India's manufacturing under-performance Part 1: New clue from multi-plants

One of the most intriguing and relatively undocumented developments of the last 20 yrs has been "multi-plants," whereby a single firm operates not one but multiple production facilities within a state

Perhaps the biggest failure of India’s development and an equally big puzzle is why India has under-performed in labour-intensive manufacturing whether measured in terms of output, exports or employment. 

For example, India’s share of global apparel exports was 3 per cent in 2005 and almost two decades later, despite China’s vacating its export space, it remained at 3 per cent.
 

Explanations abound on this under-performance. One of the prominent ones relates to plant size, in particular that Indian plants are kept too small by regulations, especially labour laws. In our latest research, we discover a new clue that speaks to this aspect of plant size but focusing not on why plants are small but why they are not large, very large.

One of the most intriguing and relatively undocumented developments of the last 20 years has been “multi-plants,” whereby a single firm operates not one but multiple production facilities within a state. Last week’s Economist magazine spotlighted Shahi Exports, India’s largest garment exporters whose CEO Harish Ahuja spoke of the compulsion to “split big investments across a number of factories”.

The multi-plant phenomenon has expanded over the last two decades and today they account for a large share of non-managerial employment and output of private manufacturing plants.
 

Focusing on labour-intensive industries and large plants (defined as those employing more than 200 workers), the share of multi-plants rose from 15 per cent in 2001 to 30 per cent of total plants in 2022, and from 17 per cent to 43 per cent in terms of total non-managerial employment.

Why is the phenomenon of multi-plants significant? For three reasons: They change our understanding of firm size and its evolution; they are associated with lower productivity; and they shed light on how regulations and labour markets work.

The Annual Survey of Industries (ASI), one of the key sources of data on India’s manufacturing sector, is susceptible to misrepresenting the size of these firms. The ASI allows firms with multiple plants in a state (but not across states) to file joint returns if they have at least two plants, each with 100-plus employees. This leads to an overestimation of the size of plants, especially larger ones — what economists call the “right tail”. As a result, official statistics often suggest that India’s largest manufacturing plants are bigger than they actually are.

For instance, consider the employment profile of one of India’s most successful exporting firms, and largest employers. Table 1 shows that in one state where it has a large presence, it has increased the number of plants from 11 to 58 between 2005 and 2023 with average employment per plant remaining broadly the same (with small variation during the intervening years). The same is true for the maximum size of these plants. But if this is not accounted for, the ASI data would suggest that the firm has one plant in the state that has increased its employment from 14,000 to 76,000 over this period and become Foxconn-like in size.

Indeed, accounting for multi-plants changes our understanding of the evolution of large firms over time. For example, important research by Bertrand, Hsieh, and Tsivanidis suggests that large plants grew in size over the last two decades in part because changes in the law and implementation allowed for greater contractualisation of the labour force. In this view, contractualisation allowed firms to circumvent some of the burdens imposed by labour laws, especially critical ones — relating to hiring and firing — which kick in above certain employment thresholds. And, it is true that the share of contract labour in total employment has increased from 22 per cent in 2001 to 41 per cent by 2022 (the latest year for which ASI data is available).

But here is the striking new finding: Despite massive contractualisation and indeed despite the massive deregulation of trade and other policies that has taken place since 2000 (reduction in tariffs, elimination of small-scale sector reservations, etc) we estimate (making some assumptions described in our paper) that large plants have not become any larger. As Table 2 shows, regardless of whether we think of plant size in terms of the number of employees or as the employment share accounted for by large plants they have not become larger and may even have become slightly smaller.

Consider now the costs of multi-plants. We find that in labour-intensive industries, multi-plants are about 5 per cent less productive on average than single plants of equivalent employment size (after controlling for the sector and state of operation) with this wedge rising as the number of employees rises. This finding illustrates the benefits of scale. Five hundred workers are more productive if they work in one plant than if they are split into two plants of 250 workers each.

So, this pattern of large firms not becoming larger and remaining low productive is consistent with India’s stagnating export performance in labour-intensive manufacturing (where perhaps scale matters substantially). Consider the comparison between Bangladesh and India in the apparel sector. While India initially had the upper hand, Bangladesh’s market share in global apparel exports skyrocketed from 2.5 per cent to 8 per cent over two decades, while India’s share stagnated at around 3 per cent.

A closer examination of plant size data offers a possible explanation. Correcting for mismeasurement, we find that in 2013 Bangladeshi apparel plants were consistently larger than their Indian counterparts across all metrics. For instance, the 95th percentile plant in Bangladesh was about 40 per cent larger than a similar plant in India. This size advantage is critical, as larger plants in Bangladesh are far more efficient, exporting 95 per cent of their output compared to just 37 per cent for Indian plants. This significant size differential may be one of several reasons why Bangladesh outpaced India in the global apparel market — larger plants lead to greater efficiency and stronger export performance, providing Bangladesh with a competitive edge.

We have shown that the phenomenon of multi-plants sheds light on the sobering facts relating to firm size; and multi-plants are also less efficient which affects India’s manufacturing competitiveness. But they also profoundly affect our understanding of some aspects of India’s regulatory regime, especially the debate about labour laws and plant size. We turn to this tomorrow.

Abhishek Anand is Visiting Professor, Madras Institute for Development Studies (MIDS), Arvind Subramanian is Senior Fellow, Peterson Institute for International Economics, and Naveen Thomas is Professor, OP Jindal Global University. This piece is based on their latest MIDS paper: https://www.mids.ac.in/assets/doc/WP_244.pdf.

Tomorrow: Part II