Definition and Calculation of the Vertical Spillover Variables

Supplementary Materials

Definition and calculation of the vertical spillover variables

In the paper, we follow Lenaerts and Merlevede (2012) when calculating vertical spillover variables. To date, most of the literature has used the definitions introduced by Javorcik (2004). Here we highlight the difference between both approaches and show that our findings also hold when we make use of the Javorcik (2004) definition.

The starting point for the calculation is the same. The definition of the horizontal spillover variable reproduced as in equation (A.1) is exactly the same as in equation (1) in the main text.

where Yit is the output produced by firm i in year t. HRjt is industry j’s share of output that is produced by foreign firms. Foreign firms are identified by Fit which is the share of foreign participation in firm i in year t. HRjt is then combined with input-output coefficients obtained from input-output tables to calculate the vertical spillover variables. The difference between Lenaerts and Merlevede (2012) and Javorcik (2004) arises here. Javorcik (2004) excludes inputs sold within the firm’s own industry j since these are already accounted for HRjt. However, Lenaerts and Merlevede (2012) argue that intermediate trade within the firm’s own industry j (i.e. the diagonal elements in the IO-tables for intermediate consumption) should be included in the definition of the vertical spillover variables because these diagonal elements still refer to supplier-client relationships (which thus give rise to a different type of spillover effect than competitive relations). Intuitively, one can also understand this by imagining two input-output tables for the same economy where the first uses a broad industry aggregation and the other a more detailed industry aggregation. The ‘aggregated’ IO-table will contain a lot more intermediate trade within the firm’s own industry that gives rise to vertical spillovers.

The Javorcik (2004) definition of the backward spillover variable, BK_JAVjt, is the following:

as before γjkt is the proportion of industry j’s output supplied to sourcing industry k at time t. The γs are calculated from time-varying IO-tables for intermediate consumption. The forward spillover variable FW_JAVjt is defined as:

where the IO-tables reveal the proportion δjlt of industry j’s inputs purchased from upstream industries l.

Table A.1 repeats the analysis of Table 4 in the main text using the Javorcik (2004) definitions for the backward and forward spillover variables. Comparing Table 4 and Table A.1 confirms the findings of Lenaerts and Merlevede (2012): the use of the Javorcik (2004) definition of vertical spillovers tends to make the horizontal spillover variable significant and the backward spillover insignificant. It also tends to decrease the backward spillover effect as point estimates of coefficients are comparable, while the average value of the backward spillover variable is smaller when the diagonal is excluded. The main findings in Table 4 are qualitatively confirmed. The smallest and the largest foreign firms do not generate spillover effects. There is also some indication that foreign firms that employ between 10 and 50 employees generate some spillover effects in this case.

Table A.1: Spillover effects across domestic and foreign firm size categories. The Javorcik (2004) definition is used to construct the backward and forward spillover variables (instead of the method of Lenaerts and Merlevede (2012)).

OP TFP / ACF TFP
All dom.
firms / Domestic firms with average number of employees / All dom. firms / Domestic firms with average number of employees
<10 / 10-50 / 50-250 / >250 / <10 / 10-50 / 50-250 / >250
HR
<10 / -0.243 / -0.041 / -0.333 / -1.238* / -0.503 / -0.620 / -0.289 / -0.080 / -3.681** / -0.987
[0.618] / [0.630] / [0.570] / [0.716] / [0.651] / [1.607] / [1.631] / [1.596] / [1.821] / [1.451]
10-50 / 0.713 / 0.767 / 0.760 / 0.676 / 0.607 / 2.070* / 2.229* / 2.014 / 2.042* / 2.304
[0.551] / [0.597] / [0.542] / [0.413] / [0.433] / [1.223] / [1.279] / [1.276] / [1.142] / [1.538]
50-250 / 0.944*** / 1.010*** / 0.848*** / 0.828*** / 0.593*** / 3.050*** / 3.211*** / 2.966*** / 2.710*** / 1.551**
[0.268] / [0.288] / [0.248] / [0.239] / [0.225] / [0.681] / [0.720] / [0.658] / [0.764] / [0.696]
>250 / 0.091 / 0.122 / 0.036 / 0.092 / -0.029 / 0.250 / 0.273 / 0.435 / 0.010 / -0.256
[0.184] / [0.186] / [0.206] / [0.190] / [0.132] / [0.514] / [0.529] / [0.578] / [0.564] / [0.475]
BK
<10 / -0.442 / -0.477 / -0.332 / -1.074 / -0.504 / -1.665 / -1.590 / -1.690 / -3.246 / -0.070
[1.606] / [1.712] / [1.716] / [1.363] / [0.912] / [3.512] / [3.852] / [3.633] / [3.006] / [2.140]
10-50 / 2.437 / 2.276 / 3.499* / 2.724** / 1.473*** / 5.043 / 4.942 / 7.398 / 4.697 / 3.035**
[1.568] / [1.915] / [1.828] / [1.340] / [0.513] / [3.867] / [5.137] / [4.654] / [2.988] / [1.219]
50-250 / 1.667* / 1.454 / 1.306 / 2.422** / 2.414*** / 4.986* / 4.499 / 3.542 / 8.041** / 4.529**
[0.993] / [1.071] / [1.052] / [1.146] / [0.665] / [2.872] / [3.019] / [3.610] / [3.523] / [1.754]
>250 / 0.913 / 0.996 / 0.831 / 0.870 / 0.679 / 2.072 / 2.118 / 2.812 / 1.227 / 0.355
[0.724] / [0.751] / [0.746] / [0.629] / [0.423] / [1.928] / [2.025] / [2.051] / [1.464] / [1.215]
FW
<10 / -5.031 / -5.051 / -5.479 / -3.893 / -2.199 / -7.390 / -6.673 / -6.390 / -11.666* / -6.033
[3.302] / [3.427] / [3.367] / [2.928] / [2.023] / [6.836] / [7.134] / [7.241] / [6.523] / [5.734]
10-50 / 1.287 / 1.267 / 1.130 / 1.240 / -0.729 / 3.285 / 2.872 / 3.184 / 4.847* / -1.289
[1.276] / [1.306] / [1.287] / [1.126] / [1.015] / [3.452] / [3.461] / [3.552] / [2.909] / [3.303]
50-250 / -3.496
*** / -3.420
*** / -3.383
*** / -3.733
*** / -2.716
*** / -8.009
*** / -8.069
*** / -8.118
*** / -6.890
*** / -3.052*
[1.009] / [1.038] / [0.920] / [0.938] / [0.905] / [2.019] / [2.091] / [1.914] / [1.913] / [1.819]
>250 / -0.043 / 0.193 / -0.397 / -0.493 / -0.445 / -0.803 / -0.518 / -1.281 / -1.103 / -1.298
[0.622] / [0.671] / [0.588] / [0.545] / [0.470] / [1.718] / [1.952] / [1.686] / [1.214] / [1.168]
Obs. / 167,022 / 107,780 / 42,847 / 11,161 / 5,234 / 120,763 / 74,783 / 32,262 / 9,081 / 4,637
R² / 0.057 / 0.059 / 0.091 / 0.107 / 0.144 / 0.079 / 0.088 / 0.111 / 0.119 / 0.108

Spillovers from four size categories of foreign firms (micro, small, medium and larger) on all domestic firms (columns 1 and 6) and on four size categories of domestic firms (columns 2-5 and 7-10). Results for OP TFP (columns 1-5) and ACF TFP (columns 6-10). Robust standard errors in brackets. ***/**/* denotes significance at 1/5/10 percent.

The classification of firms as foreign or domestic

In the paper, we classify firms as foreign on the basis of the OECD FDI benchmark definition,taken from the OECD Glossary ( In this definition, firms are classified as foreign when there is at least a single foreign investor who owns no less than 10% of the shares. In our data on Romania, we find a substantial number of micro and small foreign firms.

One may wonder whether these firms are truly ‘small’ and whether our results are specific for Romania (which was characterized by low fixed costs of entry that are associated with engaging in FDI). In order to invest abroad, firms have to overcome the fixed entry costs. Helpman et al. (2004) point out that only the most productive firms are able to bear these costs and conduct FDI. These entry cost might have been relatively low in Romania in the sample period, which facilitates the entry of small foreign firms (which can enter more easily than on large developed market such as the US or Japan). Nevertheless, the productivity distributions across the size classes of foreign firms in our data do not differ much. Firms of all sizes can therefore be expected to enter the Romanian market, and even when the fixed entry costs go up, still firms of all size classes can enter provided that they are sufficiently productive. Our results are context-dependent but should, as a result, be generalizable to other developing economies as well.

In order to shed some more light on the issue, we consider the country of origin of the foreign investors in Romania using a more recent version of the Amadeus database (because this information is not available in the dataset we use in the paper). We compare firms from six country groups: EU15, CEEC (excluding Romania), Other European countries, US and Canada, Asia and Other non-European countries. The size distributions of foreign firms from each of these country groups are highly similar, as evidenced by Table A.2 and Figure A.1 below. The majority of the foreign firms are micro and small firms. Large firms only represent about 2% of the foreign firms in Romania. This suggests that the smaller firms indeed actually are small and that the spillover effects we detect in the paper do not appear to be driven by firms of a specific country or region of origin.

Table A.2:Overview of the country of origin of the foreign owners in Romania by size class.

Micro / Small / Medium / Large
EU-15 / 68.2% / 20.3% / 8.9% / 2.6%
CEEC-10 (minus RO) / 78.5% / 17.7% / 3.1% / 0.7%
Other Europe / 61.0% / 21.0% / 13.0% / 5.1%
US-Canada / 70.0% / 22.2% / 5.6% / 2.2%
Asia / 79.2% / 15.4% / 4.2% / 1.2%
Other Non-Europe / 70.5% / 17.8% / 9.0% / 2.7%
Total / 71.6% / 19.0% / 7.2% / 2.2%

Figure A.1: Overview of the country of origin of the foreign owners in Romania by size class.

A related issue is the degree of foreign ownership and its impact on the spillovers transmitted to domestic firms.An interesting study on this topic is the work of Javorcik and Spatareanu (2008)[1], who compare the spillover effects generated by wholly and partially foreign-owned firms. Javorcik and Spatareanu (2008) argue that especially partially foreign-owned firms can play an important role as sources of FDI spillover effects because the costs of finding a local supplier are lower for these firms and because they likely transfer less sophisticated technologies to their affiliate (which are then easier to adopt by domestic companies). Their results indeed indicate that positive backward spillover effects are mainly associated with partially foreign-owned firms, and for these firms the negative horizontal spillover effects detected are also smaller. This implies that a higher degree of foreign ownership is not necessarily reflected in more positive spillover effects. For that reason, we use a 10% cutoff of foreign participation to classify firms as foreign (as is common in the spillover literature). However, the get a better understanding of how the ownership structure of firms may affect results, we use a more stringent definition of at least 50% of foreign participation to categorize a firm as foreign. This 50% threshold ensures that the owner of the firm indeed is the decision maker (who can decide to import intermediates or convince their original suppliers to also invest abroad). We performed a similar analysis as in Table 4 but now use the 50% threshold. The results of this analysis are presented in Table A.3 below and are fully in line with the findings in Table 4. Again, spillover effects are generally transmitted by medium-sized foreign firms. Note that most foreign firms in Romania are more than 50% foreign-owned.

Table A.3:Spillover effects from majority foreign-owned firms across domestic and foreign firm size categories.

OP TFP / ACF TFP
All dom. firms / Domestic firms with average number of employees / All dom. firms / Domestic firms with average number of employees
<10 / 10-50 / 50-250 / >250 / <10 / 10-50 / 50-250 / >250
HR
<10 / 0.772 / 1.128 / 0.481 / -0.728 / -0.815 / 0.478 / 0.903 / -0.083 / -1.364 / -0.979
[1.219] / [1.295] / [1.123] / [1.131] / [1.037] / [3.171] / [3.315] / [3.061] / [3.076] / [2.517]
10-50 / 0.516 / 0.658 / 0.416 / 0.307 / 0.494 / 1.295 / 1.546 / 0.951 / 1.008 / 2.159
[0.624] / [0.685] / [0.583] / [0.458] / [0.421] / [1.383] / [1.480] / [1.455] / [1.158] / [1.528]
50-250 / 1.125** / 1.240** / 1.023* / 0.568 / 0.455 / 2.911** / 3.253*** / 2.711** / 1.667 / 0.637
[0.517] / [0.514] / [0.530] / [0.538] / [0.391] / [1.174] / [1.175] / [1.209] / [1.219] / [1.092]
>250 / 0.198 / 0.214 / 0.175 / 0.175 / 0.086 / 0.801 / 0.725 / 1.124* / 0.536 / 0.594
[0.195] / [0.206] / [0.216] / [0.205] / [0.139] / [0.564] / [0.619] / [0.623] / [0.502] / [0.366]
BK
<10 / 1.161 / 1.206 / 1.434 / 0.303 / -0.287 / 1.130 / 1.149 / 1.466 / 0.272 / -0.259
[1.239] / [1.296] / [1.314] / [1.153] / [0.977] / [2.950] / [3.185] / [3.202] / [2.469] / [2.201]
10-50 / 0.471 / 0.100 / 0.917 / 1.000 / 0.665 / 1.967 / 1.397 / 3.983 / 2.457 / 1.100
[1.356] / [1.574] / [1.406] / [1.164] / [0.446] / [3.526] / [4.403] / [4.044] / [2.641] / [1.046]
50-250 / 2.373*** / 2.275** / 2.380*** / 3.019*** / 2.548*** / 8.698*** / 8.404*** / 9.006*** / 8.873*** / 6.398**
[0.839] / [0.886] / [0.805] / [0.814] / [0.885] / [2.415] / [2.494] / [2.631] / [2.635] / [2.747]
>250 / -0.082 / -0.029 / -0.182 / -0.106 / 0.008 / -0.773 / -0.767 / -0.737 / -0.031 / -0.818
[0.515] / [0.531] / [0.571] / [0.488] / [0.401] / [1.525] / [1.608] / [1.719] / [1.320] / [1.126]
FW
<10 / -3.072 / -3.558 / -2.894 / -1.535 / 0.640 / -3.417 / -3.943 / -1.034 / -5.561 / -1.749
[2.693] / [2.829] / [2.430] / [2.607] / [2.216] / [6.094] / [6.367] / [5.874] / [6.002] / [5.100]
10-50 / -0.048 / 0.006 / -0.331 / 0.289 / -1.778
** / -0.348 / -0.437 / -0.952 / 1.378 / -5.899
**
[1.247] / [1.396] / [1.129] / [0.933] / [0.835] / [2.759] / [2.889] / [2.958] / [2.439] / [2.875]
50-250 / -3.124
*** / -3.075
*** / -3.196
*** / -3.228
*** / -2.096
** / -7.373
*** / -7.481
*** / -7.616
*** / -6.534
*** / -2.412
[0.983] / [0.987] / [0.955] / [0.953] / [0.813] / [1.974] / [1.956] / [2.010] / [2.068] / [1.866]
>250 / -0.036 / 0.104 / -0.326 / -0.189 / -0.214 / -0.545 / -0.249 / -1.198 / -0.624 / -0.696
[0.563] / [0.595] / [0.532] / [0.534] / [0.441] / [1.431] / [1.651] / [1.342] / [0.980] / [0.981]
Obs. / 167,022 / 107,780 / 42,847 / 11,161 / 5,234 / 120,763 / 74,783 / 32,262 / 9,081 / 4,637
R² / 0.053 / 0.056 / 0.086 / 0.100 / 0.137 / 0.077 / 0.086 / 0.109 / 0.113 / 0.108

Spillovers from four size categories of majority-owned foreign firms (micro, small, medium and larger) on all domestic firms (columns 1 and 6) and on four size categories of domestic firms (columns 2-5 and 7-10). Results for OP TFP (columns 1-5) and ACF TFP (columns 6-10). Robust standard errors in brackets. ***/**/* denotes significance at 1/5/10 percent.

Additional robustness tests of our main estimation results

The empirical approach we take in this paper is used extensively in the literature on FDI spillover effects. We follow the guidelines that Havranek and Irsova (2011, 2013) describe as ‘best practice’. As detailed in the manuscript, ‘best practice’ involves using firm-level data, a productivity (TFP) measure that accounts for the endogeneity of firms’ input choices, several industry-level controls and also requires estimating in first-differences. This empirical procedure therefore comprises two steps: a first step in which an unbiased estimate of TFP is obtained and a second step in which TFP is related to the spillover variables. In this second step, the equation that relates TFP to the lagged FDI spillover variables is first-differenced after which time, industry and region dummies are introduced in the equation. This approach allows one to remove all fixed effects and other time-invariant factors and to further capture unobserved factors that may impact the growth of total factor productivity (for more details also see Javorcik, 2004 and Haskel et al., 2007). The latter is particularly important for transition countries such as Romania. However, as a further robustness test, we re-ran our estimations in Table 4 without re-introducing the industry and region dummies (since one may argue that these should already be captured by taking first-differences). The results of this robustness test are reported in Table A.4 below. These results confirm our previous conclusions as they clearly reveal that positive horizontal and backward and negative forward spillover effects appear to arise especially from medium-sized foreign firms. Furthermore, domestic firms of all size classes benefit from these effects.

Table A.4: Spillover effects across domestic and foreign firm size categories. Estimations performed excluding region and industry dummies from the regressions.

OP TFP / ACF TFP
All dom. firms / Domestic firms with average number of employees / All dom. firms / Domestic firms with average number of employees
<10 / 10-50 / 50-250 / >250 / <10 / 10-50 / 50-250 / >250
HR
<10 / 0.359 / 0.548 / 0.099 / -0.729 / -0.312 / -1.146 / -1.104 / -1.631 / -2.501 / -0.834
[1.393] / [1.439] / [1.316] / [1.376] / [1.043] / [3.491] / [3.641] / [3.449] / [3.427] / [2.588]
10-50 / 0.440 / 0.581 / 0.307 / 0.355 / 0.526 / 0.659 / 0.834 / 0.230 / 0.778 / 2.315
[0.656] / [0.709] / [0.631] / [0.520] / [0.474] / [1.425] / [1.493] / [1.562] / [1.240] / [1.568]
50-250 / 1.029** / 1.120** / 0.908* / 0.576 / 0.712* / 2.442** / 2.619** / 2.347* / 1.366 / 1.162
[0.510] / [0.510] / [0.523] / [0.532] / [0.373] / [1.150] / [1.144] / [1.261] / [1.278] / [1.182]
>250 / -0.096 / -0.087 / -0.131 / -0.132 / -0.063 / -0.007 / -0.040 / 0.234 / -0.292 / -0.161
[0.250] / [0.248] / [0.261] / [0.243] / [0.164] / [0.663] / [0.675] / [0.721] / [0.584] / [0.549]
BK
<10 / -0.595 / -0.499 / -0.590 / -1.281 / -0.818 / -2.171 / -1.844 / -2.206 / -3.351 / -1.299
[1.560] / [1.578] / [1.673] / [1.404] / [0.969] / [3.484] / [3.578] / [3.703] / [3.033] / [2.064]
10-50 / 0.009 / -0.495 / 0.306 / 0.734 / 1.009* / 1.382 / 0.540 / 2.610 / 2.210 / 1.731
[1.310] / [1.523] / [1.491] / [1.256] / [0.603] / [3.520] / [4.483] / [4.196] / [2.761] / [1.302]
50-250 / 2.175** / 2.180** / 2.211** / 2.809*** / 1.860*** / 6.179** / 6.552** / 5.861* / 6.888** / 3.062*
[0.856] / [0.886] / [0.908] / [0.944] / [0.682] / [2.494] / [2.569] / [3.048] / [2.992] / [1.652]
>250 / 0.469 / 0.562 / 0.240 / 0.304 / 0.431 / 1.113 / 1.135 / 1.182 / 0.923 / 0.353
[0.596] / [0.614] / [0.617] / [0.531] / [0.382] / [1.718] / [1.802] / [1.818] / [1.379] / [1.086]
FW
<10 / -1.000 / -1.260 / -0.919 / -0.377 / 0.328 / 2.275 / 1.699 / 4.443 / 0.257 / 0.442
[2.628] / [2.765] / [2.425] / [2.493] / [1.932] / [6.185] / [6.630] / [6.073] / [5.511] / [4.900]
10-50 / 0.518 / 0.546 / 0.443 / 0.520 / -1.345 / 2.328 / 2.460 / 2.159 / 2.611 / -4.266
[1.146] / [1.157] / [1.206] / [0.992] / [0.920] / [3.211] / [3.184] / [3.373] / [2.774] / [3.134]
50-250 / -3.094
*** / -3.095
*** / -3.136
*** / -3.257
*** / -2.361
** / -6.240
*** / -6.501
*** / -6.484
*** / -5.455
** / -2.247
[1.081] / [1.092] / [1.049] / [1.035] / [0.935] / [2.249] / [2.302] / [2.286] / [2.192] / [1.994]
>250 / 0.090 / 0.209 / -0.222 / -0.260 / -0.213 / 0.083 / 0.193 / -0.509 / -0.274 / -0.625
[0.635] / [0.653] / [0.633] / [0.625] / [0.476] / [1.583] / [1.768] / [1.605] / [1.167] / [1.048]
Obs. / 167,022 / 107,780 / 42,847 / 11,161 / 5,234 / 120,763 / 74,783 / 32,262 / 9,081 / 4,637
R² / 0.042 / 0.047 / 0.065 / 0.076 / 0.101 / 0.058 / 0.071 / 0.081 / 0.082 / 0.064

Spillovers from four size categories of foreign firms (micro, small, medium and larger) on all domestic firms (columns 1 and 6) and on four size categories of domestic firms (columns 2-5 and 7-10). Results for OP TFP (columns 1-5) and ACF TFP (columns 6-10). Regressions do not include industry and region dummies. Robust standard errors in brackets. ***/**/* denotes significance at 1/5/10 percent.

Absorptive capacity interaction across size classes

This section presents the results of the interaction with absorptive capacity (AC) for different size classes of domestic firms. Table A.5 reports these results for the OP TFP definition, Table A.6 shows these results for the ACF TFP definition.

Table A.5: Absorptive capability as determinant factor (OP TFP).

All dom. firms / Domestic firms with average number of employees
<10 / 10-50 / 50-250 / >250
level / AC-
inter / level / AC-
inter / level / AC-inter / level / AC-
inter / level / AC-
inter
AC / -0.018
*** / -0.033
*** / -0.082
*** / -0.311
*** / -0.400
***
[0.004] / [0.004] / [0.020] / [0.034] / [0.049]
HR
<10 / 0.770 / -0.503 / 0.828 / -0.129 / 2.286* / -4.461
*** / 0.431 / -2.557 / -1.731 / 3.687
[1.283] / [0.361] / [1.384] / [0.436] / [1.174] / [1.374] / [2.037] / [3.428] / [1.952] / [4.490]
10-50 / 0.582 / -0.422
** / 0.686 / -0.233 / 0.492 / -0.800 / -0.303 / 1.784 / 0.965* / -1.087
[0.610] / [0.190] / [0.664] / [0.233] / [0.655] / [0.972] / [0.567] / [1.254] / [0.512] / [1.398]
50-250 / 0.983
** / 0.430
*** / 0.878
* / 0.806
*** / 2.559
*** / -3.270
*** / 2.747
*** / -4.818
*** / 0.985 / -0.752
[0.484] / [0.161] / [0.480] / [0.192] / [0.595] / [0.776] / [0.800] / [1.542] / [0.790] / [1.708]
>250 / -0.023 / -0.138 / -0.011 / -0.137 / -0.108 / -0.109 / 0.143 / -0.487 / -0.393 / 0.864
[0.211] / [0.092] / [0.212] / [0.091] / [0.259] / [0.358] / [0.420] / [0.823] / [0.246] / [0.526]
BK
<10 / 0.385 / -0.727** / 0.414 / -0.609* / 0.013 / 0.120 / -1.301 / 1.130 / -0.496 / -0.189
[1.582] / [0.351] / [1.630] / [0.346] / [1.825] / [1.615] / [2.012] / [2.569] / [1.221] / [3.403]
10-50 / 0.112 / 1.135*** / -0.405 / 1.085*** / -0.808 / 5.009** / 2.094 / -2.911 / 0.277 / 3.629
[1.232] / [0.399] / [1.494] / [0.390] / [1.619] / [2.456] / [2.248] / [4.595] / [0.669] / [3.159]
50-250 / 2.328*** / 0.025 / 2.444*** / 0.082 / 1.907* / 0.230 / 0.480 / 1.283 / 1.662** / 0.238
[0.783] / [0.142] / [0.805] / [0.134] / [1.133] / [1.037] / [1.242] / [1.379] / [0.672] / [1.377]
>250 / 0.568 / -0.461* / 0.605 / -0.748** / 0.262 / 0.374 / -0.270 / 5.828** / 0.193 / 0.829
[0.587] / [0.238] / [0.603] / [0.293] / [0.832] / [1.659] / [0.751] / [2.387] / [0.668] / [2.379]
FW
<10 / -2.568 / 1.658** / -2.115 / 0.174 / -4.809
** / 6.441
*** / -4.202 / 10.305
** / 0.761 / -1.615
[2.406] / [0.758] / [2.540] / [0.824] / [2.103] / [2.066] / [3.206] / [4.200] / [3.277] / [7.116]
10-50 / 0.649 / 0.136 / 0.591 / 0.195 / 0.311 / 0.489 / -0.542 / 2.070 / -2.400 / 3.245
[1.194] / [0.285] / [1.248] / [0.300] / [1.334] / [1.016] / [1.372] / [2.646] / [1.479] / [3.508]
50-250 / -3.020
*** / -0.411
* / -3.017
*** / -0.320 / -4.352
*** / 2.307
** / -3.622
*** / -0.287 / -2.751
* / 1.309
[0.946] / [0.213] / [0.983] / [0.214] / [0.940] / [1.172] / [1.228] / [2.206] / [1.484] / [2.960]
>250 / 0.384 / -0.330* / 0.333 / 0.084 / -0.627 / 1.043 / -0.780 / 1.448 / 0.063 / -0.650
[0.587] / [0.196] / [0.609] / [0.182] / [0.668] / [0.912] / [0.719] / [1.389] / [0.625] / [1.339]
Obs. / 167,022 / 107,780 / 42,847 / 11,161 / 5,234
R² / 0.073 / 0.084 / 0.119 / 0.174 / 0.190

Spillovers from four size categories of foreign firms (micro, small, medium and larger) on all domestic firms (columns 1-2) and on four size categories of domestic firms (columns 3-10). Results for OP TFP. Robust standard errors in brackets. ***/**/* denotes significance at 1/5/10 percent.

Table A.6: Absorptive capability as determinant factor (ACF TFP).

All dom. firms / Domestic firms with average number of employees
<10 / 10-50 / 50-250 / >250
level / AC-
inter / level / AC-
inter / level / AC-inter / level / AC-
inter / level / AC-
inter
AC / -0.345
*** / -0.275
*** / -0.428
*** / -0.477
*** / -0.686
***
[0.019] / [0.018] / [0.037] / [0.066] / [0.111]
HR
<10 / 0.600 / -2.708 / 0.614 / -1.891 / -0.577 / 0.274 / 0.072 / -9.358 / -2.021 / 8.476
[3.221] / [1.710] / [3.379] / [1.625] / [3.416] / [2.888] / [3.713] / [7.937] / [2.982] / [11.501]
10-50 / 0.904 / -0.363 / 1.205 / -0.185 / 0.146 / 0.303 / 1.630 / -3.807 / 2.929 / -2.677
[1.467] / [0.920] / [1.562] / [0.881] / [1.701] / [1.860] / [1.191] / [2.561] / [1.996] / [6.952]
50-250 / 3.635
*** / -1.712
*** / 3.955
*** / -1.841
*** / 2.871
** / 0.843 / 2.095 / -0.901 / 1.180 / -0.529
[1.155] / [0.636] / [1.174] / [0.597] / [1.285] / [1.540] / [1.349] / [1.073] / [1.159] / [3.854]
>250 / 0.071 / -0.379 / 0.092 / -0.546* / 0.197 / 0.145 / -0.090 / -0.336 / -0.680 / 2.891
[0.572] / [0.385] / [0.599] / [0.293] / [0.693] / [0.676] / [0.726] / [1.194] / [0.761] / [1.876]
BK
<10 / -1.059 / 1.531 / -0.807 / 0.766 / -0.981 / 2.822 / -3.683 / 8.975** / -0.784 / 5.100
[3.572] / [1.229] / [3.783] / [1.154] / [4.114] / [3.127] / [3.484] / [4.314] / [2.802] / [9.030]
10-50 / 1.149 / 4.676** / -0.207 / 6.165*** / 3.310 / 2.122 / 0.527 / 8.843 / 1.873 / 1.251
[3.374] / [1.878] / [4.369] / [1.757] / [4.248] / [3.558] / [3.142] / [5.951] / [1.609] / [3.867]
50-250 / 6.388*** / -0.240 / 6.458*** / -0.768 / 7.795*** / 2.172 / 9.009*** / -0.419 / 2.474 / 2.788
[2.219] / [0.813] / [2.332] / [0.823] / [2.791] / [1.590] / [2.977] / [1.560] / [1.910] / [2.756]
>250 / 1.399 / 1.821* / 1.766 / 1.270 / 0.666 / -3.573 / 0.820 / -4.825 / -0.473 / 9.783
[1.653] / [1.015] / [1.738] / [0.802] / [1.958] / [2.884] / [1.497] / [3.365] / [1.397] / [7.240]
FW
<10 / -1.850 / 1.931 / -1.765 / 1.500 / 0.809 / -1.185 / -4.239 / 3.465 / 0.652 / -18.101
[5.799] / [3.025] / [6.216] / [3.002] / [5.858] / [4.856] / [6.168] / [11.879] / [5.583] / [16.927]
10-50 / 1.497 / 2.055 / 1.204 / 2.115 / 2.686 / -1.528 / 2.795 / 1.151 / -5.166 / 9.754
[3.197] / [1.318] / [3.229] / [1.364] / [3.569] / [2.776] / [2.857] / [3.145] / [3.742] / [10.561]
50-250 / -7.355*** / 0.780 / -7.603*** / 0.993 / -7.651*** / 1.927 / -6.638*** / 1.368 / -0.899 / -7.403
[2.034] / [1.001] / [2.051] / [0.965] / [2.093] / [1.916] / [2.194] / [2.379] / [2.265] / [6.146]
>250 / 0.036 / 0.125 / 0.482 / -0.452 / -0.820 / 1.365 / 0.511 / -3.139* / -0.762 / 1.609
[1.519] / [0.953] / [1.681] / [0.948] / [1.610] / [1.815] / [1.291] / [1.760] / [1.368] / [2.749]
Obs. / 120,763 / 74,783 / 32,262 / 9,081 / 4,637
R² / 0.101 / 0.104 / 0.130 / 0.143 / 0.130

Spillovers from four size categories of foreign firms (micro, small, medium and larger) on all domestic firms (columns 1-2) and on four size categories of domestic firms (columns 3-10). Results for ACF TFP. Robust standard errors in brackets. ***/**/* denotes significance at 1/5/10 percent.

1

[1] Javorcik, B. S. and Spatareanu, M. (2008). To Share or Not To Share: Does Local Participation Matter for Spillovers from Foreign Direct Investment? Journal of Development Economics, 85(1-2):194–217