Restrictions on firm size in India

Mohan (2002) lists the industries in which the Indian government restricted the size of firms.

Used by Hsieh and Klenow (2009) to show that industries with size restriction have more variation in the marginal revenue product at the plant level.

Timor-Leste Transparency Portal

http://www.transparency.gov.tl/public/index

It doesn't seem to work properly yet, but maybe in the near future...

Vietnam Statistical Yearbooks

Data on provincial population in 1985-2000 and on total state investment flows by province in 1976-1985 from Vietnam Statistical Yearbooks are used by Miguel and Roland (2011).

Political Terror Scale

Political Terror Scale measures the degree (in 5-point scale) of human rights abuse in each country annually for the period 1976-2009, based on two sources: Amnesty International and US State Department.

Papers using this dataset are listed here. In addition, Qian and Yanagizawa (2009) investigate what causes the divergence between the two reports, Amnesty International and US State Department.

Agricultural research expenditure in Africa since 1971

IFPRI's Agricultural Science and Technology Indicators (ASTI) Initiative collected agricultural research expenditures for 27 sub-Saharan African countries since 1971, downloadable at www.asti.cgiar.org/data.

Used by Block (2010), to explain the agricultural total factor productivity in sub-Saharan Africa.

World Development Indicators (WDI)

The data is downloadable at http://go.worldbank.org/6HAYAHG8H0 if you or your institution has subscribed.

If you use Stata, you need to transform the WDI data structure into the country-year panel format. P. Wilner Jeanty's ado file WDIRESHAPE may be useful for this purpose.

The List of variables.

The World Bank's region classification puts Djibouti into Middle East & North Africa rather than sub-Saharan Africa. More details on the World Bank's country classification is found here.

For the treatment of countries that broke up, merged, etc., see under the heading of "Country Notes" in the Users Guide.

Data series on tax revenue goes back only to 1990 in the 2005 edition while the data for 1970-1989 was also available in the 2004 edition. This is either because the data before 1990 "have been determined to be questionable" or because "revisions to time-series cause breaks in series which cannot be reconciled by linking or other estimation techniques." See WDI's FAQ.

Data series on general government final consumption expenditure as % of GDP in the 2005 edition seems to be calculated in the current term, not in the constant term. Therefore, change in this series may result from different inflation rates between government consumption goods and the others.

Chinese provincial panel data

The CHINA STATISTICAL DATABASES available at China Data Online of the China Data Center at University of Michigan provide various annual statistics at the provincial level since 1949 and monthly macroeconomic statistics since 1998. Subscription required. For what data is available and how much subscription costs, see this page.

The Comprehensive Statistical Data and Materials on 50 Years of New China, published by the China Statistical Press in 1999, provides a provincial panel data for the years 1949-1998. The CD-ROM version was published in 2004. The data on grain production is used by Meng, Qian, and Yared's working paper entitled "The Institutional Causes of China's Great Famine, 1959-61".

To obtain the real GDP growth for each province, look at what they call the GDP Index. This is the ratio (in percentage) of the current year GDP to the previous year GDP with price being constant. If you subtract 100 from this index, therefore, you obtain the real GDP growth. Real GDP itself is never published. Only nominal GDP can be directly obtained.

China: government expenditure, growth, poverty, and infrastructure, 1952-2001, is compiled by IFPRI researchers. Available for free of charge.

Difang caizheng tongji ziliao (Local Public Finance Data) provides extra-budgetary revenue at the provincial level. According to Zhan (2009), this statistics is an indicator of excessive taxation and bribe extortion by local governments.

Jin, Qian, and Weingast (2005) compiled the provincial panel data (1970-1999) on the provincial government budgetary revenue, budgetary expenditure, extra-budgetary revenue, and extra-budgetary expenditure from various sources including China Statistical Yearbooks.

In his study on the impact of the Household-Responsibility System (HRS) on crop production, Lin (1992) compiled the provincial panel data from 1970 through 1987 on the gross value of crops produced, cultivated land, farm labor force, number of tractors, number of draft animals, gross weight of fertilizer consumed, the ratio of teams converted to the HRS, and so forth from various sources including National Agricultural Statistics for the 30 Years Since the Founding of the People's Republic of China, 1949-1979 (Beijing: State Statistical Bureau, 1980) and China Agricultural Yearbooks.

Luxembourg Income Study

Luxembourg Income Study provides micro data on gross and net income, mostly for developed countries but also for Latin American middle-income countries such as Brazil (2006), Colombia (2004), Guatemala (2006), Mexico (1984, 1989/1992, 1994/1996, 1998/2000/2002, 2004), Peru (2004), and Uruguay (2004) and for former socialist countries such as Czech Republic (1992, 1996, 2002, 2005), Estonia (2000, 2005), Hungary (1991, 1994, 1999, 2005), Poland (1986, 1992, 1995, 1999, 2004), Romania (1995/1997), Russia (1992, 1995, 2000), Slovak Republic (1992, 1996, 2005), and Slovenia (1997, 1999, 2004). The data for Taiwan (1981, 1986, 1991, 1995/1997, 2000, 2005) is also available.

Used by Milanovic (2000) to test the median-voter hypothesis.

Income tax

Baunsgaard and Keen (2005) compile the panel data on income tax revenues as opposed to trade tax revenues for 111 countries over more than 25 years.

Used by Besley and Persson (2009).

Peter et al. (2008) compile a 1981-2005 panel data on the national personal income tax system for 189 countries.

Social Security Programs throughout the World

US Social Security Administration reports on the design of Social Security programs (old age pension) across countries. The 1995 data is used by Mulligan and Sala-i-Martin (1999).

Social security spendings

ILO's Cost of Social Security annually reports public spending on social security across countries.

Mulligan, Gil, and Sala-i-Martin (2002) compile annual public spending on old age pensions from 1960 to 1990 with its Data Appendix showing the data availability of each country-year cell.

Government Finance Statistics (GFS)

The IMF Government Finance Statistics (GFS) is a well-known source of government expenditure data. But it's difficult to understand for the first time users. Here's some help:

List of countries (as of December 2005), provided by ESDS International (UK's data distributor)

There was a change in the data classification system in 2001. Data during 1972-1989 is available in the old format (GFSM 1986 format) while data since 1990 is available in the new format (GFSM 2001 format). Classification comparison between the two systems is available in

Tobias M. Wickens (2002) "CLASSIFICATION OF GFSM 1986 DATA TO THE GFSM 2001 FRAMEWORK," Government Finance Statistics Manual 2001 Companion Material.

Each variable has 13-letter series code in the GFSM 1986 format. It's surprisingly difficult to find the definition of this code, but here it is:

First 3-digit numbers: country code
The next 5 letters: subject code
The 9th letter: version code
The 10th letter: publication code
The final 3 digit numbers: Partner-country code (... if the data refers to a single country)

For the subject code, the first letter is always 8 for GFS data. The second letter is

0 - deficit/surplus of total financing
1 - revenue or grants
2 - expenditure
3 - lending minus repayments
4 - domestic financing
5 - foreign financing
8 - domestic debit, total debt (the next position in the time series
distinguishes between the two)
9 - foreign debt

The third and fourth letters are classification distinctions (I don't know what they are). The final letter is

. - central government budgetary accounts
H - central government consolidated accounts
O - central government extrabudgetary accounts
3 - central government social security funds
J - state or province governments
L - local governments

Health expenditures

World Health Organization began collecting cross-country health expenditure data since 2000. Its World Health Report publishes the figures each year. As it continuously updates the data each year, the latest report must be consulted.

The Report offers data on public and private health expenditures, external resources for health (ie. foreign aid, including NGOs, on health), out-of-pocket expenditure (ie. expenditures directly paid by households), and private prepaid plans (ie. private insurance spending).

Government health expenditure data is available as % of GDP, as % of total government expenditures, in per capita terms in US dollars, and in per capita terms in international dollars (ie. PPP adjusted).

For years 1998-2002, World Development Indicators 2005 and World Health Report 2005 Statistical Annex (see tables 5 and 6) provide the data.

For year 1997, World Health Report 2004 Statistical Annex (see tables 5 and 6) offers the data.

For year 1995-6, World Health Report 2002 Statistical Annex (see table 5) offers the data.

For years 1990-1996, see

Eduard Bos, Vivian Hon, Afkiko Maeda, Gnanaraj Chellaraj, and Alexander Preker (1999) Health, Nutrition, and Population Indicators: A Statistical Handbook (Washington, D.C.: World Bank)

See Table 2. Pages 70-71, however, says, "the time series data should be interpreted with caution" because "data for each country are not necessarily from the same source." Some data in this book are reproduced in World Health Report 1999, Annex Table 1).

For 1990, see

C.J.L. Murray, R. Govindaraj, and P. Musgrove (1994)
"National Health Expenditures: A Global Analysis"
Bulletin of the World Health Organization, 72(4): 623-637.

This paper was first written as a background paper for the World Development Report 1993, and, according to Bos et al. (1999, p. 6), "The World Bank's World Development Report 1993: Investing in Health was the first major effort to compile health expenditures for all developing countries." Bos et al. (1999) is "essentially an expanded update of that 1993 work" (Ibid.).

The paper's annex provides a table of public health expenditures, private health expenditures, aid flows for health, and total health expenditures (as % of GDP, in 1990 US$, and in 1990 US$ per capita) with clear indications of countries for which the data is predicted by regression analysis. It's probably best avoided to use the same data published in World Development Report 1993 (Table A.9) and in World Health Report 1995 (Annex 3 Table A1) because these reports cite the working paper version.