An Empirical Test of the Infant Industry Argument: Comment

The debate on the role of government intervention in imperfectly competitive markets is still unresolved because of insufficient data. And even with data availability, difficulties in interpretation still arise. For example, the study of Krueger and Tuncer (1982) concluded that more-protected Turkish firms or industries have no systematic tendency to have had higher growth of output per unit of input than less-protected firms or industries. They arrived at this conclusion by just presenting the data; no formal statistical tests were employed. However, if simple correlations were run, the data from Turkey actually indicates a positive relationship between increased protection and higher productivity growth. It is possible to show that protected industries in Turkey did achieve greater productivity gains during the sample period.

Krueger and Tuncer (1982) used three different measures of protection: two effective rates of protection (ERPs) that use tariffs, quotas and input-output tables to compute the effective protection of an industry; and domestic resource costs (DRCs). They also presented two measures of productivity growth: rate of growth of output per unit of input for the firm level and rate of growth of output per unit of input for the industry level.

The data of Krueger and Tuncer (1982) consisted of a sample of 92 firms and an industry sample of two-digit manufacturing industries in the private sector of the Turkish economy. The data covered the period 1963-1976. Estimates were generated for three separate inputs: labor, capital, and material inputs. It appears that the industry aggregates are more reliable than the firm sample.

With few data points and missing observations, one data point may influence the results. Some analysts also highlight that rubber products, which exhibit the highest level of protection and productivity growth, may lead to spurious results. Others argue that if rubber is removed in the data, the petroleum sector must also be removed since it is also an outlier. The petroleum industry had the highest ERP₂ but the its rate of growth of productivity was eight times more negative than any other entry among industry measures of productivity growth. The data may show that protected industries must also be exposed to domestic competition to ensure good performance.

The tests were repeated with and without the three outliers (apparel and footwear, rubber, and petroleum). In four out of six cases, the correlation coefficient is positive, suggesting that higher protection is associated with higher productivity growth. Concentrating on the aggregate industry results, one would find that in two out of three cases, there is a statistically significant and positive relationship between protection and productivity growth.

The sample was divided between protected and traditional industries. Results from a paired t-test show that the difference in means is statistically significant for the sample that excludes the outliers (restricted sample). For the original firm-level and aggregate industry level (full sample), the lack of statistical significance could support the conclusions of Krueger and Tuncer. It also suggests that more data points are needed to yield better results.

After a regression of productivity growth on protection was run, results show that there is a positive and statistically significant relationship between higher DRC’s and higher industrial productivity.

Public sector plants play a major role in promoting infant industries. When this sector is included, protection and productivity growth shares an even stronger and statistically significant relationship. The difference in means is statistically significant for both the full and restricted sample.

The statistical significance of the positive relationship varies with the type of statistical test employed. But regardless of the test, high protection is significantly correlated with high growth rates of productivity at the industry-wide level. Also, splitting the data into protected and unprotected sectors reveals higher productivity growth in protected industries. Hence, there is a tendency for more protected sectors to exhibit higher productivity growth.

Some may argue that this association may be due to the pro-cyclical nature of productivity increases.  Protection may stimulate growth, which in turn is positively associated with production gains. It is also possible that Turkish officials protected the more promising industries. In any case, the gain from protection seems to have been short-lived. Productivity growth in Turkey declined significantly during the 1960’s and 1970’s but Krueger and Tuncer’s claim that protection did not elicit growth in output per unit of input does not seem to be supported with evidence. Further research is required.

Source:

Harrison, Ann, “An Empirical Test of the Infant Industry Argument: Comment.” American Economic Review (September 1994), Volume 84, Number 4, pp. 1090-1095.