The Performance of Industrial Policy: Evidence from Korea

Abstract

This paper first shows that Korea implemented industrial policy properly, promoting infant industries rather than mature ones. The paper then shows that infant industries promoted by industrial policy have matured over time, as well as grown faster than mature industries not promoted by industrial policy. However, this happened as industrial policy was being lifted, rather than as it was being implemented. The paper also shows that, although industrial policy may pay off more easily than previously thought, Korean industrial policy fails to pay off because it distorted the price mechanism too severely and for too long. The analysis of the paper suggests that industrial policy in latecomer countries, when implemented to address market failure, should be much more moderate than what Korea implemented.

Keywords:

• Industrial policy
• infant industry
• cost-benefit analysis
• protection
• subsidy

JEL CLASSIFICATIONS :

• O25
• F13
• L52

Notes

¹There is one study that calculates the welfare effects of protecting a particular industry: the US tinplate industry in the late 19th century (Irwin, 2000). However, an empirical study about the industrial policy covering all of the manufacturing industries has not been conducted.

²This may be complicated by the existence of a policy loan that was important but not related to industrial policy, i.e. loans for exports. Loans for exports remained neutral across the promoted and non-promoted industries, but they may have affected the difference in the average cost of borrowing if its composition differed between the promoted and non-promoted industries. The composition of the loans for exports should have depended on the composition of exports themselves. The share of promoted industries in total manufacturing exports was lower than 50% from 1970 to 1978 (shown later). The lower average cost of borrowing for promoted industries, in spite of a lower proportion of loans for exports, means that the credit rationing for loans other than exports favored the promoted industries. It is less clear that credit rationing for industrial policy accounted for the lower average cost of borrowing for promoted industries after 1978 because of the higher export share of promoted industries. Anyway, all credit rationing was phased out from the 1980s, and this should be responsible for the shrinking difference in the average cost of borrowing.

³To represent the unit cost, the ERP should be calculated for total sales rather than domestic sales only. Kim and Hong (1982) provide the ERP data for both domestic and total sales, but Hong (1997) only for domestic sales. We thus calculate the ERP for total sales from Hong's (1997) data as the weighted average between domestic sales and exports, using domestic sales in world market prices (that is, (domestic sales)/(1+NRP _i (t))) and exports as weights. All ERP figures are calculated by the Corden method. Using the Balassa method does not alter the conclusion of the paper.

⁴ Equation (3) reduces the comparison of unit cost between domestic producers and foreign producers into a comparison of the unit cost between promoted (or non-promoted) industries and the whole manufacturing sector. They are actually equivalent to each other because what matters in both of them is efficiency in the allocation of domestic resources (Bruno, 1972).

⁵One may wonder how the average cost of borrowing could be used as a proxy of the cost of capital for the 1970s when there was a large amount of subsidies because of credit rationing. However, credit rationing to some particular firms or industries meant that firms or industries unable to access official credit had to resort to borrowing from the curb market, which carried much higher interest rates than the market-clearing one. Therefore, for the economy as a whole, it is questionable whether credit rationing lowered the average cost of borrowing.

⁶Owing to short-term fluctuations in net profitability, UC _N (t) is larger than one in 1970 and 1971. However, the average of UC _N (t) from 1970 to 1973 is 0.958, and the average of UC _N (t) from 1970 to 1979 is 0.930. This suggests that the non-promoted industries were mature industries in the 1970s.

⁷The fact that the net benefit is negative under so many favorable assumptions (some of them extremely favorable) may offset some problems with the data, such as interpolating the ERP figures and using the ‘average cost of borrowing in the whole manufacturing sector’ as a proxy of the cost of capital. The data problem is unlikely to undermine the conclusion that industrial policy fails to pay off.

⁸It has been reported that President Park, the authoritarian ruler who pushed the ‘heavy and chemical industry’ drive, devoted 30 to 40% of his time to that task from 1973 (Stern et al., 1995, p. 17).