Krishna G. Iyer, Alicia N. Rambaldi, and Kam Ki Tang, "Efficiency Externalities of Trade and Alternative Forms of Foreign Investment in OECD Countries", Journal of Applied Econometrics, Vol. 23, No. 6, 2008, pp. 749-766. All files are ASCII (text) files in DOS format. They are zipped in the file irt-files.zip. Unix users should use "unzip -a". The data for the variables used in the empirical analysis of this article are drawn from a variety of sources, namely: Penn World Tables Version 6.1 (PWT 6.1), Penn World Tables Version 5.6 (PWT 5.6), World Development Indicators 2006 (WDI), SourceOECD, International Financial Statistics of IMF (IFS), and Barro and Lee human capital series. The variables used are listed below. Notations, where used in the article, are shown in parenthesis: Output, Source: PWT 6.1 Capital, Source: PWT 5.6 and 6.1 Labour, Source: WDI Domestic R&D (DRD), Source: SourceOECD Human Capital (HC), Source: Barro and Lee Foreign Direct Investment Inflows (FDII), Source: IFS Foreign Direct Investment Outflows (FDIO), Source: IFS Foreign Portfolio Investment Inflows (FPII), Source: IFS Foreign Portfolio Investment Outflows (FPIO), Source: IFS Other Foreign Investment Inflows (OFII), Source: IFS Other Foreign Investment Outflows (OFIO), Source: IFS Trade Openness (TOP), Source: WDI Financial Market Development (FMD), Source: SourceOECD Relative R&D (RRD), Source: SourceOECD Other variables include time trend and interaction variables constructed by multiplying some variables listed above. Models: All estimations have been obtained with Frontier 4.1. For Frontier 4.1, see http://www.uq.edu.au/economics/cepa/frontier.ht Four empirical models are presented in the article. There is one main model and three nested ones (A, B and C). All models have two components: a production function and an inefficiency function. For each model, 3 files are required: a data file (*.dta), a command file (*.ins), and an output file (*.out). For the main model, the files are mm.dta, mm.ins and mm.out. For nested model A the files are nma.dta, nma.ins and nma.out For nested model B the files are nmb.dta, nmb.ins and nmb.out For nested model C the files are nmc.dta, nmc.ins and nmc.out The production function for the main model is now detailed (nested models will be discussed later). Production Function: Output is Real GDP in natural logarithms. Inputs include capital, labour, DRD and HC. Capital is net of residential capital and in natural logarithms. PWT 5.6 was used to filter out residential capital from total capital and perpetual inventory method was applied to derive capital stock estimates from investment flows. See appendix of the paper for further details. Labour is number of people in the labour force, in natural logarithms. DRD is in natural logarithms. Perpetual inventory method was applied to derive stock estimates from R&D investment flows. See appendix of the paper for further details. HC is in percentage. Share of population with higher school attainment. Translog specification is applied. So, there are squared terms and cross products in the production function. Notably, all the production function variables have been mean differenced. This ensures that the coefficients of the first order terms can be interpreted as elasticities. Inefficiency Function: The functional forms of the inefficiency explaining variables in the main model are as follows: FDII: Ratio, FDII over GDP FDIO: Ratio, FDIO over GDP FPII: Ratio, FPII over GDP FPIO: Ratio, FPIO over GDP OFII: Ratio, OFII over GDP OFIO: Ratio, OFIO over GDP TOP: Ratio, (Exports plus Imports) over GDP HC: Percentage. Share of population with higher school attainment. FMD: Ratio, Contribution of the financial sector (direct as well as indirect) over the total value added in the economy. RRD: Ratio, Country stock of R&D over OECD stock of R&D. Stock of R&D estimated using the perpetual inventory method on R&D investment data. FDIIxHC, FDIIxRRD, FDIIxFMD, FPIIxFMD, OFIIxFMD are derived as simple multiplications of the two variables. For example, FDIIxHC is FDII multiplied by HC. The ordering of the variables in the mm.dta files corresponds with order in Table 4, column Main model. For ease of understanding, mm dta and ins files are now discussed: mm.dta This file has 380 rows and 39 columns (A-AM in an excel sheet). From the article, it is apparent that this is a panel data set covering 20 countries over 19 time periods (years). Rows Rows 1-20 are the first year observations for the 20 countries, Rows 21-40 are the second year observations for the 20 countries and so on. Rows 361-380 are the last or nineteenth year observations for the 20 countries. Columns Column1(A): Country Code Column2(B): Year Column3(C): Output Column4(D): Capital Column5(E): Labour Column6(F): DRD Column7(G): HC Column8(H): Time (i.e., year) Column9(I): Capital Squared Column10(J): Labour Squared Column11(K): DRD Squared Column12(L): HC Squared Column13(M): Time Squared Column14(N): Capital x Labour Column15(O): Capital x DRD Column16(P): Capital x HC Column17(Q): Capital x Time Column18(R): Labour x DRD Column19(S): Labour x HC Column20(T): Labour x Time Column21(U): DRD x HC Column22(V): DRD x Time Column23(W): HC x Time Column24(X): FDII Column25(Y): FDIO Column26(Z): FPII Column27(AA): FPIO Column28(AB): OFII Column29(AC): OFIO Column30(AD): TOP Column31(AE): HC Column32(AF): FMD Column33(AG): RRD Column34(AH): FDIIxHC Column35(AI): FDIIxRRD Column36(AJ): FDIIxFMD Column37(AK): FPIIxFMD Column38(AL): OFIIxFMD Column39(AM): Time (i.e, year) mm.ins file is self-explanatory. mm.out mm.out is a file generated by the software programme Frontier 4.1. Details on how to run this programme can be found in http://www.uq.edu.au/economics/cepa/frontier.htm Nested Models The nested models A, B and C are similar to the main model. The nested models have the identical production function as the main model. So the production function is not discussed in the context of the Nested models. The inefficiency vectors are different (see Table 4 of the article). This is now discussed. Nested Model A nma.dta Rows Rows are identical to mm.data Columns In nested model A, FPII, FPIO, OFII, OFIO, FPIIxFMD and OFIIxFMD are not included. The data file is adjusted such that: Column26(Z): TOP Column27(AA): HC Column28(AB): FMD Column29(AC): RRD Column30(AD): FDIIxHC Column31(AE): FDIIxRRD Column32(AF): FDIIxFMD Column33(AG): Time Note: Column33(AG) is the last column. nma.ins file is self-explanatory. nma.out nma.out is a file generated by the software programme Frontier 4.1. Nested Model B Rows Rows are identical to mm.data Columns In nested model B, FPII, FPIO, OFII, OFIO, TOP, FPIIxFMD and OFIIxFMD are not included. The data file is adjusted such that: Column26(Z): HC Column27(AA): FMD Column28(AB): RRD Column29(AC): FDIIxHC Column30(AD): FDIIxRRD Column31(AE): FDIIxFMD Column32(AF): Time Note: Column32(AF) is the last column. nmb.ins file is self-explanatory. nmb.out nmb.out is a file generated by the software programme Frontier 4.1. Nested Model C Rows Rows are identical to mm.data Columns In nested model C, HC, FMD, RRD, FDIIxHC. FDIIxRRD, FDIIxFMDD, FPIIxFMD and OFIIxFMD are not included. The data file is adjusted such that: Column31(AE): Time Note: Column31(AE) is the last column. nmc.ins file is self-explanatory. nmc.out nmc.out is a file generated by the software programme Frontier 4.1. Please address any questions to: Krishna G Iyer Department of Economics and Finance Massey University, Palmerston North, New Zealand Email: K.Iyer [AT] massey.ac.nz