Industrialization and manufacturers' inventories in India

SW P 1 47 This paper is prepared for staff use and is not for publication. The views expressed are those of the author and not necessarily those of the Bank. INTERNATIONAL BANK FOR RECONSTRUCTION AN.D DEVILDPHENT IThT.aLTIOXAL DEVEIOPNENT ASSDCIATION Development Economics Staff Working Paper No. 147 March 1973 Induatrialization and Manufacturers' Inventories in India This paper examines the influence of structural changes in manufacturing activity which occur during the industriali- zation process, on the relative size and composition of in- ventory, by the stages of fabrication. The manufacturing activity is divided into four pairs of industrial groups, according to the durability of products, end use of products, origin of supply of purchased materials, and the manufacturing procbss. A simple accelerator model is used to test several hypotheses. The data, annual and mostly undeflated, relate to the large scale manufacturing sector of Inaia, 1946 (or 1950) -1963. Value added is found to be a good proxy for output in forecasting inventory investment for a long-term plan where detailed output targets are not available. The industrial groupings help to explain the almost unchanging composition of total inventory for the entire manufacturing sector. A negative trend in evident in the purchased materials inven- tory relations of fast growing industrial groups, which also possess higher inventory-output coefficie1ts. This paper indicates the significance of planning for inventories in nedium- and long-term development plans. Prepared by Vinod Prskash* Industry Division Demelopunt Econcmics Department *The bulk of statistical work for thi:. paper was com.pleted before the author Joined the Bank.  Contents Srmary. . . . . . . . . . . . . . I. IntroductionI . . . . . . . . . . . . . . . . . . . . . . . II. Fonmulation of the 1Hypotheses S>mbols . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Econcmietric Problems .... . . . . . . . . . . . . .. 3 Hypotheses ..... . . . . . . . ........... . 4 III. Significance and Limitations of Industrial Groups . . . . . 6 IV. Results Determinants .............. . 11 Industrial Groups ..... . . . . . . . . . . . . . . . . 14 Table I - Growth Rates and Percentage Shares of Industrial Orcups in Value Added and Total Inventory . . . . 15 Table 2 - Selected Regression Results . . . . . . . . . . . 16 Table 3 - Inventory Output/Value Added Ratios and Inventory Composition . . . . . . . . . . . . 17 Appendix - Results of Homogeneity Tests . . . . . . . . . . 21 Selected References . . . . . . . . . . . . . .. . .. .. 23  Sumrary and Conclusions i. Several structural changes in the manufacturing activity of a developing econoimy occur during the industrialization process. The manufacturing sector is classified here into four pairs of industrial groups, according to the duraoilizy of products, end use of products, origin of supply of purchased materials, and the manufacturing process. A simple accelerator model is used to detent ne the iniluence of various factors on the relative size and composition of inventory by the stages of febrication. Several h.potheses are tested using the ainnual, mostly undeflated, data relating to large scale manuiacturing sector of India, 1946 (or 1950) - 1963. rne time-dependence of errors, multicollinearity, and heteroscedasticisy are considered; and the simultaneity and errors ir variables a'e by-passed. ii. The present analysis leads to many interesting results though ;.t times their implications challenge comprehension. Output and sales are found to be most important determinants of inventory level, as could be expected. They are closely followed by value added. Thus value added is likely to be a good Proxy for output in forecasting inventory invesument for a long-tenn plan -whcre detailed output targets are not available. Of the two long-term factors -- secular trend, and average establishnert size -- the former was found much mole effective. This contrast may be somewhat deceptive due -o the limitacions of data. Among the two short-term factors -- interest rate, and price index -- the former obtained only a few significant coefficients but with a perverse positive sign. The coefficients of the latter were usuaily in-ignificant, and the sign of significant coefficients were positive for the finished products and negative for the purchased materials inventories. iii. Analysis of four types of industrial groups shows that the fast growing groups -- durables, capital goods, mineral-based, and continuous process indus:ries -- possess higher marginal total inventory-output coefficients Than the slow growing ones -- non-durables, consumer goods, agro-based, and discontinuous process industries. Consequently, we are inclined to infer that relatively more resources for inventory investment will be needed by the Indian manufacturing sector, if industrial priorities remain unchanged. The industrial groupings also help to explain the almost unchanging composition of the total inventory for the entire manufacturing sector. iv. The inventory composition of the groaps differs widely by inventory components. Purchased materials inventory, generally the largest component, is considerably higher in the fast growing groups than in the others. A strong negative trend in the purchased materials inventory relations is evident in the durables and mineral-based industry. The capital goods industry, otherwise sharing the common characteristics of the fast growing groups, for some reason does not show such a trend in the purchased materials inventory relations in the fourteen year period, though the  - ii - negative trend is evident from the eighteen year period. The coefficients of fziished products irventory relations generally support the hypothesis of a negative trend. Sbtilarly, the wor.--in-process inventory-output coefficients and components valida,-e the hypotheses, although the extent of observed distinctions among the industrial groups was not fully anticipated.  I. Introduction 1. The indu.strialization process is characterized by various structural changes in manufacturing activity of a developing economy. Some of these changes are revealed by tb9 changing composition of four types of industrial Product groups derived on the basis of: (a) the durability of products; durables versus non-durables, (b) the end-use of products; capital versus consumer goods, (e) the origin of purchased materials; mineral-based versus agro-based, and (d) tho process of manufacture; continuous versus diszontinuous. / The influence of the structural changes on the inventory-output (or inventory- value added) relationships in manufacturing industry is analyzed here. Further, manufacturers' inventories are distinguished by the three stages of fabrication, that is, purchased materials, work-in-process, and finished products; and the influence of these structural changes on the composition of manufactarers' inventories is analyzed in this paper. This study, however, does not consider inventories of manufactured products or industrial raw materials in the distributive or any other sectors of an economy excert manufacturing. 2. Several scholars nave worked on these aspects of the Indian economy ZJ, 14, 17, 197. Sen's study, which is probably the most comprehensive, estimates working capital requirements in India during the Third Five Year Plan. It also discusses many conceptual problems involved in the identification and measurement of inventory investment. The study conducted by Krishnamurty and Sastry is the rost recent and detailed. It is basically a disaggregated study of the Census of Manufacturing Industries, 19h6-1958, but it is often supplemented by the Reserve Bank of India's statements on finances of public limited companies, 1950-1962. Since the findings were generally inconclusi e, the authors emphasized the necessity of further research. 3. The present study is macro-econo-ic, intertemporal, and related to the large scale manufacturing sector of India,, 1950 (or 1946) to We can visualize many more tiypes of industrial groups,' azd in any type we ca8 hare more than two groups. Practical consideratio m led to the election of the abovenentioned classification. Any such grouping of industries mlli, however, represent the reality only approximately, especially when the original industrial classification has aggregated industries like chemicals and chemical products, general and electrical engineering. Nonetheless, the changing ccmposition of the manufacturing sector is empirically discersiible, and many interesting in-erences can be drawm, though at times on weak statistical bases.  -2- 1963. 1/ Even the annual and unde':. .1 data for this short pe.iod involved meshing the two sources: first, su 'o.'ing aata for 29 industries f;r the period 1946-1958, and second, supplying data for moro than 200 industries covering the period 1959-1963. Tgis a.nalysis uses the simple accelerator hypoth;ses, 2/ where inventory total c: anyone of its three components by the stages of fabricationi is a cidpendent variable, and there are seven explanatory variables (though seldom more than three ',i any single hypothesis). The manufacturing sector is disaz-gcgated i.lto the above- mentioned four pairs of industrial groups, arnd the data were usually undef.ated. j The first four years (1946-1949) representing the pre-planning period were often relegated due to more frequent failures in the homogeneity tests, non-availability of work-in process inventory data, and high non-response in the first two years. The loss of precious four degrees of freedom was the main consideration in going back to 1946. / Flexible accelerator hypotheses and ratio hypotheses, used by the present author in an earlier study /T87, do not generally lead to inore statistically significant results.  and procedures /3, 4, 9, ,37, including a rule of thumb, have been rigorously followed, though it shculud e recognized t-.,at econometric theory is still not well advanced to treat these statistical proble-ms simultaneously, particularly when there are only a few time-series observations. IIypotheses 5. The acceleration principle in its simDlest form in the present context postulates that the manufacturcrs succeed in maintainir,g their inventory total or its components, at an equilibrium level Y* which is linearly related to actual output level X1 or sales Xo in an. industry. (1) Y* = bo + b, (XO or XI ) e, b1> 0 (2) Y* = bo + b1 Xh + e ,b>