World Bank Reprint Series: Number 284 REP284 December 1982 Anne 0. Krueger and Baran Tuncer An Empirical Test of the Infant Industry Argument FILE copy Reprinted with permission from American Economic Review, vol. 72, no. 5 (December 1982), pp. 1142-52. World Bank Reprints No. 250. J. B. Knight and R. H. Sabot, "From Migrants to Proletarians: Employment Experience, Mobility, and Wages in Tanzania," Oxford Bulletin of Economics and Statistics No. 251. M. Louise Fox, "Income Distribution in Post-1964 Brazil: New Results," Journal of Economic History No. 252. Nizar Jetha, "The Welfare Cost of Taxation: Its Meaning and Measure- ment," Bulletini for International Fiscal Documentation No. 253. Larry E. Westphal, "Fostering Technological Mastery by Means of Selective Infant-Industry Protection," Trade, Stability, Technology, and Equity in Latin America No. 254. Gershon Feder, "On Exports and Economic Growth," Journal of Development Economics No. 255. Mohan Munasinghe, "Third World Energy Policies; Demand Manage- ment and Conservation," Energy Policy No. 256. Keith Marsden and Alan Roe, "The Political Economy of Foreign Aid: A World Bank Perspective," Labour and Society No. 257. James A. Hanson, "Contractionary Devaluation, Substitution in Production and Consumption, and the Role of the Labor Market," Journal of International Economics No. 258. Christiaan Grootaert, "The Conceptual Basis of Measures of House- hold Welfare and Their Implied Survey Data Requirements," The Review of Inco7me and Wealth No. 259. Guy Pfeffermann and Richard Webb, "Poverty and Income Distribu- tion in Brazil," The Review of Income and Wealth No. 260. Pradeep K. Mitra, "A Theory of Interlinked Rural Transactions," Journal of Public Economics No. 261. David L. Lindauer and Richard H. Sabot, "The Public/Private Wage Differential in a Poor Urban Economy," Journal of Development Economics No. 262. J. B. Knight and R. H. Sabot, "Labor Market Discrimination in a Poor Urban Economy," Journal of Development Studies No. 263. Carl Dahlman and Larry Westphal, "Technical Effort in Industrial Development: An Interpretative Survey of Recent Research," The Economics of New Technology in Developing Countries No. 264. Michael Bamberger, "The Role of Self-Help Housing in Low-Cost Shelter Programs for the Third World," Built Environment No. 265. Bela Balassa, "The Adjustment Experience of Developing Economies after 1973," IMF Conditionality No. 266. Bela Balassa, "Outward Orientation and Exchange Rate Policy in Developing Countries: The Turkish Experience," The Middle East Journal No. 267. Dipak Mazumdar, "Segmented Labor Markets in LDCs," American Economic Review No. 268. Stephen P. Heyneman and William A. Loxley, "The Effect of Primary- School Quality on Academic Achievement across Twenty-nine High- and Low-Income Countries," The Amierican Journal of Sociology An Empirical Test of the Infant Industry Argument By ANNE 0. KRlIJcr,EI AND BAIRAN TIJNCT-R* Since World War 11, many developing Interestingly, the debate has been entirely countries have provided high levels of pro- theoretical. There has been virtually no sys. tection for newly established industries. tematic examination of the empirical rele. These policies have generally been followed vance of the infant industry argument. This on the grounds that new industries are "in- is remarkable in light of the importance of fants," and that dynamic factors will come the question, and the fact that thirty years' into play to insure later economic efficiency.' evidence or more has accumulated in a num- At a theoretical level, the infant industry ber of countries. Even if there are conditions exception to the proposition that free trade under which dynamic factors and externali. is optimal has always been noted.2 Skep- ties in an infant industry might warrant in. tics have centered their misgivings on two tervention, that does not prove that those grounds: I) they have questioned whether conditions are in fact met. In the last analy. protection through the trade -regime would sis, defense of infant industry protection must achieve the goals of infant industry protec- rest on empirical grounds: do the long-run tion; 3 and 2) they have pinpointed the com- benefits justify the short-run costs of starting bination of "dynamic factors" and "external- up an initially high-cost infant? ities" that would have to arise to justify It is the purpose of this paper to develop a infant industry intervention and questioned test of whether infant industry criteria are the empirical likelihood of such circum- satisfied, and then to apply the test to one stances. developing country, Turkey. Section I sets forth the infant industry argument and nec- *University of Minnesota and the World Bank, re- essary conditions for there to be a valid case spectively. We wish to thank the National Science for intervention. Section 1I then examines Foundation for research support, grant no. SOC77- the various ways in which the necessary dy- 25776. Wu tre also griteful to the Turkislh Industrial iiimic externaiilities mighlt niaiiifest ilini- D)evelopment hlank for il, assistance in the project. We have benefited greatly from the comments and sugges- selves. This immediately suggests a simple tions of W. M. Corden and James M. Henderson. Ebbe empirical test. Section III then presents re- Yndgaard, Claus Vastrup, Martin Paldam, and other sults of the test for Turkey. A final section members of the faculty at Aarhus University provided contains some concluding observations and helpful comments and suggestions at an early stage of suggestions for further work. The data and the research, as did colleagues at Bogazici Unuversity, do including especially Maxwell Fry. Members of the Trade detans of procedures for estmation are given and Development Workshops at the University of in the Appendix. Chicago made many useful comments on an earlier draft of this paper. Zafar Ahmed, Roger Johnson, tnci I. IlTe Infant Industry Argument Mubarek, Lale Tezel, and Paitoon Wiboonchutikula provided research assistance at various stages of the project. It Is smplest to state the mfant mdustry The "infant industry argument" is also the basis on case in naive form, and then to consider the which developing countries are excepted from some conditions under which it would be valid. provisions of the GATr. The basic argument, crudely put, is that: 2See Paul Samuelson, (1958, ch. 8), for a discussion A. Some newly established activities are of the history of thought with regard to the optimality of . . c free trade. The optimum tariff argument is irrelevant to initially t igh cost relative to established for- the concerns of this paper and therefore not considered eign enterpnses and it requires time for them here. to become competitive. 3See Robert Baldwin (1969) for an excellent state- B. It does not pay any individual en- ment of the reasons why, even in the presence of infant industry considerations, tariff protection might fail to trepreneur to enter an Infant idustry at free correct the assumed market imperfection. trade prices; but 1142 VOL. 72 NO. 5 KRUEGER AND TUNCER: INFANT INDUSTRYARGUMENT 1143 C. The industry, if developed, would be the activity initially. Thus, the presence of economic enough to permiit a reasonable rate externalities is necessary in order to show of return on the initial losses; and therefore that private activity will not generate the D. The industry requires a temporary optimal development of infant industries in a period of protection or assistance during market-oriented economy. Whether..the ex- which its costs will fall enough to permit it to ternalities are at the individual industry level, survive international competition without or rather at the level of the entire industrial assistance. sector is an open question, discussed further The first proposition is essentially that in Section II. Clearly, whatever infant in- costs of a new activity may initially be high. dustry assistance is provided should be pro- Reasons put forth as to why they might vided to the industry, group of industries, or be high include learning by doing and the sector generating the externalities at a rela- possibility that there are "linkages" between tively uniform rate. Different levels of pro- industries set forth by Albert Hirschman tection to different activities would be war- (1958). If the latter, then in the early stages ranted only if the sector containing the of development, the absence of complemen- more-protected activity were expected to ex- tary activities or small size of the industrial perience greater cost reductions than the sector of an economy might constitute a less-protected sector.5 reason why all industrial activities would ini- The third proposition asserts that the losses tially be high cost. associated with an initial period of high costs The reasons why there might be learning must be recovered (with interest) at a later by doing are numerous. Workers might re- date, although not by the individual en- quire a period of training. There might be an trepreneur starting up the activity. In es- initial shake-down period as the activity be- sence, the costs of production of those ben- came operational. Management itself might efiting by the development of the activity gain in experience. These possibilities have must fall enough to repay the initial losses been neatly encapsulated by Kenneth Arrow and to provide a reasonable rate of return on (1962) in the notion that output per unit of those losses (since resources could otherwise input might increase as cumulative output have been allocated to unprotected activities within a given line of activity (the plant? the with incremental international value-added). firm? the industry? the entire industrial sec- The fourth proposition is really a logical tor?) increased.4 Regardless of which reason consequence of the first three. All analysts is put forth, an essential feature of the infant have been willing to concede that if the industry notion is that a new activity will first three propositions were valid, some initially be high cost, but that unit costs will form of assistance (and intervention with a decline over time. laissez-faire outcome) to the externality-gen- The second proposition is that, while costs erating activity is warranted. However, pro- will decline, they will do so in a way that tection would never be first best (contrasted individuals initially starting the activity will with a production subsidy), and might not not reap the full rewards. Otherwise, there even achieve its intended purposes, as argued would be no case for protection: if start-up by Baldwin. costs are high but the activity is economic, it For the present, the important aspect of would pay an individual entrepreneur to in- the infant industry case seems summarizable cur those costs in order to reap later benefits. in the proposition that, in order for it to be For there to be a case for intervention, there empirically valid, a necessary (but not suffi- must be positive externalities from the de- cient) condition is that costs in (temporarily) velopment of an infant activity which accrue to individuals other than those undertaking 51t might be contended that a more-protected Sctor would generate greater externalities accruing to other 'See Baldwin as to other mechanisms by which exter- activities. The unresolved question in that case is why nalities have been said to affect profitability of start-up protection should be temporary. If it were not tem- of new activities. porary, infant industry considerations do not apply. 1144 THE AMERICAN ECONOMIC REVIEW DECEMBER 1962 assisted or protected industries should have Clearly, the change in i's costs is fallen over time more rapidly than costs in nonprotected or less-protected industries. (2) dC, = 2dW,V,, + d V,,W,, This interpretation coincides with the case i where it is assumed that prices in the rest of the world are given, and do not change over and the change in cost per unit of output is time due to differential rates of technical change in the rest of the world.6 If there were 3 C, dWj WjVj, C, also technical change in the corresponding (3) d Wj C, X, industries abroad, the infant industry case would need to be reformulated to state that intervention would be warranted only if unit + v , I C1 _ dX C costs were expected to decline more rapidly + V C, X, X, X, in the infant industry than in the mature industry abroad (with the same qualifica- Denoting the share of the jth factor in total tions as above regarding externalities and costs in industry i by a,y, equation (3) can be recovery of the initial investment with inter- rewritten: est). For the purposes of this paper, it is assumed that world prices are 'iven, so that (4) a decline in costs in one industry at a more c XI dWj dV dX raipi(d rate thn in Lnother constituites nore d________ aj '+I Vi dX of an infant industry case. ' ' j J j i i 11. A Simple Empirical Test Thus, the proportionate change in costs per unit of output in the ith industry represents There are two ways that one industry's the share-weighted sum of changes in input costs per unit of output (or value-added) can prices plus the share-weighted sum of factor change relative to another's: either its share- inputs less the rate of change of output. Let weighted inputs per unit of output must fall C, represent the proportionate rate of change more (or rise less) than the other's, or the in costs per unit of output. Contrasting relative price of the factor it uses relatively changes in costs between the ith and the kth intensively in production must fall. industry yields This can readily be seen as follows. Define dW the total cost, C, of the ith industry as (5) C,-ek= (a, - akJ)W (I) Cj= Wi Vi d|,} V,j_dX,| where W is the reward to the jth factor of production and VJ' is the quantity of the jth dVjk dXk factor employed in the ith industry. L2akiv- - The first term on the right represents the 61t might be anticipated that the world price of a change in relative costs due to changing rela- particular commodity would rise over time (for reasons tive input prices. Clearly, that relative input other than technical change), and therefore a presently prices may change in the process of growth is uneconomic industry might become economic. This poricesnma chang infn the str proesfrothisn. would not constitute a case for infant industry protec- not grounds for infant industry protection. tion, however, because: (i) it is not clear that because the priee in the future will rise, activity now is eco- nomic; and (i) there is no reason for intervention since 7As a practical matter, changes in relative factor those undertaking the activity now will benefit from the prices do not scem to be a major source of relative cost higher future price, changes. Turkey had a very large change in relative VOL. 72 NO. 5 KRUEGER AND TUNCER: INFANT INDUSTRYARGUMENT 1145 If there is to be a dynamic cost reduction, it industry grounds and its costs have fallen must be reflected in a difference between the relative to k, it will be judged that there were two bracketed terms on the right-hand side. some dynamic factors in industry i that may We are thus led to the straightforward have warranted intervention (although there proposition that if there are dynamic factors is no presumption whatsoever that interven- warranting intervention, they will be re- tion was optimal). Passing the test is a neces- flected in a difference in the two right-hand sary condition for there to have been an terms of equation (5). Define, now,8 infant industry. It is not sufficient to prove that infant industry protection was war- dA4 dX, dV), ranted because: (i) the industry might have (6) - _ X -, a,, V * developed anyway; (ii) the rewards may all Ai XI j 'i i have gone to the entrepreneurs in the in- dustry; (iii) the reduction in costs might Substituting (6) into (5), and dropping the have come about for reasons other than ex- first term as irrelevant for infant industry ternalities; or because (iv) the reduction in purposes, costs was not sufficient to provide an ade- quate rate of return on earlier losses. It might dAk_ dA, not have been optimal because an alternative 7e ek Ak - A, intervention instrument or a lower level of protection might have achieved the same or In order for infant industry considerations to better results with lower costs. have warranted intervention in favor of in- If, however, costs in industry i did not fall dustry i, costs per unit of output must have relative to industry k, clearly protection was fallen more in i than in k. Equation (7) not warranted.9 It is in this sense that a shows that a necessary condition for this to contrast of rates of growth of output per unit occur is that inputs per unit of output de- of input between more- and less-protected crease more rapidly in industry i than in industries constitutes a test for the empirical industry k. As formulated, this unit cost validity of the infant industry argument. reduction could come about because of tech- Before proceeding to the empirical results, nical change, the overcoming of indivisibili- two questions remain. A first question is the ties, the realization of scale economies, or for time period over which infant industry con- genuine infant industry reasons. siderations might warrant intervention. The This, then, is the empirical test. Should second pertains to the range of activities over industry i have been protected on infant which the test should be carried out. The first is the simpler question, since all that is required is a period sufficiently long ractor prices due to government Intervention in the so that, if cost reductions were not incurred, labor market, and yet the changes in costs that these it could reasonably be concluded that the could have induced seem relatively small. See Table A2 costs of protection would not in all likeli- for the calculations. It has also been suggested that hood be recovered. Since the Turkish data commercial policies themselves might induce changes in relative factor prices and that this is a dynamic factor pertain to a period of thirteen years (and that should be considered. However, if commercial longer), we note simply that with a real rate policies cawsed changes in relative factor prices, they of return of 10 percent,'0 the present value of would clearly increase costs in protected industries and thus tend to weaken whatever dynamic case there was for intervention. 9This does not prove that there might not bave been 8The dA/A is nothing other than the conventional an infant industry case. It is conceivable that incentives formula for total factor-productivity growth, which is other than those created by the forms of protection thc rate of growth of output less the sharc-wcighted ratc actually used might have induced entrcprcneurs to en- of growth of inputs per unit of output. For present gage in cost-reducing activities. See our 1981 paper for purposes, however, the assumptions necessary to justify an analysis of the effects of the trade regime on incen- usc of dA/A as a measure are far wcaker than thosc tives in Turkey. necessary for a total factor-productivity growth interpre- "5Most observers would put the real ratc of return in tation. Turkey at a number substantially higher than this. 1146 TIHE AMERICAN ECONOMIC REVIEW DECEMBER 1982 cost savings ten years' hence is less than 40 observed to have experienced a relatively percent of the anticipated amount. It seems high rate of growth of output per unit of lotiibltfil wilctlicr plolection lfor at periodl of input in contrast to tlc rest of the econloImy more than ten years, with no beginning of a (in contrast to mature industrial economies), reduction in costs, could conceivably conilc CoIIIparison of output per uniit of input tiidcr thl lictidinig of justified infainit industry aicross countries is inherently difficult, but protection. nonetheless can provide a partial check on The second question is the more difficult. the plausibility of this possibility. It will be recalled that the infant industry argument presumes both dynamic factors and III. Results externalities. The test described above is straightforward in evaluating for the. pres- As already mentioned, Turkey has pro- ence of dynamic factors, but does not indi- vided protection, on infant industry grounds, cate to which units it might apply. Since to a variety of new industries. Protection has protection is granted at different rates to been largely automatic because the authori- different industries, it seems natural to sup- ties have generally prohibited imports of any pose that the relevant i and k to contrast good once domestic production began." would be different industries subject to dif- Rates of effective protection have been fairly ferent levels of protection. Having done so, a high, and estimates must be based upon direct higher dA /A would be required for a more- price comparisons rather than upon tariff protected industry than for a less-protected schedules. The best available estimates are one to satisfy the infant industry test. Most given in Table I below. proponents of the infant industry argument The details of data sources and procedures seem to adopt this notion that the benefits for estimating output per unit of input are are external to the firm but internal to the given in the Appendix. Here, only three industry. This would appear to imply that points need to be noted. First, there are two rates of growth of output per unit of input sets of estimates available: one from a sam- should be higher for the industry than for ple of 92 firms and the other for two-digit new firms (or new investments of existing manufacturing industries in the private sec- firms). It is also possible, however, that ex- tor of the Turkish economy. Secondly, the ternalities spread across new entrants, and main thrust of import substitution activity do not affect more traditional firms within (on infant industry grounds) in Turkey was industries. In that event, one would expect during the early and mid-1960's. The two- output per unit of input to grow more rapidly digit industry data cover the period 1963-76, in newly established firms or activities than while data for individual firms cover at least in preexisting ones. that period when the firms were already in Both of these possible relationships imply existence but shorter periods in some in- that the relevant unit for externalities to be stances when the firms started operation in recaptured is somewhere within a given, pro- the late 1960's. Finally, since much of the tected, industry. While this seems the most import-substitution process consists of re- plausible infant industry interpretation (and placing imported inputs with domestic ma- the one used here), some might argue that terials, estimates were generated for three the benefits of new industries are spread separate inputs: labor, capital, and material across the entire industrial sector, and are inputs.'2 not centered in the protected industries Table I gives the main findings. The Ap- themselves. One might be skeptical of the pendix gives sources and procedures and the argument, on the grounds that it is hard to see why different levels of temporary protec- "One would anticipate that, in the absence of cx- tion should be accorded to different in- pected monopoly power from entering a given tine of dustries unless their own costs would fall activity first, the automatic protection mechanism would differentially. But if the relevant source of provide an incentive for the more economic among the import-competing industries to be developed first. externalities is the entire industrial sector, '2No data were available with which to estimate the industrial sector as a whole should be changes in skills of the labor force. VOL. 72 NO. 5 KRUEGERAND TUNCER: INFANTINDUSTRYARGUMENT 1147 TABLE I-EFFECTIVE RATES oF PROTECTION AND RATES Op GROWTH OF OUTPUT PER UNIT OF INPUT Rate of Growth of Output/Input Industry ERP, ERP2 DRC Firm Sample rlndustry Food Products 13 n.a. 18 .25 .16 Fur and Leather Products 14 -24 -15 n.a. -1.17 Wood and Cork Products 16 58 -13 -3.34 -.55 Furniture and Fixturcs 16 n.a. n.a. n.a. -.56 Nonmctallic Mineral Products 23 -27 1 1.61 .72 Textiles 42 -23 12 .72 .84 Apparel and Footwear 42 47 n.a. 5.24 4.10 Metal Products 57 140 682 -.05 1.61 Chemicals 60 200 21 -.04 .46 Electrical Machinery 63 113 36 5.76 1.41 Paper and Products 72 105 97 n.a. 1.55 Rubber Products 77 N-IVA 279 n.a. 4.27 Basic Metals 80 113 14 2.21 -.93 Nonelectric Machinery 142 132 36 n.a. .62 Petroleum Refining n.a. 236 n.a. n.a. -8.80 Transport Equipment 209 134 131 n.a. .94 All Manufacturing 1.91 1.84 Source: ERP1: Ozfirat estimates given in Krueger, Table IX-2; ERP2: Baysan estimates given in his Table 1, p. 126. DRC: Krueger, Table VlIl-I. Notes: I) Beverages and Tobacco are not reported here due to lack of a measure of effective protection; both are traditional. Estimated rates of growth of output per unit of input are 4.31 percent annually for Beverages and 5.97 percent for Tobacco. 2) All rates of growth are continuous natural rates. N-IVA denotes negative international value-added. underlying data on rates of growth of out- the same industry at different points in time. puts and inputs from which these estimates Even more important is the consideration were derived. that there are import-substitution industries The first three columns of Table I give within "traditional" sectors (such as syn- three different estimates of sectoral protec- thetic textiles), and "traditional" activities tion all pertaining to the late 1960's. The first within import-substitution sectors (such as are based on sectoral averages computed by copper processing within basic metals). Thus, the State Planning Organization for 1968. the variability reflects the underlying reality Sectors are listed in order of increasing pro- that levels of protection differ widely even tection based on these estimates. The second within particular industries.'4 are based on input-output tariff data ad- Despite the wide variability, the three sets justed for the estimated additional protection of estimates together provide a fairly good accorded by import quotas and prohibitions. indication of the height of protection in the The third are domestic resource-cost esti- mid- to late 1960's, and its differential across mates taken from a sample of firms. While industries. Essentially, the first seven are all the last are most closely based on price com- regarded as traditional industries within parisons, they suffer from the drawback that Turkey; the last nine are regarded as the levels of protection vary so much within each import-substitution sectors.'5 The latter are sector that sampling error is probably fairly large.'3 This variability stems partly from the fact that an import-licensing regime inher- 14While data in nominal terms are availablc for three- ently provides varying levels of protection to and four-digit industries, no appropriate price deflators or detailed estimates of effective protection rates are available. "See Krueger, where the variance in the estimated '5The import substitution industries generally expe- sectoral means was also calculated. rienced more rapid growth of output. See Table Al. 1148 THEAMERICAN ECONOMIC REVIEW DECEMBER 1982 those that were encouraged in the early higher rate of growth of output per unit of 1960's, and were the focus of Turkey's im- input than their corresponding industries, but port-substitution polices on infant industry surely the difference is well within the margin grounds. The positive rates of protection for of error of the calculations. the traditional sectors probably did little Finally, there is the question as to whether more than offset currency overvaluation in externalities could have been realized else- the late 1960's; the Turkish lira was devalued where in the manufacturing sector. Here, the by 66 percent in 1970. only way of judging is to evaluate the esti- The last two columns in Table I give mated rate of growth of output per unit of estimated rates of growth of output per unit input in the manufacturing sector as a whole. of input. The firm sample column 'gives That, in turn, involves a comparison of the estimated rates for the sample of firms. As rate realized in Turkey with that in other can be seen, there were some industries for countries. Because data are not entirely com- which no firm data were available. In some parable, and because the estimates are resid- instances (such as petroleum refining) this uals and therefore subject to fairly wide was because the activity is undertaken pri- margins of error, such comparisons are nec- marily by one large firm; in other instances, essarily extremely hazardous. Estimates typi- there were simply no firms in' the sample cally range from 3-4 percent for developed data. Rates for the industry cover the years and other developing countries."6 Despite 1963-76, the period for which State Institute problems of comparability, it hardly seems of Statistics data are available. plausible that differences in measurement As can be seen, there is no systematic account for the lower figure in Turkey. tendency for more-protected firms or in- To see just how low the estimated rates of dustries to have had higher growth of output growth of output per unit of input are, con- per unit of input than less-protected firms sider the following. Suppose a firm initially and industries. Two industries-apparel and experienced a 50 percent cost disadvantage footwear and rubber products-appear to (i.e., required 50 percent effective protection). have experienced relatively rapid growth of Output per unit of input would have to grow output per unit of input. Apparel and 4 percent annually more rapidly than in other footwear is a traditional industry in Turkey, industries in order for it to be able to survive and its medium rate of effective protection without protection ten years hence."7 This, reflects currency overvaluation and the nega- however, would provide no return on the tive protection to textiles, rather than posi- initial loss. It is thus an underestimate of the tive nominal protection directed toward ap- differential in growth of output per unit of parel and footwear. Rubber products is a input that would be necessary to warrant sector with a sizable traditional component protection of the infant industry. and import-substitution activities consisting For the Turkish case, when all manufac- primarily of tire production. This latter activ- turing was experiencing increased output per ity was extremely high cost, as reflected both in Tercan Baysan's estimate that interna- 16Edward Chen (1977) estimated rates of manufac- tional value added was negative and in a tunng total factor productivity growth of 2.29, 3.47, very high DRC estimate. No firms producing 3.50, and 3.75 percent for Hong Kong, South Korea, rubber products were in the sample, so only Taiwan and Singapore, respectively, for the 1960's. a sectoral rate is available. Estimated rates for developed countries include 3.5 per. There is likewise no apparent tendency for cent for Norway (V. Ringstad, 1971), 3.66 for Japan (Mieko Nishimizu and Charles Hulten, 1978), 3.75 for the new activities, as reflected by the firm Italy (Vittorio Conti and Renato Filosa, 1979), and 2.9 data, to have experienced rates of growth of percent for the United States (John Kendrick, 1976). In output per unit of input systematically higher some of these cases, quality adjustments have been or lower than the industry to which they made to estimated inputs. Using unadjusted data would -.longed Thus, the externality argument raise those estimates, making the contrast with Turkey belonged. Thus, the extemnality argument even sharper. does not seem borne out by the data: if 7This number is found by solving I, = 10 (I - r)' for anything, sample firms experienced a slightly r when l,=.667, 10= I and t equals 10. VOL. 72 NO. 5 KRUEGER AND TUNCER: INFANTINDUSTRYARGUMENT 1149 unit of input at a rate of 1.8 annually, this centive structure, output per unit of input would imply that industries experiencing 50 might have grown more rapidly in some, or percent protection should increase output per possibly even all, Turkish industries.'9 What unit of input at a rate of at least 5.8 percent can be concluded is that, at least in the annually. More concretely, consider rubber Turkish case, protection did not elicit the products, the "best" Turkish case. If the low sort of growth in output per unit of input on estimate of the ERP for rubber products is which infant industry proponents base their accepted, the output per unit of input would claim for protection. have to grow at 7.34 percent annually in order for their costs to fall enough for them APPENDIX to become competitive in ten years. At their existing rate of growth of output per unit of The major thrust of import substitution input, it would require twenty-three years for into new industries occurred in the early and them to become competitive."8 Even that mid-1960's. For the period 1963-76, there calculation takes the low estimate of the rate are industry-level data available with which of protection, allows for no return on the to estimate inputs and outputs. A Census of investment over the twenty-three-year inter- Manufacturers and Annual Survey of In- val, and is for the two-digit industry with the dustries provides detailed data on number of highest estimated rate of growth of output employees, wage bill, value of purchased in- per unit of input. Obviously, for "infant" puts, value of output, investment made by industries such as paper where outputs per firms, and number of firms for private activi- unit of input grew at less than the average ties within each industrial sector employing rate of all manufacturing, there can never be ten or more employees. These data, com- a "catch up" as long as existing relative rates bined with estimates of capital stock pro- are maintained. vided by the State Planning Organization and appropriate price deflators20 form a data IV. Conclusions set from which it is possible to infer the behavior of inputs and outputs for two-digit This paper has attempted two things: to manufacturing industries in the private sec- develop an empirical test for the validity of tor in Turkey.2' Since much of the import- the infant industry argument; and to use that substitution process consists of replacing im- test on Turkish data. The test is simple and ported inputs with domestic materials, three straightforward: input per unit of output inputs were separately estimated: labor, must fall more rapidly in more protected capital, and material inputs. industries if there is to be any rationale for A second set of data is at the firm, rather infant industry protection. In the Turkish than Industry, level. It covers those firms case, there was no such tendency over the which received loans from the Turkish In- period covered. dustrial Development Bank. For them, data The fact that protected Turkish industries were available on a variety of their attributes did not experience rapid increases in output (size, date of inception, precise composition per unit of input is sufficient to prove that of output, etc.) and also for annual invest- protection was not warranted. It does not, however, prove that there were no infant '9See our earlier paper for an attempt to trace the industries. It might be that the trade regime links between the growth of output per unit of input and itself provided the wrong incentives. It is at the ebbs and flows of the trade regime. least possible that, under an alternative in- of Wholesale price indices were available for outputs ofeach two-digit industry. These data were then used, in conjunction with the Turkish input-output tables, to obtain a weighted input price for each sector's purchases. "'This calculation is based on the assumption that The same price deflators were used for two-digit in- output per unit of input in rubber products continues to dustries and for the firm data described below. grow at 4.27 percent annually while in the entire manu- 21 In our earlier paper the behavior of the private and facturing sector it continues to grow at 1.84 percent public sectors is analyzed and contrasted, and a fuller annually. description of the data is given. 1150 TIlEAMERICAN ECONOMICR EVIEW DECEMBER 1982 TABLE Al-UNDERLYINO GROWTH RATES OF OUTPUT AND INPUT FOR PRIVATE SECTOR INDUSTRIES AND SAMPLE FiuBms Industnes' 1963-76 Growth Rate of: Sample Firm Growth Rate of: Labor Capital Real Output Labor Capital Real Output Food Products 4.6 14.2 7.7 7.5 13 3 9.5 Beverages 18.0 14.2 22.5 - - - Tobacco -3.3 -.3 5.7 - - - Textiles 4.6 13.3 11.0 6.2 14.1 10.2 Apparel and Footwear 23.3 13.3 28.3 1.9 8.0 6.8 Wood and Cork Products 10.0 14.2 12.6 9.1 27.9 15.8 Furniture and Fixtures 4.7 14.2 6.6 - - - Paper and Products 16.7 26.0 23.7 - - Chemicals 8.4 15.4 15.1 4.2 12.4 12.0 Rubber Products 5.1 13.3 16.8 - - Fur and Leather Products 7.8 17.0 8.6 - - - Petroleum and Coal Products 28.1 60.5 33.7 - - Nonmetallic Minerals 8.0 16.7 15.3 5.0 7.4 7.4 Basic Metals 18.1 25.3 21.5 7.6 14.9 15.8 Metal Products 7.8 13.1 11.8 9.0 17.1 13.3 Nonelectrical Machinery 15.2 17.6 17.9 - - Electrical Machunery 12.5 20.1 19.8 Transport Equipment 22.7 30.5 30.1 Nore: All rates of growth are continuous natural rates, computed by running a logarnthmic regression of each vanable on time. ments, annual labor force and wage bill, lowance for price level changes in their capital annual purchases of raw materials and inter- stock in the context of a relatively high rate mediate goods and inventory changes, sales, of inflation. Depreciation rates were esti- profits, depreciation, and so on. Altogether, mated from American engineering data found there are 91 firms for which data were avail- in W. R. Park,23 and then scaled to equal the able on a reliable basis for a period of more State Planning Organization's estimate of the than five years.22 Most new investments of average rate for all manufacturing. Invest- these firms were undertaken in response to ment deflators available from the State Pltin- ituecintives provided by the trude rcginic, ul- niibg Orgutniwution were first cmployed t) thotugh some were in more traditional in- convert nominal invesiment into conslant- dustries. Since there was credit rationing in price estimates of additions to capital stock. Turkey, there is some presumption that bor- Investment in a given year was treated as rowers from the Industrial Development becoming effective capital only at the begin- Bank were firms of above-average quality, ning of the following year.24 Period t -l's according to the criteria used by the Bank capital stock was depreciated, and then real for its lending. On the basis of these data, it was possible "it is an interesting question whether one should to compute an estimated capital stock for a priori expect depreciation rates to be lower or higher each firm using perpetual inventory tech- in Turkey than in the United States. On one hand, niques. Doing so was judged better than cheaper labor should encourage more maintenance and using balance sheet estimates (which were thus a longer economic life. On the other hand, poor also available) since the latter made no al- and irregular materials quality, irregular supplies of electric power, and workers with less experience in the care of equipment might tend to the opposite result. 22interviews were held with more than a quarter of 24For some older firms, data were not available from the firms, which provided a check on the reliability of inception. In those cases, initial balance sheet data were the data, and also, provided additional information on converted to an estimate of real capital stock based on characteristics of firms and their management. knowledge of the firm's history and starting date. VOL.72 NO.5 KRUEGERANDTUNCER:INFANTINDUSTRYARGUMENT 1151 investment in t-I was added to obtain TABLE A2-COMPUTATION OF POSSIBLE CHANOES capital stock in period t. IN RELATIVE COSTS DUE TO CHANGES In addition, data from the firms could be IN RELATIVE FACrOR PRICES directly used for the number of workers. Wage Share of. Rate of Purchased inputs, adjusted for inventory Industry Value-Added Cost Change changes, were deflated to yield an estimate of material inputs. Finally, for some firms a Food Products .264 -0.4 physical indicator of homogeneous output Beverages .101 -2.0 (for example, tons of cement) was available Tobacco .183 -1.3 Textiles .368 0.3 and used to indicate output. For others, it Apparel and Footwear .521 1.7 proved preferable to take deflated sales ad- Wood and Cork Products .485 1.4 justed for inventory change as the measure Furniture and Fixtures .381 1.5 of output. Papcr and Products .372 .4 Thus, for both firms and industries, data Rubhcr Products .274 - .5 were available on materials inputs, outputs, Fur and Leather Products .484 1.4 labor inputs, and capital stock inputs, along Petroleum and Coal .016 -2.8 with the shares of the respective factors in Nonmetallic Minerals .326 -0.0 the value of output. For purposes of estimat- Basic Metals .201 - 1.2 Metal Products .384 0.5 ing changes in input per unit of output, the Nonelectric Machinery .320 -0.1 average share (over the life of the firm and Electrical Machinery .332 0.0 for thc entire 1963-76 period for two-digit Trnnsport Equipmcnt .562 3.7 industries) of labor, material, and capital was All Manufacturing .245 0.0 used. T'his procedure was judgcd superior to employing a Divisia index because of the Nores: Real wages are estimated to have risen at a volatility of shares from year to year.23 continuous rate of 6 percent from 1963 to 1976, based Table AI provides the estimates of rates of on avcerage weekly eanings covered under social in- surance. The median share of labor was 33 percent. growth of labor and capital inputs and out- Relative cost changes were computed be weighting these puts. All rates are computed for the period rates of change by each industry's actual (1968) shares. 1963-76 at the industry level. For firms, For the median industry's costs to have remained con- rates were computed over the period for stant, capital costs would have had to decline at a which data were available, and a minimum continuous rate of 2.95 percent. of five years. In some instances, firm data span a period of twenty years, but some newer import substitution activities did not REFERENCES start until the late 1960's; data on these firms are also included in the estimates. Arrow, Kenneth J., "The Economic Implica- Table A2 gives data on the wage share in tions of Learning by Doing," Review of each two-digit industry and computes the Economic Studies, June 1962, 29, 155-73. maximal rate of change in relative costs that Baldwin, Robert E., "The Case Against In- could have been associated with the very fant-Industry Tariff Protection," Journal steep increase in real wages that occurred in of Political Economy, May/June 1969, 77, Turkey during the period. The real wage 295-305. increase was the result of labor legislation Baysan, Tercan, "Economic Implications of and did not reflect underlying labor market Turkey's Entry into the Common Market," conditions: urban unemployment was rising unpublished doctoral dissertation, Univer- rapidly during most of the period. sity of Minnesota, 1974. Chen, Edward K. Y., "Factor Inputs, To- tal Factor Productivity and Economic "Initial eslimates, hased on Divisia indices, yielded Growth: The Asian Case," Developing occasionally bizarre results. For example, for firms or E J 3 19 industrics sifrfering losscs, the cnpital '%hare was ncgn- conomies, une, 77, 15, 121-43. live, and firms with hcavy investment were calculated to Conil, Vtltorlo and F;losNa, Renalo, "A Disag- have increased output per unit of input! gregate Analysis of Accumulation, Produc- 1152 THE AMERICAN ECONOMIC REVIEW DECEMBER 1982 tivity, and Labor Costs in the Manufactur- 1955-71," Review of Economics and Statis- ing Industry" in Economic Papers, 2, tics, August 1978, 60, 351-61. Rome: Bank of Italy Research Depart- Park, W. R., Cost Engineering Analysis, New ment, June 1979. York: John Wiley and Sons, 1973. Hirschman, Albert O., The Strategy of Eco- Ringstad, V., Estimating Production Functions nomic Development, New Haven: Yale and Technical Changes from Microdata, University Press, 1958. Norway: Statistisk Sentralbyra, 1971. Kendrick, John W., "Productivity Trends and Samnuelson, Paul A., Foundations of Economic Prospects," in U.S. Economic Growth from Analysis, Cambridge: Harvard University 1976 to 1986, Washington: Joint Eco- Press, 1958. nomic Committee of Congress, October 1, Turkish Republic, State Institute of Statistics, 1976. Census of Manufacturing Industries, An- Krueger, Anne O., Foreign Tradi Regimes kara, 1963; 1970. and Economic Development: Turkey, New , , Annual Survey of Manufac- York: Columbia University Press, 1974. turing Industries, Ankara, various years to -_____ and Tuncer, Baran, "Growth of Factor 1976. Productivity in Turkish Manufacturing In- _ , State Planning Organization, Turkiye dustries," unpublished paper, 1981. Imalat Sanayiinde Dermaye ve Isgucu Nishinizu, Mieko and Hulten, Charles R. "The (Capital and Labor in Turkish Manufac- Sources of Japanese Economic Growth: turing Industry), Ankara, December 1977. I No. 269. James R. Follain, Jr., Gill-Chin Lim, and Bertrand Renaud, "Housing Crowding in Developing Countries and Willingness to Pay for Additional Space: The Case of Korea," Journal of Development Economics No. 270. Bela Balassa, "Policy Responses to External Shocks in Sub-Saharan African Countries," Journal of Policy Modeling No. 271. Jaime de Melo and Sherman Robinson, "Trade Adjustment Policies and Income Distribution in Three Archetype Developing Economies," Journal of Development Economics No. 272. J. B. Knight and R. H. Sabot, "The Role of the Firm in Wage Determination: An African Case Study," Oxford Economic Papers No. 273. William G. Tyler, "The Anti-Export Bias in Commercial Policies and Export Performance: Some Evidence from Recent Brazilian Experi- ence," WeltWirtschaftliches Archiv No. 274. Ron Duncan and Ernst Lutz, "Penetration of Industrial Country Markets by Agricultural Products from Developing Countries," World Development No. 275. Malcolm D. Bale, "Food Prospects in the Developing Countries: A Qualified Optimistic View," The American Economic Review (with Ronald C. Duncan) and "World Agricultural Trade and Food Security: Emerging Patterns and Policy Directions," Wisconsin International Law Journal (with V. Roy Southworth) No. 276. Sweder van Wijnbergen, "Interest Rate Management in LDCs," Journal of Monetary Economics No. 277. Oli Havrylyshyn and Iradj Alikhani, "Is There Cause for Export Optimism? An Inquiry into the Existence of a Second Generation of Successful Exporters," Weltwirtschaftliches Archiv No. 278. Oli Havrylyshyn and Martin Wolf, "Recent Trends in Trade among Developing Countries," European Economic Review No. 279. Nancy Birdsall, "Fertility and Economic Change in Eighteenth and Nineteenth Century Europe: A Comment," Population and Development Rcview No. 280. Walter Schaefer-Kehnert and John D. Von Pischke, "Agricultural Credit Policy in Developing Countries," translated from Handbuch der Landwirtschiaft und Ernahrung in den Entwicklungslandern (includes original German text) No. 281. Bela Balassa, "Trade Policy in Mexico," World Development No. 281a. Bela Balassa, "La politica de comercio exterior de Mexico," Comercio Exterior No. 282. Clive Bell and Shantayanan Devarajan, "Shadow Prices for Project Evaluation under Alternative Macroeconomic Specifications," The Quarterly Journal of Economics No. 283. Anne 0. Krueger, "Trade Policies in Developing Countries," Hand- book of International Economics Issues of the World Bank Reprint Series are available free of charge from the address on the bottom of the back cover. THE WORLD BANK Headquarters: U 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Telex: WUI 64145 WORLDBANK RCA 248423 WORLDBK Cable address: INTBAFRAD WASHINGTONDC European Office: 66, avenue d'Ina 75116 Paris, France Telephone: (1) 723-54.21 Telex: 842-620628 Tokyo Office: Kokusai Building 1-1, Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (03) 214-5001 Telex: 781-26838 The full range of World Bank publications, both free and for sale, is described in the Catalog of Publications; the continuing research program is outlined in Abstracts of Current Studies. The most recent edition of -each is available without charge from: PUBLICATIONS SALES UNIT THE WORLD BANK 1818 H STREET, N.W. WASHINGTON, D.C. 20433 U.S.A. ISSN 0253-2131
Группа Всемирного банка · Journal Article
An empirical test of the infant industry argument
Открыть оригинал документа
Полный текст размещён на сайте публикующей организации. lawenc.com индексирует метаданные и ведёт на официальный источник.
Полный текст
Основные сведения
Организация
Группа Всемирного банка
Тип документа
Journal Article
Страна
Турция
Источник
Всемирный банк