See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/242342219

An econometric analysis of some major manufacturing industries:

A case study

Article in Managerial Auditing Journal · August 2004

DOI: 10.1108/02686900410543895

CITATIONS READS
11 1,043

3 authors:

Mohammad Zakir Hossain M. Ishaq Bhatti

Sultan Qaboos University La Trobe University
60 PUBLICATIONS 306 CITATIONS 200 PUBLICATIONS 1,724 CITATIONS

SEE PROFILE SEE PROFILE

Mohammad Zulficar Ali

Patuakhali Science and Technology University, Patuakhali, Bangldesh
17 PUBLICATIONS 47 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Islamic Mutual Fund View project

Obeid EBS Paris, France View project

All content following this page was uploaded by Mohammad Zakir Hossain on 06 October 2015.

The user has requested enhancement of the downloaded file.

 The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at
www.emeraldinsight.com/researchregister www.emeraldinsight.com/0268-6902.htm

MAJ
19,6 An econometric analysis of some
major manufacturing industries
A case study
790
Md Zakir Hossain and M. Ishaq Bhatti
Department of Operations Management and Business Statistics,
College of Commerce and Economics, Sultan Qaboos University,
Muscat, Oman, and
Md Zulficar Ali
Department of Statistics, Shaikh Bourhanuddin College, Dhaka, Bangladesh
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

Keywords Econometrics, Functional analysis, Production processes, Labour, Capital,

Bangladesh
Abstract Recently, Bhatti and Khan considered various functional forms to see their suitability in
terms of efficiency and competency in assessing production process based on agricultural data.
This paper reviews some models recently used in the literature and selects the most suitable one for
measuring the production process of 21 major manufacturing industries in Bangladesh. In
particular, the paper estimates and tests the coefficients of the production inputs for each of the
selected manufacturing industries using Bangladesh Bureau of Statistics annual data over the
period 1982-1983 through 1991-1992.

1. Introduction
Just recently, Bhatti (2002) and Bhatti and Khan (2000) considered various functional
forms to see their suitability in terms of efficiency and competency in assessing
production process based on Bangladesh agricultural data. This paper reviews some of
the production processes being used in agriculture (see for example, Bhatti (2002, 2001),
Bhatti and Khan (2000), Khan et al. (1998), Bhatti et al. (1998), Wu and Bhatti (1994),
Hoque (1991)) and revisits in applying in manufacturing industries to see which model
is the best among those already considered.
In recent years it has been common practice to use sophisticated models to solve
simpler problems in applied economics and other social sciences in order to increase
their chances of selling the models in high tech, competitive and computing rich
environments, due regard having been given to the trade off between sophistication
and simplicity in utilizing the selected models in terms of computing time and cost
effectiveness. For example, Bhatti (1993), Baltagi (1996), Bhatti and Owen (1996), Bhatti
(1997), Bhatti et al. (1998), Hoque (1991) among others have used linear regression
models to measure the Cobb-Douglas (C-D) type production processes.
Bhatti (1993) demonstrated that a high block-wise and/or industry-wise
heteroscedasticity exists between all industries while using Bangladeshi data. Hoque
(1991) used the survey data of Bangladesh to resolve the relationship between farm size
and efficiency. Hoque estimated two C-D-type production functions both ordinary least
Managerial Auditing Journal squares (OLS) with fixed and random coefficients. Wu and Bhatti (1994) tested various
Vol. 19 No. 6, 2004
pp. 790-795
q Emerald Group Publishing Limited
0268-6902
The authors would like to thank the anonymous referees for their valuable suggestions on an
DOI 10.1108/02686900410543895 earlier version of this paper.
 functional forms, constant elasticity of substitution (CES), variable elasticity of Major
substitution (VES) (see in detail the work of Hojman (1986)) and C-D against each other
in terms of paired and recommended use of C-D-type production function, while
manufacturing
analyzing a data set from a least developed country (LDC) such as Bangladesh. Bhatti industries
and Owen (1996) used Chinese data set to gauge production process in selected Chinese
provinces in order to understand the pattern of agro-based industries’ production.
Their findings were in-line with that of Hoque’s and Wu and Bhatti’s findings. Bhatti 791
et al. (1998) measured the agricultural productivity in Shanghai region of China using a
complex econometric model, whereas Bhatti and Khan (2000) defused the idea of
complexities in econometric analysis of production functions due to the gain in terms of
computing time and cost benefits.
Bangladesh belongs to the group of LDC countries of the world, which has been
characterized by very low level of economic activities for the last few decades. This is
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

due to the fact that the vast population of Bangladesh cannot fully or even partially be
employed in different sectors of the country. As a result, the country depends
significantly on foreign aid, grants and loans for materialization of its year-to-year
development plan, education policy, family planning and other fiscal programs. Mainly
an agrarian economy, Bangladesh depends heavily on a few export items such as jute
products, fish, prawns, shrimps, hides, skins, leather and leather manufacture, tea,
ready-made garments, etc. On the other hand, it imports quite a large variety of items
both for private and government consumption as well as capital goods for its export
industries. As a result, Bangladesh has been experiencing quite a negative balance of
trade since its inception. This clearly indicates a sign of the undevelopment,
impoverishment and weaker economy of the country.
In order to overcome such an unwelcome situation of the Bangladesh economy, at
least partially, it needs to go for mass industrialization throughout the country. This is
because the development of an economy mainly depends on the industrial policy of the
country. It is important to mention here that the policy for industrialization of a
country must be well planned, well defined and well thought out. In this regard, our
view is to analyze an important production function for some major manufacturing
industries of Bangladesh through which we can get some indications about the
industrial policies or at least the nature of the production inputs used in the production
process. This is due to the fact that a production function provides a base for the
analysis of the relations between costs and volumes of output as well as serving as a
base for the theory of demand of firms for factors of production (Singh et al., 1999).
Nowadays businessmen as well as industrialists are very much concerned about the
theory of firm in order to take correct decisions regarding what items, how much and
how to produce them. Each of these decisions are directly related with the cost
considerations, markets situations where the firm is to be operated and internal
organization of the firm (Harbury, 1968). Here the factor “firm” is very important due to
the fact that it is the basic unit of production in producing goods and services such as
transporting, financing, wholesaling and retailing using the factors of production such
as labour and capital (Intriligator, 1980).
A production function generally expresses the technical relationships between
physical quantities of inputs and outputs. Thus, an analysis of a production function
helps us in making decision so far as production of goods and services is concerned. In
this paper we analyze the C-D production function because of its enormous economic
importance in making different policies in economic field. Moreover, it is the most
commonly used function for different sectors of economy of a country. As mentioned
 MAJ earlier, the annual BBS time series data have been employed to estimate the function.
In Bangladesh, the Bangladesh Bureau of Statistics (BBS) is the only source of
19,6 published data and no data can be generated through a direct method, so, we had to use
these data. Moreover, we could not use the latest time series data simply because the
BBS data are not up to date. The rest of this paper is organized as follows. Section 2
briefly outlines the theoretical concepts of the C-D production function. Results and
792 discussion have been presented in section 3. Section 4 concludes the paper.

2. The model
Due to various reasons stated above, the use of C-D model is preferred for the LDC
countries and hence we use this particular type of production function model in our
study. This model is due to C.W. Cobb and D.H. Douglas who developed a production
function in the late 1920s which is often known as C-D production function and it is
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

mathematically expressed as:

P ¼ AK a L b u: ð2:1Þ
Where, P ¼ output, K¼ capital input, L ¼ labor input and A ¼ constant, a and b are
positive parameters and u is the disturbance term. Here a and b represent the
elasticities of output with respect to capital and labor, respectively. Taking log of both
sides of equation (2.1), the function transforms to a log-linear form as:
lnP ¼ lnA þ alnK þ blnL þ lnu: ð2:2Þ
The underlying assumptions to be satisfied for the production function (2.1) and are
found in Intriligator (1980). The sum of the elasticities, a þ b, provides the returns to
scale of the firms in question. That means if a þ b ¼ 1, the production operates under
constant returns to scales, if a þ b . 1 or a þ b , 1, the production operates under
increasing or decreasing returns to scale, respectively. Often this function has the
underlying assumption of constant returns to scale.
Using time series or cross-section data to equation (2.2), we can obtain the parameter
estimates that will determine inherent nature of the firm, such as wage policies,
inter-sector comparisons, substitutability, degree of homogeneity, etc. We have fitted
the C-D function to the time series data of the 21 major manufacturing industries of
Bangladesh by the OLS method.

3. Results and discussion

The estimates of C-D production function such as the capital and labour elasticities, the
returns to scale
parameters, the standard errors, the economy of an industry
(g ¼ 1= a þ b ); Theil’s R 2 and the Durbin-Watson d statistics are summarized in
Table I. The estimated results clearly demonstrate that the manufacturing industries of
Bangladesh generally seem to underlie the case of increasing returns to scale, i.e. a unit
increase in either capital or labor input cases output to grow at a more than 1 percent
rate. Of the 21 industries, 14 exhibit increasing return to scale and the rest show
decreasing return to scale. We found no industry with constant return to scale.
In order to
check whether an industry is economy or not, we calculate an index
g ¼ 1= a þ b . We observe that the major manufacturing industries like drugs,
printing, pottery, industrial chemicals, paper, food manufacturing, wood, iron and
steel, fabricated metal, wearing apparels except foot wear, textiles, glass products,
plastic and electrical machinery indicate good economy (since g , 1) according to their
 Return
Major
Capital to manufacturing
elasticity Labour scale industries
Industry a elasticity a þ b SE (a) SE (b) g ¼ 1=ða þ bÞ R2 d

Food manufacturing 0.6499 0.7026 1.3524 0.2158 0.2804 0.7394 0.9872 1.7094
Textiles 0.9101 0.2127 1.1227 0.1691 0.2174 0.8906 0.9793 2.4422 793
Leather 0.7290 0.2523 0.9813 0.2999 0.3334 1.0191 0.9689 2.2952
Footwear 0.3019 0.5674 0.8693 0.1548 0.2074 1.1504 0.9468 1.7905
Wood 0.8689 0.3964 1.2653 0.1064 0.5207 0.7903 0.9612 2.007
Furniture 0.2815 0.3050 0.5865 0.7542 0.9192 1.7049 0.8171 1.2092
Paper 0.9464 0.4705 1.4169 0.1825 0.1692 0.7057 0.9627 1.6589
Printing 0.3756 1.3346 1.7102 0.3724 0.5504 0.5847 0.9874 1.9157
Industrial chemicals 0.3790 1.0672 1.4462 0.0444 0.1984 0.6915 0.9814 2.6055
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

Plastics 0.9963 0.1054 1.1017 0.4936 1.0984 0.9077 0.8656 1.0915

Drugs 0.1238 2.7823 2.9062 0.0584 5.6183 0.3441 0.7793 2.3199
Pottery 0.2947 1.2970 1.5917 0.1303 0.2728 0.6282 0.9368 2.6516
Iron and steel 0.3715 0.8294 1.2008 0.7601 0.2327 0.8327 0.9611 2.8158
fabricated metal 0.5817 0.6137 1.1954 0.3185 0.2612 0.8365 0.8933 2.0845
Non-electrical
machinery 0.1283 0.0917 0.2200 0.6305 0.0496 4.5451 0.5066 1.6072
Mineral products 0.4539 0.3933 0.8472 0.1131 0.1191 1.1803 0.9258 3.0039
Transport equipment 0.1851 0.7497 0.9349 0.1698 0.4814 1.0696 0.8746 2.0426
Glass and glass
products 0.3055 0.8008 1.1063 0.0580 0.3819 0.9039 0.8420 1.2784
Other chemical
products 0.5883 0.1011 0.6394 0.2773 0.1017 1.5639 0.9839 2.3934 Table I.
Electrical machinery 0.1186 0.9753 1.0939 0.0947 0.1671 0.9141 0.9337 2.0594 Estimation of the
Wearing apparel Cobb-Douglas production
except footwear 0.2361 0.9159 1.1521 0.2381 0.2171 0.8681 0.9981 2.3995 function

contribution. The remaining industries such as leather, transport equipment, footwear,

mineral products, other chemical products, furniture and non-electrical machinery
demonstrate the diseconomy of scale. By improving the efficiency of these industries
we can ensure a better economy for them.
Except non-electrical machinery all other industries exhibit very high value of
Theil’s R 2. That means for almost all industries considered in this study, the estimated
regression lines fit the data very well. Out of the 21 industries, food manufacturing,
leather, footwear, wood, paper, printing, drugs, fabricated metal, transport equipment
and electrical machinery show no positive or negative autocorrelation. Of the
remaining industries, however, some depict low degree of positive and negative
autocorrelation and in some cases inconclusive results prevail.
In Table II we estimate the rates of change of output for a 1 percent increasing in
capital and labor input, respectively. Keeping labor input constant, we observe that 1
percent increase in the capital input leads to on the average about 1 percent increase in
the output for plastic, 0.95 percent increase in the output for paper and 0.91 percent
increase in the output for textile. We arrange the other industries according to the order
of percentage increased in output in the table. Again, keeping capital input constant,
we observe that 1 percent increase in the labor input leads to on the average about
2.7823 percent increase in the output for drugs, 1.3346 percent increase in the output for
 MAJ Percent Percent
19,6 1 percent increase in capital increased 1 percent increase in labor increased
input for in output input for in output

Plastics 0.9963 Drugs 2.7823

Paper 0.9464 Printing 1.3346
794 Textiles 0.9101 Pottery 1.2970
Wood 0.8689 Industrial chemicals 1.0672
Leather 0.7290 Electrical machinery 0.9753
Food manufacturing 0.6499 Wearing apparel except footwear 0.9159
Fabricated metal 0.5817 Iron and steel 0.8294
Other chemical products 0.5389 Glass and glass products 0.8008
Mineral products 0.4539 Transport equipment 0.7497
Industrial chemicals 0.3790 Food manufacturing 0.7026
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

Printing 0.3756 Fabricated metal 0.6137

Iron and steel 0.3715 Footwear 0.5674
Glass and glass products 0.3055 Paper 0.4705
Footwear 0.3019 Wood 0.3964
Pottery 0.2947 Mineral products 0.3933
Furniture 0.2815 Furniture 0.3050
Wearing apparel except footwear 0.2361 Leather 0.2523
Table II. Transport equipment 0.1851 Textiles 0.2127
A rate of change of output Non-electrical machinery 0.1283 Plastics 0.1054
for a given change in Drugs 0.1238 Other chemical products 0.1011
capital and labor input Electrical machinery 0.1186 Non-electrical machinery 0.0953

printing and 1.297 percent increase in the output for pottery. As before, we arrange the
other industries according to the order of increased in output in the same table.

4. Conclusion
The C-D production function has been used to estimate the 21 manufacturing
industries of Bangladesh. This is due to the vast literature which support using simple
C-D production function as compared to CES and VES functional formats. See for
example Wu and Bhatti (1994). The BBS annual time series data over the period
1982-1983 through 1991-1992 have been used in this study. The overall results
generally indicate that for most of the selected industries the C-D function fits the data
very well in terms of labor and capital elasticities, return to scale measurements,
standard errors, economy of the industries, high value of R 2 and reasonably good
Durbin-Watson statistics. The estimated results suggest that the manufacturing
industries of Bangladesh generally seem to underlie the case of increasing returns to
scale. Of the 21 industries, 14 exhibit increasing return to scale and the rest show
decreasing return to scale. We found no industry with constant return to scale. Keeping
labor input constant, we observe that 1 percent increase in the capital input leads to on
the average about 1 percent increase in the output for plastic, 0.95 percent increase in
the output for paper and 0.91 percent increase in the output for textiles. We arrange the
other industries according to the order of percentage increased in output in the table.
Again, keeping capital input constant, we observe that 1 percent increase in the labor
input leads to on the average about 2.7823 percent increase in the output for drugs,
 1.3346 percent increase in the output for printing and 1.297 percent increase in the Major
output for pottery.
manufacturing
References industries
Baltagi, B.H. (1996), Econometrics Analysis of Panel Data, John Wiley, New York, NY.
Bhatti, M.I. (1993), “Efficient estimation of random coefficient models based on survey data”,
Journal of Quantitative Economics, Vol. 9 No. 1, pp. 99-110. 795
Bhatti, M.I. (1997), “A UMP invariant test for testing block effects: an example”, Far East Journal
of Theoretical Statistics, Vol. 1 No. 1, pp. 39-50.
Bhatti, M.I. (2001), “Environmetrics analysis of complex surveys: a model-based approach”,
Complexity, Vol. 6 No. 5, pp. 41-55.
Bhatti, M.I. (2002), “On cluster effects in mining complex econometrics data”, in Bozdogan, H.
(Ed.), Statistical Datamining and Knowledge, Chapman & Hall, London.
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

Bhatti, M.I. and Khan, I. (2000), “Complex versus simplex model: an example of estimation using
Bangladeshi agricultural data”, Journal of Statistics & Management Systems, Vol. 3 No. 3,
pp. 295-304.
Bhatti, M.I. and Owen, D. (1996), “An econometrics analysis of agricultural performance in
Sichuan, China”, Asian Profile, Vol. 26 No. 6, pp. 443-57.
Bhatti, M.I., Khan, I.H. and Czerkawski, C. (1998), “Agricultural productivity in Shanghai region
of China: an econometric analysis”, Journal of Economic Sciences, Vol. 1 No. 2, pp. 1-12.
Harbury, C. (1968), Workbook in Introductory Economics, Pergamon Press, New York, NY.
Hojman, D.E. (1986), “The external debt contribution to output, employment, productivity and
consumption”, Economic Modeling, Vol. 3, pp. 53-71.
Hoque, I. (1991), “An application and test of random coefficient model in Bangladesh
agriculture”, Journal of Applied Econometrics, Vol. 6, pp. 77-90.
Intriligator, M.D. (1980), Econometric Models, Techniques and Applications, Prentice-Hall,
New York, NY.
Khan, I.H., Bhatti, M.I. and Czerkawski, C. (1998), “Indicator analysis approach to medium-term
employment: Australian case study”, Journal of Economic Studies, Vol. 2 No. 9, pp. 9-32.
Singh, S.P., Parashar, A.K. and Singh, H.P. (1999), Econometrics and Mathematical Economics,
S. Chand & Company, New Delhi.
Wu, X.P. and Bhatti, M.I. (1994), “Testing for block effects and misspecifications in regression
models based on survey data”, Journal of Statistical Computation and Simulations, Vol. 50,
pp. 75-90.

Further reading
Bangladesh Bureau of Statistics (1981), Statistical Yearbook of Bangladesh, Ministry of Planning,
Government of Bangladesh, Dhaka.
Bangladesh Bureau of Statistics (1986), Statistical Yearbook of Bangladesh, Ministry of Planning,
Government of Bangladesh, Dhaka.
Bangladesh Bureau of Statistics (1990), Statistical Yearbook of Bangladesh, Ministry of Planning,
Government of Bangladesh, Dhaka.
 This article has been cited by:

1. Md. Moyazzem Hossain, Tapati Basak, Ajit Kumar Majumder. 2013. Application of Non-Linear Cobb-
Douglas Production Function with Autocorrelation Problem to Selected Manufacturing Industries in
Bangladesh. Open Journal of Statistics 03, 173-178. [CrossRef]
2. Md. Moyazzem Hossain, Ajit Kumar Majumder, Tapati Basak. 2012. An Application of Non–Linear
Cobb-Douglas Production Function to Selected Manufacturing Industries in Bangladesh. Open Journal
of Statistics 02, 460-468. [CrossRef]
3. Mohammad Zakir Hossain, Khalid Said Al‐Amri. 2010. Use of Cobb‐Douglas production model on some
selected manufacturing industries in Oman. Education, Business and Society: Contemporary Middle Eastern
Issues 3:2, 78-85. [Abstract] [Full Text] [PDF]
Downloaded by Sultan Qaboos University At 01:06 06 October 2015 (PT)

View publication stats