Technological performance, economic performance and behaviour: a study of Argentinean firms during the 1990s.

(14) Equations {1/3} and {2/3} describe the theoretical latent model. Although we do not observe the latent variables (prod_inn)* and (proc_inn)*, we observe the categorical prod_inn and proc_inn. Thus, for instance, when we observe prod_inn=2 we could infer that the latent variable (prod_inn)* falls somewhere between threshold parameters that, together with the a coefficients are estimated by maximum likelihood.

(15) A Random Effects Model estimated by GLS is consistent and efficient if the time-invariant cross-section specific effect is uncorrelated with the explanatory variables (Crespi, 2002) that is if , one can test for that using a Hausman test that specifies this equation as the null hypothesis of the test. Thus, when the test is not rejected one is allowed to use random effects models (GLS). Otherwise, a fixed effect models should be adopted. However, the panel available used here has a very large N and a small T, with T being so small that makes fixed effect models unfeasible. Then, I needed to stick to GLS results even thought the Hausman test was not passed (see Table 3). As a consequence the coefficients may be biased and/or inconsistent. Having said this, it is worth noting that the sizes and signs of the coefficient do not vary dramatically across the three different specifications of Model 1 (Table 3). The stability in coefficients arguably enhances the robustness of the results.

(16) These results may be explained by the way productivity was measured here. It might be the case that using sales to measure productivity biases the association against process and towards product and management, given that the latter could be claimed to be more market-driven-types of activities.

(17) This latter result might be highly influenced by the way of defining (and measuring) the variables.

VALERIA ARZA

SPRU (Science and Technology Policy Research)

University of Sussex

Freeman Centre, Falmer

Brighton, United Kingdom TABLE 1: INDICATORS ON GDP VOLATILITY

Annual

cumulative Annual Standard Country Period rate average rate Deviation Argentina 1930-2002 2.24 2.28 5.09 US 1930-2002 3.58 3.53 5.25 Argentina 1950-2002 2.00 2.11 5.11 US 1950-2002 3.37 3.49 2.43 Argentina 1950-2000 2.41 2.49 4.77 US 1950-2000 3.45 3.58 2.43

Negative

Coefficient of growth Country Variation probability Argentina 2.24 33% US 1.49 22% Argentina 2.42 35% US 0.69 13% Argentina 1.91 32% US 0.68 14% Source: Author's calculation based on Argentinean National Institute of Statistics and Financial Statistics from the IMF. TABLE 2.1: SUMMARY STATISTICS FOR ALL FIRMS, PERIOD 1992 & 1996 (THOUSANDS OF PESOS)

All firms with information for both periods

Difference

Year 1992 Year 1996 1996-1992

mean median mean median mean Sig N FIRMS 1,533 1,533 Price Index 98.42 106.41 7.99

IPIB 1993=100 Total Sales (S) 22710 6762 30767 7668 8057 *** Sales of Goods 20678 6283 27461 6816 6783 ***

Produced inisde

the firms (SFG) Sales of Unmodified 2031 0 3305 0 1274 *

Goods (SUG) Sales over 98 64 127 80 29 ***

employment (S/EMP) Total Investment (IFA) 887 66 1499 90 613 *** Investment over 7.58% 0.97% 13.04% 1.19% 5.46%

sales (IFA/S) R&D (RD) 32.77 0.00 47.98 0.00 15 * R&D intensity (RD/S) 0.16% 0.00% 0.20% 0.00% 0.04% Export coefficient (E/S) 6.43% 0.00% 8.96% 0.40% 2.53% *** Import coefficient (M/S) 8.04% 0.85% 9.87% 2.44% 1.83% ** Openess coefficient 14.48% 3.94% 18.83% 8.43% 4.35% ***

((EX+IM)/S) Sales of Unmodified 6.66% 0.00% 8.34% 0.00% 1.68% **

Goods over

Sales (SUG/S) Import of Final 1.48% 0.00% 2.02% 0.00% 0.54% **

Goods over

Sales (MFG/S) Import of Capital 1.48% 0.00% 1.91% 0.00% 0.43%

Goods over

Sales (MKG/S) Product innovation 1.81 2.00 1.81 2.00

Results: Likert

scale 0-5 Process innovation 1.12 1.00 1.12 1.00

Results: Likert

scale 0-3 Innovation in 3.64 4.00 3.64 4.00

Management

Results: Likert

scale 0-7

All firms with

information

for both periods

Percentual

Difference

1996/1992

mean N FIRMS Price Index 8.1%

IPIB 1993=100 Total Sales (S) 35.5% Sales of Goods 32.8%

Produced inisde

the firms (SFG) Sales of Unmodified 62.7%

Goods (SUG) Sales over 29.4%

employment (S/EMP) Total Investment (IFA) 69.1% Investment over 72.0%

sales (IFA/S) R&D (RD) 46.4% R&D intensity (RD/S) 25.0% Export coefficient (E/S) 39.3% Import coefficient (M/S) 22.8% Openess coefficient 30.0%

((EX+IM)/S) Sales of Unmodified 25.2%

Goods over

Sales (SUG/S) Import of Final 36.5%

Goods over

Sales (MFG/S) Import of Capital 29.1%

Goods over

Sales (MKG/S) Product innovation

Results: Likert

scale 0-5 Process innovation

Results: Likert

scale 0-3 Innovation in

Management

Results: Likert

scale 0-7 *** Significant at 1%; ** Significant at 5%; * Significant at 10%. Source: Author's calculation based on The Survey TABLE 2.2: SUMMARY STATISTICS ACROSS FIRM TYPOLOGIES, AVG. 1992-1996 (THOUSANDS OF PESOS)

Nationality

National Foreign Difference

mean median mean median mean Sig N FIRMS 1316 313 Total Sales (S) 17140 5162 69763 28582 -52623 *** Sales over employ- 100.16 63.71 184.77 126.72 -85 ***

ment (S/EMP) Total Investment 938 71 2747 828 -1809 ***

(IFA) Investment over 10.36% 1.54% 4.72% 2.05% 5.64%

sales (IFA/S) R&D (RD) 25.43 0.00 127.60 0.00 -102 *** R&D intensity 0.20% 0.00% 0.20% 0.00% 0.00%

(RD/S) Export coeffi- 6.65% 0.00% 12.51% 3.67% -5.86% ***

cient (E/S) Import coeffi- 6.50% 1.29% 17.59% 12.90% -11.09% ***

cient (M/S) Openess coefficient 13.16% 4.66% 30.10% 22.81% -16.94% ***

((EX+IM)/S) Sales of Unmodified 7.20% 0.00% 10.34% 0.88% -3.14% ***

Goods over

Sales (SUG/S) Import of Final 1.01% 0.00% 4.86% 0.00% -3.85% ***

Goods over

Sales (MFG/S) Import of Capital 1.33% 0.00% 2.36% 0.58% -1.03% **

Goods over

Sales (MKG/S) Product innovation 1.62 1.00 2.50 2.00 -0.89 ***

Results: Likert

scale 0-5 Process innovation 1.01 1.00 1.54 2.00 -0.53 ***

Results: Likert

scale 0-3 Innovation in 3.33 4.00 4.73 5.00 -1.40 ***

Management

Results: Likert

scale 0-7

Group Affiliation

Independent Group Difference

mean median mean median mean Sig N FIRMS 1194 415 Total Sales (S) 14710 4423 63025 20833 -48314 *** Sales over employ- 98.91 63.09 167.61 118.78 -69 ***

ment (S/EMP) Total Investment 733 64 2623 590 -1890 ***

(IFA) Investment over 10.62% 1.50% 5.05% 2.14% 5.57%

sales (IFA/S) R&D (RD) 21.12 0.00 109.25 0.00 -88 *** R&D intensity 0.20% 0.00% 0.22% 0.00% -0.02%

(RD/S) Export coeffi- 6.86% 0.04% 10.49% 2.47% -3.63% ***

cient (E/S) Import coeffi- 6.73% 1.30% 14.25% 8.59% -7.52% ***

cient (M/S) Openess coefficient 13.59% 4.91% 24.75% 17.19% -11.16% ***

((EX+IM)/S) Sales of Unmodified 7.66% 0.00% 7.97% 0.00% -0.31%

Goods over

Sales (SUG/S) Import of Final 1.18% 0.00% 3.42% 0.00% -2.24% ***

Goods over

Sales (MFG/S) Import of Capital 1.26% 0.00% 2.26% 0.53% -1.00% **

Goods over

Sales (MKG/S) Product innovation 1.58 1.00 2.39 2.00 -0.82 ***

Results: Likert