Sucro-Econ Plan SCI Vol. 20. No. 3. pp 135-143, 1986 0038~l21/86 $3.00 + .I0

Pnnted in Great Bntam 0 1986 Pergamon Journals Ltd.

INAPPROPRIATE INDUSTRIES AND INEFFICIENT RESOURCE-USE IN BANGLADESH: SOME EVIDENCE

FROM INPUT-OUTPUT ANALYSIS

MOHAMMAD ALAUDDIN and C. A. TISDELL Department of Economics, University of Newcastle, NSW 2308, Australia

(Received 14 September 1985)

Abstract-A 47-sector input-output table was used to consider the appropriateness of Bangladeshi industries given their labour-capital requirements and relative availability of labour and capital in the economy. We examined the efficiency of industries in terms of needed production to satisfy final demand given relative factor intensities. The findings suggest that agricultural industries are most appropriate to Bangladesh. A case, in terms of attaining the highest returns on capital, can be made for directing capital to agricultural and agroindustries as well as for avoiding (subject to qualifications) construction and energy industries. It seems that a number of urban located industries with inappropriate factor proportions have been developed. There is a case on productivity grounds for expanding Bangladeshi investment in industries based on living resources.

1. INTRODUCr’ION

This paper uses the latest available 47-sector input- output table [I] to consider the appropriateness of Bangladeshi industries given their labour/capital re- quirements and the relative availability of labour and capital in Bangladesh [cf. 21. Bangladesh, like all LDCs, is characterized by a low ratio of available capital to labour. For Bangladesh to maintain or expand indus- tries requiring higher ratios of capital to labour than available overall in the economy can add to unem- ployment or underemployment and reduce output available to satisfy final demand. Furthermore, where industries have similar relative factor intensities some may be more efficient than others in supplying pro- duction to satisfy final demand. This aspect is also ex- amined for Bangladesh.

The analysis is based upon a Leontief open input- output model. Basically, it is assumed that the tech- nology of each industry is such that inputs are required in fixed proportions and that constant returns to scale prevail. These are limiting assumptions but may hold at least for small changes in the economy. The usual limitations of static input-output analysis apply and one has to be cautious in interpreting the results. Nev- ertheless, it is still possible to obtain significant pointers from an input-output overview which takes account of production interdependence in the economy as a whole.

The static Leontief model underlying this paper is summarized in the Appendix. The model assumes two primary inputs (labour and capital) and allows for in- terdependencies between industries in production. The measurement of the effect of a change in sectoral final demand should take into account both its direct and indirect impact. Sometimes the indirect input require- ments for the expansion of an industry exceed the extra input required directly in the industry. It is therefore important not to neglect indirect effects.

The paper proceeds first of all by discussing the data used for determining output, capital and labour re- quirements (both direct and indirect) for a unit ex- pansion in final demand of each of the Bangladeshi industries specified in the 1976-77 input-output table.

It then indicates how capital and labour availability have been estimated for Bangladesh. This is followed by the presentation of empirical results. After discussing the degree of dependence of each of the industries or sectors on inputs from outside the industry or sector, the factor-intensities of different industries and their efficiency in producing a unit of final demand are out- lined. In comparing these intensities with ratios of available labour and capital, industries with inappro- priate factor proportions can be identified. Further- more, industries that appear to be relatively inefficient in supplying final demand can be identified.

2. THE BASIC DATA

For the purpose of the present analysis the latest available 47-sector input-output table [I] has been used. It relates to 1976-77 and provides data on sectoral output and capital coefficient vectors. The methods used to estimate the capital coefficients are discussed in [3]. By multiplying sectoral outputs by the capital coefficients, we obtain the estimates of the amount of capital used in each of the sectors. These estimates of the quantity of capital used in each of the sectors are set out in Table 1 together with output levels. The aggregate amount of capital stock employed in the economy in 1976-77 is estimated to be Taka 223,658 million.

The employment figures used to derive sectoral la- bour coefficients were estimated by Alauddin [4] from a number of sources. This was necessary because es- timates did not accompany the 1976-77 input-output table. Without going into the controversies surrounding conceptual and measurement issues, labour employ- ment is defined to include only those persons who were actually engaged in various sectors of the economy. Labour requirements for the agricultural crop sectors were estimated by using per acre labour requirement data for various crops from [5]. This required a con- sideration of distribution of acreage by variety of crops and use of irrigation by techniques, and details of these variables were obtained from [6]. Estimation of labour requirements for livestock, forestry and fishery sectors posed special problems since information was not

SFPI 2c:i.s 135 ,

136 M. ALAUDDIN and C. A. TISDELL

Table I. Output, capital stock and employment by sector in Bangladesh, 1976- 1977

x sector Name Rice 01

02 03 04

05

06 07

08

09

10

11

12

13

14

15 16

17

16

19

20

21

22

23 24

25

26

27

28

29

30

31

32

33

34 35

36 37

38

39

40

41

42

43

44

45

46

47

Wheat

Jute

Cot t “11 ‘Tea Other crops Livestock

Fisheric!s

Forestry

sugar

Edible oil

Salt

Tobacco

Other food Cotton yarn

Cloth: millmadr

Cloth' handloom

.Jute textile

Paper Leather

Fertilizer

Pharmaceuticals

Other chemicals Cement

Basic metals

Metal products

Machinery

Transport equipment Wood

Miscellaneous industries

Urban hous.ebuildinE Rural housebuilding Non-residential building Construction: elec. & gas Construction: transport other construction Petroleum Electricity Gas Transport service Trade service Housing service Health Education Public administration Banking and insurance Other services

Code Output Capital Employment used (million Taka) (000 man y r‘s) RICE-~-- 33700.1 32470.143 6910.272 WHEAT- JUTE----

COTTON- TEA-----

0'rHRCHDI'

LIVSTOCK

FISHRIES

FORESTRY

SUtiAK

EDlBLOIL SALT ~~~

TOBACCO-

O'THRFOUL) COTTNYRN

XILLCLTH

HANDCLTH

JUTETXTI. PAPER-- -

LEATHER FERTILZR

PHRMACEU

OTHRCHEM

CEMENT BSICMETL

METLPROD

MACHINRY

TRANSEQI'

WOOD--

MISCINDS

URBHOUSE

RURHOUSE

NONRESBL

CONSELtiS

CONSTRAN

OTHRCONS

PETRLEUM

ELCTRCTY GAS__-_-

TRANSERV TRADSERV

HOUSSERV HEALTH--

EDUCATN-

PUBLICAD

BANKINSR

OTHRSERV

fi94. I 684.668 fi5.508

3218.6 1373.377 732.512

1.1 0.394 0.366 650.0 586.430 24.48(1

11197.4 1700.885 827.966 8941.5 939.752 1143.560 9232.7 fi670. 'J7fi 631.960 1956.5 484.038 111.872 34JO.I 5T.l.fj02 104 ,983 1221.3 157.792 23.511 649.6 23.789 74.812

2379.1 97.305 20.965 2681.6 106.996 110.523 1562.5 1092.188 44.905 725.1 422.153 16.885

4812.0 1275.180 200.000 2995.5 2814.572 184.062 604.1 833.054 8.992

3400.3 775.608 70.780 893.5 2256.981 5.042 870.2 1194.089 Il.475 1492.5 2048.009 29.659 .4 50 9 1173.061 2 305

2319.4 1065.532 8.504 968.8 1000.770 28.249 912.9 1129.075 14.071 173 I 454.838 8.920 894.1 132.237 43.076

5059,:i 3930.064 155.031 2575.4 217.106 74.840 2640.9 101.411 67.720 1296.4 153.105 37.120 1183.1 4 6 fi 2 5 45.760 1209.2 52 358 53.360 986.O 25.537 121.200

2026.9 1185 736 0.456 555.5 5322.134 4.663 162.4 1268.376 5.137

7052.3 22506.709 400.000 10039 7 15597.678 990.000

10219.7 85163.820 1196.900

1253.6 1959 610 103.100

2191.3 3307.548 281.100

3835.4 5774.194 369.800

1578.4 2776.406 70.000 4974.1 10534.646 659.000

Total : <2365_1.750 16095.124

Source: As mentioned in the text.

available from the previously mentioned sources. The 1976-77 employment figures for these sectors were es- timated by taking the 1975-76 estimates of Clay and Khan [7, p. 211 and projecting these forward using FAO [8] growth rates.

The derivation of employment figures for the sectors within the nonagricultural complex involved approx- imation and arbitrariness to a certain degree. This was unavoidable because employment data according to the sector classification scheme used in the input-out- put table were not directly available from any official source. Various bodies [such as 9. Table 6.2; 10 p. 4431 provide employment data on the basis of broad cate- gories of economic activity. e.g. industry (large and medium; small and cottage), construction, services. transport, finance and banking, etc. Industrial em- ployment was estimated to be in the vicinity of 1.2 million man-years [9; 1 1, p. IO]. It consists of em- ployment in large and medium firms (about 0.4 mil-

lion); in handloom industries (0.2 million); and in small and cottage industries (about 0.6 million). The Census of Manufacturing Industries employment figures (for large and medium) available from Bangladesh Bureau of Statistics [6, p. 169 and pp. 324 ff.] were used as they broadly conform with the sector classification of the input-output table. Small and cottage industry fig- ures were allocated to different industries according to the proportion of their value added except for those industries where small and cottage industry category was known to be nonexistent. These included sectors such as fertilizer, petroleum products, cement. and ba- sic metal [I. pp. 35-361. For mill-made cloth, jute tex- tiles. cotton yarn. no small and cottage component was added as it was included in the handloom sector. For the other nonagricultural sectors, construction (0.4 million): trade (0.99 million); finance and banking (0.07 million); electricity and gas (.O I million); trans- port (0.4 million): services (2.6 I million): employment

Inappropriate industries and inefficient resource-use 137

Table 2. Direct requirements of output, capital and labour as percentage of total (direct and indirect) averaged by broad sector category in Bangladesh, 1976-77*

sector category output Capital Labour

Agriculture 83.18 ( 9.04) 76.97

crop 79.67 ( 8.86) 74.29

Non-Crop 90.21 ( 4.39) 82.31

7.64) 92.98 ( 5.34)

7.47) 91.57 (5.99)

5.43) 95.79 ( 2.48)

-

Industry 69.05 (10.67) 63.92 ( 9.98) 67.83 (16.44)

Agro-Inds 73.72 (10.64) 59.35 ( 7.84) 72.48 (20.24)

Others 63.90 ( 8.45) 68.95 ( 9.97) 62.70 ( 9.46)

Construction and Energy 52.93 (18.43) 47.22 (26.27) 57.54 (24.61)

Services 80.92 (12.32) 87.95 (10.47) 91.70 ( 6.58)

ALL SECTORS 70.09 (16.31) 66.45 (19.44) 74.01 (21.00)

*Standard deviations in parentheses

Note: Agriculture (Sectors 01-09, Crop 01-06. Non-Crop 07-09) Industry (Sectors 10-30. Agro-Inds lo-18,20.29.

Others 19.21-28.30) Construction and Energy (Sectors 31-39) Services (Sectors 40-47)

Table3.Capital-labourratios(directandindirectrequirements)forfinaldemandbysectorsoftheBan~adesh economy 1976-77

t Sector K/L * Sector K/L Ratio Ratio

38 ELCTRCTY 0.2339 13 TOBACCO- 0.0176 39 GAS----- 0.2049 41 TRADSERV 0.0174 37 PETRLEUM 0.0945 47 oTHRSERV 0.0162 24 CEMENT-- 0.0774 16 MILLCLTH 0.0159 21 FERTILZR 0.0715 08 FISHRIES 0.0146 42 HOUSSERV 0.0666 02 WHEAT--m 0.0136 40 TRANSERV 0.0605 18 JUTETXTL 0.0126 IY PAPER--- 0.0558 44 EDUCATN- 0.0126 25 BSICMETL 0.0507 14 OTHRFOOD 0.0120 27 MACHINRY 0.0438 17 HANDCLTH 0.0115 22 PHRMACEU 0.0387 11 EDIRI.OIl, 0.0113 26 METLPROD 0.0373 20 LEATHER- 0.0106 28 TRANSEUI' 0.0369 10 SUGAR--m 0.0103 46 BANKINSR 0.0366 29 WOOD------ 0.0088 30 MISCINDS 0.0299 36 OTHRCONS 0.0080 23 OTHRCHEM 0.0298 09 FORESTRY 0.0078 05 TEA-m--- 0.0291 15 COTTNYRN 0.0076 33 NONRESBL 0.0275 06 OTHRCROP 0.0069 31 URBHOUSE 0.0255 01 RICE---- 0.0059 35 CONSTRAN 0.0251 12 SALT---- 0.0052 34 CONSELGS 0.0250 03 JUTE---- 0.0043 43 HEALTH--- 0.0236 04 COTTON-- 0.0035 45 PU81.1CAD O.OlY6 07 LIVSTOCK 0.0029 32 RURHOUSE 0.0188

* Capital-(employed) Labour Ratio 0.0139 ** Capital-(available)Labour Ratio O.OOQZa *** Capital-(available)Labour Ratio o .oo8ob

a Using conversion factors. b Without using conversion factors.

138 M. ALAUDDIN and C. A. TISDELL

figures were allocated for different component sectors according to the proportion of their value added in the broad category of economic activity to which they be- longed. The amount of labour employed in each sector is set out in Table I. The total amount of labour ac- tually employed in the Bangladesh economy in 1976- 77 is estimated to be 16.095 million man-years.

The total available supply of labour exceeds that which is actually employed. Bangladesh population figures for 1976-77 [ 12, p. 601 have been utilized to estimate total available labour supply. The size of the civilian labour force has been calculated assuming a 33.9 percent participation rate [7, p. 91. The figure so derived has been converted into standard units on the basis of the age and sex composition of the labour force in the 1974 census [ 12, p. 851. A conversion factor of 0.50 has been used for labourers in the lo-14 and over

55 age groups. Female workers have been converted into standard man-units using a conversion factor of 0.75. Taking all these into account the total available labour for 1976-77 is estimated to be about 24.4 1 mil- lion man-years. It should be noted that other available estimates (e.g., [7, 91) do not take into account the age and sex composition of the labour force. Despite ar- bitrariness in our assumption concerning the conver- sion into standard man-units, this aspect cannot be ignored. If no conversion factor is used, the total amount of available labour in 1976-77 is (using Clay and Khan’s suggested participation rate of 33.9 percent) 28.03 million man-years. Undoubtedly it is difficult to estimate precisely the total amount of labour in the economy willing to work. The two estimates could provide the lower and upper limits of the actual labour available for work. However, available labour is con-

,

:

TOTAL LABOUR (MAN - YEARS )

i : ZJ 0

t . Livstock

.Jute cottrx

. paper

l Cement

. Petrleum

. Gas

. Elctrcty , 1 I .

Fig. I. Total labour (man years) and capital (million of taka) required per million Taka of final demand, by sectors of the Bangladesh economy, 1976-77. The slope of line OB represents the ratio of unemployed labour to capital and the slope ofthe line OA represents

the ratio of the available labour to capital.

Inappropriate industries and inefficient resource-use 139

siderably in excess of that actually employed. Our es- timates using conversion factors indicate involuntary employment or underemployment equal to at least one-third of the total available labour force but this rises to over 40 percent if conversion-units are not used. The basic argument is not affected by the precise pro- portion of the labour force unemployed or underem- ployed. Clearly, some industries are inappropriate or inefficient anyway, given the labour-capital availability in Bangladesh, and these industries can be identified.

3. EMPIRICAL OBSERVATIONS

Input-output analysis is well placed as a technique to highlight production interdependencies between sectors of an economy, and the 1976-77 input-output table for Bangladesh indicates that some sectors depend heavily on resources from other sectors whereas others show little dependence on other sectors. As Table 2 indicates, industries based directly on living resources (agriculture, fisheries and forestry) and the service in- dustries show little relative dependence on other in- dustries for inputs in producing output to satisfy final demand. They also require relatively little employment of labour and capital in other sectors. On the other hand, the construction and energy industries are rel- atively heavily dependent on other sectors for inputs whereas manufacturing industries show an interme- diate degree of dependence. However, sectoral linkages or dependence (or independence) of sectors are not the main concern in this paper.

The main purpose of this paper is to consider the appropriateness of various industries to Bangladesh in the light of capital-labour ratios required in the in- dustries and the availability of capital and labour in the economy. In addition, the relative efficiency of various sectors are discussed.

Table 3 sets out the estimated capital-labour ratios of the sectors of the Bangladesh economy in descending order of capital-labour requirement per unit of direct and indirect requirements. It is very difficult to deter- mine precisely which industries are inappropriate in terms of their capital-labour ratios, however, the ratio of capital to employed labour in the economy is esti- mated to be 0.0139. Industries requiring capital and labour in a higher ratio than this certainly appear to be inappropriate. However, available labour exceeds that employed. If such labour happened to be em- ployed, the capital-labour ratio would be 0.0092 using conversion factors or 0.0080 if conversion-units are not used. In any case sectors such as electricity, gas, petroleum, cement, fertilizer, housing and transport services, paper and basic metals involve higher capital- labour ratios in terms of techniques used than seem appropriate to factor availability in the economy [ 131.

In Fig. I, inappropriate industries are those for which the slope of the ray from the origin to their (capital, labour) combination is less than the slope of the line OB, which represents the ratio of the employed labour and capital in the economy. However in terms of la- bour-availability (using lower limit) all of those indus- tries falling below the line OA are less than appropriate. Nevertheless, given unavoidable problems in the qual- ity of the data it may be wise to take a conservative view and only regard those industries requiring quite high capital-labour ratios as being inappropriate.

It is useful to classify the sectors according to broad categories and consider the capital-labour ratios for these broad categories. The results for broad categories are shown in Table 4. These figures are derived using ratios in Table 3 and are unweighted averages. The aggregate averages are presented in Table 6. The table indicates that capital-labour ratios are lowest in agri-

Table 4. Average capital-labour ratios by broad sector category for Bangladesh, 1976-77

sector Category Capital-Labaur Ratio (Million Taka per man-year) Mean St. Deviation

Agriculture ,010 ,008 crop ,011 ,010

Non-Crop ,008 .006

Industry ,028 ,022

Agro-Inds ,011 ,003

Others ,047 ,017

Construction and Energy ,074 ,066

services ,032 ,021

ALL SECTORS .034 .045

Note: Agriculture (Sectors 01-09. Crop 01-06. Non-Crop 07.-09) Industry (Sectors 10-30. Agro-Inds 10-18.20.29.

Others 19,21-28.30) Construction and Energy (Sectors 31-39) Services (Sectors 40-47)

140 M. ALAUDDIN and C. A. TISDELL

culture and highest in construction and energy. Broadly speaking agricultural sectors and agroindustries are most appropriate to Bangladesh’s resource availability.

The efficiency of industries when one proceeds at an aggregative level is difficult to determine. Neverthe- less, one can identify industries in Bangladesh which are more efficient than others in satisfying a unit of final demand, account being taken of direct and in- direct effects. In Fig. I, for example, the capital and the labour requirements (direct plus indirect) of the various sectors per unit of final demand satisfied are shown. Any industry located northeast of another is relatively inefficient. The most efficient industries if one follows Farrell’s approach 1141 would be those falling on the apparent envelope of points closest to the axes in Fig. 1. Industries such as livestock, salt, forestry, tobacco, basic metal and gas would seem to be located on such an efficiency “isoquant”. Of course, Farrell’s problem is rather different in that he is con- cerned with frontier production functions and effi-

ciency within industries [cf. 141. In the present context, the quality of the data is such that we cannot categor- ically identify the most efficient industries in satisfying final demand. However. we can with a fair degree of confidence identify some that are relatively inefficient.

It can be seen from Fig. 1 that electricity production is relatively inefficient compared to gas production. Local natural gas is the source of most gas production. The jute textile industry is relatively inefficient com- pared to the hand cloth and mill cloth industries which are less efficient than the leather industry even though all of these industries employ labour and capital in proportions that are roughly appropriate to resource availability in Bangladesh. The wood working industry is even more efficient again. The relative inefficiency of the jute textile industry may be a reflection of the fact that it is nationalized and x-inefficiency occurs [ 161. Industries such as fisheries and forestry production seem relatively efficient users of capital and labour. In agricultural sectors requiring high labour to capital ra-

Table 5. Total (direct and indirect) capital and labour requirements per million Taka of final demand, Bangladesh, 1976-77

It Sector Capital Labour

Name (million Taka) (man-years) 38 ELCTRCTY 14.7699 63.137 42 HUUSSEf(V 8.5894 129.065 39 GAS-mm-- 8.1783 39.909 37 PETRLEUM 7.2516 76.741 24 CEMENT-m 5.9033 76.279 21 FERTILZR 4.7148 fi6.780 40 TRANSERV 4.5351 74.995 19 PAPER- ~- 4.2063 75.391 22 PHRMACEU 3.2724 84.474 30 MlSCINDS 3.0339 101.543 23 OTHRCHEM 2.9549 99.317 43 HEALTH- 2.9196 123.573 16 MILLCLTH 2.8768 180.704 27 MACHINRY 2.8441 64.985 18 JUTETXTL 2.6119 222.320 35 CONSTRAN 2.7064 107.618 33 NONRESBL 2.692G 97.747 31 URBHOUSE 2.4916 97.620 45 PUBLICAD 2.4498 125.236 26 MI:TL.f'RGD 2.4059 64.437 46 BANKINSR 2.2894 62.614 28 TRANSEQP 2.2747 61.576 15 COTTNYRN 2.2611 297.166 47 OTHRSERV 2.182$1 134.526 34 CONSELGS 2.1616 86.296 17 HANDCLTH 2.1238 184.244 25 BSICMETL 1.9966 39.393 05 TEA-m--- 1.9610 67.298 41 TRADSERV 1.8735 107.702 02 WHEAT---- 1.7606 129.467 44 EDUCATN- 1.6855 133.554 32 RURHOUSE I .fiOl'J 86.326 20 LEATHER- 1.3846 130.334 01 RICE- ~~ 1.3840 236.296 08 FISHRIES 1.3702 94.137 14 OTHRFOOD 1.3588 113.087 04 COTTON-- 1.3329 375.520 36 OTHRCONS 1.2987 162.791 03 JUTE-m-- 1.1366 264.813 11 EDIBLOIL 1.1167 98.829 10 SUGAR-m- 1.0477 101.550 13 TOBACCO- 0.9218 52.249 29 WOOD---- 0.8163 92.861 06 OTHRCROP 0.7175 103.894 12 SALT---- 0.5681 110.106 09 FORESTRY 0.5336 68.290 07 LIVSTOCK 0.4785 166.182

Inappropriate industries and inefficient resource-use 141

tios, cotton production appears least efficient. Jute, rice, livestock, other crops and forestry are more efficient. These industries show an ascending order of efficiency in using labour and capital. Other crude comparisons of efficiency can be made from Fig. 1.

4. POSSIBLE POLICY IMPLICATIONS AND LIMITATIONS OF THE ANALYSIS

The analysis suggests that certain industries are rel-

atively efficient (e.g. fisheries). This indicates that there

is a case for expanding these industries. Also, it seems that most agricultural industries are quite appropriate to Bangladesh in terms of their capital to labour re- quirements and many are relatively efficient.

If one considers capital to be the factor that is ba- sically in short supply in Bangladesh (the limiting factor to production) then it is instructive to consider the capital requirements (direct and indirect) of the sectors of the Bangladesh economy for fulfilling one unit of final demand. Table 5 shows the sectors in terms of descending capital requirements. Table 6 shows capital requirements by sector groupings. Note that agricul- tural industries have a low capital requirement per unit of final demand satisfied. It could be argued that these industries should be given priority in investment.

It might be noted that if capital (K) required per unit of final demand is plotted against the capital/la- bour ratio (K/L) for each sector, that the scatter dia- gram in Fig. 2 emerges. A linear regression of the two variables gives the best fit line as

K = 1.0182 + 52.1724 (K/L).

(R* = 0.8403, t-value = 15.387)

As the K/L ratios of industries rise, there is a tendency for the amount of capital required per unit of final demand to increase. This would imply that industries with low K/L ratios tend (in terms of final demand satisfied) to give a higher per unit return on capital invested than those sectors with high K/L ratios. Thus in terms of attaining the highest returns on capital, a case can be made for directing the limited amount of capital towards industries with low K/L requirements per unit of final demand. From Table 6 these can be seen to the “agricultural” industries and agroindustries. It can also be seen that construction and energy in- dustries ought on this basis be avoided as investment outlets, except to the extent necessary to support fa- voured industries.

However, there are a number of limitations of the analysis to be kept in mind. We do not live in a two- factor world. All labour is not necessarily homogeneous and easily substituted. Furthermore, land or natural resources are important inputs in production from liv- ing resources. Expansion in agricultural, fishing and forestry industries may therefore be subject to decreas- ing returns. Despite this, at least over a small range there seems to be a case for expanding these industries.

Again, production in a particular sector either may be by a variety of techniques or by machinery of dif- ferent vintages. Much capital is already embodied. It is possible that only new investment is free. Conse- quently. incremental output-capital ratios or capital requirements may differ from those estimated from existing capital stock. These limitations are inherent in input-output analysis and are a further reason for caution.

Table 6. Total (direct and indirect) capital and labour requirements per million Taka of final demand

by broad sector category, Bangladesh, 1976-77*

sector category Capital

(million Taka)

Labour

(man-years)

K/L

Ratio

Agriculture

crop

Non- Crop

1.166

1.382

0.794

_

0.521) 167.32 (105.02) 0.0071

0.443) 196.22 (116.67) 0.0070

0.500) 119.54 ( 50.73) 0.0066

Industry 2.426 (1.363) 110.36 ( 62.60) 0.0220

Agro- Inds 1.572 (0.811) 143.95 ( 70.52) 0.0109

others 3.367 (1.233) 73.42 ( 18.53) 0.0459

Construction

and Energy 4.798 (4.475) 90.97 ( 33.82) 0.0527

Services 3.316 (2.308) 111.41 ( 27.77) 0.0298

ALL SECTORS 2.794 (2.554) 117.77 ( 67.39) 0 _ 0237

* Standard deviation in parentheses

Note: Agriculture (Sectors 01-09. Crop 01-06. Non-Crop 07-09)

industry (Sectors 10-30. Agro-Inds 10~18.20.29.

Others 19.21-28.30)

Construction and Energy (Sectors 31-39)

Services (Sectors 40-47)

142 M. ALAUDDIN and C. A. TLSDELL

‘r .

1 I I

0.05 0.10 0.15 0.20

Capital - Labour Ratio C million taka I man year)

0.2 ‘5

Fig. 2. Relationship between the capital-labour ratios ofsectors ofthe Bangladesh economy indicated in Table 3 and their capital requirements (both direct and indirect) per million

taka of final demand for 1976-77.

The analysis, despite its limitations, indicates sectors

that might be given preference in Bangladesh for ex- pansion in terms of the appropriateness of their K/L

ratio and their efficiency. It seems as though there has been urban bias in the past; that is, a number of urban- located industries with inappropriate factor proportions (some of which are relatively inefficient) have devel- oped [cf. 171. There seems to be a case on productivity grounds for expanding investment in several industries based on living resources. On the other hand, it could be argued that measures which encourage urbanization and modernization help to reduce population growth rates and should be preferred from the point of view of long-run strategy [ 18, 191. For a contrary view see [20]. A further complication is that rural “bias” if it is associated with greater trade in primary products may bring about structural dependence, economic risks and problems of sustainability [cf. 2 I-231.

1.

2.

3.

4.

5.

6.

REFERENCES

Bangladesh Planning Commission, The Structure of the Bangladesh Economy: An Input-Output Analysis Back- ground papers of the Second Five Year Plan of Bangla- desh, Volume I, Dhaka (1980). R. S. Eckaus. The factor proportions problem in under- developed areas. .4m. Econ. Rev 45, 539-565 (1955). Bangladesh Planning Commission, Estimalion of Gross Fixed Capital formation in Bangladesh: Methodology and A Preliminary Eslimate jtir 1976-77 Background Papers of the Second Five Year Plan of Bangladesh, Volume II, Dhaka ( 1980). M. Alauddin, Identification of key sectors in the Bangla- desh economy: A linkage analysis approach. Applied Economics (forthcoming). Bangladesh Planning Commission, A Study on All Agri- cultural Crops. Second Plan/Perspective Plan Study Re- port Series (undated). Bangladesh Bureau of Statistics. Statistical Year Book of Bangladesh Bangladesh Bureau of Statistics, Dhaka (1982).

7.

8.

9.

10.

1 I.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

E. J. Clay and M. S. Khan, Agricultural employment and underemployment in Bangladesh. Agricultural Economics and Rural Social Science Papers (4), BARC, Dhaka (1977). FAD, Working Paper on Agricultural Employment and Appropriate Technology. Perspective Study of Agricul- tural Development in Bangladesh Policy Analysis Divi- sion, FAO, Rome (1975). Bangladesh Planning Commission, The Second Five Year Plan: 1980-I 985, Dhaka (1980). Bangladesh Bureau of Statistics, Bangladesh Population Census 1974, National Volume Bangladesh Bureau of Statistics, Dacca ( 1977).

S. A. L. Reza, Trade, output and employment: A case study of Bangladesh. Bangladesh Development S’ludles 6, 1-26 (1978). Bangladesh Bureau of Statistics, Statistical Year Book qf Banglade.th 1979 Bangladesh Bureau of Statistics, Dacca ( 1979). A. R. Khan, Capital intensity and ethciency of factor use. Pakistan Developmenl Review 10, 232-263 (1970). M. J. Farrell, The measurement of productive efficiency. J. R. Slatis. Sot. 120, 253-281 (1957). F. Forsund and L. Hjalmarsson, Technical progress and structural change in Swedish cement industry I955- 1979. Economefrica 51, 1449-1467 (1983). J. Harris and M. Moore (eds.), Development and the ru- ral-urban divide. Special Issue J. Development Studies 20, I-165 (1984). M. G. Kibria and C. A. Tisdell, Operating capital patterns in jute weaving in Bangladesh. J. Deve/opmmt Economics 18, 133-152 (1985). J. Bongaarts, A framework of analyzing the proximate determinants of fertility. Populawn and Development Reviena 4, 220-234 (1978). K. Davis and J. Blake. Social structure and fertility: An analytical framework. Economics Deve/opment and Cub tural Change 4, 62-74 (I 956). M. Nag, How modernization can also increase fertility. Current Anthropology 21, 571-587 (1980). Economic Commission for Latin America, Economic Survey oflalin America I949 United Nations, New York (1951).

Inappropriate industries and inefficient resource-use 143

22. R. Prebisch, Commercial policy in the underdeveloped countries. Am. Econ. Rev. 49,25 l-273 (I 959).

23. C. A. Tisdell and T. I. Fairbaim, Subsistence economies and unsustainable development and trade: Some simple theory. J. Development Studies 20, 227-241 (1984).

24. C. S. Ahammed and R. W. Herdt, Measuring the impact of consumption linkages on the employment effects of mechanization in Philippine rice production. J. Devel- opment Studies 20, 242-255( 1984).

25. R. Krishna, Measurement of the direct and indirect em- ployment effects of agricultural growth with technical change. In: Agriculture in Development Theory (Edited by L. G. Reynolds). Yale University Press, New Haven pp. 297-326 (1975).

APPENDIX: THE INPUT-OUTPUT

FRAMEWORK USED

Input-output analysis provides a formulation of general equilibrium in terms of structural interdependencies of a net- work of intermediate transactions as well as final products to consumers. The measurement of any effect of change in the sectoral final demand should take into account both its direct as well as indirect impact (see, e.g. [24 and 251). The indirect effects of any change in one sector can be traced in that sector itself as well as in other sectors in the economy. We use the open static Leontief model to determine the total (direct and indirect) output and input (labour and capital) requirements to satisfy a unit increase in sectoral final demand. The basic model is as follows:

(a) Total output requirements

Define Q = [I -A]-’ = [Q,] (I)

where A is the matrix of technical coefficients. Then G Q,, ,

measures total (direct and indirect) output required to satisfy a unit increase in the final demand for the output of sector j.

Now Q = [I - A]-’ = [I + A + AZ + . . . + A”]

= [I + A] + i: A*. (2) 8=2

The first term in (2) refers to the direct effect, while the second to the indirect effect, Thus for the jth sector, direct effect (say a;) of a unit change in final demand is given by

a,=[1 + Ca,l (3)

where a, is an element of A.

(b) Total capital requirements

Define K = [k][I - A]-’ = [K,j] (4)

where k is the diagonalized matrix of capital coefficients. Then C K,, measures total (direct and indirect) capital required to

sustain a unit increase in the final demand for the output of sector j. For the jth sector, direct capital requirement (say @,) of a unit increase in final demand is given by

0, = ]k, + C &I. (5)

(c) Total labour requirements

Define L = [[][I - A]-’ = [L,,] (6)

where i is the diagonalized matrix of labour coefficients. Then C L, = total (direct and indirect) labour input re-

quired to sustain a unit increase in the final demand for the output of the jth sector. The direct requirement of labour input in this sector (say 7,) of a unit increase in final demand is given by

r, = ]I, + C La,,]. (7)

The indirect output and input requirements can be derived as residuals from their respective totals once the direct effects are known.