Attribution of Changes in Vietnam’s Labor Productivity

Abstract

This study examines the change in labor productivity in Vietnam by means of a Fisher index decomposition and attribution analysis. The results can be summarized as follows. First, the aggregate labor productivity is decomposed into pure labor productivity and structural change from 2007 to 2019. All of the aggregate labor productivity, pure labor productivity, structural change, and interaction terms have increased by 69.83%, 36.74%, 24.20%, and 8.89%, respectively. Second, the percentage change in labor productivity is attributed to 20 sub-industries by pure labor productivity and structural change. The sum of the multi-period attribution of pure labor productivity and structural change shows that the manufacturing industry positively dominates (15.84%) and plays a key role in economic development. The positive pure labor productivity and structural change in the manufacturing industry imply that the structural bonus hypothesis does hold in the industry. The findings also indicate that pure labor productivity, especially in the service industry, should be improved to sustain economic growth.

1. Introduction

As the United Nations’ [1] Sustainable Development Goals take into account economic growth as being a high priority around the world, attaining and sustaining high economic growth are the two most challenging economic goals for most developing countries. By accomplishing noteworthy economic growth, Vietnam has realized at least one of these goals. Since fully integrating into the global economy by joining the World Trade Organization (WTO) in 2007, the Vietnamese economy has grown at a compound annual growth rate (CAGR) of 6.18% from 2007 to 2019. It has experienced positive economic growth: it experienced 2.91% growth in 2020, whereas most countries experienced a sharp decline because of the ongoing COVID-19 pandemic. As a result, per capita GDP (constant 2010 US$) increased 1.85 times from US$1145.15 in 2007 to US$2082.24 in 2019, reaching the status of a lower-middle-income country in 2010 (World Bank [2]). Motivated by its economic success, Vietnam established an ambitious long-term development plan to become a high-income or upper-middle-income country by 2035 (World Bank and Ministry of Planning and Investment of Vietnam [3]).

In order to reach its planned status as a high-income country, Vietnam must maintain a more than 7% annual growth rate (Fukuoka [4]). However, Vietnam previously experienced slow economic growth, 5.55%, during the post-global financial crisis recovery phase between 2011 and 2014. Even its economic performance in 2020 (2.91%) and 2021 (2.58%) during the COVID-19 pandemic should be taken into account as outliers. Vietnam has already deviated from its sanguine planned path, except for a 7.08% growth rate in 2018 and 7.02% in 2019 (General Statistical Office of Vietnam [5]) This raises concerns about the sustainability of Vietnam’s economic growth and its ability to reach the upper level. Many experts have suggested that Vietnam may be trapped in the early stages of being a middle-income country or is at least on the verge of falling into such a trap (World Bank [6]; Ohno [7]; Klinger-Vidra and Wade [8]). Typically, Vietnamese economic growth has largely relied on inputs together with the inflows of foreign direct investment (FDI) and a plentiful low-cost labor force. Ohno [7] diagnosed the slowdown of Vietnamese economic growth as investment-driven growth with little improvement in productivity. In addition, the dichotomized FDI invested and domestic sectors make the spillover effects from FDI invested firms to domestic firms unfasten, which decreases the efficiency of investment. Wages that are rising faster than labor productivity continue to challenge the comparative advantage of Vietnam. Consequently, the economic growth—that to a great extent depends on production factor accumulation without being in tune with productivity—will eventually reach its limit.

This study aims to investigate the sustainability of economic growth in Vietnam by focusing on labor productivity growth from 2007 to 2019. This study makes several contributions. First, it employs index decomposition analysis (IDA) as a new approach to examine Vietnamese labor productivity. Second, this study expands the existing aggregate labor productivity measurement by adding disaggregated measurements across 20 industries. Existing studies on Vietnamese labor productivity have typically focused on aggregate labor productivity and productivity of industries in broad industrial classifications. It is essential to apply sectoral analysis in order to figure out the labor productivity of Vietnam as a developing country, in which the economic structure changes drastically, and the role of industry varies accordingly. Third, the impact of WTO accession on Vietnamese labor productivity has been much less frequently studied. This study covers the period 2007–2019, which is the period of the Vietnamese economy’s full participation in the global economy by joining the WTO.

By looking at the sustainability of Vietnamese economic growth, this study formulates three hypotheses. The first hypothesis is that labor productivity growth has contributed to overall economic growth. In other words, Vietnamese economic growth stems from both accumulations of labor input factors and labor productivity growth. The second hypothesis is that the manufacturing industry dominates service industries in terms of labor productivity growth. The third hypothesis is that there is a positive relationship between economic growth and structural change based on labor reallocation from lower value added per labor input industries to higher value added industries. This hypothesis is called the structural bonus hypothesis.

This study is organized as follows. Section 2 is a literature review, and Section 3 introduces the methodology of the Fisher index decomposition. Section 4 explains the data and presents empirical results. Section 5 provides a summary and a conclusion.

2. Literature Review

Economic growth stems from increases in factors of production and productivity growth. It is a stylized fact that productivity plays a key role in sustainable long-run economic development and is a main determinant of per capita growth (Krugman [9]). As an economy of a country grows, it relies more on increasing productivity rather than factors of production due to the diminishing effectiveness (Owyong [10]; Kim et al. [11]). Weil [12] also demonstrates that productivity growth is the most important factor in explaining the differences in economic growth rates among countries. Economic growth rates are composed of two factors: productivity growth and factor growth. The former, interestingly, is 65% and the latter only 35% of the economic growth rate. Nordhaus [13]) Hartwig [14], and Oh and Kim [15] show that total factor productivity growth has led to an increase in real value added.

To have a sustainable economic growth large enough to reach a high-income country as planned, Vietnam is required to have not only the proportionated accumulation of production factors such as a sufficient pool of qualified labor forces and investment in physical capital, as well as human capital from domestic sources and foreign direct investment (FDI), and, more importantly, management of the efficiency of an economy that includes an increase in productivities. In examining sustainable economic growth, labor productivity, among various productivity measures, is widely used because it is clearly identified, and data on it are more readily available (Sargent and Rodriguez [16]).

There are a number of existing studies that discuss the importance of labor productivity on economic growth. Several studies investigate the contribution of labor productivity to Vietnam’s economic growth. Breu et al. [17] estimated that an increase in labor productivity within some sectors contributed to 34% of the GDP growth from 2005 to 2010. Ohno et al. [18] observed that the trend in labor productivity was closely related to Vietnam’s economic growth from 1991 to 2005, pointing out that labor productivity stagnation during the period 1996 to 2012 resulted in the slowdown of economic growth. Dong et al. [19] identified the contributions of different sectors to economic development between 1996 and 2017. Dong et al. [20] examined the relationship between labor productivity and the global competitiveness of Vietnam. Le et al. [21] examined the effect of FDI and human capital on labor productivity in Vietnam.

The key elements affecting labor productivity are physical capital, institutional infrastructures, human capital, and technology changes (ILO [22]). In the standard microeconomic theory, wage and employment correspond to the productivity of labor in the short run. In the endogenous growth theory (Romer [23]), human capital and endogenous technological development enhance labor productivity, which is recognized as the main driver of sustainable long-run economic growth. More recently, labor market integration improves labor productivity and the resilience to market disturbances, which are relevant to the sustainability of the economy (Oostendrop [24]; Moro et al. [25]).

Along with these elements, the path of economic development entails structural changes in the economy. As Lewis [26] highlighted in the initial literature of growth theory, the economy in the process of economic growth has reallocated a substantial share of employment from the traditional agricultural sector to the modern manufacturing and service sectors because of productivity differentials between the dichotomized sectors. It is observed that the movement from low productivity sectors to high productivity sectors enhances the overall productivity and plays an important contributor to economic growth (Baumol [27]). There are ample studies on the labor productivity analysis incorporating the structural changes (Timmer and Szirmai [28]; McMillan and Rodrik [29]; McMillan et al. [30]; Lee and Mckibbin [31]; Konte et al. [32]).

Following a typical path of growth as a developing country, Vietnam has experienced a rapid structural change in the labor market, accompanied by a corresponding fast-growing economy. The contribution of agriculture, forestry, and fishing to GDP has declined distinctly from 24.5% in 2000 to 18.7% in 2007 to 13.9% in 2019. On the other hand, the share of the industry and service sectors has steadily increased correspondingly (World Bank [2]) As a result, the employment structure also has shifted toward the industry and service sectors from agriculture, forestry, and fishing sector. The contribution of employment in the agriculture, forestry, and fishing sectors has dramatically reduced from 52.9% in 2007 to 34.5% in 2019 (General Statistical Office of Vietnam [5]) Together with a young and abundant working-age population, capital inflows from FDI and export-oriented growth strategy have contributed to changes in labor productivity (McCraig and Pavcnik [33] examined labor allocation across the informal and formal sectors in Vietnam). Driven by these changes, there is a need to scrutinize the effect of structural change on labor productivity along with pure labor productivity. A decomposition analysis can be applied to gain meaningful insights into labor productivity in Vietnam.

Many studies of decomposition analysis on individual countries or a group of countries have been conducted, including Beebe and Haltmaier [34] on the U.S., Peneder [35] on the OECD, Lee and Mckibbin [31] and McMillan and Rodrik [29] on Asian countries, McMillan et al. [30] on African countries, and Te and Dong [36], Ohno et al. [18], Intrarakumnerd and Sunami [37], and Dong et al. [19] on Vietnam. These studies on labor productivity decomposition were based on the shift-share decomposition method. However, Balk [38] demonstrates that the shift-share analysis does not pass the global monotonicity test, the linear homogeneity test, and the time-reversal test. Instead, IDA, which is based on the index number theory, passes the three tests. This puts IDA in the lead within the recent trend in the index number practice of using the superlative index (The definition of a superlative index is from Diewert [39]. However, the original definition is rather complicated and indirect since it is based on the two different concepts of flexible, functional form, and exact index number. Barnett and Choi [40] provide an easier and more intuitive definition: the superlative index can provide a second-order approximation to an arbitrary aggregator function) and chain computation (considering that index numbers are a discrete approximation of the continuous-time Divisia index, the chain index is the preferred method). That is why this study adopts the IDA approach.

3. Materials and Methods

3.1. Data

This study measures labor productivity as real value added divided by employment (This is measured as persons engaged, which is the sum of employees, self-employed, and unpaid family workers). The sample period is from 2007 to 2019, and the data frequency is annual. The source of employment data is the General Statistical Office of Vietnam [5]. The real GDP data is taken from the Asia Development Bank [41]. This study calculates Vietnam’s labor productivity in terms of Vietnam’s unit of currency (Vietnamese Dong: VND), which is intended to eliminate the effect of changes in currency exchange rates. The number of industries is 20.

3.2. Fisher Index Decomposition of Aggregate Labor Productivity

Let $L$ be the total number of workers and $Y$ be the total real value added of sectors. Let us assume that there are N sectors, and $L_i$ $Y_i$ are, respectively, the number of workers and the real value added of sector i. The sectoral labor productivity is defined as $P_i=Y_i/L_i$ and labor share $S_i=L_i/L$ for sector i. The aggregate labor productivity $P=Y/L$ can be written as the following identity:

$$P={\displaystyle \sum _{i=1}^N\frac{L_i}{L}}\frac{Y_i}{L_i}={\displaystyle \sum _{i=1}^N{S}_i{P}_i}$$

(1)

Taking a logarithmic differentiation (denoted by a hat (^) over the variables) for both sides of (1), we obtain the initial identity of IDA,

$$\widehat{P}={\displaystyle \sum _{i=1}^N{w}_i\left({\widehat{S}}_i+{\widehat{P}}_i\right)},\hspace{1em}\hspace{1em}{w}_i=\frac{Y_i}{Y}$$

(2)

Integrating and taking the exponents of both sides, we obtain the prototype IDA of the aggregate labor productivity,

$$\frac{P_T}{P_0}=S_{0,T}\times P_{0,T},$$

(3)

where

$$S_{0,T}\equiv \exp \left({\displaystyle {\int }_0^T{\displaystyle \sum _{i=1}^N{w}_i{\widehat{S}}_{i}dt}}\right),\mathrm{and}$$

(4)

$$P_{0,T}\equiv \exp \left({\displaystyle {\int }_0^T{\displaystyle \sum _{i=1}^N{w}_i{\widehat{P}}_{i}dt}}\right),$$

(5)

The term $S_{0,T}$ is the Divisia structural change index that aggregates the labor share change when sectoral labor productivities are all zero, i.e., ${\widehat{P}}_i=0,\hspace{0.17em}(i=1,2,\cdots ,N)$. In the same manner, the other term $P_{0,T}$ is the Divisia pure labor productivity index that aggregates the sectoral labor productivity changes, i.e., ${\widehat{S}}_i=0,\hspace{0.17em}(i=1,2,\cdots ,N)$.

The prototype Divisia index numbers can be approximated by certain discrete time-chained index numbers. A recent trend in IDA, as well as in other applications, is to use superlative index numbers, such as the Fisher index. Some examples are the statistical agencies in the U.S. and Canada, which have adopted the Fisher index to compile real GDP. In the IDA, the logarithmic mean Divisia index has been preferred among numerous decomposition techniques. It has two versions, the Montgomery–Vartia index, and the Sato–Vartia index. Choi and Ang [42] chose the latter, considering its linear homogeneity required by the attribution analysis in the next stage. Since the logarithmic mean indexes are based on logarithmic change, a transformation is required, such as Reinsdorf et al. [43] or Balk [44].

• Single-period Fisher index decomposition

Through the Fisher index decomposition, we express the change in the aggregate labor productivity in (3) between two consecutive times [t − 1, t] in the following form:

$$\frac{P_t}{P_{t-1}}=S_{t-1,t}^F\times P_{t-1,t}^F,$$

(6)

where $S_{t-1,t}^F$ is the structural change index and $P_{t-1,t}^F$ is the pure labor productivity index. The terms on the right-hand side of (6) are computed using sectoral data with functional forms as follows:

$$S_{t-1,t}^F\equiv \sqrt{\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t-1}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t-1}}}\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t}}}},$$

(7)

$$P_{t-1,t}^F\equiv \sqrt{\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t-1}}}\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t-1}}}}.$$

(8)

2.

Multi-period Fisher index decomposition

The multi-period Fisher decomposition of aggregate labor productivity index can be defined using the chain method as follows,

$$\frac{P_T}{P_0}=S_{0,T}^F\times P_{0,T}^F,\hspace{1em}\hspace{1em}\because \frac{P_T}{P_0}={\displaystyle \prod _{t=1}^T\frac{P_t}{P_{t-1}}}={\displaystyle \prod _{t=1}^T\left(S_{t-1,t}^F\times P_{t-1,t}^F\right)}={\displaystyle \prod _{t=1}^T{S}_{t-1,t}^F}\times {\displaystyle \prod _{t=1}^T{P}_{t-1,t}^F}$$

(9)

where the right-hand side is the cumulative product of single-period indices as below:

$$S_{0,T}^F\equiv {\displaystyle \prod _{t=1}^T{S}_{t-1,t}^F}$$

(10)

and

$$P_{0,T}^F\equiv {\displaystyle \prod _{t=1}^T{P}_{t-1,t}^F}$$

(11)

3.3. Attribution Analysis of the Fisher Index Decomposition

The growth rate of an index is just the index number minus unity. Using the growth rates of indexes, the single-period Fisher index decomposition (6) can be rewritten as follows:

$$\left(1+r_{t-1,t}\right)\equiv \left(1+r_{t-1,t}^S\right)\times \left(1+r_{t-1,t}^P\right),$$

(12)

where we use the definitions as below:

$$r_{t-1,t}\equiv \frac{P_t}{P_{t-1}}-1,$$

(13)

$$r_{t-1,t}^S\equiv S_{t-1,t}^F-1,$$

(14)

$$r_{t-1,t}^P\equiv P_{t-1,t}^F-1.$$

(15)

Rearranging Equation (12), we obtain the identity of the attribution analysis:

$$r_{t-1,t}\equiv r_{t-1,t}^S+r_{t-1,t}^P+r_{t-1,t}^S\times r_{t-1,t}^P.$$

(16)

The above identity is directly comparable with the identity (17) of the shift-share analysis in Beebe and Haltmaier [34], Peneder [35], and Casler [45] in the notation of this paper,

$$\frac{P_t}{P_{t-1}}-1={\displaystyle \sum _{i=1}^N\frac{Y_{i,t-1}}{Y_{t-1}}}\left(\frac{S_{i,t}}{S_{i,t-1}}-1\right)+{\displaystyle \sum _{i=1}^N\frac{Y_{i,t-1}}{Y_{t-1}}}\left(\frac{P_{i,t}}{P_{i,t-1}}-1\right)+{\displaystyle \sum _{i=1}^N\frac{Y_{i,t-1}}{Y_{t-1}}}\left(\frac{S_{i,t}}{S_{i,t-1}}-1\right)\left(\frac{P_{i,t}}{P_{i,t-1}}-1\right)$$

(17)

where $Y_{i,t}$ is the real value added of sectors and $Y_t$ is the sum of sectoral values added, and the terms on the right-hand side of (17) correspond to $r_{t-1,t}^S$, $r_{t-1,t}^P$ and $r_{t-1,t}^S\times r_{t-1,t}^P$, respectively.

For the fixed-weight Laspeyres index, the attribution analysis is trivial since the Laspeyres index is an arithmetic mean index. After the Fisher chain index was introduced to the national accounting practices in 1996, several methods were proposed to transform the Fisher index into an arithmetic index. The solution was found in van IJzeren [46], which is being used in the U.S. and Canada.

Similarly, Reinsdorf et al. [43] devised a method to transform a geometric mean index such as the Sato–Vartia index into an arithmetic mean index. On the other hand, Balk [44] provided a transformation in the opposite direction. Combining the two studies, Choi and Ang [42] derived a definition for the multi-period attribution analysis. However, it may be preferable to use the superlative Fisher index, which has an alternative arithmetic mean form.

1.

Single-period attribution of Fisher index decomposition

The two Fisher indexes of pure labor productivity and structural change have alternative (van Ijzeren) arithmetic mean forms, presented as follows:

$$P_{t-1,t}^F\equiv \sqrt{\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t-1}}}\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t-1}}}}\equiv \frac{{\displaystyle \sum _{i=1}^N\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t}}}{{\displaystyle \sum _{i=1}^N\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t-1}}},$$

(18)

$$S_{t-1,t}^F\equiv \sqrt{\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t-1}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t-1}}}\frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}{P}_{i,t}}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}{P}_{i,t}}}}\equiv \frac{{\displaystyle \sum _{i=1}^N{S}_{i,t}\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)}}{{\displaystyle \sum _{i=1}^N{S}_{i,t-1}\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)}},$$

(19)

For each index number, there is a corresponding growth rate. The attribution analysis is based on the additive decomposition of arithmetic mean indexes, as follows:

$$r_{t-1,t}^P\equiv P_{t-1,t}^F-1={\displaystyle \sum _{i=1}^N\frac{\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t-1}}{{\displaystyle \sum _{i=1}^N\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t-1}}}\left(\frac{P_{i,t}}{P_{i,t-1}}-1\right)},$$

(20)

$$r_{t-1,t}^S\equiv S_{t-1,t}^F-1={\displaystyle \sum _{i=1}^N\frac{\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)S_{i,t-1}}{{\displaystyle \sum _{i=1}^N\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)S_{i,t-1}}}\left(\frac{S_{i,t}}{P_{i,t-1}}-1\right)},$$

(21)

where $r_{t-1,t}^P$ and $r_{t-1,t}^S$ are the growth rates of pure productivity and structural change during the period [$t-1,t$]. Now the identities of a single-period attribution analysis can be written as:

$$r_{t-1,t}^P={\displaystyle \sum _{i=1}^N{c}_{t-1,t}^{P,i}},$$

(22)

$$r_{t-1,t}^S={\displaystyle \sum _{i=1}^N{c}_{t-1,t}^{i,S}}.$$

(23)

where $c_{t-1,t}^{P,i}$ and $c_{t-1,t}^{S,i}$ are the contributions of the $i$th component to the growth rate of productivity and structural change $r_{t-1,t}^P$ and $r_{t-1,t}^S$, respectively, during the period [$t-1,t$],

$$c_{t-1,t}^{P,i}\equiv \frac{\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t-1}}{{\displaystyle \sum _{i=1}^N\left(S_{i,t-1}{S}_{t-1,t}^F+S_{i,t}\right)P_{i,t-1}}}\left(\frac{P_{i,t}}{P_{i,t-1}}-1\right),$$

(24)

$$c_{t-1,t}^{S,i}\equiv \frac{\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)S_{i,t-1}}{{\displaystyle \sum _{i=1}^N\left(P_{i,t-1}{P}_{t-1,t}^F+P_{i,t}\right)S_{i,t-1}}}\left(\frac{S_{i,t}}{P_{i,t-1}}-1\right).$$

(25)

2.

Multi-period attribution of Fisher index decomposition

According to the definition of chained pure labor productivity index given by (11), the following is an identity:

$$\begin{array}{cc}{r}_{0,t}^P& \equiv P_{0,t}^F-1=P_{0,t}^F-P_{0,0}^F=\left(P_{0,t}^F-P_{0,t-1}^F\right)\hspace{0.17em}+\left(P_{0,t-1}^F-P_{0,t-2}^F\right)+\cdots +\left(P_{0,1}^F-P_{0,0}^F\right)\hspace{1em}\because P_{0,0}^F\equiv 1\hfill \\ & =P_{0,t-1}^F\left(P_{t-1,t}^F-1\right)\hspace{0.17em}+P_{0,t-2}^F\left(P_{0,t-1}^F-1\right)+\cdots +P_{0,0}^F\left(P_{0,1}^F-1\right)\hfill \\ & ={\displaystyle \sum _{s=1}^t{P}_{0,s-1}^F{r}_{s-1,s}^P},\hspace{1em}\because \hspace{0.17em}{r}_{s-1,s}^P\equiv P_{s-1,s}^F-1\hfill \end{array}$$

(26)

Inserting the definition of a single-period attribution (22) into (26), we obtain the following definition of the multi-period attribution analysis of pure labor productivity in the Fisher chain indices:

$$r_{0,t}^P={\displaystyle \sum _{s=1}^t{P}_{0,s-1}^F{\displaystyle \sum _{i=1}^N{c}_{s-1,s}^{P,i}}\hspace{0.17em}}\equiv {\displaystyle \sum _{i=1}^N{c}_{0,t}^{P,i}},$$

(27)

where

$$c_{0,t}^{P,i}\equiv {\displaystyle \sum _{s=1}^t{P}_{0,s-1}^F}{c}_{s-1,s}^{P,i}.$$

(28)

The value $P_{0,s-1}^F{c}_{s-1,s}^{P,i}$ in (28) is the contribution of the $i$th component during the period [$s-1,s$] evaluated from time 0. Using exactly the same procedure, we obtain the following definition of the multi-period attribution analysis of structural change in Fisher chain indices:

$$r_{0,t}^S\equiv {\displaystyle \sum _{i=1}^N{c}_{0,t}^{S,i}}$$

(29)

where

$$c_{0,t}^{S,i}\equiv {\displaystyle \sum _{s=1}^t{P}_{0,s-1}^F}{c}_{s-1,s}^{S,i}.$$

(30)

4. Results and Discussion

4.1. Fisher Index Decomposition of Aggregate Labor Productivity

Table 1. Fisher index decomposition of Vietnam’s labor productivity change at the aggregate level (Unit: %).

Year	Single-Period Analysis (Base = Previous Year)				Multi-Period Analysis (Base = 2007)
	Pure	Structure	Interaction	Total	Pure	Structure	Interaction	Total
2008	2.53	0.27	0.01	2.81	2.53	0.27	0.01	2.81
2009	0.33	2.23	0.01	2.57	2.87	2.51	0.07	5.45
2010	−0.15	3.75	−0.01	3.59	2.71	6.35	0.17	9.23
2011	1.16	2.30	0.03	3.49	3.90	8.80	0.34	13.05
2012	1.40	1.63	0.02	3.06	5.36	10.58	0.57	16.50
2013	3.92	−0.08	0.00	3.84	9.49	10.49	0.99	20.97
2014	4.65	0.24	0.01	4.90	14.58	10.75	1.57	26.90
2015	4.27	1.60	0.07	5.95	19.48	12.53	2.44	34.45
2016	3.47	2.20	0.08	5.74	23.62	15.00	3.54	42.17
2017	4.79	1.24	0.06	6.09	29.54	16.43	4.85	50.83
2018	3.75	2.12	0.08	5.95	34.40	18.90	6.50	59.80
2019	1.74	4.46	0.08	6.28	36.74	24.20	8.89	69.83

illustrates the results of the Fisher index decomposition of aggregate labor productivity change in Vietnam. The total multi-period labor productivity index indicates a cumulative sum over the time period increased from 1 in 2007 as the base year to 1.6983 in 2019. As shown in the table, the total labor productivity increases 69.83% over a 13-year period. This result is consistent with the first hypothesis that labor productivity growth has contributed to overall economic growth in Vietnam.

Meanwhile, the single-period total labor productivity index denoting a year-over-year increase shows that it improved the most in 2019 (6.28%), followed by 6.09% in 2017 and 5.95% in 2018. This implies that the growth of labor productivity has accelerated in recent years, which is consistent with the results of Ohno et al. [18] and Intrarakumnerd and Sunami [37], but inconsistent with Dong et al. [19], which finds that aggregate labor productivity in 2011–2015 was higher than in 2016 and 2017.

By decomposing aggregate labor productivity, the value of multi-period pure labor productivity increased from 1 in 2007 to 1.3674 in 2019, meaning that pure labor productivity improved by 36.74%, as shown in the table during the period. The increase in single-period pure labor productivity shows that it increased the most in 2017, followed by 2014 and 2015. In the case of the structural change index, the value of the multi-period index increased from 1 in 2007 to 1.2420 in 2019, implying that it improved 24.20% during the time span. It is worth noting that pure labor productivity contributes 52.61% to total labor productivity improvement, which is greater than the contribution from the structural change accounting for 34.66% during the period studied. The single-period structural change index increased the most in 2019, followed by 2010 and 2011.

In the early years of accession to the WTO, the effect of restructuring the overall economy from a comprehensive liberalization process seems to have played a major role in improving aggregate labor productivity. However, the contribution has reversed since 2013, except in 2019. In the same manner, Intrarakumnerd and Sunami [37] pointed out that, applying the shift-share decomposition analysis, the contribution of labor productivity growth from the structural change was 53.5% of the aggregate labor productivity growth from 2005 to 2010. However, during the period of 2010 to 2018, the contribution of the pure labor productivity component to total labor productivity was 66.8%. Te and Dong [36] also mentioned that structural change played a dominant role in the growth of total labor productivity, but pure labor showed a negative contribution during the period of 2008 to 2012.

According to the results, the global financial crisis did not negatively affect either pure labor productivity or structural change at an aggregate level. Both indices were positive in 2008 and 2009. This can be explained by the fact that, in the initial phase of WTO accession, Vietnam was not fully integrated into global financial markets, so the external global financial crisis did not have as serious of an impact.

The interaction term also shows improvement, and its multi-period contribution from 2007 to 2019 is 8.89%, which denotes a relatively insignificant contribution compared to those of structural and pure labor productivity. Our results of the interaction term are generally consistent with other studies (Ohno et al. [18]; Intrarakumnerd and Sunami [37]; Dong et al. [19]). The positive interaction term can be interpreted as employment structural change, which moves according to a change in productivity. The positive interaction term is also considered to be a sectoral spillover effect of technological innovation in a progressive industry (progressive industry refers to those whose productivity growth rate is above the economy’s average). By comparing structure and interaction terms, the structural change index is always larger than the interaction term.

4.2. Attribution of the Fisher Decomposition of Labor Productivity by Industry

Like Choi and Ang [42] and Choi and Oh [47], this study attributes the percentage of changes in the estimated pure labor productivity index quantitatively to the 20 industries in Vietnam. The detailed results provide meaningful insights into understanding the role and contribution of industry to labor productivity growth for the rapidly growing economy of Vietnam. One of the advantages of attribution analysis is calculating the multi-period contribution by industry.

Table 2. Multi-period contribution of pure labor productivity (base = 2007; unit: %).

Year	in1	in2	in3	in4	in5	in6	in7	in8	in9	in10	in11	in12	in13	in14	in15	in16	in17	in18	in19	in20	sum
2008	0.62	−0.15	0.65	0.02	0.11	−0.27	0.40	0.07	−0.18	−0.03	0.17	0.64	−0.15	−0.01	0.18	0.21	0.13	−0.17	0.24	0.02	2.53
2009	0.76	0.62	−0.18	0.32	0.14	0.09	1.25	0.17	−0.49	−0.07	−0.03	−0.65	−0.33	−0.02	0.45	0.22	0.21	−0.22	0.62	0.01	2.87
2010	3.92	2.37	−2.54	0.88	0.11	0.10	−2.80	0.92	−0.23	−0.11	0.54	−2.72	−0.09	0.00	1.07	0.58	0.22	−0.16	0.64	0.01	2.71
2011	4.61	2.49	−1.30	0.96	0.23	−0.21	−2.35	1.22	−0.52	−0.08	−0.12	−3.64	−0.02	0.00	1.32	0.66	0.19	−0.17	0.61	0.04	3.90
2012	5.04	2.69	−0.31	1.71	0.27	−0.13	−2.26	1.24	−0.58	−0.04	−0.06	−5.14	−0.12	−0.04	1.43	0.78	0.26	−0.14	0.71	0.05	5.36
2013	5.57	3.38	0.53	1.92	0.32	0.19	−1.92	1.58	−0.51	0.01	−0.08	−5.07	0.00	−0.04	1.57	0.94	0.34	−0.13	0.82	0.06	9.49
2014	6.34	4.24	1.54	2.27	0.36	0.68	−1.16	1.80	−0.45	0.03	−0.01	−5.23	0.11	−0.04	1.68	1.09	0.44	−0.11	0.91	0.08	14.58
2015	7.99	5.96	0.89	2.49	0.33	1.54	−0.30	1.55	−0.53	0.03	0.13	−5.23	0.29	−0.07	1.92	1.30	0.39	−0.02	0.75	0.06	19.48
2016	9.19	6.20	1.87	2.65	0.31	0.98	0.68	1.98	−0.28	0.24	0.17	−6.12	0.54	0.02	2.13	1.53	0.44	0.09	0.93	0.07	23.62
2017	10.51	5.54	3.82	3.45	0.39	1.26	1.53	2.03	0.04	0.32	1.28	−7.05	0.52	−0.02	2.44	1.53	0.68	0.11	1.08	0.10	29.54
2018	12.36	5.96	5.77	3.49	0.36	1.61	1.87	2.33	−0.06	0.55	1.21	−7.76	0.41	−0.03	2.74	1.61	0.61	0.23	1.02	0.10	34.40
2019	14.51	6.10	5.35	3.12	0.33	1.70	3.14	0.92	1.47	0.47	0.45	−8.55	0.30	−0.02	3.40	2.09	0.66	0.31	0.90	0.09	36.74
Rank	1	2	3	6	15	8	5	10	9	13	14	20	17	19	4	7	12	16	11	18

Note: Rank = rank by the value of 2019. Note: in1: Agriculture, forestry, and fishing, in2: Mining and quarrying, in3: Manufacturing, in4: Electricity, gas, stream, and air conditioning supply, in5: Water supply, sewage, waste management and remediation activities, in6: Construction, in7: Wholesale and retail trade; repair of motor vehicles and motorcycles, in8: Transportation and storage, in9: Accommodation and food service activities, in10: Information and communication, in11: Financial, banking and insurance activities, in12: Real estate activities, in13: Professional, scientific and technical activities, in14: Administrative and support service activities, in15: Activities of the Communist Party, socio-political organizations; public administration and defense; compulsory security, in16: Education and training, in17: Human health and social work activities, in18: Arts, entertainment and recreation, in19: Other service activities, in20: Activities of households as employers; undifferentiated goods and services-producing activities of households for own use.

demonstrates the multi-period contribution of pure labor productivity by industry. The last column, “sum,” results in the same value as the total pure labor productivity percentage change at an aggregate level in Table 1 as a matter of course. The base year is 2007, and the results are measured based on Equation (28).

There are several findings of interest in the results. First, all industries, except the real estate activities industry and administrative and support service activities industry, in the multi-period contribution of pure labor productivity in 2019 are positively growing. The Vietnamese real estate activities industry experienced a boom in the early years of the WTO accession, between 2007 and 2008, due to high expectations of ongoing economic growth and massive inflows of FDI to the real estate activities industry. However, right after Vietnam’s entry into the WTO, Vietnam faced the global financial crisis. As has been mentioned earlier, this was less severe at an aggregate level, but the real estate activities industry was one of the more significantly influenced sectors. As a result, the abnormally high labor productivity of the real estate activities industry during the boom period declined sharply (−9.6%) because of the rapid inflows of labor and the stagnation of the real estate market during the period, as presented in

Table A1. Labor productivity by industry in Vietnam (Unit: Million VDN/person).

Industry	2007	2013	2019	CAGR (2007–2019, %)
Agriculture, forestry, and fishing	15.5	20.2	30.6	5.8
Mining and quarrying	681.2	931.0	1198.5	4.8
Manufacturing	55.9	57.9	71.1	2.0
Electricity, gas, stream, and air conditioning supply	442.8	751.0	953.3	6.6
Water supply, sewage, waste management, and remediation activities	87.3	150.1	153.0	4.8
Construction	48.1	49.5	59.8	1.8
Wholesale and retail trade; repair of motor vehicles and motorcycles	46.3	38.8	57.1	1.8
Transportation and storage	49.8	72.9	63.9	2.1
Accommodation and food service activities	45.1	38.1	57.7	2.1
Information and communication	98.3	97.9	133.4	2.6
Financial, banking, and insurance activities	477.5	475.7	513.6	0.6
Real estate activities	2199.6	1064.5	657.1	−9.6
Professional, scientific, and technical activities	158.3	155.5	186.4	1.4
Administrative and support service activities	49.1	44.9	46.8	−0.4
Activities of the Communist Party, socio-political organizations; public administration and defense; compulsory security	27.0	47.3	75.5	8.9
Education and training	27.1	38.9	53.8	5.9
Human health and social work activities	49.8	66.8	81.2	4.2
Arts, entertainment, and recreation	91.7	74.9	114.3	1.8
Other service activities	38.7	62.8	65.5	4.5
Activities of households as employers; undifferentiated goods and services-producing activities of households for their own use	15.7	23.5	26.6	4.5
Total	40.3	48.7	68.4	4.5

Note: CAGR = compound annual growth rate.

.

Second, the agriculture, forestry, and fishing industry exhibits the highest increase (14.51%). It is worth mentioning that the labor productivity of the industry is the lowest over the time period but has the highest growth, showing an upward trend in Table A1. The contributions of the industry have dominated all other industries since 2010. Although the share of the industry in the overall economy decreased, the industry has benefited the most since its WTO accession and maintains its importance in Vietnam’s economy as a main source of food and raw materials, as a major exporter, as well as being the largest employer across industries. Labor movement from the agriculture, forestry, and fishing industry to the manufacturing industry during this period might help improve labor productivity, which is a typical economic growth pattern in developing countries (The effect of structural change will be discussed later).

Third, the manufacturing industry presents the third-highest increase (5.35%) in the multi-period contribution of pure labor productivity for the period studied. The table shows manufacturing industry dominates service industries in terms of labor productivity growth, which is the second hypothesis. This implies the importance of the manufacturing industry’s pure labor productivity to enhance total labor productivity. This result is consistent with the well-known fact that export-oriented manufacturing-led growth is a key driver of successful economic growth in Vietnam. In the cases of Korea and China, which are relevant since Vietnam has been following a similar economic growth path, labor productivity in the manufacturing industry grew the fastest during the period of rapid economic growth (Ohno et al. [18], p. 57). Consequently, the manufacturing sector played a leading role in economic growth in both countries. Meanwhile, the manufacturing industry in Vietnam might not reach that point (Breu et al. [17]; Intrarakumnerd and Sunami [37]). As in Table A1, labor productivity of manufacturing was the 11th lowest in 2007, but it moved up to ninth in 2019, even with the third-highest increase among 20 industries. This may have been caused by the fact that it relies on low skill value-added manufacturing and also engages in a low value-added global value chain compared to other ASEAN countries and China (Breu et al. [17]). Vietnam has already made significant efforts to boost labor productivity but still has much room for improvement.

Table 3. Multi-period contribution of structural change (base=2007; unit: %, %p).

Year	in1	in2	in3	in4	in5	in6	in7	in8	in9	in10	in11	in12	in13	in14	in15	in16	in17	in18	in19	in20	Sum
2008	−0.24	−0.56	0.53	0.19	−0.09	0.08	0.08	0.14	0.42	0.10	0.19	−0.48	0.16	0.02	−0.09	−0.09	−0.08	0.20	−0.19	−0.02	0.27
2009	−0.54	−0.84	1.35	0.08	−0.10	0.21	−0.14	0.02	0.89	0.18	0.69	0.90	0.37	0.04	−0.24	−0.02	−0.12	0.29	−0.50	0.00	2.51
2010	−1.39	−1.76	1.40	−0.06	0.01	1.27	0.41	−0.11	1.07	0.28	1.13	3.83	0.32	0.06	−0.37	0.05	0.07	0.35	−0.22	0.01	6.35
2011	−1.90	−1.92	1.76	0.11	−0.07	1.34	0.63	−0.24	1.52	0.31	2.07	4.83	0.30	0.08	−0.51	0.08	0.16	0.38	−0.14	−0.01	8.80
2012	−2.37	−1.90	1.71	−0.28	−0.08	1.30	1.26	−0.07	1.70	0.34	2.18	6.28	0.46	0.14	−0.49	0.07	0.14	0.39	−0.19	−0.02	10.58
2013	−2.66	−2.79	1.85	−0.21	−0.08	1.28	1.52	−0.04	1.77	0.38	2.56	6.26	0.44	0.16	−0.44	0.11	0.14	0.42	−0.17	−0.03	10.49
2014	−2.87	−3.49	2.03	−0.10	−0.08	1.21	1.55	−0.08	1.87	0.46	2.83	6.55	0.43	0.19	−0.34	0.15	0.13	0.46	−0.15	−0.03	10.75
2015	−4.12	−4.55	4.41	0.17	−0.01	1.10	1.64	0.22	2.09	0.56	3.14	6.70	0.36	0.25	−0.37	0.14	0.27	0.44	0.10	0.00	12.53
2016	−5.06	−5.23	5.69	0.56	0.06	2.33	1.62	0.12	2.08	0.46	3.61	7.77	0.24	0.19	−0.34	0.13	0.32	0.39	0.06	0.00	15.00
2017	−5.85	−5.38	6.70	0.27	0.05	2.70	1.75	0.45	2.04	0.48	3.13	8.83	0.37	0.26	−0.39	0.35	0.21	0.44	0.04	−0.01	16.43
2018	−7.01	−6.11	7.57	0.72	0.11	3.03	2.34	0.46	2.38	0.36	3.71	9.66	0.56	0.30	−0.54	0.46	0.35	0.39	0.17	−0.01	18.90
2019	−8.69	−6.19	10.49	1.52	0.20	3.65	2.23	0.94	2.40	0.53	4.98	10.62	0.77	0.33	−1.01	0.25	0.41	0.38	0.38	0.02	24.20
Rank	20	19	2	7	16	4	6	8	5	10	3	1	9	14	18	15	11	12	13	17
Employment share
2007	52.9	0.7	12.5	0.3	0.2	5.2	10.9	3.0	2.4	0.4	0.4	0.1	0.4	0.3	3.7	3.3	0.9	0.3	1.6	0.4	100.
2019	34.5	0.4	20.7	0.4	0.3	8.4	13.3	3.6	5.0	0.6	0.9	0.6	0.6	0.7	2.7	3.6	1.1	0.5	1.9	0.4	100
Diff	−18.5	−0.3	8.1	0.1	0.1	3.2	2.4	0.6	2.6	0.2	0.5	0.4	0.2	0.3	−1.0	0.3	0.3	0.2	0.2	0.0

Note: Rank = rank by the value of 2019. Diff = difference (2019–2007), %p.

illustrates the multi-period contribution of structural change (base = 2007) and employment share by industry. The formula for this calculation is defined similarly by Equation (30) of the multi-period contribution of productivity change. In particular, among the 20 industries, the real estate activities industry has the highest positive attribution (10.62%), followed by the manufacturing industry (10.49%) and the financial, banking, and insurance activities industry (4.98%). The highest multi-period contribution of structural change in the real estate activities industry is mainly caused by an increase in employment from 53.9 thousand persons in 2007 to 303.4 thousand persons in 2019, denoting a rapid influx of labor from other industries, perhaps mainly from the agriculture industry (−8.69%). Domination of the manufacturing industry over service industries except for the real estate activities industry supports the second hypothesis as in the multi-period contribution of pure labor productivity.

As can be observed in Table 3, the employment share of the manufacturing industry has increased from 12.5% in 2007 to 20.7% in 2019, resulting in an increase of 8.1%p, the highest among the 20 industries. The highest multi-period contribution of structural change in the manufacturing industry in both 2008 and 2009 also implies that the manufacturing industry helped overcome the global financial crisis. The positive pure labor productivity and positive structural change in the manufacturing industry imply that the structural bonus hypothesis does hold in the Vietnam manufacturing industry. This hypothesis postulates a positive relationship between structural change and economic growth based on labor reallocation from industries that are comparatively low to those with a higher value added per labor input (Peneder [35], p. 428).

Wholesale and retail trade and the repair of motor vehicles and motorcycles in Vietnam are industries that are characterized as highly fragmented and dominated by individual owners operating low valued-added businesses. Accordingly, the labor productivity of the industry is 57.1 million VND per person, ranked 16th lowest in 2019, but the share of employment increased from 10.9% in 2007 to 13.3% in 2019, which is ranked third following agriculture, forestry and fishing, and manufacturing industries, as shown in Table A1. The multi-period contribution of pure labor productivity and structural change in the wholesale and retail trade industry and the repair of motor vehicles and motorcycles industry has increased by 3.14% and 2.23%, respectively. As noted, these are still characterized by a low level of labor productivity. It is necessary to initiate a concerted action plan with relevant stakeholders to boost labor productivity in those industries.

Financial services could play a critical role in economic growth. In this sense, it is worth exploring the growth of labor productivity in the financial, banking, and insurance activities industry. This industry increased by only 0.45% in multi-period pure labor productivity, with positive and negative values of variation over the entire period as displayed in single-period pure labor productivity (

Table A2. Single-period contribution of pure labor productivity intensity (base = previous year; unit: %).

Year	in1	in2	in3	in4	in5	in6	in7	in8	in9	in10	in11	in12	in13	in14	in15	in16	in17	in18	in19	in20	Sum
2008	0.62	−0.15	0.65	0.02	0.11	−0.27	0.40	0.07	−0.18	−0.03	0.17	0.64	−0.15	−0.01	0.18	0.21	0.13	−0.17	0.24	0.02	2.53
2009	0.13	0.75	−0.81	0.30	0.03	0.35	0.82	0.10	−0.30	−0.04	−0.19	−1.26	−0.18	−0.01	0.26	0.01	0.07	−0.05	0.37	−0.01	0.33
2010	3.07	1.70	−2.29	0.55	−0.02	0.01	−3.93	0.72	0.26	−0.04	0.55	−2.02	0.24	0.02	0.61	0.35	0.01	0.06	0.01	0.00	−0.15
2011	0.68	0.11	1.20	0.08	0.11	−0.30	0.43	0.30	−0.29	0.03	−0.64	−0.90	0.07	0.00	0.24	0.08	−0.04	−0.01	−0.02	0.02	1.16
2012	0.41	0.19	0.96	0.72	0.04	0.08	0.09	0.02	−0.06	0.04	0.05	−1.44	−0.09	−0.04	0.11	0.11	0.08	0.03	0.10	0.02	1.40
2013	0.51	0.65	0.80	0.20	0.05	0.31	0.33	0.32	0.06	0.04	−0.01	0.06	0.11	0.00	0.14	0.15	0.08	0.01	0.10	0.01	3.92
2014	0.70	0.79	0.92	0.32	0.04	0.45	0.69	0.20	0.06	0.02	0.06	−0.14	0.10	0.00	0.10	0.14	0.09	0.02	0.09	0.01	4.65
2015	1.44	1.50	−0.56	0.19	−0.03	0.75	0.75	−0.21	−0.07	0.00	0.12	−0.01	0.16	−0.03	0.21	0.19	−0.04	0.08	−0.14	−0.02	4.27
2016	1.01	0.20	0.82	0.13	−0.02	−0.47	0.82	0.35	0.21	0.18	0.04	−0.75	0.21	0.08	0.17	0.19	0.04	0.10	0.15	0.01	3.47
2017	1.07	−0.53	1.58	0.65	0.06	0.23	0.69	0.04	0.26	0.06	0.89	−0.75	−0.02	−0.04	0.25	0.00	0.19	0.01	0.12	0.03	4.79
2018	1.43	0.32	1.51	0.03	−0.02	0.27	0.26	0.23	−0.08	0.18	−0.06	−0.55	−0.08	0.00	0.24	0.07	−0.05	0.09	−0.05	0.00	3.75
2019	1.60	0.11	−0.31	−0.28	−0.02	0.07	0.94	−1.05	1.14	−0.06	−0.56	−0.59	−0.08	0.01	0.49	0.35	0.04	0.06	−0.09	−0.01	1.74

Note: in1: Agriculture, forestry, and fishing, in2: Mining and quarrying, in3: Manufacturing, in4: Electricity, gas, stream, and air conditioning supply, in5: Water supply, sewage, waste management and remediation activities, in6: Construction, in7: Wholesale and retail trade; repair of motor vehicles and motorcycles, in8: Transportation and storage, in9: Accommodation and food service activities, in10: Information and communication, in11: Financial, banking and insurance activities, in12: Real estate activities, in13: Professional, scientific and technical activities, in14: Administrative and support service activities, in15: Activities of the Communist Party, socio-political organizations; public administration and defense; compulsory security, in16: Education and training, in17: Human health and social work activities, in18: Arts, entertainment and recreation, in19: Other service activities, in20: Activities of households as employers; undifferentiated goods and services-producing activities of households for own use.

). On the contrary, it increased by 4.98% in a multi-period contribution of structural change, which is the third highest among 20 industries. It can be interpreted that the industry has expanded quantitatively, which is associated with rapid economic growth, but uncertainty and risk in the industry make labor productivity stagnant and volatile.

On the other hand, the agriculture, forestry, and fishing industry results in the lowest negative multi-period (2007–2019) contribution of structural change (−8.69%), followed by the mining and quarrying industry (−6.19%). The negative multi-period contributions of structural change might emanate from decreases in the employment share of the two industries (−18.5%p and −0.3%p, respectively). It is a natural result of the labor reallocation from the two industries to the manufacturing and service industries.

5. Conclusions

This study analyzed the growth of labor productivity in Vietnam by means of the Fisher index decomposition and attribution analysis. Using the official employment data from the General Statistical Office of Vietnam and value-added data from ADB’s Key Indicators for Asia and the Pacific 2020 for the period of 2007–2019, this study decomposed the labor productivity into pure labor productivity, structural change, and interaction terms, and then attributed the percentage change in labor productivity to 20 sub-industries. This study also examined both single-period and multi-period contributions of pure labor productivity and structural change.

The Fisher index decomposition of aggregate labor productivity has shown that the total labor productivity, pure labor productivity, structural change, and interaction terms increased by 69.83%, 36.74%, 24.20%, and 8.89%, respectively, from 2007 to 2019. The dominance of multi-period pure labor productivity over multi-period structural change is gradually increasing, which means pure labor productivity plays a prominent role in the progressive growth of labor productivity. In the early years of the WTO entry, the contribution of structural change is a dominant component in improving aggregate labor productivity, but as the effect of the accession has internalized, the contribution of pure labor productivity has been a main driver of aggregate labor productivity.

The multi-period attribution of pure labor productivity by industry illustrates that the agriculture, forestry and fishing industry, mining and quarrying industry, and manufacturing industry are dominant contributors to the growth of aggregate labor productivity. This suggests that the improvement in these industries’ pure labor productivity is critical in enhancing aggregate labor productivity. The multi-period attribution of structural change by industry presents that, among the 20 sub-industries, the agriculture, forestry, and fishing industry and mining and quarrying industry have negative growth rates. The positive pure labor productivity and structural change imply that the structural bonus hypothesis does hold in the industry. However, the agriculture, forestry and fishing industry, mining and quarrying industry, and the activities of the Communist Party, socio-political organizations, public administration and defense, and compulsory security industry have opposite results: positive pure labor productivity and negative structural change. The administrative and support service activities industry also has polarizing results, with negative pure labor productivity and positive structural change. The opposite signs between the two indices in those industries mean that the structural bonus hypothesis does not hold.

Even though Vietnam has already made remarkable economic growth, it lags behind other ASEAN countries in terms of labor productivity (APO [48]). There are a variety of challenges that Vietnam is facing in its efforts to boost labor productivity in order to fill in these gaps. As our empirical findings indicate, to strengthen pure labor productivity growth, it is essential to improve the quality of labor via training and education, as many other studies have proposed; for example, Dong et al. [20]. In addition to this, the government should encourage and support firms in upgrading their technology and innovating in ways to achieve high-quality labor. Homlong and Springler [49] and Nguyen [50] investigated the positive impacts of FDI on labor productivity in Vietnam. To take full advantage of FDI in the era of the fourth industrial revolution, it is necessary to focus on targeted sectors to attract FDI, such as IT, biotechnology, software development, and tourism (Anh and Hoa [51]). The Vietnamese government has taken a leading role in implementing the relevant policies to foster the growth of labor productivity. The Vietnamese government has issued laws and regulations to enhance productivity (Ohno et al. [18]; Intrarakumnerd and Sunami [37]). To meet this challenge effectively, the government provides support with well-targeted strategies and monitors goals to achieve productivity-led growth for sustainable development.

In this study, because of the limitations of the data, the manufacturing industry was analyzed as one aggregated sector. However, the manufacturing industry includes a variety of sectors associated with different and unique features of labor productivity. Therefore, an analysis of sub-categorized manufacturing industries—depending on data availability—would provide more insights; this could be a topic for future research. Another direction for future research would be to study the impacts of labor productivity and spillover effects between the FDI and domestic sectors caused by the volatility of the global economy and changes in the global value chain. Future studies may address how the COVID-19 pandemic specifically impacted Vietnam’s economy and whether this poses an obstacle to increasing labor productivity and/or attaining sustainable development and labor market (sustainable) integration; for example, Bloom et al. [52] and Gulseven et al. [53]. Furthermore, studying the relationship between the labor market and productivity with sustainability is worthwhile. Nowadays, sustainability as a macroeconomic objective poses significant challenges to traditional economic approaches. For example, how does this come into play regarding labor productivity, working hours, wages determination, etc.?