世界宏观数据库链接总结（Markus Eberhardt）

InequalityThe [color=#089c9 !important]World Income Inequality Database (WIID2) (~1960-2006) compiled by UNU WIDER has Gini coefficients for 167 countries, although the time-series element of the dataset varies considerably. The website states: "WIID2 consists of a checked and corrected WIID1, a new update of the Deininger & Squire database from the World Bank, new estimates from the Luxembourg Income Study and Transmonee, and other new sources as they have became available." Quintile and decile income share data is also contained in the dataset.

The other standard source for inequality data is the [color=#089c9 !important]Estimated Household Income Inequality Data Set (EHII) prepared by theUniversity of Texas Inequality Project, where you can also find other data on inequality. Particularly noteworthy here is the [color=#089c9 !important]UTIP-UNIDO dataset which calculates the industrial pay-inequality measures for 156 countries from 1963-2003. It has a total of 3,554 observations based on the UNIDO Industrial Statistics, thus representing a very large cross-section dimension and containing annual data.

The Luxembourg Income Study (LIS) is a cross-national Data Archive and a Research Institute located in Luxembourg. The[color=#089c9 !important]LIS archive contains two primary databases.  The LIS Database includes income microdata from a large number of countries at multiple points in time, starting from the early 1980s. The newer LWS Database includes wealth microdata from a smaller selection of countries. Both databases include labour market and demographic data as well.

Facundo Alvaredo, Tony Atkinson, Thomas Piketty and Emmanuel Saez at Oxford, PSE and Berkeley have created the [color=#089c9 !important]World Top Incomes Database. "The world top incomes database aims to providing convenient on line access to all the existent series. This is an ongoing endeavour, and we will progressively update the base with new observations, as authors extend the series forwards and backwards. Despite the database's name, we will also add information on the distribution of earnings and the distribution of wealth. Around forty-five further countries are presently under study." This is very much work in progress.

A 2005 IMF [color=#089c9 !important]working paper by Garbis Iradian ([color=#089c9 !important]Deputy Director, Africa/Middle East at the Institute of International Finance, Washington) provides inequality data for 82 countries over the period 1965–2003 (the data is averaged over periods of three to seven years). The data is constructed from household surveys.

The dataset on income inequality compiled by Klaus Deininger and Lyn Squire for the World Bank is one of the most commonly used data to investigate any links between inequality and growth at the macro level. The [color=#089c9 !important]data distributes unevenly for 138 countries and over the period of 1890-1996 (but much shorter and sporadic for the vast majority of countries).  For some countries this is not merely the Gini, but also cumulative quintile shares, available for download in Excel format.

The World Bank Development Economics Department has developed the Global Income Distribution Dynamics ([color=rgb(0, 137, 201) !important]GIDD), the first global CGE-microsimulation model. "The GIDD takes into account the macro nature of growth and of economic policies and adds a microeconomic—that is, household and individual—dimension to it. The GIDD includes distributional data for 121 countries and covers 90 percent of the world population." The data cover the period 1992 to 2005 (survey year), although most observations are in 2000-2002 and 2005. "The GIDD database is not a mere compilation of secondary cross-country inequality indices. Instead, it is an actual presentation of a truly global income distribution based entirely on household survey data. Additionally, the GIDD global income distribution data includes information on the conditional distribution of important household income determinants like education, age, household size, among others." There is extensive documentation for the data on the website, together with a research agenda and recent work by the group. Download is in excel spreadsheet or as a Stata file [This is used in a recent [color=rgb(0, 137, 201) !important]paper on trade in agriculture and global poverty by Bussolo, De Hoyos and Medvedev (all World Bank) in World Economy Vol.34(12), December 2011.]

The Society for the Study of Economic Inequality (ECINEQ) has [color=rgb(0, 137, 201) !important]links to a number of datasets for the analysis of inequality. These include the Cross-National Equivalent File (CNEF) which contains equivalently defined variables for the British Household Panel Study (BHPS), the Household Income and Labour Dynamics in Australia (HILDA), the Korea Labor and Income Panel Study (KLIPS) (new!), the Panel Study of Income Dynamics (PSID), the Swiss Household Panel (SHP), the Canadian Survey of Labour and Income Dynamics (SLID), and the German Socio-Economic Panel (SOEP).



Sectoral Data (i): AgricultureThe UN Food and Agriculture Organisation (FAO) provides [color=#089c9 !important]FAOSTAT, which is divided into ProdSTAT, TradeSTAT, ResourceSTAT, etc. This is the primary source for cross-country data on agricultural production and trade. Data usually starts in 1961, although this varies greatly by variable and dataset.

The [color=#089c9 !important]Earth Trends database at the World Resources Institute also has (among other links) the complete FAO data for agricultural production, land, inputs, etc. from 1961 onwards. UPDATE August 2012: Sadly, the WRI no longer seem to host this dataset.

[color=#089c9 !important]A Cross-Country Database for Sector Investment and Capital (1967-1992) is a unique World Bank dataset which provides investment in fixed capital stock for agriculture, as well as a capital stock variable created by the authors - Al Crego, Rita Butzer, Yair Mundlak, and Don Larson. This version provides data from 1967 to 1992 for (up to) 63 developing and developed countries. Manufacturing is also covered separately. New-ish: The World Bank team has also created an expanded version of this, which goes up to the year 2000, albeit only for 30 countries. As far as I know the latter dataset is not in the public domain, although Rita Butzer told me they are keen to get people to use it (so just email one of the aforementioned authors). A link to a STATA version of the 'old' dataset is [color=#089c9 !important]here.

The World Bank recently completed a big data compilation exercise for [color=#089c9 !important]Distortions to Agricultural Incentives, with a team of researchers headed by Kym Anderson providing various Estimates of Distortions to Agricultural Incentives (1955-2007). A core database provides data for Nominal Rates of Assistance to producers (NRAs), together with a set of Consumer Tax Equivalents (CTEs), for farm products and a set of Relative Rates of Assistance to farmers in 75 focus countries. Note that the variable 'border price' (bp) does however not represent the... how can I say this... 'border price', but a hypothetical producer price in the absence of distortions (domestic producer price divided by (1+NRA) and expressed in USD). The border price (fob) is not contained in the main datafile but can be found in the individual country spreadsheet (rows 37-39 for primary products, or 44-46 for lightly processed products). I am grateful to Kym Anderson and Ernesto Valenzuela for clarification; they also point to an [color=#089c9 !important]alternative data reporter at Adelaide University where they are both based. See also next entry.

The Global Trade Policy Analysis group at the AgEcon Department of Purdue University provides a number of datasets related to trade but also climate change and geography. "The GTAP Data Base is a [color=#089c9 !important]fully documented, publicly available global data base which contains complete bilateral trade information, transport and protection linkages among [color=#089c9 !important]113 regions for all [color=#089c9 !important]57 GTAP commodities for a single year (2004 in the case of the GTAP 7 Data Base)." Single academic user licenses for GTAP 7 are $520, but a large number of free datasets (including summaries of GTAP, Social Accounting Matrix [SAM] extraction, the Global [bilateral] FDI Dataset, [color=#089c9 !important]Project on Bilateral Labor Migration, CO2 emissions and utilities related to the Distortions of Agri Incentives project) can be found [color=#089c9 !important]here.

[color=#089c9 !important]HarvestChoice, a collaboration between IFPRI and researchers at the University of Minnesota, "generates knowledge products to help guide strategic investments to improve the well-being of poor people in sub-Saharan Africa through more productive and profitable farming." A vast number of datasets related to agricultural production, markets, demography, climate, etc. is available for download from their website. A lot of emphasis is placed on spatial?GIS data with further tools for map-making etc available on the website. There are also a wealth of publications and policy briefs on all topics related to production, R&D and innovation in agriculture (includes analysis of US data).

The OECD has a dedicated database for [color=rgb(0, 137, 201) !important]PSE (Producer and Consumer Support Estimates) which covers OECD member states as well as a small number of Eastern European 'Emerging' Economies and the BRICS countries for 1986 to 2008. A recent[color=rgb(0, 137, 201) !important]paper by Kym Anderson compares and contrasts the methodology applied in his own work (see previous entry) constructing measures of agricultural production and trade distortions with that of the OECD.

The International Center for Tropical Agriculture (CIAT) has produced a [color=rgb(0, 137, 201) !important]Crop Atlas of the World: "The map shows derived estimates of the spatial distribution and productivity of crops for 10-km grids using a novel allocation approach involving the fusion of sub-national crop production statistics. The values in this digital [map] are the number of harvested hectares within each 10 km grid cell. This data includes area harvested in multiple season (therefore this is NOT the physical harvested area, but rather the total area harvested) [...] The sub-national crop production data comes from agricultural censuses and surveys and has scaled values, so as to obtain national production estimates that were compatible with the annual average FAO national crop statistics for 1999-2001. The prototype crop distribution database used in this study is available from the authors upon request but is currently being regenerated using newer and additional data sources (including revisions based on expert validation) and an enhanced allocation algorithm." If you have Google Earth you can look at these data maps.

The Agricultural Market Access Database ([color=rgb(0, 137, 201) !important]AMAD) is a collection of available public data on WTO market access in agriculture. It contains data for over 50 countries. After registration, all files can be downloaded for free (self-extracting zip files) and there is documentation on how to do this.

The UN body which covers trade and investment, UNCTAD, has created a snazzy website that combines all of its statistical databases: [color=rgb(0, 137, 201) !important]UNCTADstat has lots of data on trade (merchandise, services), FDI flows and stocks (inward FDI from 1970!), external finance (incl. remittances), labour force/employment, global commodity price indices (from 1960!) as well as some more recent rubrics such as the creative and information economies and maritime transport (from around 2000).

A Cross-Country Database For Sectoral Employment And Productivity In Asia And Latin America (1950-2005) by Marcel P. Timmer and Gaaitzen J. de Vries at the Groningen Growth and Development Centre. This is a balanced panel with data on agriculture, mining, manufacturing, construction, public utilities, retail and wholesale trade, transport and communication, finance and business services, other market services and government services. The sample comprises 10 East Asian Countries (nope, not China) and 9 Latin American ones. Variables covered in the data set are value added, output deflators, and persons employed... but not investment (or capital stock). Now part of the socalled [color=rgb(0, 137, 201) !important]10-sector database.

The Food and Agricultural Policy Research Institute (FAPRI), with research centers at the Center for Agricultural and Rural Development (CARD) at Iowa State University and the Center for National Food and Agricultural Policy (CNFAP) at the University of Missouri-Columbia, provides [color=rgb(0, 137, 201) !important]data on commodities and agricultural policy (at the product level) for a large number of developing and developed countries (time-series dimension differs widely across variables, products and countries).

The Food and Agriculture Organisation (FAO) of the UN publishes [color=rgb(0, 137, 201) !important]AquaStat which represents a "global information system on water and agriculture, developed by the Land and Water Division. The main mandate of the programme is to collect, analyze and disseminate information on water resources, water uses, and agricultural water management with an emphasis on countries in Africa, Asia, Latin America and the Caribbean." A bit more specifically, the main Aquastat database reports 70 variables under the headings 'Land use and population', 'Climate and water resources', 'Water use (by sector and by source)', 'Irrigation and drainage development' and 'Environment and health' for 5-year intervals from 1958-1962 onwards for a large number of countries. Other databases include the excellent 'Geo-referenced database on dams' and data on 'River sediment yields'. The data can be exported in CSV format.

CIMMYT, which stands for International Maize and Wheat Improvement Center (didn't you know?) have currently got three separate datasets on their [color=rgb(0, 137, 201) !important]website. First, some statistical price series for wheat, maize, sorghum, barley, rice, oil, fertilizers, and freight rate for wheat. This is a good dataset to act as reference global market price, since there are monthly observations for e.g. CIF Rotterdam price, but coverage varies a lot. Another interesting dataset is for agricultural production in Mexico: Agricultural information (1980-2008) related to planted area, harvested area, production, and production value of 657 permanent and seasonal crops, per cycle and regime (entries are in Spanish, but it's not too difficult to guess what 'valor ($)' means... Finally, they report the FAO data but you can pick alternative regional aggregation. [Thanks to [color=rgb(0, 137, 201) !important]Doug Gollin for these links]

The Center for International Earth Science Information Network ([color=rgb(0, 137, 201) !important]CIESIN) at Columbia's Earth Institute provides dozens of datasets under the headlines of Agriculture, Biodiversity & Ecosystems, Climate Change, Economic Activity, Environmental Assessment & Modeling, Environmental Health, Environmental Treaties, Indicators, Land Use (LU)/ Land Cover (LC) and LU/LC Change, Natural Hazards, Population, Poverty, Remote Sensing for Human Dimensions Research. The overarching theme for all datasets is environment and climate (change). Since not all data are accessible from the website there's a separate page for[color=rgb(0, 137, 201) !important]downloadable data.

The PBL Netherlands Environmental Assessment Agency provides the [color=rgb(0, 137, 201) !important]History Database of the Global Environment(interestingly, the acronym is HYDE). HYDE presents (gridded) time series of population and land use for the last 12,000 years ! It also presents various other indicators such as GDP, value added, livestock, agricultural areas and yields, private consumption, greenhouse gas emissions and industrial production data, but only for the last century.

[color=rgb(0, 137, 201) !important]Rural and Urban Education (i.e. rural = proxy for agricultural population) data (1960-1985) by C Peter Timmer is available in Chapter 29, 'Agriculture and economic development', of the Handbook of Agricultural Economics, Volume 2, Part 1, 2002, Pages 1487-1546. The link above is for the IDEAS RePec entry of this article: this is a copyrighted publication, but if you have access to the Handbook through your library you can easily copy the data. The coverage is exclusively for developing countries (N=65), and the data offers average years of schooling per person over the age of 25 for the rural and non-rural areas. OECD data on the same topic should allow for the inclusion of developed countries in the analysis.

[color=rgb(0, 137, 201) !important]International Land Quality Indexes (1987) by Willis Peterson covers relative land and cropland quality for 126 countries. The data is a single cross-section. The link is for the University of Minnesota, Department of Agriculture and Applied Economics paper (Staff Paper P87-10, 1987), which can be downloaded from the excellent [color=rgb(0, 137, 201) !important]AgEcon website at the same institution.

Still with the topic of arable land, the FAO Statistics Division also provides [color=rgb(0, 137, 201) !important]Gini data for land holdings as well as the Number and Area of Holdings by Tenure of Holdings. These data are decadal (1970/1980/1990).

The Köppen-Geiger Climate zones, documented in the [color=rgb(0, 137, 201) !important]Geography Datasets by John Gallup, Andrew Mellinger, and Jeff Sachs mentioned above are obviously a good resource for investigations of global agricultural production.

IFRPI offers acess to a number of [color=rgb(0, 137, 201) !important]geospatial datasets on agricultural production systems and agroeco-systems.

Louis Putterman at Brown University has compiled an [color=rgb(0, 137, 201) !important]Agricultural Transition Year Data Set which provides estimates for "the year when the first significant region within each of 165 present-day countries underwent a transition from reliance mainly on gathered wild and hunted food sources to reliance mainly on cultivated crops (and livestock)." This data is very much in line with the long-run growth theory work coming out of Brown.

[color=rgb(0, 137, 201) !important]AgroMaps at the FAO also has extensive geospatial data with relevance to agriculture.

The [color=rgb(0, 137, 201) !important]Agricultural Science and Technology Indicators (ASTI) are provided by IPRI. These are agricultural R&D indicators for developing countries only, with varying time-series coverage (earliest time around 1970, most recent up to around 2002). The data is split by institutional category (Higher Education, Private, Public Sector, NFP, government agencies) and provides numbers on researchers and R&D expenditure on agriculture.

For data covering agriculture R&D in developed countries check out the Science & Technology section of the UNESCO[color=rgb(0, 137, 201) !important]database.

Sectoral Data (ii): Manufacturing[color=rgb(0, 137, 201) !important]Trade, Production and Protection (TPT) was compiled by Alessandro Nicita and Marcelo Olarreaga at the World Bank. This provides data on 28 manufacturing sub-sectors for 1976-2004, compiled from UNIDO IndStat and COMTRADE - since neither of these two are freely available, the Nicita-Olarreaga dataset is quite a find. Although the files are quite hefty in size, the bilateral trade data should be interesting. Note also the (static) Input-Output matrices. Apart from the WB paper describing the dataset (see link on the TPT website), it helps reading [color=rgb(0, 137, 201) !important]this paper by Tetsuo Yamada of UNIDO.

$$ The Centre D'Etudes Prospectives et D'Information Internationales (CEPII) provides [color=rgb(0, 137, 201) !important]TradeProd, the Trade, Production and Bilateral Protection Database used in Mayer & Zingago (2005). This data is linked to the aforementioned Nicita & Olarreaga database and covers bilateral trade data for 1980-2004, manufacturing production data to 2004, and protection data to 2001.

$$ Said [color=rgb(0, 137, 201) !important]UNIDO IndStat4 database is now available in its 2008 version, providing number of establishments, employment, wages and salaries, output, value added, gross fixed capital formation (so that capital stock can be constructed via the perpetual inventory method) and number of female employees at the 4-digit-level of ISIC (Rev. 3!). This means data from 1990 onwards for 151 manufacturing categories. There is a second dataset contained at the 4-digit-level of ISIC (Rev. 2!), which covers 81 manufacturing sectors from 1980 onwards.
The [color=rgb(0, 137, 201) !important]UNIDO IndStat3 database from 2006 covers 29 manufacturing sectors in 181 countries from 1963-2004 at the 3-digit-level of ISIC (Rev. 2). The data covers the same variable as the previous dataset. This is the dataset of choice if you're running production functions since it has reasonable investment coverage from around 1970.
[color=rgb(0, 137, 201) !important]UNIDO IndStat2 2009 is out for 23 subsectors across 161 countries (1963-2007). Naturally, they do not collect data for the sectoral deflator, so it is necessary to go to the UN Common Database or individual country statistic offices to collect this information.

A Cross-Country Database For Sectoral Employment And Productivity In Asia And Latin America (1950-2005) by Marcel P. Timmer and Gaaitzen J. de Vries at the Groningen Growth and Development Centre. Described in the agriculture section above. This is now part of the [color=rgb(0, 137, 201) !important]10-sector database.

The [color=rgb(0, 137, 201) !important]Rural and Urban Education data (1960-1985) by C Peter Timmer described in the agriculture section above could be applied to manufacturing as well (non-rural education to proxy for education in manufacturing).

[color=rgb(0, 137, 201) !important]Occupational Wages around the World for 161 occupations in over 150 countries from 1983 to 2003, compiled by Richard Freeman and Reemco Oostendorp. Available at the NBER website (in STATA or ASCII format).

Sectoral Data (iii): ServicesThe World Bank’s [color=#089c9 !important]Services Trade Restrictions Database provides comparable information on services trade policy measures for 103 countries, five sectors (telecommunications, finance, transportation, retail and professional services) and key modes of delivery. "Compared to the vast empirical literature on policies affecting trade in goods, the empirical analysis of services trade policy is still in its infancy. One major constraint has been inadequate data on policies affecting services trade. Our limited knowledge of the pattern of services policy contrasts with the importance of services. Today, some 80 percent of GDP in the United States and the European Union originates from services, and the proportion is well over 50 percent in most countries, industrial and developing alike."

Trade Flows, Trade Protection/Policy and Globalisation[color=rgb(0, 137, 201) !important]The World Integrated Trade Solution (WITS) by the World Bank is not a dataset but a software which enables the use of COMTRADE, TRAINS (UNCTAD), IDB & CTS (WTO). The link is for the software - this only works if the user subscribes to the above dataset(s).

Mitch Abdon of the Stata Daily blog recently suggested a way of [color=rgb(0, 137, 201) !important]combining the UN ComtradeTools and Stata. Comtrade is the International Merchandise Trade Statistics (IMTS) of the UN, which records item-level trade for all countries in the world and contains around 1.8 bn observations from 1962 onwards. Access to this data is free, but for technical reasons a maximum of 50,000 observations per query (even more reason to use the Stata Daily application). Having installed the[color=rgb(0, 137, 201) !important]software which allows one to download Comtrade data (registration/subscription required for access) there are a number of simple steps to pull this data directly into Stata and save it. In fact, the entire process is run from within Stata once everything is installed. Since I had some minor trouble setting up and getting this tool to work I've written a [color=rgb(0, 137, 201) !important]simple Stata 10 do-file with additional information.

The European Commission's eurostat [color=rgb(0, 137, 201) !important]COMEXT database covers trade data from 1988 to 2009 (monthly or annual) for trade with the EU or its member countries. There are some restrictions on the maximum number of cells that can be downloaded, though. You may be better off going to COMTRADE and using the help by Mitch Abdon of the Stata Daily blog to [color=rgb(0, 137, 201) !important]combine the UN ComtradeTools and Stata:having installed the [color=rgb(0, 137, 201) !important]software which allows one to download Comtrade data (registration/subscription required for access) there are a number of simple steps to pull this data directly into Stata and save it. The entire process is run from within Stata once everything is installed. Since I had some minor trouble setting up and getting this tool to work I've written a [color=rgb(0, 137, 201) !important]simple Stata 10 do-file with additional information. [the COMEXT data is used in a recent [color=rgb(0, 137, 201) !important]ECB paper by Gabor Pula and Daniel Santabárbara]

[color=rgb(0, 137, 201) !important]UNCTAD Statistical Databases compiled by the UN Conference on Trade and Development cover items such as the World Investment Report (WIR) which has data on Foreign Direct Investment and Transnational Corporations. It also has Commodity Price Statistics, data on world trade in 'creative products', ICT statistics and TRAINS (mainly tariff data). Access is free as far as I know, but you need to register. It's a bit of a struggle to get through the menus, so basically look out for 'Interactive Database' on the side-menu, since this will offer access to the Beyond 20/20 database (at least in the case of FDI). The[color=rgb(0, 137, 201) !important]UNCTAD Handbook of Statistics now has been updated as well (is this subscription-based?).

Update September 2010: UNCTAD has now created a snazzy website that combines all of its statistical databases:[color=rgb(0, 137, 201) !important]UNCTADstat has lots of data on trade (merchandise, services), FDI flows and stocks (inward FDI from 1970!), external finance (incl. remittances), labour force/employment, global commodity price indices (from 1960!) as well as some more recent rubrics such as the creative and information economies and maritime transport (from around 2000).

The Global Trade Policy Analysis group at the AgEcon Department of Purdue University provides a number of datasets related to trade but also climate change and geography. "The GTAP Data Base is a [color=rgb(0, 137, 201) !important]fully documented, publicly available global data base which contains complete bilateral trade information, transport and protection linkages among [color=rgb(0, 137, 201) !important]113 regions for all [color=rgb(0, 137, 201) !important]57 GTAP commodities for a single year (2004 in the case of the GTAP 7 Data Base)." Single academic user licenses for GTAP 7 are $520, but a large number of free datasets (including summaries of GTAP, Social Accounting Matrix [SAM] extraction, the Global [bilateral] FDI Dataset, [color=rgb(0, 137, 201) !important]Project on Bilateral Labor Migration, CO2 emissions) can be found [color=rgb(0, 137, 201) !important]here.

The World Bank's [color=rgb(0, 137, 201) !important]Temporary Trade Barriers Database (TTBD) website hosts newly collected, freely available, and detailed data on more than thirty different national governments’ use of policies such as antidumping (AD, 1980s-2010), global safeguards (SG, 1995-2010), China-specific transitional safeguard (CSG, 2002-2010) measures, and countervailing duties (CVD, 1980s-2010). The information provided here in this detailed database will cover over 95% of the global use of these particular import-restricting trade remedy instruments. Information is provided in excel files on a country-by-country basis, given the amount of detail provided for each county. The website also features research reports and meta-information. Chad P. Bown seems to be the person in charge.

Kristian Skrede Gleditsch at Essex University provides data estimates of [color=rgb(0, 137, 201) !important]trade flows between independent states (1948-2000) and GDP per capita of independent states (1950-2004) for around 150 countries. These data files are too large to be opened in excel, but Kristian provides some tips on how to proceed.

The World Shipping Register provides free access to their [color=rgb(0, 137, 201) !important]World Sea Ports database. For each country its ports' longitude, latitude and time zone are provided, for some port the maximum draft is also provided. Given the geospatial information this data could be used to calculate distance to closest port.

$$ The Centre D'Etudes Prospectives et D'Information Internationales (CEPII) [color=rgb(0, 137, 201) !important]BACI (Italian for kiss, if I'm not mistaken; wonder whether the French knew that) dataset aims to provide the most disaggregated international trade database (more than 5,000 products) for the largest number of countries (over 200) and years (from 1995 to 2005, with updates to follow). It took me a while to realise that there is a disclaimer: "Files by year of BACI data for the period 1995-2005 are available for researchers already subscribing to the United Nations COMTRADE database. Users of BACI have to testify that their organisation is fully licensed COMTRADE to download BACI." So no baci after all.

The [color=rgb(0, 137, 201) !important]TPT database by Nicita & Olarreaga mentioned above has data on manufacturing trade, including bi-lateral trade flows (1976-2004).

The FAO provides trade data for agricultural goods in its [color=rgb(0, 137, 201) !important]TradeSTAT database. This can also be accessed via the World Resources Institute link - both of these can be found in the Agriculture section of this webpage.

The World Bank’s [color=#089c9 !important]Services Trade Restrictions Database provides comparable information on services trade policy measures for 103 countries, five sectors (telecommunications, finance, transportation, retail and professional services) and key modes of delivery. "Compared to the vast empirical literature on policies affecting trade in goods, the empirical analysis of services trade policy is still in its infancy. One major constraint has been inadequate data on policies affecting services trade. Our limited knowledge of the pattern of services policy contrasts with the importance of services. Today, some 80 percent of GDP in the United States and the European Union originates from services, and the proportion is well over 50 percent in most countries, industrial and developing alike."

A team at the World Bank comprising Tolga Cebeci, Ana M. Fernandes, Caroline Freund, and Martha Denisse Pierola have come up with the [color=#089c9 !important]Exporter Dynamics Database. This presently covers around 45 developed and developing countries, covering mainly 2003-2009 but also the 1990s for some countries. "It allows for cross-country comparisons of exporters based on factors such as size, survival, growth, and concentration. More countries will be added as the database expands. Until now, most databases focus not on exporting firms, but on the aggregate flow of goods across borders based on countries or products." Melitz will be happy!

Matthew Ciolek at Australian National University edits the site for the [color=#089c9 !important]Old World Trade Routes (OWTRAD) Project: "This site supports online research in the field of dromography and provides a public-access electronic archive of geo/chrono-referenced data on land, river and maritime trade routes of Eurasia and Africa during the period 10,000 BCE - circa 1820 CE." The files are published in CSV, MapInfo and Google Earth (KML) formats, downloadable by region. There's also a link to the Trade Routes Resources [color=#089c9 !important]blog [via Masa Kudamatsu's [color=#089c9 !important]DevEconData blog]

The IADB website hosts the [color=#089c9 !important]data used in the work on trade intensity and business cycles by César Calderón, Alberto Chong and Ernesto Stein (2006, JIE). From the abstract: "Using annual information for 147 countries for the period 1960-99 we find that the impact of trade intensity on business cycle correlation among developing countries is positive and significant, but substantially smaller than that among industrial countries. Our findings suggest that differences in the responsiveness of cycle synchronization to trade integration between industrial and developing countries are explained by differences in the patterns of specialization and bilateral trade."  

The World Bank recently completed a big data compilation exercise for [color=#089c9 !important]Distortions to Agricultural Incentives, with a team of researchers headed by Kym Anderson providing various Estimates of Distortions to Agricultural Incentives (1955-2007). A core database provides data for Nominal Rates of Assistance to producers (NRAs), together with a set of Consumer Tax Equivalents (CTEs), for farm products and a set of Relative Rates of Assistance to farmers in 75 focus countries. Note that the variable 'border price' (bp) does however not represent the... how can I say this... 'border price', but a hypothetical producer price in the absence of distortions (domestic producer price divided by (1+NRA) and expressed in USD). The border price (fob) is not contained in the main datafile but can be found in the individual country spreadsheet (rows 37-39 for primary products, or 44-46 for lightly processed products). I am grateful to Kym Anderson and Ernesto Valenzuela for clarification; they also point to an [color=#089c9 !important]alternative data reporter at Adelaide University where they are both based.

The OECD has a dedicated database for [color=rgb(0, 137, 201) !important]PSE (Producer and Consumer Support Estimates) which covers OECD member states as well as a small number of Eastern European 'Emerging' Economies and the BRICS countries for 1986 to 2008. A recent[color=rgb(0, 137, 201) !important]paper by Kym Anderson compares and contrasts the methodology applied in his own work (see previous entry) constructing measures of agricultural production and trade distortions with that of the OECD.

The Agricultural Market Access Database ([color=rgb(0, 137, 201) !important]AMAD) is a collection of available public data on WTO market access in agriculture. It contains data for over 50 countries. After registration, all files can be downloaded for free (self-extracting zip files) and there is documentation on how to do this.

[color=rgb(0, 137, 201) !important]Proximity in Product Space and Diversification Strategies, compiled by Valentino Piana, is not a standard trade dataset but contains data emerging from research by C. A. Hidalgo, B. Klinger, A.-L. Barabasi, and R. Haussman, published in Science under the title [color=rgb(0, 137, 201) !important]The Product Space Conditions the Development of Nations (subscription required?). The basic idea in the paper is that the product space is like a forest, and each product is a tree, with the fanciest (highest value-added) products concentrated in the centre. Countries/producers in this product space are then like monkeys that jump from tree to tree... I'd prefer if they'd taken squirrels. Using 4-digit level trade data for a large number of countries (not sure what the time-horizon is) the authors establish the probability of a country producing and exporting good x, given that it produces and exports good y. Data at the above link includes 775-by-775 matrix of revealed proximities between products among other things.

World Bank data on [color=rgb(0, 137, 201) !important]Trade and Import Barriers has been collated by Francis Ng. Most significant time-series dimension among these is for [color=rgb(0, 137, 201) !important]Trends in average applied tariff rates in developing and industrial countries, 1981-2007.

The [color=rgb(0, 137, 201) !important]NBER-UN Trade Data (1962-2000) is available at the Center for International Data, UC Davis website. This data, constructed by the Bobs Feenstra and Lipsey with co-authors (see [color=rgb(0, 137, 201) !important]NBER paper #11040 for details), is organised by 4-digit ITC (Rev.2). Primacy is given for the mirror data (i.e. trade flows as reported by importing country). Wide country coverage. Users must agree not to resell or distribute the data for 1984-2000.

The [color=rgb(0, 137, 201) !important]homepage of Andrew Rose, an economist at Haas Business School, UCB, has tonnes of data on trade, including bi-lateral trade data from 1950-1999, which can be found in the file associated with the paper “Do We Really Know that the WTO increases Trade?”. This dataset also contains a small number of geographical variables and lots of information on regional trade agreements, GATT/WTO, etc. Note that the country codes are the IMF International Financial Statisticss codes, a list of which can be found [color=rgb(0, 137, 201) !important]here.
The dataset associated with "Does the WTO Make Trade More Stable?" seems to have a slightly better coverage, but with somewhat less variables.
Various measures for 'Remoteness' of a country (linked to GDP and distance measures) are contained in the panel data link associated with the "Currency Unions and International Integration" paper - coverage is 1960-1996 for up to 210 countries. He has a more up-to-date version of the same variable to 2000 somewhere on his website but I haven't been able to find it yet.

The World Bank compiles the [color=rgb(0, 137, 201) !important]World Trade Indicators, covering 299 indicators for 210 countries from 1995 to 2007.