Sources of Total Factor Productivity in South Asia

Abstract

Using World Development data for the period of 1991–2020, this article examines the total factor productivity (TFP) measured by Malmquist productivity index (MPI) and its components’ technological progress and technical efficiency change (TEC) for five South Asian countries, namely India, Pakistan, Nepal, Bangladesh and Sri Lanka. TFP growth is calculated using the MPI along with panel data of those countries. This article adds onto the existing literature by estimating the technological progress and TEC for these five South Asian countries. Further, the article also emphasises differences in the TFP growth among these countries. The study finds that TFP of all the South Asian countries are cyclical in nature, highly volatile and full of ups and downs during the whole study period. TFP is growing in standard technology-based economy (India), which is the consequence of technological progress. On the other hand, among labour-intensive economies, Sri Lanka is performing better as TFP in Sri Lanka is improving at a good rate, while Pakistan’s TFP is also growing steadily. Here, both the countries’ TFP growth is driven by positive technological change. Bangladesh is observing TFP decline as Bangladesh is the worst performer among five nations in South Asia despite the fact that technological change is still progressing. Nepal’s TFP is improving as well, but at a slower pace. Except Bangladesh, efficiency change shows progressive trend in the remaining countries.

JEL Classification: D24, F02, F62, O14

Keywords

Total factor productivity Malmquist productivity index technological change efficiency change

Introduction

According to traditional growth theories, aggregate production is the output, which is the combination of human resources and physical capital, along with the gains in production that can be attributed by increase in inputs along with a residual portion that cannot be explained by other measurable tangible inputs like labour and physical capital, commonly known as ‘total factor productivity’ (TFP) (Dutz & O’connel, 2013). Inadequate savings, inadequate expenditure in physical capital and less skilled human capital, and insufficient growth in TFP are the major causes behind the slow growth of aggregate production. The crucial fact is that differences in those tangible and intangible inputs explain less than half of the per capita income variation (Jones & Romer, 2010).

Developing economies are not only because they have less investment in physical and human capital than their counterparts but also because the production entities of those nations are incapable of making good use of tangible factors such as physical capital stock, labour force and other raw materials. In addition, they are also inefficient at integrating them with complementary knowledge-based intangible inputs.

TFP growth is influenced by the assessment of other factor inputs. TFP growth will focus on changes in capital usage when capital factors are quantified in terms of the capital stock. Again, TFP improvement will represent the absorbing and utilising capacity of the labour force when labour inputs are quantified using the labour force (Nombulelo, 2007).

TFP gains are important for any emerging countries because they allow themselves to shift their acquired physical and human capital to more productive sectors such as science, technology and higher education, which can help them to sustain economic growth and development (Kavita, 2021).

The concept of TFP receives a special attention in South Asia, as most of the studies consistently note the issue of the productivity gap in this part of the world and recommend policy steps to enhance the TFP (Adnan et al., 2019; Srinivasan, 2005). Since most of the nations have impressively consistent economic growth, TFP gains will assist those nations to maintain a more sustainable growth in the future (David et al., 2014).

Productivity growth comes from two sources: technological advancement and changes in technical efficiency. A study of these sources is essential for understanding the factors that cause productivity stagnation and for implementing suitable actions to boost productivity at the firm, industry and government levels.

Studies related to South Asia coherently address the issue of deficiencies of factor productivity in this region and suggest policy measures to improve it (Adnan et al., 2019; Srinivasan, 2005). Thus, the sources of productivity attract a lot of attention for South Asia from two perspectives: (a) the productivity of domestic firms of those countries are generally far behind from that of multinational companies (MNCs). Since productivity gain is the most convenient way to acquire knowledge and technology, those South Asian countries can make the best use of it to imitate technology or innovate and (b) as South Asian countries have persistently high economic growth rates, this productivity gains can help to sustain this growth for long term (David et al., 2014). To fulfil those motives, South Asian countries have liberalise their trade since the 1990s.

So far, not many studies have been conducted about the sources of productivity in South Asia. But there are few analyses conducted by countries that exist (Manjappa & Mahesha, 2008, on India; Dutz & O’connel, 2013, on Srilanka; Amjad & Awais, 2016, on Pakistan; and Sinha, 2017, on Bangladesh),¹ though they focus merely on TFP. These studies also claim productivity as the leading obstacle for the consistent growth and economic development in the near future. Previous literature also convey that most of the cross-country differences (either per capita income difference or growth difference) can be explained by productivity growth (Li & Liu, 2005). Thus, there is enough room to find the sources of productivity at the cross-country level.

Keeping this gap in mind, in our study, we attempt to estimate the productivity gains in five South Asian economies, including Bangladesh, India, Pakistan, Sri Lanka and Nepal² using the data set of 1991–2020. Productivity growth can be measured using the so-called distance-function-based output-oriented Malmquist productivity index (MPI) under constant returns to scale (CRS). We use data envelopment analysis (DEA) to measure TFP and then decompose TFP into two parts: technical efficiency change (TEC) and technology change (TC).

The remaining study is divided into the following sections: the second section describes productivity measurement. The third section deals with the data sources and estimation procedure. The fourth section discusses the estimation results, and the fifth section provides the concluding remarks.

Productivity Measurement

The most frequent way for estimating productivity improvements is to use MPIs, which were developed by Caves et al. (1982). Few years later, Simar and Wilson (1998) used MPI approach to analyse the productivity of the industrialised countries. According to the MPI approach, TFP can rise not only as a result of technological advancement (moving of the production frontier) but also as a result of increased technical efficiency (catching up). The Malmquist index, according to Färe et al. (1996), has three key advantages.

It does not, for starters, assume profit maximisation or expense minimisation. Second, it does not necessitate knowledge of input and output pricing. Finally, if the researcher has panel data, productivity changes can be decomposed into two parts (TEC, or catching up, and technical change, or changes in best practices). The need to compute distance functions is the biggest disadvantage. The DEA technique, on the other hand, can be utilised to tackle this issue.

Another significant benefit of the Malmquist index is that it distinguishes between shifts in the frontier (TC) and gains in efficiency relative to the frontier (efficiency change, TEC), two mutually exclusive and exhaustive sources of TFP.

MPI is a DEA, which is static in nature, and accounts for the movement of the production frontier over time as it examines the productivity of a firm or an industry in relation to industries with best practices in a given year. It can offer light on the mechanism of productivity change since it can decompose productivity increase into TEC and technological improvement (Kumar & Russell, 2002).

Caves, Christensen and Diewert (CCD) first proposed the DEA-based MPI technique in 1982, and Fare, Grosskopf, Lindgren and Roos (FGLR) and Fare, Grosskopf, Norris and Zhang (FGNZ) empirically tested it in 1992 and 1994. Several versions of MPI have been produced since then.

Data Sources and Methodology

The present research is based on data accessed from secondary sources. We utilise the data set of five South Asian countries from 1991 to 2020. For the calculation of TFP, TEC and TC, we use real GDP as output and physical capital stock and labour force as inputs. The real GDP data are derived from the World Development Indicators—WDI (World Bank, 2021). According to the previous literature (Danquah & Amankwah-Amoah, 2017), the total labour force is calculated using the economically active population, which is defined as those aged between 15 and 64, and is derived from the WDI (World Bank, 2021). As a capital input, WDI (World Bank, 2021) data on real gross fixed capital formation are utilised.

Figure 1 shows basic linear frontier lines at time t and t + 1, which are defined by CRS, for one input and one output. Linear frontier lines exhibit CRS. The MPI based on CRS distance functions is widely accepted as it measures productivity change accurately, while variable returns to scale (VRS)-based distance functions might be biased (Johnes, 2006).³

Figure 1.

Source: Hossain & Oh (2019).

The frontier technology line and the distance from that line are important concepts in understanding MPI. P^t and P^t ^{+ 1} are frontier technology lines of t and (t + 1) year, respectively (Figure 1). These frontier lines refer to the most productive and efficient production technology. Points under those lines are considered as less productive and less efficient as compared to the points that are on frontier lines. Point M (x^t, y^t) and Point N (x^t ^{+ 1}, y^t ^{+ 1}) are two such production cases or entities below frontier lines (Figure 1).

Now the question is which point is more productive? Suppose point N produces more output than M, but that does not mean that point N is more productive than point M because one entity may use more input/output than another. Thus, reference is needed in determining productivity performance. Here, two technological frontier lines (P^t, P^t ^{+ 1}) can be regarded as references. Points near the frontier line is more productive than points farther away from the frontier. Ratio of distances from the frontier line can be used as a measurement of productivity.

The value of efficiency ranges from 0 to 1, that is, while 0 means fully inefficient, 1 means completely efficient. If M or N (or any other points) is located on the frontier lines, then the entity is said to be completely efficient. Points below the frontier curves are considered as inefficient, and the farther the point is from the frontier lines, the less efficient it will be. We may now compare N and M’s productivity levels based on their distance from frontier technologies. The progress or regress of two points’ productivity is determined by the ratio of their efficiency levels using distance from the frontier.

To measure productivity growth and its two parts: TC and TEC,⁴ the famous non-parametric MPI has been utilised to a great extent. Many previous studies also use MPI methodology to measure productivity growth at both the macro and micro levels. Fare et al. (1994) were the first to use the index. Later, Rao and Coelli (1999) applied it successfully. Further, Headey et al. (2010) used this index to measure agricultural productivity. We also employ the output-based MPI methodology at the macro level to measure productivity growth and its components for the above-mentioned five South Asian nations, where output measured by real GDP and two raw materials are labour and physical capital stock.

We derive MPI utilising DEA, which measures the change of productivity over time. In the case of MPI, input or raw materials and output values are not necessary to measure TFP and can be represented by the shifts in terms of efficiency and technical change. Therefore, TFP is the product of TEC and TC. Additionally, TEC also has two parts: pure technical efficiency (PTEC) and scale efficiency change (STEC). While PTEC basically means the organisational skills and efficiency, STEC measures the degree of efficiency to the highest level.

Using MPI through DEA, we can measure both the productivity change and its two parts, which show the efforts to catch up and the shift of frontiers. According to MPI, TFP change is the ratio of the distance of each of the two points from a technology leader between the two time periods. MPI can be calculated as the ratio of productivity growth in two certain points by measuring the distance from each of that point to the leader technology.

We can write the output-oriented Malmquist productivity change index in the following way:

M_{0} (x^{t + 1}, y^{t + 1}, x^{t}, y^{t}) = \frac{D_{0}^{t + 1} (x^{t + 1}, y^{t + 1})}{D_{0}^{t} (x^{t}, y^{t})} {[\frac{D_{0}^{t} (x^{t + 1}, y^{t + 1})}{D_{0}^{t + 1} (x^{t + 1}, y^{t + 1})} \frac{D_{0}^{t} (x^{t}, y^{t})}{D_{0}^{t + 1} (x^{t}, y^{t})}]}^{\frac{1}{2}} = TEC \times TC

(1)

here

$D_{0}^{t}$ is the distance function related to the efficiency frontier at time period $t; D_{0}^{t + 1}$ is the distance function related to the efficiency frontier at time period $(t + 1); y^{t}$ is the amount of country’s output of the analysis at period $t; x^{t}$ is the country’s input amount at period $t; y^{t + 1}$ is the country’s amount of output at period $(t + 1); x^{t + 1}$ is the country’s amount of input at period $(t + 1); D_{0}^{t} (x^{t}, y^{t})$ measures the distance of the country $(x^{t}, y^{t})$ to the efficient frontier at period $t; D_{0}^{t + 1} (x^{t + 1}, y^{t + 1})$ measures the distance of the country $(x^{t + 1}, y^{t + 1})$ to the efficient frontier at period (t + 1); $D_{0}^{t} (x^{t + 1}, y^{t + 1})$ is one of the mixed period that measures the distance of the country $(x^{t + 1}, y^{t + 1})$ to the efficient frontier at period t; and $D_{0}^{t + 1} (χ^{t}, y^{t})$ is another mixed period that measures the distance of the country $(x^{t}, y^{t})$ to the efficient frontier at period (t + 1).

Equation (1) expressed the two parts of MPI. The expression inside the bracket provides the form of technological change and the shift in frontier measured in two different time periods (t and (t + 1)). The part outside the bracket representing the efficiency change of the same two time periods and productivity is the product of efficiency change by TC. MPI is measured on the basis of fixed returns to scale. M₀ > 1 represents the progress of productivity, while M₀ < 1 refers to the reduction of productivity, and M₀ = 1 means no change.

Even if two nations possess an identical amount of labour and capital, they still differ in terms of aggregate production. These differences can exist within a country over different time periods and across nations at a certain period of time. The phenomenon responsible for this variation is commonly known as change in TFP. TFP signifies the efficiency with which various factors like physical and human capital and labour are employed.

Many researchers thought that most of the disparities in per capita income between nations can be explained by variations of those factor inputs: skilled labour force and capital intensity. But growth accounting theory claims the changes in TFP as the major cause of cross-country income differential. Researchers may think that differences in tangible input physical capital and intangible input human capital account for less than half of the income variation. But it is the differences in technology that account for the major portion of the disparity (Aubhik, 2009). TFP accounts more than half (50%–70%) of the income variation (Hall & Jones, 1999; Klenow & Rodriguez-Clare, 1997). Hall and Jones (1999) found that variation in income resembles the change in TFP, while correlation between the two is more than 80%.

Empirical Results

Data limitations are always a matter of concern in South Asia. The variable ‘labour’ consists of a number of labourers only from the formal sector, while a large informal sector remains uncovered, which comprises about 40% of GDP and two-thirds of the total labour force. Thus, the following tables and figures do not necessarily show the complete picture of South Asia. But still, it can provide a substantial number of findings for the discussion to proceed on.

The year-basis estimates of TFP change and its two parts are reported in Table 1, which shows overall growth in productivity. The growth of TFP has been found in 22 out of 30 years, and the rest 8 years exhibit productivity revert. Among those TFP progressive years, the rate ranges from 3.2% (in 1991) to 0.27% (in 1993), while productivity regress is varied between 0.01% and 2.1%. The TFPC index has the lowest value of −2.3% in 2006 and 2017 and a high of 0.17% in 2018 (Table 1).

Table 1.

Yearly Position of Total Factor Productivity Change, Technical Efficiency and Technology Change in Five South Asian Countries from 1991 to 2020.

Year	TFPC	TEC	TC
1991	1.0317	1.02164	1.00996
1992	1.01424	0.99772	1.01654
1993	1.0027	1.02602	0.97758
1994	1.0202	1.02774	0.99294
1995	1.01132	1.00928	1.00226
1996	1.00764	1.0387	0.97102
1997	1.0123	1.00592	1.00664
1998	0.99062	0.96598	1.0262
1999	1.00694	0.94156	1.07038
2000	1.02676	1.02484	1.00212
2001	0.98276	0.96934	1.01376
2002	1.01698	0.96026	1.0597
2003	1.0029	1.00462	0.99828
2004	0.99586	0.98992	1.00602
2005	0.99396	0.99396	0.96738
2006	0.97652	1.0201	0.95778
2007	1.00896	0.98426	1.02504
2008	1.00692	1.01732	0.99
2009	1.02748	0.97456	1.05508
2010	1.00964	0.95296	1.06058
2011	1.02968	0.98044	1.0508
2012	1.023	1.01106	1.01194
2013	1.00712	1.00184	1.0055
2014	1.01724	0.98136	1.03628
2015	0.98596	1.01186	0.9741
2016	0.9969	1.00324	0.99362
2017	0.97652	0.98762	0.98828
2018	0.99832	1.00172	0.99672
2019	1.02424	0.9716	1.05542
2020	1.02344	1.0043	1.01914

Source: The author.

Note: The values of the three aforementioned indicators refer to the change over the previous to the current year. For example, the value of 1992 indicates the changes over the period 1991–1992 and so on.

From the individual country perspectives, India has experienced moderate TFP progress (0.65%), and Sri Lanka holds the top-most position in South Asia with a TFP growth of 2.4% followed by Pakistan (1.5%) and Nepal (0.402%), while Bangladesh has a negative growth of 1.14% (Figure 3). Most of the countries’ TFP growth were contributed considerably by TC. TEC of most of the countries remains constant or shows negative trend. The mean indices reported in the last row of Table 1 are the averages of all years in South Asia.

Figure 2.

Source: The author.

Figure 3.

Source: The author.

Throughout the study period, we saw that TFP increase in India fluctuated (Figure 2). India’s average TFP growth was positive (0.64), although it was quite low. During the years 1991–2020, the economy suffered on average negative TFP growth, meaning that technological progress (TC) had occurred. TFP growth mainly began in India from 1991 when the Indian economy went through a broad-based reform process (Saha, 2014). During the 1980s, significant changes were import liberalisation, the increase of export incentives and liberal access to finance and foreign exchange. Significant pro-market changes such as (a) the elimination of industrial licensing, (b) the liberalisation of foreign direct investment (FDI), (c) the elimination of import licensing, (d) the liberalisation of important sectors like telecommunications and (e) major financial sectors would require another decade (Kalpana et al., 2006). The rise in TFP growth for the Indian economy from 1991 to 2020 can be linked to a number of changes in macroeconomic conditions that are generally favourable to productivity growth.

The Indian economy registered a large increase in private sector lending, 24.61% during 1991–2000 and 39.35% during 2001–2008. Furthermore, gross fixed capital formation rose to 98% during 1981–1990, while the growth was 642% during 1991–2020. Again, there has been a significant improvement in health, as seen by the increase in average life expectancy at birth from 56 years between 1981 and 1990 to 62.64 years between 1991 and 2019. Similarly, average FDI inflows as a percentage of GDP increased more than threefold from 0.46% in 1991–2000 to 1.59% in 2001–2019, and average trade share of GDP increased significantly from 22.09 % in 1991–2000 to 41.3 % in 2001–2019. Again, the share of agriculture value added as a percentage of GDP has decreased considerably, from 31.71% in 1981 to 15.97% in 2018. Aside from that, reliance on foreign aid has decreased substantially, from 0.28% in 1981 to 0.192% in 2019.

This productivity increase of India is fully attributed to the positive technological change of 0.75%, while the TEC is zero. Though India has never maintained constant TFP growth over a long period of time, India’s success story of TFP is largely contributed by its moderately stable FDI flow, which is highest in South Asia. Still India largely falls behind China in receiving FDI as China, with a considerably larger economy than India, competes for investment due to sheer size. Furthermore, two economic shocks impeded FDI flow: the Asian crisis in 1997–1998 and the global financial crisis in 2007–2008. Both significantly hurt the Indian economy, limiting India’s potential to catch up and, therefore, resulting in a widening of the gap with the leading countries.

The source of this technological progress originates from capital-intensive sector.⁵ On the other hand, in the case of labour-intensive sector,⁶ low growth rate of TEC is offset by the large decline of technological change (Manjappa & Mahesha, 2008). The surge of growth of capital-intensive industry is characterised by the introduction of delicensing, concentration of large flow of FDI and import of advanced technology. The result implies that India has substantial innovative power in South Asia, which is backed up by the significant FDI flow in their major export-oriented sector like services. Thus, India is regarded as the standard technology-based economy in South Asia against the remaining four labour-intensive countries.

Moreover, India as well as other South Asian countries do not become successful in structural transformation of sectors. Most of those economies are still agriculture-oriented. Though the industrial sector has contributed more than agricultural sector (except Pakistan and Nepal), complete industrial base has not been built up yet.

India is the only country in South Asia characterised with both quality- and quantity-based export growth (Brunner & Cali, 2006). Thus, India’s export growth is led moderately by quality export but largely by quantity export with comparatively lower prices. Indian manufacturing has maintained sufficient levels of international standard but, it lacks quality of international standard. Though minimum standard of technology prevails in few capital-intensive sectors such as capital machinery, chemical, iron and steel, the standard is lower than that of East Asian counterparts. Diversification of export is always questionable in South Asian countries. According to the statistics, a country with FDI equal to 20% of GDP can diversify their export successfully. All the South Asian countries’ FDI ratio remains far below than this standard.

A study shows that India and Sri Lanka are performing well to diversify their products, while Bangladesh remains steady and Pakistan shows a downward trend in this case (Javed & Kasif, 2016). India ranks 58th in global competitiveness index (GCI) and achieves good score in almost all the subfields. Moreover, India performs better in innovation, entrepreneurship and trade openness index, which help to upgrade productivity growth.⁷ The information and communications technology (ICT) sector has successfully adopted technologies as it is the pioneer exporting sector of India, which attracts a significant amount of FDI.

In comparison to other South Asian countries, India has the highest capital intensity, with service and manufacturing being the most capital-intensive sectors, while service is the most FDI-receiving sector, followed by manufacturing. As a result, India’s capital intensity, along with the foreign technology it receives through FDI, makes it a technology-based economy of the minimal level.

Unlike India, Bangladesh is characterised with negative TFP growth rate (−1.14%), which is contributed by the negative technical efficiency (−1.4%), partly offset by very low growth of technological change (0.46%). Due to the decline in TEC induced by the lack of heavy industry, Bangladesh is the only country in this study to see TFP regression, as it has been declining constantly from 1993 to 2010 (Figure 3). After 2012, the reduction of TFPC is largely driven by the few incidents in the manufacturing sector (Hossain & Oh, 2019).⁸

TFP growth has slowed in Bangladesh, reflecting not just a stagnation of resources in sectors where productivity growth has been weaker (e.g., agriculture) but also diminishing productivity growth within sectors that now account for the majority of employment and economic activity. Overall, TC has been positive during the entire period. TFP growth has been slowed by the negative TEC. A developing country’s TEC must be positive, especially if growth is likely to be boosted by foreign investments. Thus, Bangladesh is characterised with low catching-up ability but moderate innovative strength. Investors are turned off by negative TEC. Increased efficiency can be achieved by properly allocating resources and reducing wasted inputs (Nugawela, 2019).

The Bangladesh economy is dominated by agriculture and manufacturing (particularly ready garments [RMG] sector). Both are labour-intensive sectors. Agriculture accounts for 13.32% of GDP, while RMG covers 12.36% of the GDP and 80% of total export income (BBS, 2020). About half of the population is directly engaged in agriculture, thus making it highly labour concentrated. Again, the government spends more than 8% of total expenditure on agriculture, which is one of the highest in the world. Still, TFP in the agricultural sector is comparatively low because over-usage of inputs increases the cost of agricultural output and declines the proportion of output to input that causes the TFP to fall (Zubayer, 2019).

Similarly, competitiveness of the RMG sector is very low when compared to other parts of the world. For example, productivity of this sector in Bangladesh is 77%, which is 92%, 90% and 88% in India, Vietnam and Pakistan, respectively. Again, minimum wage is US$69 in Bangladesh against US$71, US$78 and US$79 in India, Vietnam and Pakistan, respectively. Unskilled labour and low capital to labour ratio are the major reasons behind this situation (Hossain & Oh, 2019).

Though Bangladesh has achieved the highest ever growth rate of 8.2% in 2019 it has performed very poorly in terms of competitiveness. According to the GCI report,⁹ Bangladesh is ranked 113th among 152 countries in 2018, much lower than other counterparts like India and Vietnam.¹⁰ Export flow in Bangladesh is based on quantity rather than quality export. Bangladesh’s export is dominated by the low cost and low-quality garments sector. The Garment sector covers 84% of the total export. According to the six-digit HS code, 68% of total export earnings of the garment sector come from 10 varieties of RMG goods, whereas the statistics are 46% and 42% in the case of India and Vietnam.¹¹

On the other hand, there is a very few capital-intensive industries, including cement, chemical and light engineering. Despite a large population, TEC is declining due to mass inefficient less skilled labour. Due to the irrelevance of education and employment, the deterioration of higher education quality, insufficient allocation in the education sector, and ignorance of the information, science, technology, and engineering sectors, the current system is unable to develop qualified human capital. The ratio of total investment and FDI to GDP has grown quite well over the past 30 years, but it remains significantly lower than other regions of the world. Total investment climbed by 69% (17.68% in 1991 to 29.9% in 2020), while FDI increased by 84% (0.758 in 1991 to 1.659 in 2019).

In this study, TFP decline is entirely due to TEC’s negative growth, while TC remains relatively positive and partially offsets TFPC’s negative trend. The result implies that, though investment has increased slowly over the past three decades, it has been primarily inefficient, yielding modest returns.

All the South Asian countries (except India) have considered garments as their prime exporting products. Despite a large number of population, they are not able to utilise their inputs appropriately. Lack of sufficient complementary knowledge¹² can be one of the reasons. Moreover, insufficient backward linkage industry is a very crucial factor. For example, though garments constitute the export base in Bangladesh (and Sri Lanka), the necessary machineries and even cotton have to be imported from India and China.

Though most of the previous studies found negative TFP trend in Pakistan, in our study, we find TFP progress that can be characterised by the long-term reform process taken by the Pakistan government according to the prescription of the International Monetary Fund (IMF) and World Bank. Pakistan recoded the highest TFP growth rate (Figure 4) in South Asia due mainly to favourable technological change (1.5%). Despite the fact that innovative strength appears to be improving, it is actually being bolstered by the unaffected financial sector amid the Asian and global financial crises.

Figure 4.

Source: The author.

This TFP growth is bolstered by few noteworthy factors. One of them is FDI, which started rising since 2004 and reached at its peak (3.2% of GDP) in 2008, which amounted to about US$5.15 billion in absolute value. Compared to 2004 statistics, FDI in 2008 increased by 443%, which was completely led by privatisation procedure and greenfield investment (Khan & Khan, 2011). At that period, FDI concentrated mainly on telecommunication, finance, oil and gas sectors.

External forces operated as key external variables in conjunction with the boom in the manufacturing sector, which was aided by a surplus in power supplies (Afia et al., 2017). Despite the fact that investment in Pakistan is low, it remained modest until 2007, owing to increased domestic demand (Amjad & Awais, 2016).¹³ TFP growth grew before 2008 as a result of increased demand and new investment in cutting-edge technologies.

Since human capital and arable land are not included as inputs in our measurement of TFP, these two finally add to the residual, TFP, which may contribute to its positive growth. Pakistan also improved their ranking in GCI (moved from the position of 92nd in 2008 to 123rd in 2011) and ‘ease of doing business index’ (moved from the position of 85th in 2008 to 105th in 2011). However, low FDI for the majority of the study period due to lack of foreign investors’ trust and high input costs such as electricity, water and land, as well as investment in less productive labour-intensive industries, result in TFP reduction, partially offsetting TFP gain. Moreover, Pakistan’s economy relies heavily on the agricultural sector, accounting for 20% of GDP, but productivity of this sector is half of the other sector (Faiz, 2020). Labour-intensive technologies and export-oriented manufacturing sector have largely determined economic growth. Unlike India, Pakistan’s exports are neither quality-based nor quantity-based. Export diversification index is also showing a negative trend. All those factors weaken the TFP gain to a certain extent.

Like other South Asian countries, the TFP curve is volatile in Sri Lanka (Figure 5). TFP is increasing moderately right from the beginning and reached at the top position with the growth rate of 3.93% in 1999 (Figure 5). This enormous upsurge is boosted up by the unprecedented growth of FDI in 1997 when it grew from 0.43% in 1996 to 2.85% of GDP in 1997 (World Bank, 2021). Despite a 30-year long civil war and internal political upheaval, the overall TFP growth is impressive indeed (2.3%), led by TC progress, which is attributed to the huge FDI flow in the most productive sectors. Sri Lanka’s FDI concentrated on real estate, port and telecommunication sector. This is a consequence of a series of initiatives on reforms taken during the 1980s by the Sri Lankan government after opening the economy in 1977. After 2010, TFP starts increasing again due to the structural reconstruction programme (Konara, 2013).

Figure 5.

Source: The author.

During 1980–2016, TFPC of Sri Lanka is responsible for 45% of output growth, while the remaining 55% ascribed to capital and labour (Kumari & Tang, 2019). Although the TFPC for Sri Lanka is high, it is extremely volatile, ranging from -4.4% in 1998 to 3.9% in 1999, as seen in Figure 5. The significant volatility reflects Sri Lanka’s inability to maintain TFPC at a consistent level as a result of policy changes, domestic riots or exogenous shocks.

TFP growth during 1991–1995 was boosted between by a huge influx of FDI into the ready-made garments sector as a result of the ‘multi-fibre agreement’. Due to the rise of domestic violence, including a bomb attack on the central bank in 1996 and another explosion in Kandy in 1998, TFPC hit its lowest point in 1998, with a negative growth of −4.4% (Fernando, 1999). The negative impacts of Asian financial crisis, as well as a bomb explosion at one of Sri Lanka’s airports, compelled TFPC to decline slowly between 2000 and 2004. During the rest of the period, TFP growth remains quite consistent except in 2013. This was largely supported by the government-financed infrastructure investments in seaports, airports, electricity power generation and roads and highways. TFPC improvement is also the outcome of this large-scale capital formation.

Several characteristics stand out in this growth pattern of TFP depicted in Figure 5. During the 30 years of the study period, Sri Lanka experienced higher TFP growth with moderately high TC. This indicates that the factor inputs (capital and labour) are utilised more efficiently. The quality of labour improved through acquired knowledge, skill, experience and capacity of technology adoption. The efficiency of capital can be determined with the employment of high-tech equipment and innovation through research and development.

Nepal is a landlocked developing country. TFP in Nepal has also fluctuated over time (Figure 6). But Nepal managed to have a noteworthy TFP growth backed up by the technological progress and positive efficiency change.

Figure 6.

Source: The author.

TFP growth in capital-intensive industries was much lower during the post-liberalisation period (after 1990s), while industries with foreign investment tended to have higher productivity growth, and export-intensive industries had poor productivity performance. It can be said that trade liberalisation has only exacerbated the productivity of capital-intensive sectors, possibly because of incapacity with foreign competence production, and productivity growth has been observed to be much higher in large-scale and foreign-invested industries. At the same time, capital productivity growth in labour-intensive industries (such as readymade garments, carpets and textiles) is higher in Nepal (Regmi, 2006). Generally, low-tech sectors should have the lowest labour productivity, but in Nepal, high-tech manufacturing, shockingly, has lower labour productivity than mid-high-tech and mid-low-tech industries (Maliani, 2017). Conversely, high-tech, mid-high-tech, and mid-low-tech industries had positive capital productivity, on average, which is lower than that of low-tech manufacturing, meaning that the former has more value added with less investment. In summary, TFPC is higher in years with higher capital productivity, showing that TFPC is more tied to capital than labour.

Conclusion

In this study, we assess productivity changes and its two sources for five developing countries from 1991 to 2020. Using the non-parametric technique of DEA-type Malmquist index. This model assisted us in separating the contributions of technical change and efficiency change from TFP change.

Therefore, in summary, TFP is growing in standard technology-based economy (India), which is the consequence of technological progress. On the other hand, among labour-intensive economies, Sri Lanka is performing better as TFP in Sri Lanka is improving at a good rate, while Pakistan’s TFP is also growing steadily. Here, both the countries’ TFP growth is driven by positive technological change. Bangladesh is observing TFP decline as Bangladesh is the worst performer among five nations in South Asia despite the fact that technological change is still progressing. Nepal’s TFP is improving as well, but at a slower pace. Except Bangladesh, efficiency change shows progressive trend in the remaining countries.

TFP of all the South Asian countries are cyclical in nature, highly volatile and full of ups and downs as it rises suddenly, stays for a while and drops largely again, reflecting the inconstancy of any policy shift. South Asian countries have significantly higher capital productivity in capital-intensive industries, which accelerates technological progress. technical inefficiency in labour-intensive industries somewhat offsets TC progress.

The findings of this study will be of great interest to policymakers. Because productivity is the primary driver of GDP growth, emerging countries in South Asia may benefit the most from high productivity. Furthermore, productivity sources like TC and TEC are crucial for developing countries. Though South Asian nations are known for their labour-intensive technology, their performance in raising TEC is not always good, and their TC evaluation is sometimes less inefficient.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

Notes

ORCID iD

Sanjoy Kumar Saha

References

Adnan

, Chowdhury

, & Girijasankar

(2019). Foreign direct investment and total factor productivity in South Asia. Theoretical and Applied Economics , 2, 105–120.

Afia

, Ejaz

, & Musleh

(2017). An assessment of Pakistan’s export performance and the way forward (PIDE-Working Papers). Pakistan Institute of Development Economics.

Amjad

, & Awais

(2016). Pakistan’s productivity performance and TFP Trends, 1980–2015: Cause for real concern. The Lahore Journal of Economics , 21, 33–63. https://doi.org/10.35536/lje.2016.v21.isp.a3

Aubhik

(2009). Accounting for cross-country differences in income per capita. Business Review , (Q1), 11–18.

Bangladesh Bureau of Statistics (BBS). (2020). Statistical yearbook of Bangladesh . Bangladesh Bureau of Statistics, Ministry of Planning.

Brunner

H. P.

, & Cali

(2006). The dynamics of manufacturing competitiveness in South Asia: An analysis through export data. Journal of Asian Economics , 17(4), 557–582.

Caves

D. W.

, Christensen

, & Diewert

(1982). The economic theory of index numbers and the measurement of input, output, and productivity. Econometrica , 50, 1393–1414.

David

, Tan

, & Sadeghi

(2014). Attracting Foreign Direct Investment: What can South Asia’s lack of success teach other developing countries? South Asia Economic Journal , 15(2), 133–174, https://doi.org/10.1177/1391561414548939

Danquah

, & Amankwah-Amoah

(2017). Assessing the relationships between human capital, innovation and technology adoption: Evidence from Sub-Saharan Africa. Technological Forecasting and Social Change , 122, 24–33. https://doi.org/10.1016/j.techfore.2017.04.021

10.

Dutz

, & O’Connell

(2013). Productivity, innovation and growth in Sri Lanka: An empirical investigation (Policy Research Working Paper, no. 6354). World Bank. https://doi.org/10.11588/xarep.00003763

11.

Faiz

(2020, June 17). What happened to productivity across sectors in Pakistan over the years? Macro Pakistani. https://macropakistani.com/productivity-in-pakistan/

12.

Färe

, Grosskopf

, Norris

, & Zhang

(1994). Productivity growth, technical progress, and efficiency change in industrialized countries. American Economic Review , 84, 66–83.

13.

Färe

, Grifell-Tatjé

, Grosskopf

, & Lovell

(1996). Biased technical change and the malmquist productivity index. The Scandinavian Journal of Economics , 99(1), 119–127. https://dx-doi-org.web.bisu.edu.cn/10.1111/1467-9442.00051

14.

Fernando

(1999). Sri Lanka in 1998: Political stalemate and economic drift. Asian Survey , 39(1), 185–190. https://doi.org/10.2307/2645608

15.

Hall

, & Jones

(1999). Why do some countries produce so much more output per worker than others? Quarterly Journal of Economics , 114, 83–116.

16.

Headey

, Alauddin

, & Rao

(2010). Explaining agricultural productivity growth: An international perspective. Agriculture Economics , 41, 1–14.

17.

Hossain

, & Oh

(2019). Changes in total factor productivity and technical efficiency of Bangladesh RMG industry focused on Rana Plaza tragedy. Business Economics Research [경영경제연구], 41(4), 145–169.

18.

Jajri

(2007). Determinants of total factor productivity growth in Malaysia. Journal of Economic Cooperation , 28, 41–58.

19.

Javed

, & Kashif

(2016). Impact of export composition on economic growth in South Asia (MPRA Paper 71519). University Library of Munich.

20.

Johnes

(2006). Efficiency and productivity change in the English higher education sector from 1996/97 to 2002/03 (Working Papers 575369). Lancaster University Management School, Economics Department.

21.

Jones

, & Romer

(2010). The New Kaldor facts: Ideas, institutions, population, and human capital. American Economic Journal , 2(1), 224–245.

22.

Kavita

(2021, January 15). Enhancing factor productivity to drive growth in Timor-Leste . Asian Development Bank.

23.

Kalpana

, Kumar

, Rajan

, Subramanian

, & Tokatlidis

(2006, February). India Apos; S patterns of development: What happened, what follows . https://ssrn.com/abstract=883073

24.

Khan

, & Khan

(2011). Foreign direct investment and economic growth in Pakistan: A sectoral analysis (PIDE-Working Papers). Pakistan Institute of Development Economics.

25.

Klenow

, & Rodriguez-Clare

(1997). The neoclassical revival in growth economics: Has it gone too far? In National Bureau of Economic Research Macroeconomics Annual. MIT Press.

26.

Konara

M. P.

(2013). FDI in Sri Lanka: Determinants and impacts (PhD thesis). University of York.

27.

Kumar

, & Russell

(2002). Technological change, technological catch-up and capital deepening: Relative contributions to growth and convergence. American Economic Review , 92(3), 527–548.

28.

Kumari

, & Tang

(2019). Sri Lanka’s sources of growth: The application of primal and dual total factor productivity growth accounting approaches (Economics Discussion/Working Papers 19-05). The University of Western Australia, Department of Economics.

29.

, & Liu

(2005). Foreign direct investment and economic growth: An increasingly endogenous relationship. World Development , 33(3), 393–407.

30.

Maliani

P. K.

(2017). Productivity and growth of manufacturing industries in Nepal (PhD dissertation).

31.

Manjappa

, & Mahesha

(2008). Productivity performance of selected capital-intensive and labor-intensive industries in India during reform period: An empirical analysis. ICFAI Journal of Industrial Economics , 5(4), 57–65

32.

Nombulelo

(2007). Sri Lanka’s sources of growth (Working Papers 2007/225). International Monetary Fund.

33.

Nugawela

T. G.

(2019). Sri Lanka’s total factor productivity change during conflict and post-conflict periods. Staff Studies , 49(1), 1–19. http://doi.org/10.4038/ss.v49i1.4714

34.

Rao

D. S. P.

, & Coelli

(1999). Economic growth, productivity change and inequality: Methodology for the assessment of economic performance of nations . University of New England/Department of Econometrics/Centre for Efficiency and Productivity Analysis.

35.

Regmi

U. R.

(2006). Output growth, total factor productivity, and technical efficiency in Nepalese manufacturing industries. The Economic Journal of Nepal , 114(29), 85–92.

36.

Saha

(2014). Total factor productivity trends in India: A conventional approach. The NEHU Journal , 12(1), 95–106.

37.

Simar

, & Wilson

(1998). Productivity growth in industrialized countries (Discussion paper no 9810). Universite Catholique de Louvain.

38.

Sinha

(2017). Long-term growth scenarios for Bangladesh (Policy Research Working Paper No. 7952). World Bank. https://ssrn.com/abstract=2905442

39.

Srinivasan

(2005). Productivity and economic growth in South Asia and China. The Pakistan Development Review , 44(4), 479–503. https://www-jstor-org.web.bisu.edu.cn/stable/41260729

40.

World Bank. (2021). World development indicators . World Bank.

41.

Zubayer

(2019, October 18). Boosting the low productivity in agriculture sector in Bangladesh. Daily Financial Express . https://thefinancialexpress.com.bd/views/boosting-the-low-productivity-in-agriculture-sector-in-bangladesh-1571412384