The view from outside Dhaka
Our short-sighted Dhaka-centric official focus has led to uneven industrial development and increasing regional disparities, argues Syed Akhtar Mahmood.
Cityscape, Dhaka. Photo: Amirul Rajiv
Anyone driving down the Dhaka-Comilla road, as I did last summer, can't help noticing the continuous spread of factory buildings on either side of the road as one leaves the boundaries of Dhaka city. But, quite mysteriously, the industrial plants disappear once you cross the river into Comilla district.
The short journey to Comilla reveals both, the encouraging and the disappointing, aspects of Bangladesh's recent industrial growth. Industry is expanding, but in an uneven fashion. While Bangladesh is not a particularly industrialized country -- in 2000, manufacturing contributed only 15% of GDP, more or less the same share as in 1995 -- the pace of industrialization has accelerated in more recent years. More up-to-date data, when available, may indicate that Bangladesh, finally, is on its way to becoming a more industrialized country. But this rapid industrialization, even if it happens, may by-pass large parts of the country.
This concern arises from one simple fact, i.e., much of Bangladesh's industrial output is concentrated around the two major metropolitan centres of Dhaka and Chittagong. In 2000, the last year for which I could obtain district level statistics, over two-thirds (65%) of manufacturing GDP originated in just four of the 64 districts in the country.
Of these, Dhaka, and its adjacent districts Gazipur and Narayanganj, accounted for 50%, while Chittagong district contributed another 15%. The next 10 districts -- Narsingdhi, Tangail, Pabna, Sirajganj, Khulna, Jessore, Kushtia, Comilla, Brahmanbaria, and Sylhet -- contributed another 18%. This leaves us with a disturbing fact. Fifty, out of the 64, districts of Bangladesh together contribute only 17% of the country's industrial output.
The national average share of manufacturing in GDP (15%) thus hides significant regional variations. Manufacturing's contribution to district GDP varies from as low as 2%, in Lalmonirhat in northern Bangladesh, to a high of 40% in the Gazipur district adjacent to Dhaka. In as many as 26 out of the 64 districts in the country, this share is below 5 %.
Such regional variations matter. Data shows that there are strong links between the degree of industrialization and per capita incomes -- districts with higher share of manufacturing in GDP tend to have higher per capita incomes (see Table 1). This relationship is less pronounced at low levels of industrialization but becomes significant when income rises above a threshold level.
Of course, it is difficult to identify the direction of causation. Is it that districts become richer when they are more industrialized or is it the other way round, i.e., that districts which are richer to start with attract more industrial investment because more income means more demand for industrial products. I suppose both factors are at work and, if so, we are likely to see a snow-ball effect with the more industrialized regions becoming even more industrialized over time. For the more advanced districts there is a virtuous cycle in play. The rest may be caught in a trap. Thus, if we would like to see greater regional balance in Bangladesh's development, we have to start thinking about a more regionally balanced industrialization strategy. But we need to be cautious here. Perfect regional balance is neither feasible nor desirable. Some regions, reflecting their comparative advantages, will inevitably be more industrialized than others, and any attempt to artificially promote regional balance will be futile. However, it appears that the existing regional disparity in industrialization in Bangladesh is far greater than what would follow from the comparative advantages of various regions. It is no surprise then that Bangladesh's PRSP places considerable emphasis on achieving a greater regional balance in industrialization.
In order to devise strategies, policies, and operational interventions to promote regional balance, it is necessary first to obtain an idea of the regional patterns in industrialization. As would be expected, manufacturing GDP per capita is highest in the four districts (Dhaka, Gazipur, Narayanganj, and Chittagong) which account for two-thirds of the industrial output in Bangladesh. Here per capita manufacturing GDP was greater than Tk 20,000 in 2000.
They are followed by Khulna and Jessore in south-western Bangladesh and Sirajganj, Pabna, and Kushtia in western Bangladesh. Khulna has been the most important industrial center after Dhaka and Chittagong. While some of the traditional industries, especially jute, have declined, the Khulna area has seen the main concentration of the shrimp industry. The Sirajganj-Pabna area has historically been a major concentration of the handloom industry.
There are two large "non-industrialized" regions in Bangladesh where per capita manufacturing GDP was less than Tk 1,000 in 2000 (see map). The first consists of 19 contiguous districts located in the northern part of the country, starting from Sunamganj and Habiganj in the north-east, and including almost the whole of the old districts of Mymensingh and Rajshahi, and all of the north-western part of the country, i.e., the old districts of Bogra, Rangpur, and Dinajpur.
The second "non-industrialized" region is in the central-southern part of the country and consists of 15 contiguous districts covering almost the whole of the old Barisal and Noakhali districts, the entire old Faridpur district and the eastern parts of the old Kushtia, Jessore, and Khulna districts. A third area with low incidence of manufacturing is, not surprisingly, the Chittagong Hill Tracts (CHT) area to the south-east.
It is important to note the contiguity of the districts with low manufacturing activity. These are not districts randomly scattered all over Bangladesh but, with the exception of CHT, are located next to each other. This suggests that particular regional factors could be at play behind the low manufacturing activity in these areas. This subject is explored below.
What explains these patterns? Why is industry concentrated in a few areas? This is partly a result of inadequate investment in public services in the smaller cities and partly due to a regulatory system where the important decisions and approvals are made or reviewed in the major cities, in many cases in Dhaka. Proximity to the central regulatory offices thus becomes an important factor in locational decisions. Major industries want to be located where they have easy access to major infrastructure (main highways, electricity, gas, communication) and services (financial services, public agencies), and export links (Chittagong port and Dhaka airport).
Data support these assertions. The concentration of manufacturing appears to be associated with a concentration of infrastructural endowments -- districts with poorer infrastructure have lower levels of industrial activity. The five districts where manufacturing accounts for more than 25% of GDP have an infrastructure endowment (defined in terms of GDP originating in infrastructure) more than three times that of the districts where the manufacture/GDP ratio is below 5% (Table 2). Similarly, the degree of urbanization in the most industrialized districts is four times that in the least industrialized parts of Bangladesh.
Market access is another important factor. Regions with good access to markets, especially large markets, have more industry and more productive industry. This is not surprising. With a large market, firms may have greater incentive to invest in productivity enhancing measures. Accessibility indices have been constructed for each thana in Bangladesh that takes into account the ease of accessing the large market centers (essentially urban centers above a certain size) in the vicinity, weighted by the size of the centers (i.e., population).
Cityscape, Chittagong. Photo: Amirul Rajiv
As expected, the index has the highest value (signifying greatest accessibility) for Dhaka and a number of its adjacent districts. Accessibility then decreases as one moves away from Dhaka, especially towards the north-east, west and south. Large parts of Bangladesh are handicapped by poor access to large market centres. It can be shown through charts and statistical analysis that districts with better access to large markets tend to have a higher manufacturing GDP per capita than those with poorer access.
So what is the agenda for going forward? As argued at the outset, there is a case for reducing the regional imbalance in industrialization because of the strong links between industrialization and income levels. Moreover, other factors, such as maintaining ecological balance may also require a more spatially diversified pattern of industrialization.
Experience in most countries, including Bangladesh, suggest that attempts to promote geographically balanced industrialization artificially, without regard to comparative advantages, is likely to fail. This is exemplified by the Bangladeshi attempt to promote regionally balanced industrialization through widely dispersed industrial estates. There is a wide variation in the performance of the 65 industrial estates of the country.
Industrial parks located in districts with poorer access to large markets tend to perform worse than those located in districts with better access to markets (performance being defined in terms of the degree to which the parks are occupied, and the proportion of sanctioned industrial units that are in operation). In other words, efforts to promote industrialization in remote locations through tools such as industrial parks have not met with much success.
Cityscape, Rajshahi. Photo: Amirul Rajiv
Here are some preliminary ideas about how the challenge of more regionally balanced industrialization can be addressed in Bangladesh, keeping in mind the above considerations. I suggest that the first phase of regional diversification of industry may focus on five regions outside the current concentrations in Dhaka and Chittagong. Of these, three are in north-west Bangladesh, one in south-west Bangladesh, and one in eastern Bangladesh. These regions, with possible focus areas of activity, are listed in Table 3. These suggestions, which may be treated as tentative, are based on a preliminary analysis of some of the possible determinants of comparative advantages in various regions such as crop suitability, access to markets and infrastructural endowments.
These regions have relatively high degrees of market accessibility compared to other parts of the country, and hence, are likely to have larger demand for their manufactured products. The suggested specializations for the regions follow from other factors, notably the current and expected economic base for industry.
For example, two regions in the western part of Bangladesh stand out in terms of their agricultural potential. Crop suitability indices constructed for different districts in Bangladesh have the highest values for two clusters of districts: a) eight contiguous districts in the north-west corner of Bangladesh, including almost the whole of the old districts of Dinjapur and Rangpur, and the district of Jamalpur on the other side of the Jamuna; and b) the districts of Meherpur, Chuadanga, Jhenaidah, and Jessore in western Bangladesh. The old district of Rajshahi in north-west Bangladesh also has good agricultural potential.
These districts also have a high density of secondary roads; indeed some of them have the highest secondary road density in the country. Another noteworthy fact is the larger than average land holding sizes in the north-western region of the country suggesting a greater scope for contract farming as a basis for agro-processing industries.
These factors suggest that the districts of Dinajpur, Rangpur, Gaibandha, and Jaipurhat (identified as Region 1) and the districts of Nawabganj, Rajshahi, Natore and Kushtia (Region 3) could be good locations for a primarily agro-processing based industrialization.
Region 2 (the districts of Bogra, Sirajganj and Pabna on the western side of the Jamuna river and Jamalpur, Tangail and Manikganj on the eastern side) could focus on light manufacturing although there may be some scope for agro-processing activities in these areas as well. Bogra, for example, has some tradition in light engineering products, and Pabna in textiles.
Region 5 (the districts of Munshiganj, Bramanbaria, Comilla and Feni), which account for a large part of the Dhaka-Chittagong corridor, and which have a dense road network as well as a relatively good access to electricity, is another strong candidate for substantial expansion of light manufacturing activities.
Finally, Region 4 (south-western districts of Jhenidah, Jessore, and Khulna), may focus on a mix of agro-processing and light manufacturing.
As these regions emerge as important centres of industry, the current concentrations of industry, i.e., the Dhaka and Chittagong belts may transform into centres of higher-end industrial activity which require a supply of highly skilled labour and highly educated professional, technical and managerial people who, for various reasons, may want to be located near the largest cities. Less demanding light manufacturing activities may then relocate to the new regions.
This note has made a beginning in highlighting the regional pattern of industrialization in Bangladesh, and made some tentative suggestions about regions which may be the focus of a strategy of regionally balanced industrialization. Further analytic work is required to answer many questions not addressed by this note. Such work will help identify operational interventions that may promote greater regional balance in industrialization without falling into the trap of artificially promoting industrialization disregarding comparative advantages.
Many years ago, we fought hard against regional disparity within a united Pakistan. It is somewhat paradoxical, given this history, that we are turning a blind eye towards regional disparities within Bangladesh. It is time to change this.
Syed Akhtar Mahmood is an eminent Bangladeshi economist.