PDF Wrapper

The gadget spec URL could not be found

 2002 Eaton, Kortum, ECMA, Technology, Geography, and Trade  United Nations International Comparison Program

[for price level in 19 countries (50 manufactured products)]
 2010 Waugh, AER, International Trade and Income Differences  Feenstra, Robert C., Robert E. Lipsey, and Harry P. Bowen. World trade flows, 1970-1992, with production and tariff data. No. w5910. National Bureau of Economic Research, 1997.

[aggregate value of the manufactured goods]
United Nations Industrial Development Organization (UNIDO)

[for Gross manufacturing data]
 ICP benchmark data
from the Penn World Table Website

[ICP is done by the World Bank, and the PWT uses the ICP to calculate the PPP.]
 Caselli (2005)

[for capital-labor ratios and labor endowments]
   Mayer, T. & Zignago, S. (2011)
Notes on CEPII’s distances measures : the GeoDist Database

[for information on distance, colony, contiguity, etc. ]
 2013 Lu, Tao, Zhang, JIE, How Do Exporters Respond to Antidumping Investigations  China Customs data (2000–2006)  World Bank Global Antidumping Database

Bown, Chad P. (2015) "Global Antidumping Database,” The World Bank, June, available at 
 We first follow the method developed by Ahn et al. (2011) for the same data by dividing firms in our sample into trade intermediaries and direct exporters.

Specifically, trade intermediaries are identified as firmswhose names contain Chinese characters (i.e., “Jinchukou”, “Jingmao”, and “Maoyi”) with the Englishequivalent meaning of importer, exporter, and/or trading. The validity of this identification approach comes from the legacy of China's centrally-planned system and the reform strategy adopted after 1978. Specifically, to insulate the Chinese domestic market from international competition, the Chinese central government only authorized 12 stateowned enterprises to conduct exports and imports in the pre-reform era (i.e., 1949–1978). These aforementioned Chinese characters were used for easy identification and regulation. Since 1978, China has adopted a gradualism approach in liberalizing its economy, with an increasing number of firms allowed to conduct foreign trade. However, the tradition of using self-revealing names for trading corporations has continued in the post-reform era. Ahn et al. (2011) find that firms identified as trade intermediaries by this method are indeed very different from direct exporters in terms of the trading volume, product categories, and export destinations.
 Row 3 includes the interactions with the quality ladder taken from Khandelwal (2010).  
 2013 Novy, EI, Gravity Redux; Measuring International Trade Costs with Panel Data  Kee, Nicita, and Olarreaga (2009) propose a trade restrictiveness index that is based on observable tariff and non-tariff barriers. They show that tariffs alone are a poor indicator of trade restrictiveness because non-tariff barriers also provide a considerable degree of trade protection.  All bilateral aggregate trade data are taken from the International Monetary Fund (IMF) Direction of Trade Statistics (DOTS) and denominated in U.S. dollars.
 Moreover, GDP data include services that are not covered by the trade data. To get the gross shipment counterpart of GDP excluding services I follow Wei (1996) in constructing y_i as total goods production based on the OECD’s structural analysis (STAN) database.

But GDP data tend to overstate the extent of intranational trade and thus the level of trade costs because they include services. I therefore prefer to follow Wei (1996) in using merchandise production data to match the trade data more accurately.
 The production data are converted into U.S. dollars by the period average exchange rate taken from the IMF international financial statistics (IFS).  
    The following variables are taken from Rose’s (2000) data set made available on his website: the adjacency dummy, the common language dummy, the free trade agreement dummy, and the island variable. The island variable takes on the value 1 if one of the trading partners is an island and the value 2 if both partners are islands, and 0 otherwise. There is also a variable for colonial relationship.  The distance data represent great-circle distances between capital cities. They are collected from the website
 2011 Costinot, Donaldson, Komunjer, RES, What Goods Do Countries Trade; A Quantitative Exploration of Ricardo's Ideas  We use trade ‡ow data taken from the OECD Structural Analysis (STAN) Bilateral Trade Database (edition 2008) relating to the year of 1997.  The import penetration ratio (IPR) is calculated as imports in the industry divided by the sum of gross output and net imports in the industry. Fifteen exporter-industry observations have an import penetration ratio (IPR) that exceeds one, a fact that presumably reflects re-export activity whose domestic gross output content is recorded in other sectors, such as the services sector. We replace these observations with the maximum IPR value, within each exporter, among those values that are less than one.  Our producer price data are taken from the GGDC Productivity Level Database; see Inklaar and Timmer (2008) for details. In this database, raw price data observations are first collected at the plant level for several thousand products (often with hundreds of products per industry, which we interpret as varieties in the context of our model). This is only made possible due to the use of the PRODCOM system of homogeneous product descriptions within the EU and OECD. The GGDC database uses the PRODCOM system to pay particular attention to matching products in di¤erent countries in order to control for quality di¤erences. These prices are then aggregated into a unique producer price index at the industry level using output data.
 To attempt to circumvent these two potential sources of bias, we estimate Equation (18) by the method of instrumental variables, with the endogenous regressor --  productivity levels (ln ezki ) instrumented with (the log of) research and develop-ment (R&D) expenditures at the country-industry level.20 This approach follows Eaton and Kortum (2002) and Gri¢ th, Redding, and Van Reenen (2004) in modeling technology as a function of R&D activity. In doing so, our identifying assumption is that relative R&D expenditures are correlated with relative bilateral trade flows only through their impact on relative productivity, i.e. relative producer prices.  
   Data on R&D expenditure are from the Analytical Business Enterprise Research and Development (ANBERD) database collected by the OECD -- as used, for example, by Griffith, Redding, and Van Reenen (2004). Wherever possible this source aims to break down the R&D expenditures of large, multi-industry firms, as well as those of enterprise-serving research institutes, into expenditures by output product. Sixteen exporter-industry observations have missing R&D expenditure information in this data source. We interpolate these observations using the fitted values from a regression on exporter and industry fixed effects.        
 2016 Cosar, Fajgelbaum, AEJ, Internal Geography, International Trade and Regional Specialization  Our regional data of Chinese industries is aggregated up from the firm-level Annual Survey of Industrial Production conducted by the Chinese government’s National Bureau of Statistics. The Annual Survey is a census of private firms with more than 5 million yuan (about $600,000) in revenue and all state-owned firms. It covers 338 mainland prefectures and 425 manufacturing industries in four-digit CSIS Chinese classification system between the years 1998–2007.  Maritime transportation is the primary mode of shipping in external trade, while exports over land to bordering countries account for a small share (6.7 percent) of total exports (authors’ calculation using 2006 UN Comtrade data). The share of air shipments in exports to top 20 trade partners is just 17.4 percent (Harrigan and Deng 2010). While inland rivers play an important role in freight transportation in general, their export share is limited. Inland river ports, most of which are also in close proximity to the sea, constitute only 20 percent of port capacity (in terms of tonnage) suitable for international trade (authors’ calculation using data from Chinese Statistical Agency).  To construct industry-specific, labor-capital ratios in the United States, we use the publicly available NBER-CES Manufacturing Industry Database (Becker et al. 2013) and divide production worker hours by the total real capital stock for each industry.17 The NAICS 2002 classification—in which the US data is reported—can be mapped to the CSIS Chinese classification by using the official concordance between CSIS and ISIC Rev. 3 as a bridge.    
 2003 Poncet, CER, Measuring Chinese Domestic and International Integration          
 2004 Redding, Venables, JIE, Economic Geography and International Inequality  Manufacturing wage per worker: data on number of employees and wages and salaries (current price US dollars) in total manufacturing are from the UNIDO Industrial Statistics Database. Information is available for 85 countries during 1992–1996.  Relative price of machinery and equipment: data on the price of machinery and equipment and GDP in local currency units per US dollar are from Phase Vof the United Nations International Comparisons Project (United Nations, 1994). The data are available for 46 countries and are for 1985.  Number of Minerals: the total number of minerals of which a country has reserves from the list of 44 main minerals compiled by Parker (1997).  Risk of Expropriation: extent of protection of property rights, measured on a scale from 1 to 5, where a higher score indicates weaker protection of property rights. Source: Holmes et al. (1997). These data are unavailable for Central African Republic, Guatemala, Kazakhstan, Kyrgyz Republic, Madagascar, Macedonia, Mauritius, and Chad.  
   Physical Geography and Institutional, Social, and Political Characteristics: data on hydrocarbons (deposits of petroleum and natural gas) per capita, fraction of land area in the geographical tropics, prevalence of malaria, socialist rule, and the occurrence of an external war are from Gallup et al. (1998). Information is available for all 101 countries in our data set, except for the data on hydrocarbons per capita which are unavailable for Moldova and Yemen. The data can be downloaded from  International Openness: data on international openness are from Sachs and Warner (1995) and are available for 98 countries. Information is unavailable for Panama, Saudi Arabia, and Yemen (see      
 2013 Autor, Dorn, Hanson, AER, The China Syndrome; Local Labor Market Effects of Import Competition in the United States  Brandt,Van Biesebroeck, and Zhang (2012) estimatetes that from 1998 to 2007, China had average annual TFP growth in manufacturinog of  8.0 percent, compared to the Bureau of Labor Statistics's estimate ( op 3.9 percent of the United States.  To assess the effect of imports of Chinese goods on local labor markets, we need to define regional economies in the United States. Our concept for local labor markets is Commuting Zones (CZs) developed by Tolberta nd Sizer (1996), who used county-level commuting data from the 1990 Census data to create 741 clusters of counties that are characterized by strong commuting ties within CZs, and weak commuting ties across CZs. Our analysis includes the 722 CZs that cover the entire mainland United States (both metropolitan and rural areas).      

National Establishment Time-Series (NETS)