Trade Liberalization, Exports, and Technology Upgrading: Evidence on the Impact of MERCOSUR on Argentinian Firms

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Recent papers by Van Biesebroeck (2005) and De Loecker (2007) found productivity increases after firms enter export markets in Ivory Coast and Slovenia.

“recent papers in this literature like Johannes Van Biesebroeck (2005) and Jan De Loecker (2007) find increases in productivity after firms enter the export market in Ivory Coast and Slovenia, respectively.”

• cites Johannes Van Biesebroeck (2005) sourcePartly supported

  The Conclusions report that African exporters improved their productivity after entering foreign markets, and Section 2.2 lists Côte d’Ivoire among the nine countries studied. The finding is reported for the pooled African sample, however, not specifically for firms in Côte d’Ivoire. The Slovenia finding is attributed to De Loecker, not to this source. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Jan De Loecker (2007) sourceNot checked — no full text of the source

May combine two claims.

Trade liberalization can raise productivity by improving factor allocation or by inducing adoption of more advanced technologies.

“Trade liberalization can increase productivity by inducing a better allocation of production factors or the adoption of more advanced technologies.”

• cites Nina Pavcnik 2002 sourceNot checked — no full text of the source
• cites Marc Melitz 2003 sourcePartly supported

  Section 9, “The Impact of Trade Liberalization” (page unavailable), says greater exposure to trade raises aggregate productivity by reallocating market shares toward more productive firms. The “Changes in Aggregate Productivity” section mentions innovation as a channel in other models, but our copy does not establish adoption of more advanced technologies as a productivity mechanism in this paper. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Andrew B. Bernard et al. 2003 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Earlier studies by Clerides et al. (1998) and Bernard and Jensen (1999) found exporters have higher productivity than nonexporters because more productive firms become exporters, with no effect of exporting on productivity.

“The first studies by Sofronis K. Clerides, Saul Lach, and Tybout (1998) for Colombia, Mexico, and Morocco, and Bernard and Jensen (1999) for the United States find that exporters have higher productivity than nonexporters, but this is because ex ante more productive firms become exporters, while there are no effects of exporting on productivity.”

• cites Sofronis K. Clerides, Saul Lach, and Tybout (1998) sourceNot checked — no full text of the source
• cites Bernard and Jensen (1999) sourcePartly supported

  Section 4.1 reports that future exporters had higher labor productivity before they began exporting, supporting the selection explanation. Section 6 concludes that exporting shows little evidence of boosting productivity, but that is more cautious than the claim that exporting has no effect on productivity. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

May combine two claims.

Trade integration reallocates market shares toward exporters, which are the most productive firms, raising aggregate productivity.

“trade integration reallocates market shares towards exporters, the most productive firms, increasing aggregate productivity.”

• cites Nina Pavcnik 2002 sourceNot checked — no full text of the source
• cites Marc Melitz 2003 sourceSupported
• cites Andrew B. Bernard et al. 2003 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

The empirical methodology follows the literature that measures trade liberalization effects through tariff changes.

“The empirical methodology implemented in this paper follows the literature measuring the effects of trade liberalization on economic outcomes through changes in tariffs.10”

• cites [10] 10. Bernard, Andrew B., Stephen Redding, and Peter Schott. 2007. "Comparative Advantage and Hetero- geneous Firms." Review of Economic Studies, 74(1): 31-66. sourceCould not verify

  The Introduction describes a theoretical model of responses to falling trade costs, and the passage under “Decline From” mentions historical declines in tariffs and transportation costs. Neither passage describes an empirical methodology that measures liberalization effects through tariff changes; searches for tariff changes and empirical methodology found no such account in our copy. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

May combine two claims.

Between 1992 and 1996, Argentina's exports to Brazil quadrupled while exports to the rest of the world rose only 60 percent.

“between 1992 and 1996 exports to Brazil quadrupled, while exports to the rest of the world increased only 60 percent.”

May combine two claims.

New exporters were not more technology intensive than nonexporters before liberalization, but they upgraded technology faster as they entered the export market during liberalization.

“In particular, new exporters were not more technology intensive than nonexporters before liberalization, but upgrade technology faster as they enter the export market during the liberalization period.”

May combine two claims.

The reduction in Brazil's tariffs had a stronger effect on both export entry and technology upgrading in the third quartile of the firm size distribution.

“I find that the reduction in Brazil's tariffs had a stronger effect on both entry in the export market and technology upgrading in the third quartile of the firm size distribution.”

May combine two claims.

The coefficient in the change in Brazil's tariffs (βτx) is negative (-0.421) and significant (t = -5.01).

“The coefficient in the change in Brazil's tariffs ( β τ x ) is negative (-0.421) and significant (t = -5.01).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Between 1992 and 1996, Argentina's average import tariffs with respect to the rest of the world increased slightly (1 percentage point).

“As Argentina's unilateral trade liberalization occurred before the period under study, between 1992 and 1996 Argentina's average import tariffs with respect to the rest of the world increased slightly (1 percentage point).”

May combine two claims.

The binary innovation measures are available for a larger subsample of 1,310 firms that is representative by size, with about 25 percent of firms in each size quartile.

“Instead, the binary measures of technology that I use to construct the innovation indexes are available for a larger subsample of firms (1,310 firms) that is representative in terms of size, as around 25 percent of firms in the subsample belong to each size quartile.”

May combine two claims.

With the inclusion of two-digit ISIC industry dummies, the correlation between capital intensity and tariffs falls to -0.06, although the correlation between tariffs and skill intensity increases to 0.06.

“With the inclusion of two-digit ISIC industry dummies the correlation between capital intensity and tariffs falls to -0.06, although the correlation between tariffs and skill intensity increases to 0.06.”

May combine two claims.

After liberalization, new exporters become more technology intensive than firms that do not export, increasing their spending on technology per worker 0.37 log points faster between 1992 and 1996.

“In contrast, after liberalization these new exporters become more technology intensive than firms that do not export, increasing their spending on technology per worker 0.37 log points faster between 1992 and 1996.”

May combine two claims.

The coefficients in the regressions including all controls (columns 5 and 8) are -0.466 (t = -4.80) and -0.533 (t = -3.78) and imply that the average reduction in Brazil's tariffs (24 percentage points) increases the probability of entry in the export market by 11 to 12 percentage points.

“The coefficients in the regressions including all controls (columns 5 and 8) are -0.466(t = -4.80)and -0.533 (t = -3.78)and imply that the average reduction in Brazil's tariffs (24 percentage points) increases the probability of entry in the export market by 11 to 12 percentage points.”

May combine two claims.

The subsample with positive technology spending in 1992 and 1996 is not representative of the smallest firms, whereas the binary innovation measures come from a sample representative in size.

“This sample is not representative for the smallest firms, while the binary measures of technology contained in the survey are available for a representative sample.”

May combine two claims.

Brazil's tariffs are negatively correlated with a measure of capital intensity (-0.21), suggesting that Brazil protected labor-intensive industries.

“Instead, tariffs are negatively correlated with a measure of capital intensity (-0.21), suggesting that Brazil protected labor-intensive industries.”

May combine two claims.

The changes in Argentina's input tariffs with respect to the world were smaller than the output tariffs, ranging from -3 to 6 percentage points.

“Similarly, the changes in Argentina's input tariffs with respect to the world were smaller than the output tariffs, ranging from -3 to 6 percentage points.”

May combine two claims.

In levels including lagged export status, the point estimates of βτxr are smaller, but the same pattern is observed: the estimate of βτxr is largest in the third quartile and precisely estimated (βτx3 = -0.541, t = -3.49).

“Column 4 presents estimation of the equation in levels including lagged export status as a control.The point estimates of β τ x r are smaller, but the same pattern is observed: the estimate of β τ x r is largest in the third quartile and precisely estimated ( β τ x 3 = -0.541,t = -3.49).”

• AssumptionThe authors state this as an assumption: Firm size relative to the four-digit ISIC industry mean in 1992 is a proxy for initial productivity.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Using only 1992 and 1996 data yields very similar estimates, though with slightly larger standard errors.

“The first alternative gives very similar results, although standard errors are slightly bigger.”

May combine two claims.

Brazil's tariffs display rates above 30 percentage points for some unskilled, labor-intensive industries protected under the import substitution policy like toys, textiles, and rubber and also for skill-intensive industries that were protected under the new policy like domestic appliances, office accounting and computing, and the car industry.

“As a result, they display tariff rates above 30 percentage points for some unskilled, labor-intensive industries protected under the import substitution policy like toys, textiles, and rubber and also for skill-intensive industries that were protected under the new policy like domestic appliances, office accounting and computing, and the car industry.”

May combine two claims.

The estimated coefficient on Brazil's tariffs is unchanged by adding firm-level controls or by including changes in Argentina's output and input tariffs relative to the world.

“The estimated coefficient is not affected by the inclusion of firm-level controls (column 2) nor by the change in Argentina's output and input tariffs with respect to the world (columns 3 to 5).”

May combine two claims.

Argentina's m.f.n. tariffs were already low before MERCOSUR, so the baseline reduction in Argentina's tariffs for imports from Brazil averaged only 13 percentage points.

“As m.f.n.tariffs in Argentina were already low before MERCOSUR was launched, the baseline for the reduction in Argentina's tariffs for imports from Brazil was only 13 percentage points on average (Table 1).”

May combine two claims.

The coefficients for the rest of the quartiles are negative but not always statistically significant.

“The coefficients are negative but not always statistically significant.”

• AssumptionThe authors state this as an assumption: Firm size relative to the four-digit ISIC industry mean in 1992 is a proxy for initial productivity.

May combine two claims.

The baseline m.f.n. rates for Argentina's input tariff reductions with respect to Brazil were smaller than the output tariffs, with an average level of 11 percentage points in 1992.

“The baseline m.f.n.rates for Argentina's input tariff reductions with respect to Brazil were smaller than the output tariffs reported above, with an average level of 11 percentage points in 1992.”

Probably not a substantial claim: may not pass S3 (load-bearing).

When the dependent variable is changed to growth in technology spending per worker, the estimated coefficients on Brazil's tariffs are very similar to those in Table 5.

“Table E.1 in online Appendix E reports estimates of equation ( 16), replacing the growth in spending on technology by the growth in spending on technology per worker as the dependent variable.The estimates of β τ x are very similar to the ones reported in Table 5.”

May combine two claims.

The 1992 baseline tariffs average 24 percentage points, slightly below 1991 tariffs, and have similar cross-industry variation with correlation 0.97.

“These are on average 24 percentage points, slightly lower than 1991 tariffs, but reflect a similar variation across industries, as their correlation is 0.97.”

Classified as probably not a substantial claim. May combine two claims.

Panel B of Table 3 reports estimation of equation (13) where the estimated coefficient goes from -0.291 (t = 4.09) in the baseline specification in column 1 to -0.490 (t = 3.40) in column 9 where all controls are included.

“Panel B of Table 3 reports estimation of equation (13) where the estimated coefficient goes from -0.291 (t = 4.09) in the baseline specification in column 1 to -0.490 (t = 3.40) in column 9 where all controls are included.”

May combine two claims.

In the innovation-index regressions, the coefficient on the change in Brazil's tariffs is negative and statistically significant for the overall index and for each innovation type.

“The coefficient on the change in Brazil's tariffs is negative and significant for all and each type of innovation.”

May combine two claims.

When these firm characteristics are included in the regressions the coefficient does not change; thus sample selection does not seem to play an important role.

“In addition, when these firm characteristics are included in the regressions the coefficient does not change (see panel C, columns 1 and 2); thus sample selection does not seem to play an important role.”

May combine two claims.

The empirical evidence indicates that the reduction in Argentina's tariffs relative to Brazil reduced domestic sales, whereas the reduction in Brazil's tariffs did not have a significant effect on domestic sales.

“The empirical evidence suggests that the reduction in Argentina's tariffs with respect to Brazil reduced domestic sales, while the reduction in Brazil's tariffs did not have a significant effect.”

May combine two claims.

The column 2 estimate implies that the average reduction in Brazil's tariffs raises the innovation index by 0.07, which is 19 percent of the average index of 0.38.

“The estimated coefficient in column 2 implies that the average reduction in Brazil's tariffs induces an increase of 0.07 in the innovation index, which is 19 percent of the average innovation index (0.38).”

May combine two claims.

The domestic-sales estimate is robust to controls for Argentina's input tariffs relative to Brazil and industry characteristics, falls to 1.022 (t = 1.89) when Brazil's tariffs are included, and becomes insignificant when industry or firm characteristics are added.

“The point estimate is robust to the inclusion of controls for changes in Argentina's input tariffs with respect to Brazil and industry characteristics (columns 2 and 4), falls to 1.022 (t = 1.89) when changes in Brazil's tariffs are included in the regression (column 5), and becomes insignificant when in addition controls for industry or firm characteristics are included (columns 6 and 7).”

May combine two claims.

Changes in Argentina's output and input tariffs relative to the world did not have a statistically significant effect on domestic sales, unlike the change in Argentina's tariffs relative to Brazil.

“Unlike the change in Argentina's tariffs with respect to Brazil, changes in Argentina's output and input tariffs with respect to the world did not have a statistically significant effect on domestic sales.”

May combine two claims.

The third-quartile point estimate is twice the average effect for all firms and more than twice the estimates for the other three quartiles.

“The point estimate is double the size of the estimated average effect for all firms reported in Table 5 and more than double the estimated coefficient for the other three quartiles of the size distribution.”

May combine two claims.

The reduction in Brazil's tariffs induced technology upgrading in both the subsample of firms that did not export in 1992 and the subsample that did.

“To explore this issue further, I split the sample of firms in two on the basis of initial export status and show that the reduction in Brazil's tariffs induced technology upgrading in both subsamples.”

May combine two claims.

Firm-level analysis of export sales did not yield consistent results, possibly because the firm-level data combine all export destinations.

“The analysis of export sales at the firm level did not produce consistent results, possibly because the data combines all destinations.”

May combine two claims.

For the innovation index, the reduction in Brazil's tariffs again has a statistically significant positive effect only in the third quartile, with a point estimate of -0.359 (t = 2.70).

“The results on the innovation index parallel the findings with the ST measure: the reduction in tariffs induces a statistically significant increase in innovation only in the third quartile of the size distribution, where the point estimate of β τ x r is -0.359 (t = 2.70), as reported in column 4.”

May combine two claims.

In the IV estimates, the coefficient on Brazil's tariffs is larger in magnitude and the standard errors are smaller than in OLS.

“Panel B reports IV estimates, where both the magnitude of the coefficient increases and standard errors fall, suggesting that measurement error in tariffs might produce attenuation bias in the OLS results.”

May combine two claims. The quote only partly states this claim.

In the split-sample estimates for technology spending, the coefficient on Brazil's tariffs is similar to the full-sample estimate and significant in almost all specifications.

“The coefficient is similar to the one estimated for the full sample and significant in almost all specifications.”

May combine two claims.

Adding interactions between Argentina's import-tariff changes and size quartiles, or between industry characteristics and size quartiles, does not change the third-quartile coefficient on Brazil's tariffs for technology spending or innovation, which remains significant at the 1 percent level.

“The estimated coefficient of the effect of Brazil's tariffs on ST and innovation in the third quartile of the firm size distribution is not affected by the inclusion of these controls and is always significant at the 1 percent confidence level.”

May combine two claims.

For the first, second, and fourth quartiles, the estimated coefficients are negative but not statistically significant.

“The coefficients are negative but not statistically significant; thus it is possible that some firms in the first, second, and fourth quartiles were induced to upgrade technology by the reduction in Brazil's tariffs.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Changes in technology affect productivity and can also affect factor markets if new technologies use skilled labor more intensively.

“More importantly, changes in technology not only affect productivity but can have implications for factor markets if new technologies use skilled labor more intensively.”

May combine two claims.

When variable trade costs fall and not all firms export, the fractions of surviving firms that export and that use the high technology both increase.

“when variable trade costs (τ) fall, and not all firms export ( τ σ-1 f x > f ): (i) The fraction of surviving firms that export, p x = ( φ x / φ * ) -k , and the fraction of surviving of firms that use technology h, p h = ( φ h / φ * ) -k , increase.”

May combine two claims.

Old exporters upgrade technology when trade costs fall only if technology adoption costs are high relative to fixed exporting costs, which is more likely in developing countries.

“the result that old exporters upgrade technology when trade costs fall obtains only when the costs of technology adoption are high (relative to fixed exporting costs), which is more likely to be the case in developing countries.”

Classified as probably not a substantial claim. May combine two claims.

Several studies document rising relative demand for skill in developing countries during trade liberalization, leaving open whether skill-biased technological change was an endogenous response to trade liberalization.

“Indeed, several studies have documented increases in the relative demand for skill in developing countries during the trade liberalization period, 9 leaving the open question of whether skill-biased technological change might have been an endogenous response to trade liberalization.”

• cites [9] 9. Bernard, Andrew B., and J. Bradford Jensen. 2004. "Why Some Firms Export." Review of Economics and Statistics, 86(2): 561-69. sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

The model was inspired by Yeaple's (2005) insight that lower trade costs raise the share of firms that export and use the most advanced technology.

“it was inspired by the insight in Stephen R. Yeaple (2005) that a reduction in trade costs increases the share of firms that export and use the most advanced technology.”

• cites Stephen R. Yeaple (2005) sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

The benefit of technology adoption scales with revenues, while its cost is fixed.

“This is because the benefit of adoption is proportional to revenues, while its cost is fixed.”

Classified as probably not a substantial claim. May combine two claims.

This paper differs from that literature by studying technology rather than productivity and by analyzing bilateral trade liberalization's effect on technology adoption rather than exporting's effect.

“This paper differs from this literature in that the outcome of interest is technology instead of productivity, and in that it analyzes the effect of bilateral trade liberalization on technology adoption, not the effect of exporting.”

Classified as probably not a substantial claim.

A positive finding even under trade reform cannot be fully attributed to the reform, especially with simultaneous market-oriented reforms that could have helped some firms invest and enter export markets.

“The opposite finding, even in the context of a trade reform, can't be fully attributed to it, especially in the context of simultaneous implementation of other market-oriented reforms that might have made it possible for some firms to invest in productivity improvements and thus enter the export market.”

May combine two claims.

In the model equilibrium, firms sort into four groups: least productive exit, low productivity serve domestic and use old technology, medium productivity export but use old technology, most productive export and use advanced technology.

“In online Appendix A, I show that in this equilibrium firms sort into four different groups: the least productive firms (φ < φ * ) exit, the low productivity firms ( φ * < φ < φ x ) serve only the domestic market and use technology l, the medium productivity firms ( φ x < φ < φ h ) still use technology l but also export, and the most productive firms ( φ h < φ) both export and use technology h.”

Classified as probably not a substantial claim.

A main potential problem is that other reforms in the same period could have had heterogeneous effects on industries with different characteristics.

“Still, a main potential problem is that other reforms carried out in the same period could have had heterogeneous effects on industries with different characteristics.”

• AssumptionRests on an assumption: Other reforms in the same period did not have heterogeneous effects across industries that would confound the estimated tariff effects.

Probably not a substantial claim: may not pass S2 (checkable).

Focusing on technology investment isolates a particular mechanism through which firm productivity can improve.

“The first departure from previous literature, namely the focus on investment in technology as the outcome of interest, has the advantage of isolating a particular mechanism through which firm productivity can improve.”

May combine two claims.

In the model, bilateral tariff reduction raises export revenues more than it lowers domestic revenues, leading more firms to adopt the new technology.

“In this setup, a bilateral reduction in tariffs increases export revenues more than it decreases domestic revenues, inducing more firms to adopt the new technology.”

May combine two claims.

Brazil's tariffs provide a good source of arguably exogenous variation because they fell from an average of 29 percent in 1991 to zero in 1995 and varied widely across industries.

“Brazil's tariffs provide a good source of arguably exogenous variation, as they fell from an average of 29 percent in 1991 to zero in 1995, and varied extensively across industries.”

May combine two claims.

Lileeva and Trefler (2010) found US tariff reductions raised productivity only in the least productive new Canadian exporters, contrasting with Argentina where Brazil's tariff cuts induced technology upgrading mostly in the third quartile and among old as well as new exporters.

“Their finding that the reduction in US tariffs induced productivity increases in only the least productive new entrants in the export market for the case of Canada contrasts with the findings for Argentina where the reduction in Brazil's tariffs induced technology upgrading mostly in the third quartile of the firm-size distribution and not only in new but also in old exporters.”

• cites Alla Lileeva and Trefler (2010) sourceNot checked — no full text of the source

May combine two claims.

The result comes from the asymmetric effect of trade liberalization with heterogeneous firms and fixed exporting costs: domestic-only firms lose revenues while exporters gain.

“Instead, the result is due to the asymmetric effect of trade liberalization in models of heterogeneous firms with fixed exporting costs: while firms serving only the domestic market lose revenues, exporters see their revenues increase.”

May combine two claims.

A reduction in trade costs raises export revenues, leading more firms to enter export markets and upgrade technology.

“I show that a reduction in trade costs increases export revenues, inducing more firms to enter the export market and upgrade technology.”

Classified as probably not a substantial claim.

Both export entry and technology upgrading could be caused by other reforms in the same period if those reforms had heterogeneous effects across firms.

“Indeed, both entry in the export market and technology upgrading could be caused by other economic reforms undertaken in the same period if these had heterogeneous effects on firms with different characteristics.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A potential problem in the estimation of equation (13) restricted to nonexporters in 1992 is sample selection.

“A potential problem in the estimation of equation (13) restricted to nonexporters in 1992 is sample selection.”

May combine two claims.

The patterns in the data show that there is a coincidence between entry in the export market and technology upgrading but cannot establish whether it is expanded export opportunities that cause technology adoption, vice versa, or whether both are caused by a third factor.

“The patterns in the data described above show that there is a coincidence between entry in the export market and technology upgrading but can't establish whether it is expanded export opportunities that cause technology adoption, vice versa, or whether both are caused by a third factor.”

May combine two claims.

In the model, underlying productivity differences produce a sorting of firms into three groups: low productivity firms serve only the domestic market and use low technology, medium productivity firms still use low technology but also export, and the most productive firms both export and use high technology.

“In the model, underlying productivity differences produce a sorting of firms into three groups: the low productivity firms serve only the domestic market and use the low technology, the medium productivity firms still use the low technology but also export, and the most productive firms both export and use the high technology.”

May combine two claims.

As changes in tariffs are predetermined, they are unlikely to be driven by political pressures arising from the effects of liberalization in Brazil or Argentina, or by contemporaneous shocks to industrial performance.

“As changes in tariffs are predetermined, they are unlikely to be driven by political pressures arising from the effects of liberalization in Brazil or Argentina, or by contemporaneous shocks to industrial performance.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The tariff structure in 1988 was based on the tariffs implemented in 1957 under the import substitution policy.

“Berlinski et al. (2006) note that the tariff structure in 1988 was based on the tariffs implemented in 1957 under the import substitution policy.”

• cites Berlinski et al. (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The sample is representative of firms owning establishments with more than ten employees and is based on 1993 census data.

“The sample is representative of firms owning establishments with more than ten employees and is based on 1993 census data.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The new tariffs would vary between 0 and 20 percent, except for a few goods facing 30-35 percent tariff rates.

“The new tariffs would vary between 0 and 20 percent, except for a few goods facing 30-35 percent tariff rates.24”

• cites Berlinski et al. (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

If there are sunk exporting costs, current export status might depend on lagged export status, which in turn is likely to be negatively correlated with the initial level of Brazil's tariffs.

“A potential problem of the specification in equation ( 12) is that if there are sunk exporting costs, current export status might depend on lagged export status, 36 which in turn is likely to be negatively correlated with the initial level of Brazil's tariffs.”

May combine two claims.

After a new government took power in March 1990, NTBs were eliminated and tariffs were reduced gradually according to a timetable ending in January 1994.

“Instead, after a new government took power in march 1990 NTBs were eliminated and tariffs were reduced gradually according to a timetable ending in January 1994.”

• cites Berlinski et al. (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Estimating the technology-spending equation in levels over all years would create serial correlation because Brazil's tariff variation over time is entirely determined by its 1992 level, so reported standard errors would understate the true standard deviation.

“The problem with this estimation strategy is that it would induce serial correlation in the error terms, as the variation across time in Brazil's tariffs is fully determined by their level in 1992.As a result, the standard error of the estimated coefficients would understate their standard deviation, as noted by Marianne Bertrand, Esther Duflo, and Sendhil Mullainathan (2004).”

• cites Marianne Bertrand, Esther Duflo, and Sendhil Mullainathan (2004)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

First differencing eliminates time-invariant plant and sector heterogeneity.

“In this case, first differencing eliminates time-invariant plant and sector heterogeneity:”

Classified as probably not a substantial claim.

The fact that within each sector exporters and new exporters are upgrading technology faster than other firms could reflect other shocks that affect middle and high productivity firms differentially.

“Still, the fact that within each sector exporters and new exporters are upgrading technology faster than other firms could reflect other shocks that affect middle and high productivity firms differentially.”

May combine two claims.

The model predicts that in sectors where tariffs are higher the exporting cutoff is higher; thus it is likely that in sectors with high initial tariffs nonexporters are more productive than in sectors with low initial tariffs, creating a positive correlation between Brazil's tariffs in 1992 and unobserved productivity, thus biasing downwards the coefficient on the change in tariffs.

“The model predicts that in sectors where tariffs are higher the exporting cutoff is higher; thus it is likely that in sectors with high initial tariffs nonexporters are more productive than in sectors with low initial tariffs, creating a positive correlation between Brazil's tariffs in 1992 and unobserved productivity, thus biasing downwards the coefficient on the change in tariffs.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

This is not inconsistent with the model, as size is not a perfect measure of productivity, and the exporting cutoffs might differ across industries.

“This is not inconsistent with the model, as size is not a perfect measure of productivity, and the exporting cutoffs might differ across industries.”

• AssumptionThe authors state this as an assumption: Firm size relative to the four-digit ISIC industry mean in 1992 is a proxy for initial productivity.

May combine two claims.

The model predicts that the reduction in tariffs would induce firms in the middle range of the productivity distribution to enter the export market but should not affect firms in the lower and upper ranges of the distribution.

“In particular, the model predicts that the reduction in tariffs would induce firms in the middle range of the productivity distribution to enter the export market but should not affect firms in the lower and upper ranges of the distribution.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The increase in the coefficient when Argentina's output tariffs relative to Brazil are included may reflect correlation with Brazil's tariffs and an opposite-sign effect on technology adoption, though that effect is not statistically significant.

“possibly because these are correlated with Brazil's tariffs but had an effect of the opposite sign in technology adoption, although it is not statistically significant.”

May combine two claims.

The 1991 m.f.n. import tariffs of Brazil are the same for Argentina and the rest of the world and are therefore unlikely to be targeted to industry characteristics particular to Argentina, whose share of Brazil's trade was only 7.7 percent.

“Second, the 1991 m.f.n.import tariffs of Brazil are the same for Argentina and the rest of the world and are therefore unlikely to be targeted to industry characteristics particular to Argentina, whose share of Brazil's trade was only 7.7 percent.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

They are also less stable, possibly due to the omission of unobserved time-invariant industry characteristics.

“They are also less stable, possibly due to the omission of unobserved time-invariant industry characteristics.”

May combine two claims.

First-difference and lagged-dependent variable estimates have a bracketing property: if the first-difference specification in (12) is correct, then (13) will tend to underestimate the absolute value of βτx, while if the lagged-dependent variable specification in (13) is correct, then (12) will tend to overestimate the absolute value βτx.

“Still, estimation of equation ( 13) is useful because first-difference and lagged-dependent variable estimates have a bracketing property: if the first-difference specification in (12) is correct, then (13) will tend to underestimate the absolute value of β τ x , while if the lagged-dependent variable specification in (13) is correct, then (12) will tend to overestimate the absolute value β τ x .”

• AssumptionRests on an assumption: The initial level of Brazil's tariffs is negatively correlated with export status in 1992.

Classified as probably not a substantial claim.

If industries with different initial characteristics are on different trends, Brazil's tariffs could be capturing some omitted industry-level time-varying variable.

“Still, if industries with different initial characteristics are on different trends, Brazil's tariffs could be capturing some omitted industry-level time-varying variable.”

Classified as probably not a substantial claim.

Brazil's tariffs might be correlated with changes in Argentina's tariffs during this period, as long as the structure of protection was similar between the two countries in 1992.

“An additional issue concerning the use of Brazil's tariffs to measure the effect of expanded export opportunities on entry in the export market and technology upgrading is that they might be correlated with changes in Argentina's tariffs during this period, as long as the structure of protection was similar between the two countries in 1992.”

May combine two claims.

Brazil's trade policy is correlated with some industry characteristics, and omitting them could be an important source of bias.

“As discussed above, Brazil's trade policy is correlated with some industry characteristics, and omitting them could be an important source of bias.”

May combine two claims.

In this setting a reduction in variable trade costs increases exporting revenues, inducing firms in the middle range of the productivity distribution to enter the export market and upgrade technology.

“In this setting a reduction in variable trade costs increases exporting revenues, inducing firms in the middle range of the productivity distribution to enter the export market and upgrade technology.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The first reforms implemented in 1988-89 did not have significant effects on the degree of protection of the domestic industry, as NTBs, which were the main instrument of protection, were not modified.

“They argue that the first reforms implemented in 1988-89 did not have significant effects on the degree of protection of the domestic industry, as NTBs, which were the main instrument of protection, were not modified.”

• cites Berlinski et al. (2006)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The sensitivity of the innovation coefficient may reflect an opposite-sign effect of Argentina's output tariff reduction on innovation, though that estimate is only marginally significant.

“It is possible that this is due to the reduction in Argentina's output tariffs having an effect on innovation of the opposite sign, although the estimate is only marginally significant (column 3).”

Classified as probably not a substantial claim. May combine two claims.

The linearization does not respect functional form, thus estimation attempts only to recover the signs of the partial derivative of interest and to assess the economic significance of the estimated coefficients.

“This linearization does not respect functional form, thus estimation attempts only to recover the signs of the partial derivative of interest and to assess the economic significance of the estimated coefficients.”

May combine two claims.

The tariff reductions were programmed in 1991 and reach a level of zero for all industries in 1995.

“First, the tariff reductions were programmed in 1991 and reach a level of zero for all industries 32 in 1995.”

Probably not a substantial claim: may not pass S3 (load-bearing).

In Eaton and Kortum's (2001) baseline model, a larger market size is offset by greater competition from imported technologies, resulting in no effect of lower trade barriers on innovation.

“in their baseline model the effect of a bigger market size is counteracted by the increased competition with technologies embedded in imports, so that there is no effect of lower barriers to trade on innovation.”

• cites Jonathan Eaton and Samuel Kortum (2001) sourceNot checked — no full text of the source

May combine two claims.

The fact that firms already exporting in 1992 also upgraded technology implies that the response is consistent with technology upgrading being driven by increased revenues.

“This implies that firms that were already exporting in 1992 are induced to upgrade technology by the reduction in Brazil's tariffs, which is consistent with technology upgrading being driven by the increase in revenues.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

This pattern is not inconsistent with the model because firm size is an imperfect proxy for productivity and technology-adoption cutoffs may vary across industries.

“As mentioned above, this is not inconsistent with the model, as size is not a perfect measure of productivity, and the technology adoption cutoffs might differ across industries.”

• AssumptionRests on an assumption: Firm size is an imperfect proxy for productivity.

May combine two claims.

The evidence is consistent with a mechanism in which lower trading-partner tariffs raise exporter revenues and make new technology adoption profitable for more firms.

“The evidence is consistent with falling trading partner's tariffs increasing revenues for exporters and making adoption of new technologies profitable for more firms.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Standard errors are clustered at the four-digit ISIC industry level.

“Standard errors are clustered at the 4-digit ISIC industry level.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

This may be because Argentina's tariff changes relative to the world were smaller, since unilateral trade liberalization occurred before 1992.

“This might be due to the fact that the changes in Argentina's tariffs with respect to the world were smaller, as unilateral trade liberalization took place before 1992.”

• AssumptionRests on an assumption: Argentina's unilateral trade liberalization occurred before 1992.

Probably not a substantial claim: may not pass S3 (load-bearing).

The instrumental-variable estimates use Brazil's 1991 tariffs as an instrument for Brazil's 1992 tariffs to correct for measurement error in the tariff measure.

“The IV estimates use Brazil's tariffs in 1991 to instrument for Brazil's tariffs in 1992 in an attempt to correct for measurement error in Brazil's tariffs.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In Yeaple's (2005) model, firms are ex ante identical, but in equilibrium they are indifferent between exporting and adopting new technology versus serving only the domestic market with old technology.

“InYeaple (2005) firms are ex ante homogeneous, but in equilibrium all firms are indifferent between entering the export market and adopting the new technology or serving only the domestic market and using the old technology.”

• cites Yeaple (2005) sourceNot checked — no full text of the source

May combine two claims.

The survey lacks value-added measures and a sufficiently long investment series to compute productivity as a residual from an estimated production function.

“I use initial firm size measured as employment relative to the four-digit industry mean in 1992 as a proxy for productivity, as the survey does not provide for measures of value added nor a long enough series of investment that would permit calculation of productivity as a residual of an estimated production function.”

Classified as probably not a substantial claim.

In Verhoogen's model, quality upgrading is driven by the higher value that consumers in developed countries, such as the United States, place on high-quality goods.

“The mechanism generating quality upgrading in his model is the higher valuation for high quality goods of consumers in developed countries, the United States in this case.”

• cites Verhoogen (2008) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The three technology measures used-spending on technology, spending on technology per worker, and spending on technology relative to sales-yield similar results.

“As measures of technology I use spending in technology, spending on technology per worker, and spending on technology over sales; all produce similar results.”

May combine two claims.

The evidence is not conclusive but suggests that the reduction in Argentina's tariffs relative to Brazil reduced domestic sales.

“The evidence is thus not conclusive but suggestive that the reduction in Argentina's tariffs with respect to Brazil reduced domestic sales.”

May combine two claims.

The absence of a technology-spending response in the top quartile suggests those firms were already above the adoption threshold before liberalization or regardless of it.

“The finding that firms in the top quartile of the size distribution did not increase ST in response to the reduction in tariffs suggests that they were above the threshold before, or regardless of, liberalization.”

Probably not a substantial claim: may not pass S3 (load-bearing).

It is possible that some firms in the first, second, and fourth quartiles were induced to upgrade technology by the reduction in Brazil's tariffs.

“The coefficients are negative but not statistically significant; thus it is possible that some firms in the first, second, and fourth quartiles were induced to upgrade technology by the reduction in Brazil's tariffs.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The survey includes questions on whether firms introduced product or process innovations or improvements during 1992-1996, and these are used for robustness checks.

“In addition, the survey contains a series of questions asking whether the firm performed a certain category of innovation or improvement in products or production process during the period 1992-1996 that I use to perform robustness checks.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Verhoogen (2008) develops a model in which increased trade with more developed countries raises production of high-quality goods, and tests it using Mexico's 1994 devaluation.

“A similar approach was followed by Eric A.Verhoogen (2008) who develops a model where increased trade with more developed countries increases production of high quality goods and tests it in the context of Mexico's 1994 devaluation.”

• cites Verhoogen (2008) sourceNot checked — no full text of the source