"... In PAGE 31: ... This interpretation is consistent with the regressions reported in Tables 8a-b, where the results indicate that even after controlling for size and a wide range of other characteristics, enterprises that use outside auditors and employee training programs generally have higher productivity than their peers that did not. The significant association of these practices with productivity is all the more remarkable in that there is no consistent relationship (see regressions in Table9 a and 9b) between productivity and the variables reflecting the ownership structure (such as whether the firm is a partnership, publicly listed corporation, or whether its shareholders have limited liability). Given the recent emphasis by many economists on corporate governance issues, the contrast between the relative insignificance of these variables in explaining differentials in productivity across firms relative to other factors is impressive.... ..."

"... In PAGE 34: ... Although there may have been complicated processes of selection among the most capable entrepreneurs in choosing the years in which they established firms, the likelihood that less productive firms would have found it easier to survive during the periods most favorable to exports should tend to bias the results against our view. As seen in the second-stage regressions presented in Table11 , however, the use of an instrument for exporting during the first year of establishment yields the same qualitative results. After controlling for other variables, domestically-owned firms predicted -- on the basis of macroeconomic conditions during their year of establishment -- to have exported during their first year of existence were much more productive at the time of the surveys than domestically-owned enterprises that were established in years less conducive to exports.... ..."

"... In PAGE 31: ... This interpretation is consistent with the regressions reported in Tables 8a-b, where the results indicate that even after controlling for size and a wide range of other characteristics, enterprises that use outside auditors and employee training programs generally have higher productivity than their peers that did not. The significant association of these practices with productivity is all the more remarkable in that there is no consistent relationship (see regressions in Table9 a and 9b) between productivity and the variables reflecting the ownership structure (such as whether the firm is a partnership, publicly listed corporation, or whether its shareholders have limited liability). Given the recent emphasis by many economists on corporate governance issues, the contrast between the relative insignificance of these variables in explaining differentials in productivity across firms relative to other factors is impressive.... ..."

"... In PAGE 17: ... Intuitively, this results from there being relatively limited variation in the output to materials ratio over firms, because the scope for producing a unit of output without certain raw materials is limited. The difference can be seen in the contrast between Tables 4a, where the regressions use estimates of TFP based on gross output, with those in Table4 b, where the regressions are ... In PAGE 59: ... Robust t-statistics are reported within parentheses: *significant at 5%; and **significant at 1%. See the notes to Table4 . The dummy variables pertaining to the experience of the founder of the firm pertain To the question n the survey as to whether he or she had previously worked at a firm in the same industry, and who had owned that firm.... ..."

"... In PAGE 17: ... Intuitively, this results from there being relatively limited variation in the output to materials ratio over firms, because the scope for producing a unit of output without certain raw materials is limited. The difference can be seen in the contrast between Tables 4a, where the regressions use estimates of TFP based on gross output, with those in Table4 b, where the regressions are ... In PAGE 59: ... Robust t-statistics are reported within parentheses: *significant at 5%; and **significant at 1%. See the notes to Table4 . The dummy variables pertaining to the experience of the founder of the firm pertain To the question n the survey as to whether he or she had previously worked at a firm in the same industry, and who had owned that firm.... ..."

"... In PAGE 17: ... Intuitively, this results from there being relatively limited variation in the output to materials ratio over firms, because the scope for producing a unit of output without certain raw materials is limited. The difference can be seen in the contrast between Tables 4a, where the regressions use estimates of TFP based on gross output, with those in Table4 b, where the regressions are ... In PAGE 59: ... Robust t-statistics are reported within parentheses: *significant at 5%; and **significant at 1%. See the notes to Table4 . The dummy variables pertaining to the experience of the founder of the firm pertain To the question n the survey as to whether he or she had previously worked at a firm in the same industry, and who had owned that firm.... ..."

"... In PAGE 8: ...14 Note that total factor productivity measures technology broadly defined: this is the levels equivalent of the Solow residual and therefore captures everything not already included in the production function. Table1 summarizes the TFP calculation for a select sample of countries; levels of A for the complete sample are reported in the appendix. As can be seen immediately by taking... In PAGE 10: ... Hall and Jones (1996) provide a more general analysis of productivity and show that differences in TFP levels across countries largely reflect differences in government policies and institutions. The last row of Table1 reports the standard deviation of each component. From this row, one sees that differences in TFP are almost as important as differences in the capital-labor ratio across countries.... ..."

"... In PAGE 20: ... Prices and Productivity As a final point of investigation, we look specifically at how technological change as measured by national industry productivity growth affects prices of domestic industry output. The results are summarized in Table6 . This question is logically distinct from the impact of productivity on either trade volumes or trade patterns.... ..."