3” heading to understand what it is doing. This code calculates average growth rates for the 109 countries in the sample for the period 1960-1999.