Human Capital Efficiency in Manufacturing: A Data Envelopment Analysis Across Economic Activity Branches and Firm Sizes in Mexico

Abstract

In a competitive global economy, the efficient use of human capital is a key determinant of productivity, growth, and sustainable development. This study assesses the efficiency of human capital in the Mexican manufacturing sector, with a focus on three strategic subsectors: the chemical industry, the food industry, and the transport equipment industry. The purpose is to analyze how human capital—measured through training, average wages, and daily working hours—relates to firm performance across different branches of economic activity and company sizes. Firm-level data from the National Institute of Statistics and Geography (INEGI) for the period 2009–2021 are analyzed using an input-oriented Data Envelopment Analysis (DEA) with CCR and BCC models. The results reveal significant differences in human capital efficiency across branches of economic activity within each—micro, small, and medium and large—firm size. Overall, the results highlight the central role of human capital investment in enhancing firm competitiveness and advancing the sustainable development of strategic industries. Policy implications underscore the need for training and wage strategies that improve efficiency and strengthen the long-term resilience of the Mexican manufacturing sector.

1. Introduction

In an increasingly competitive economic environment, human capital (HC) has become the most valuable intangible asset of an organization, as it represents a fundamental strategic factor that, through knowledge and skills, improves productivity and efficiency. An efficient company contributes to economic growth and generates quality employment [1,2,3,4]. This link between efficiency, employment quality, and long-term competitiveness also makes HC a central dimension of social and economic sustainability.

By identifying and understanding the significant relationship between HC and organizational efficiency, companies can make more informed decisions on where and how to invest, whether in training, wages, development, creating a positive work climate, retention, talent management practices, or adapting organizational structures and processes. In this way, resources can be optimized, maximizing the contribution of HC and minimizing costs, thereby increasing the probability of success and growth [1,5].

Various studies have identified that investment in HC—through wages and training—leads to an increase in employee productivity and, consequently, that of the organization [4,6,7,8,9].

From a sustainability perspective, these practices also ensure decent work, reduce inequalities, and promote inclusive growth, consistent with the Sustainable Development Goals (SDGs), particularly SDG 8 (Decent Work and Economic Growth) and SDG 9 (Industry, Innovation, and Infrastructure) [10]. From an economic standpoint, this makes HC investment a key determinant of firm productivity and competitiveness. For this reason, investment in HC is of the utmost importance. However, the optimal allocation of resources is not a simple task, as it depends on several factors, the most important being company size (micro, SME, and large) and the specific branch of economic activity [11,12].

Company size—micro, SME, or large—has a significant relationship with technical efficiency, and each of these sizes generates contributions to the country’s economic and social development, facing different challenges and opportunities in their pursuit of efficiency and sustainable growth [13]. According to the most recent census conducted by INEGI, 4,800,157 establishments were identified, employing 27,132,927 people. Of these, 94.9% are microenterprises, 4.9% SMEs, and 0.2% large companies. The contribution to value added, however, shows an inverse pattern: 54.7% from large companies, 30.7% from SMEs, and 14.6% from microenterprises. The distribution of HC is similar: 37.2% in microenterprises, 30.7% in SMEs, and 32.1% in large companies [14]. These figures underscore the structural challenges of sustainable development in Mexico, particularly regarding equity in productivity and employment across company sizes.

The manufacturing sector, despite debates about its diminishing importance, continues to represent one of the key elements for economic and social progress, as it has demonstrated a remarkable capacity for adaptation and evolution, consolidating itself as a strategic sector for development and competitiveness within the broader framework of sustainable industrialization [15,16]. In a developing country such as Mexico, this sector is a major engine of growth, a source of employment for millions of people, and a driver of innovation and foreign investment. Moreover, exports from this sector have positioned the country as a key player in global supply chains [17,18]. Yet, challenges such as low investment in training (only 6.4% of manufacturing firms provide training) represent critical barriers to sustainability and competitiveness [19].

In Mexico, there are three main sectors of economic activity: commerce, services, and manufacturing, with manufacturing generating the highest contribution—32%—to value added, despite being the sector with the lowest percentage—12.1%—of establishments [19]. Within manufacturing, most companies are microenterprises (93.7%), followed by SMEs (5.5%) and large companies (0.8%). However, the distribution of employed HC is the opposite: large enterprises account for the largest share (58.1%), followed by SMEs (22.5%) and microenterprises (19.4%) [20].

The manufacturing industry is divided into 21 economic subsectors, of which only 3 contribute 58.9% of the total value of industrial production: (1) transport equipment manufacturing (336), (2) the food industry (311), and (3) the chemical industry (325) [21].

Despite the importance of HC in organizational performance and success, several gaps remain in the literature regarding investment in HC and its relationship with efficiency in the manufacturing sector, especially when differentiating by company size and economic activity branch (EAB). Few studies [4,22] have quantified the optimal levels of investment in HC—hours, days, training, wages—that allow companies of different sizes (micro, SME, and large) to be efficient, and no studies have been identified that analyze this by EAB. For this reason, this research focuses on the EABs of the three main manufacturing subsectors. The main objective of this paper is to analyze the efficiency of HC—measured by training, hours per day, and wages—in relation to sales in each EAB of the three main subsectors, considering micro, small, and medium-sized enterprises (SMEs) and large manufacturing companies in Mexico.

Knowing the optimal levels of HC investment by company size and EAB will inevitably generate advantages for both HC and the organization. A satisfied workforce will be more motivated and committed, which increases the likelihood of generating innovative ideas, finding creative solutions to market challenges, improving productivity and efficiency in activities, and raising the quality of products and services. At the same time, a company that invests in the development of its HC demonstrates a commitment to society, contributes to economic growth, and generates quality employment. Furthermore, it will have within its workforce a strong resource to overcome crises and challenges, ensuring long-term sustainability.

This research offers a significant and novel contribution to the literature on HC—measured through investment in training, working hours per day, and wages—and its relationship with efficiency by analyzing, for each company size, the EABs within the three main manufacturing subsectors in Mexico. Unlike other studies on HC, this work incorporates a more disaggregated approach by EABs, analyzing the interaction between investment in HC and sales of micro, SME, and large companies operating within the Mexican manufacturing industry. The methodology proposed in this paper makes it possible to identify patterns and structural differences within the three key subsectors. The existing literature often focuses on broad, industry-level analyses, overlooking the specific dynamics within EABs and the variations in HC strategies employed by firms of different sizes. In this sense, the results provide practical insights for policymakers and industry leaders, helping to design strategies that enhance firm-level efficiency while simultaneously advancing sustainable industrial development in Mexico.

This paper is organized as follows: Section 2 presents the literature review and theoretical framework. Section 3 describes the methodology, including the input and output variables under study, the CCR input-oriented model and the BCC input-oriented model, the research hypotheses, and the data used. Section 4 presents the results and analysis of human capital efficiency in each of the EABs of the three main economic subsectors for the three firm sizes —micro, SME, and large. Section 5 presents a discussion of the contributions. Finally, Section 6 draws conclusions, presents the policy implications of this research, and outlines directions for future studies.

2. Literature Review

2.1. Human Capital and the Manufacturing Sector

The common denominator in any company size and/or EAB is HC. Its correct management and development are essential to improve competitiveness, efficiency, and growth, conditions that will ensure, in the medium and long term, the success of the organization [23,24,25,26]. The concept of HC encompasses knowledge, skills, personality traits, and creativity, which are acquired through education, training, and experience [27,28]. Investment in HC is fundamental to boosting economic value through work, increasing the absorptive capacity of the company, and enabling the assimilation and application of new knowledge and technologies, and it also correlates positively with economic income and negatively with unemployment [29,30,31].

Prioritizing the development of HC will generate positive effects both for the organization and for the individual, especially when the relationship between HC and physical capital is strong [32]. Therefore, strategic investment in HC is a primary factor in companies attempting to achieve a sustainable competitive advantage in the global market [5]. It has been shown that countries that strategically invest in education and in the development of their HC are better positioned to compete in the global economy and attract foreign investment [33].

In the context of sustainable development, HC becomes particularly relevant, as it allows companies not only to improve their economic performance but also to strengthen their capacity to adopt environmentally friendly practices, innovate in sustainable production processes, and contribute to long-term industrial resilience.

The economic landscape of Mexico is significantly influenced by the manufacturing sector, which is a fundamental pillar in the contribution to gross domestic product and the generation of employment. The total production of the sector is generated mainly by three subsectors: (1) the food industry, (2) the chemical industry, and (3) the manufacture of transportation equipment [34,35]. Based on the North American Industry Classification System (NAICS), the chemical industry is made up of seven EABs: (1) manufacture of basic chemical products—3251; (2) resins, synthetic rubbers, and chemical fibers—3252; (3) fertilizers, pesticides, and other agrochemicals—3253; (4) pharmaceutical products—3254; (5) paints, coatings, and adhesives—3255; (6) soaps, cleaners, and toilet preparations—3256; and (7) other chemical products—3259. The food industry is made up of nine EABs: (1) production of animal feed—3111; (2) grain and oilseed milling and production of oils and fats—3112; (3) production of sugars, chocolates, candies and similar—3113; (4) preservation of fruits, vegetables, stews, and other prepared foods—3114; (5) production of dairy products—3115; (6) slaughtering, packing, and processing of livestock, poultry, and other edible animals—3116; (7) preparation and packaging of fish and seafood—3117; (8) production of bakery products and tortillas—3118; and (9) other food industries—3119. The manufacture of transportation equipment is made up of seven EABs: (1) automobiles and trucks—3361; (2) bodies and trailers—3362; (3) parts for motor vehicles—3363; (4) aerospace equipment—3364; (5) railway equipment—3365; (6) vessels—3366; and (7) other transportation equipment—3369 [20].

2.2. Human Capital in Micro, SME, and Large Enterprises in the Manufacturing Sector

Due to the broad spectrum of scales and operational capacities of the sector, the different sizes of manufacturing companies present distinct characteristics [36]. Microenterprises are identified by their limited size of 0 to 10 employees; low investment in plants, machinery, and research and development; resource restrictions; access to financing; technology adoption; and market reach. They rely heavily on the owner’s personal investment and informal sources of financing. Their structure is mainly family-based, with centralized decision making. They frequently encounter difficulties in attracting, retaining, and developing their personnel, conditions that highlight the importance of investing in their HC to improve their competitiveness and sustainability [37]. Their productivity is low, because of informality and the absence of economies of scale. However, they play a crucial role in local economies, through employment generation and resource utilization [38].

Small and medium-sized enterprises of 11 to 250 employees, like microenterprises, face limitations related to financial and HC. However, they have a more formalized structure, with greater investment and operating scope. They adopt intermediate technologies, their market is mainly local as well as regional, and their productivity is higher compared to microenterprises, but still with a significant gap relative to large enterprises [39,40].

Large enterprises of more than 250 employees benefit from economies of scale, globalization, and mass production that have boosted their ability to compete [41]. They have multidivisional structures that allow them to control the complexity of their operations, with their market being not only local and regional but also national and international. They have a significantly greater capacity to invest in research and development, advanced technologies, attraction of specialized HC, as well as in the pursuit of international expansion, strategic alliances, and diversification [42]. Unlike micro and SMEs, large enterprises face challenges in managing organizational complexity, adaptation to disruptive changes, as well as the pressure to maintain large-scale innovation.

Although several differences are observed among the different company sizes, the contributions of micro, SME, and large manufacturing enterprises to economic growth, innovation, and job creation are substantial. In addition, their investment in HC directly influences their ability to adopt sustainable practices, reduce environmental impacts, and strengthen long-term resilience in a highly competitive global economy.

2.3. Human Capital in Chemical, Food, and Transportation Equipment Industries

Mexico has become a global manufacturing hub, with approximately half of its exports consisting of manufactured goods, of which more than 50% are produced solely by the chemical, food, and transportation equipment manufacturing industries [21,43]. These sectors are very dynamic, characterized by their complexity, specialized knowledge, technological advances, and strict regulatory environments, conditions that lead them to require highly qualified, committed, and adaptable HC, thereby increasing their ability to overcome challenges and seize market opportunities [44].

The chemical industry is a vital component of the manufacturing sector and is responsible for transforming raw materials into basic, specialized, and final consumer chemical products. This sector has proven to be efficient in the three company sizes—micro, SME, and large—and is highly relevant to the country’s economic progress [4,22]. The manufacturing processes of the sector are generally specialized, requiring trained HC to be able to develop new products, improve production processes, and mitigate environmental risks. In addition, investment in training makes it possible to meet the growing demand for environmentally friendly products, comply with stricter environmental regulations, prevent accidents, ensure regulatory compliance, maintain product integrity, and guarantee operational excellence.

The food industry is a multifaceted sector that plays a fundamental role in the nation’s economy, culture, and public health. It is made up of diverse EABs and is characterized by the complex interaction of traditional practices and modern technologies, reflecting the different socioeconomic strata and agroclimatic zones of the country [45]. The sector is composed of a dual structure, with large export-oriented companies coexisting with microenterprises and SMEs that serve local markets. The industry constantly faces changing consumer preferences, greater health awareness, and more complex supply chains, thus reinforcing the relevance of investing in HC to guarantee food safety, improving production efficiency, and developing innovative products that meet market needs and allow organizations to maintain a competitive advantage.

The transportation equipment manufacturing industry makes substantial contributions to both the national economy and global trade networks [46]. It integrates a wide spectrum of economic activities, evidencing its diverse nature and multifaceted impact. The country’s strategic geographic location, together with cost competitiveness, has positioned Mexico as an important manufacturing hub capable of producing sophisticated technologies and serving as a key exporter [43]. To ensure that the sector maintains a competitive advantage in a rapidly evolving technological landscape, as well as improvements in production efficiency and reductions in manufacturing costs, highly qualified HC is required—capable of designing, engineering, and producing complex products and systems—which necessitates investments in professional training, training and skill development and the development of advanced engineering skills, digital literacy, and complex systems [47,48]. It is important to emphasize that the automotive industry in the country stands out as one of the most dynamic and competitive sectors worldwide [47].

DEA has been used to assess productivity performance at both the firm and sectoral levels. Many studies have applied DEA to evaluate the efficiency of manufacturing firms and industrial sectors, confirming its capacity for benchmarking heterogeneous decision making units [49,50,51]. For instance, Smriti et al. [51] used DEA to measure efficiency in manufacturing firms, while Jain et al. [49] demonstrated its usefulness for performance measurement and target setting in industrial contexts. More recent surveys further highlight that DEA has been adopted across a wide range of sectors, including banking, energy, and agriculture, thereby reinforcing its role as a standard tool for comparative performance analysis [50,52]. Nevertheless, these applications typically emphasize capital, labor, and physical inputs, while the role of human capital remains underexplored. Building on this tradition, the present research advances the literature by applying DEA to economic activity branches of the manufacturing sector and explicitly incorporating human capital variables as key inputs.

In addition, Data Envelopment Analysis (DEA) has been increasingly applied to measure efficiency in contexts like this research. DEA makes it possible to evaluate the performance of companies with different sizes and EABs, considering multiple input and output variables without requiring a specific production function. This characteristic is especially relevant in the manufacturing industry sector, where companies operate under heterogeneous conditions. In this research, the application of DEA is justified since it allows measuring the efficiency of HC investment—training, wages, and working hours—relative to sales. This provides a robust framework for analyzing micro, SME, and large enterprises across the chemical, food, and transportation equipment industries. Moreover, linking DEA to sustainability, this approach enables the identification of efficiency patterns that are not only economically relevant but also critical for designing policies that foster industrial sustainability and responsible use of human resources.

In summary, measuring the efficiency of a company is essential for its survival, as well as for maintaining operational excellence in today’s highly competitive market [53,54,55,56]. Investment in HC is critical for sustaining operational excellence, promoting research and development, and ensuring compliance with environmental regulations. Given the accelerated pace of technological advances, prioritizing training programs—as well as the development of employees’ skills, their problem-solving capacity, and adaptability to new technologies—will allow each EAB to approach efficiency. To our knowledge, there are very few studies on efficiency in the manufacturing industry across the American continent [4,22,57,58,59,60,61], and no published document is available at the EAB level using DEA models to analyze human capital efficiency. In this sense, this research not only addresses an academic gap but also provides tools to strengthen sustainable industrial development in Mexico.

Based on the above, the following hypotheses are proposed in this research:

H1. 

The economic activity branches of the chemical industry are the most efficient across the three company sizes.

H2. 

The economic activity branches of the transportation equipment manufacturing industry are the least efficient across the three company sizes.

H3. 

The economic activity branches of the efficient group have significantly higher sales than those of the low-efficiency group.

H4. 

The economic activity branches of the efficient group invest significantly more in training compared to the low-efficiency group.

H5. 

The wages of human capital in the efficient group’s economic activity branches are significantly higher compared to those of the low-efficiency group.

H6. 

The efficient economic activity branches do not show a significant difference in hours worked per day compared to the low-efficiency group.

2.4. DEA Applications in Sustainability

DEA has also proven to be a robust methodological tool for analyzing sustainability-related issues. Its flexibility in handling multiple inputs and outputs without the need for a predefined production function has made it particularly suitable for assessing energy efficiency, eco-efficiency, and broader environmental impacts in manufacturing industries [62,63]. Recent literature reviews on technical efficiency highlight that DEA has been increasingly adopted to evaluate trade-offs between economic performance and environmental objectives [56]. This enables the identification of best practices and benchmarks for sustainable operations. Consequently, DEA is therefore not only relevant for measuring human capital efficiency, as in this paper, but also for addressing the broader challenge of sustainable development in the manufacturing sector.

3. Methodology

This paper applies the DEA methodology to analyze the efficiency and productivity growth of HC investments relative to sales, focusing on the EABs within Mexico’s three main manufacturing subsectors. DEA constructs an efficiency frontier that establishes benchmarks among the decision-making units (DMUs) under study, making this methodology suitable for evaluating performance at a disaggregated level.

The purpose of DEA is to estimate the efficiency and productivity of the DMUs [64]. This yields two efficiency measures. Overall technical efficiency (OTE) is calculated under the assumption of constant returns to scale (CRS), while pure technical efficiency (PTE) is derived under variable returns to scale (VRS). A DMU reaches optimal performance when OTE and PTE match, since this indicates that scale efficiency (SE) equals 1, meaning no additional efficiency can be gained by modifying the production scale.

For the analysis, SE is used as the classification criterion. Branches with SE = 1 are categorized as efficient, while those with SE < 0.5 are classified as low-efficiency. EABs with SE values between 0.5 ≤ SE < 1 are not included in the second-stage comparisons, since the focus of this paper is on highlighting differences between the best- and worst-performing groups. wages, sales, and working hours within each firm size. Importantly, we do not compare across firm sizes; the analysis is conducted independently for each size group.

In this research, the DEA CCR input-oriented model is applied to calculate OTE [65], and the DEA BCC input-oriented model is applied to estimate PTE [66]. This orientation is justified because EABs can control their HC investments (inputs), but they cannot directly determine sales (output). Thus, input-oriented DEA models are appropriate, as they minimize input while keeping output constant [67]. An input orientation is adopted because EABs have greater managerial control over HC inputs—training, wages, and working hours—than over sales, which are influenced by external market conditions. The input-oriented DEA model provides a more realistic assessment of efficiency by focusing on minimizing resource use for a given level of output. This is consistent with previous DEA application where inputs are more manageable than outputs [65,66]. In this paper, the goal of the proposed DEA models is to evaluate sales efficiency in relation to HC investments at the EAB level within the three key manufacturing subsectors—transport equipment manufacturing (336), the food industry (311), and the chemical industry (325)—considering the differences between microenterprises, small and medium-sized enterprises (SMEs), and large companies. This provides a novel contribution to the literature on HC efficiency in the manufacturing sector by incorporating the often-overlooked dimension of EABs, offering a more detailed understanding of efficiency patterns across companies’ sizes within these strategic subsectors. Given the heterogeneity across EABs and company sizes, we developed DEA models separately for micro-sized enterprises, SMEs, and large-sized enterprises. Efficiency scores are therefore analyzed and interpreted within each size group. Micro, SME, and large-sized enterprises differ in scale, access to financial resources, capital equipment, and human resources, which makes direct cross-group comparisons misleading. For this reason, a common frontier across all size groups is not constructed. As previously mentioned, we developed input-oriented DEA models to understand the efficiency patterns within each size group. While meta-frontier DEA and Technology Gap Ratio analysis would allow for benchmarking across heterogeneous groups, such cross-group comparisons are beyond the scope of this paper and remain a valuable avenue for future research.

PTE is obtained from the DEA BCC model under VRS, which separates managerial efficiency from SE. Scale efficiency (SE) is the ratio of OTE/PTE. An SE equal to 1 indicates that a DMU operates at the most productive scale size, whereas SE < 1 indicates scale inefficiency, which may correspond to either IRS or DRS. This enables us to determine whether inefficiencies are due to managerial practices or to suboptimal scale of operations.

Following Banker, Charnes, and Cooper [66], RTS is determined using the DEA BCC model. A DMU is classified as CRS when OTE = PTE = 1. Otherwise, RTS is obtained by examining the sum of the intensity variables (λ): Σλ < 1 indicates IRS, Σλ > 1 indicates DRS, and Σλ = 1 indicates CRS. In boundary cases where RTS cannot be clearly assigned due to numerical tolerances, we report a dash (-) in the respective table.

3.1. CCR Input-Oriented Model

The CCR input-oriented model is specified through Equations (1)–(5), following the standard formulation presented in [68].

$$\min OT{E}_0={\theta }_0-\epsilon \left({\displaystyle \sum _{i=1}^m{s}_i^{-}}+{\displaystyle \sum _{r=1}^s{s}_r^{-}}\right)$$

(1)

ST

$${\displaystyle \sum _{j=1}^n{\lambda }_j{x}_{ij}}+s_i^{-}={\theta }_0{x}_{i0} i=1,2,\dots ,m$$

(2)

$${\displaystyle \sum _{j=1}^n{\lambda }_j{y}_{rj}}-s_r^{+}=y_{r0} r=1,2,\dots ,s$$

(3)

$$s_i^{-},s_r^{+}\ge 0$$

(4)

$${\lambda }_j\ge 0 j=1,2,\dots ,n$$

(5)

where

DMU0	unit under analysis	[-]
xij	quantity of input i used by DMUj	[units]
xio	quantity of input i used by the DMU0	[units]
yrj	quantity of output r delivered by DMUj	[units]
xro	quantity of output r delivered by the DMU0	[units]
λj	relationship importance between DMUj and the DMU0	[-]
${\theta }_o^{*}$	DMU0 optimal OTE	[-]
$s_i^{-}$	input slack variable	[units]
$s_r^{+}$	output slack variable	[units]

3.2. BCC Input-Oriented Model

Equations (6)–(11) describe the BCC input-oriented model, as presented by [68].

$$\min PT{E}_0={\theta }_0-\epsilon \left({\displaystyle \sum _{i=1}^m{s}_i^{-}}+{\displaystyle \sum _{r=1}^s{s}_r^{-}}\right)$$

(6)

ST

$${\displaystyle \sum _{j=1}^n{\lambda }_j{x}_{ij}}+s_i^{-}={\theta }_0{x}_{i0} i=1,2,\dots ,m$$

(7)

$${\displaystyle \sum _{j=1}^n{\lambda }_j{y}_{rj}}-s_r^{+}=y_{r0} r=1,2,\dots ,s$$

(8)

$${\displaystyle \sum _{j=1}^n{\lambda }_j}=1$$

(9)

$$s_i^{-},s_r^{+}\ge 0$$

(10)

$${\lambda }_j\ge 0 j=1,2,\dots ,n$$

(11)

3.3. Data

The information used in this research was obtained from the Annual Survey of the Manufacturing Industry (EAIM), which reports data for each EAB within the three main subsectors, considering micro, SME, and large manufacturing firms [69]. The period analyzed covers the years 2009 to 2021.

The input variables are as follows.

• Annual average investments in training (TI): expenses made by firms to train their workers, including payments to internal or external instructors, training materials, and contributions to educational institutions (scholarships).
• Annual average wages (W): total compensation in cash or in kind, both ordinary and extraordinary, before taxes, paid to employees, including salaries, social benefits, and profit-sharing, whether calculated per working day or per task performed.
• Annual average working hours per day (H): number of daily working hours directly dedicated to the production process of the establishment.

The output variable is as follows:

• Average sales per year (S) income obtained from the production of goods and services.

The database was built and refined through the following steps.

• Step 1. Establishments with 1 to 10 full-time employees, 11 to 250 employees, and more than 250 employees were classified as micro, SMEs, and large enterprises, respectively [70]. As a result, three separate datasets were constructed: one for micro, one for SMEs, and one for large enterprises.
• Step 2. In each dataset, observation units that reported zero values in employees, training, wages, or sales were excluded from the sample.
• Step 3. Every dataset included four variables: three inputs—TI, W, and H—and one output—S.
• Step 4. To reduce the high variability present in the data, all variables were segmented into quartiles. Observations were grouped as follows: Group 1 = first quartile (x < 25%), Group 2 = second and third quartiles (25% ≤ x ≤ 75%), and Group 3 = fourth quartile (x > 75%). Since the coefficient of variation (CV) of the mean values in Groups 1 and 3 was greater than 1, only Group 2 was considered in the analysis, as its CV values were consistently below 1.

After refining the data and applying the DEA models, the analysis for each company size (micro, SMEs, and large firms) followed the same procedure.

Firms were classified into four categories according to their scale efficiency (SE): (1) efficient (SE = 1), (2) high efficiency (0.75 < SE < 1), (3) medium efficiency (0.5 ≤ SE ≤ 0.75), and (4) low efficiency (SE < 0.5).

For each EAB, the averages of sales, wages, training investment, and hours/day in the efficient group were compared with those of the low-efficiency group.

Within the efficient group, the input averages of each EAB were also compared against one another.

Table 1. Average values of input and output variables for each economic activity branch in the three main manufacturing subsectors, segmented by micro, SME, and large enterprises.

	Micro-Sized Enterprise					SMEs					Large-Sized Enterprise
EAB	n	H 1	W 2	TI 3	S 4	n	H 1	W 5	TI 3	S 4	n	H 1	W 5	TI 3	S 4
3111	5	8.32	223,760	12,200	11,917	228	8.09	13,775	68,430	209,156	106	8.01	76,481	253,321	423,355
3112	16	7.95	2,056,625	104,687	545,700	372	8.11	15,080	72,637	174,013	147	8	83,153	331,054	1,250,817
3113	11	7.9	995,363	318,000	221,837	231	8.12	13,128	52,949	82,912	371	8.02	126,490	272,038	1,012,259
3114	—	—	—	—	—	258	8.34	13,084	58,578	110,813	217	8.04	79,553	285,706	564,318
3115	6	10.99	455,666	11,500	30,237	228	8.16	10,012	36,438	83,985	140	8.06	101,407	248,454	921,669
3116	18	8.21	490,235	24,111	81,697	257	8.2	10,791	44,301	81,415	191	8.11	76,274	338,560	818,618
3117	—	—	—	—	—	112	8.2	9266	70,681	87,863	50	8.09	66,496	223,200	508,304
3118	14	8.37	779,769	8250	5693	257	8.14	9722	38,993	56,756	257	8.09	123,248	524,896	1,271,256
3119	12	7.95	1,712,750	26,666	24,142	318	8.16	15,354	76,173	132,321	190	7.98	97,299	552,032	884,395
3251	14	8.19	1,154,142	27,000	112,709	332	8.04	21,424	84,092	203,514	147	8.09	142,934	561,857	862,896
3252	8	8.72	550,375	309,500	615,114	170	8.04	25,923	143,326	285,695	71	8.02	93,353	355,451	549,593
3253	5	7.72	3200	308,250	651,174	103	8.05	19,079	153,198	190,863	38	7.66	121,300	847,842	1,036,295
3254	17	7.66	1,627,352	252,647	760,378	214	8.07	23,814	152,710	157,853	288	8.05	151,929	770,502	997,410
3255	11	8.26	365,000	57,000	14,060	239	8.18	18,003	93,678	147,381	85	8.11	105,334	431,647	736,614
3256	12	8	941.166	53,833	41,641	185	8.24	20,331	118,361	184,714	183	8.12	99,480	584,235	1,017,458
3259	8	8.28	1,280,500	26,000	64,225	175	8.18	12,894	62,517	137,902	89	8.03	121,944	336,205	886,033
3361	—	—	—	—	—	34	8.14	25,551	172,441	217,154	100	8.36	700,948	3,241,300	33,768,747
3362	10	8.649	1,206,600	63,100	171,449	153	8.38	16,793	74,922	111,307	105	8.58	116,509	418,095	598,051
3363	107	8.199	1,404,383	36,416	54,421	119	8.25	18,795	95,627	115,199	2091	8.2	152,527	567,922	719,092
3364	7	7.888	1,835,000	65,428	181,877	88	8.31	23,367	139,773	85,097	104	8.24	192,302	1,009,221	604,503
3365	8	8.056	3,072,857	60,125	8,092,420	24	8.02	22,467	118,542	226,792	32	8.9	377,466	2,092,844	1,563,000
3366	5	8.758	6,239,200	757,600	3,688,548	36	8.35	12,223	43,806	165,218	26	8.29	140,906	434,462	1,087,952
3369	4	9.796	576,750	3000	3831	38	8.19	10,322	27,575	93,852	56	8.33	156,244	533,375	979,277

¹ H = annual average working hours per day [hours/day]; ² W = annual average wages [MXN]; ³ TI = annual average investments in training [MXN]; ⁴ S = average sales per year [MXN mn]; ⁵ W = annual average wages [MXN mn].

presents the mean values for each EAB by firm size, which were used for statistical analysis and for the DEA efficiency evaluation.

4. Results and Analysis

Table 2. Micro, SME and large enterprises’ OTE, PTE, SE and returns to scale.

	Micro-Sized Enterprise				SMEs				Large-Sized Enterprise
EAB	OTE	PTE	SE	RTS 1	OTE	PTE	SE	RTS	OTE	PTE	SE	RTS 1
3111	0.20	1.00	0.20	IRS	1.00	1.00	1.00	CRS	0.16	1.00	0.16	IRS
3112	1.00	1.00	1.00	IRS	0.81	1.00	0.82	IRS	0.36	1.00	0.36	IRS
3113	0.34	1.00	0.34	IRS	0.46	1.00	0.46	IRS	0.36	1.00	0.36	IRS
3114	—	—	—	—	0.58	0.97	0.60	IRS	0.19	0.99	0.19	IRS
3115	0.50	1.00	0.50	IRS	0.62	1.00	0.62	IRS	0.36	1.00	0.36	IRS
3116	0.65	1.00	0.65	—	0.53	0.99	0.54	IRS	0.23	0.99	0.23	IRS
3117	—	—	—	—	0.62	1.00	0.62	IRS	0.22	1.00	0.22	IRS
3118	0.13	1.00	0.13	—	0.42	1.00	0.42	IRS	0.23	0.97	0.24	IRS
3119	0.17	1.00	0.17	—	0.61	0.99	0.61	IRS	0.19	0.99	0.19	IRS
3251	0.80	1.00	0.80	—	0.91	1.00	0.91	IRS	0.15	0.97	0.15	IRS
3252	1.00	1.00	1.00	—	1.00	1.00	1.00	CRS	0.15	0.99	0.15	IRS
3253	0.85	0.99	0.86	—	0.79	1.00	0.79	IRS	0.18	1.00	0.18	IRS
3254	1.00	1.00	1.00	—	0.58	0.99	0.58	IRS	0.14	0.96	0.14	IRS
3255	0.08	1.00	0.08	—	0.62	0.99	0.63	IRS	0.16	0.98	0.17	IRS
3256	0.16	1.00	0.16	—	0.73	0.98	0.75	IRS	0.21	0.97	0.22	IRS
3259	0.47	0.98	0.48	—	0.71	0.99	0.72	IRS	0.25	0.99	0.25	IRS
3361	—	—	—	—	0.76	0.99	0.77	IRS	1.00	1.00	1.00	CRS
3362	0.53	0.93	0.57	—	0.50	0.96	0.52	IRS	0.14	0.92	0.15	IRS
3363	0.29	0.98	0.29	—	0.48	0.98	0.49	IRS	0.12	0.96	0.13	IRS
3364	0.53	1.00	0.53	—	0.31	0.97	0.32	IRS	0.07	0.93	0.07	IRS
3365	—	—	—	—	0.87	1.00	0.87	IRS	0.09	0.86	0.10	IRS
3366	—	—	—	—	1.00	1.00	1.00	CRS	0.24	0.96	0.25	IRS
3369	0.24	1.00	0.24	—	0.90	1.00	0.90	IRS	0.18	0.95	0.19	IRS

¹ RTS classification follows [70]. “—” indicates cases where the model could not assign IRS or DRS due to boundary or numerical issues.

presents the DEA results for the three firm sizes. The analysis provides the average levels of technical efficiency across the EABs within the three main manufacturing subsectors—transport equipment manufacturing (336), the food industry (311), and the chemical industry (325). These results allow us to compare the relative performance of micro, SME, and large companies in terms of their HC investments in TI, W, and H relative to S. Beyond measuring efficiency, these findings also reveal how differences in HC allocation can influence sustainable industrial development. Identifying efficient EABs not only highlights successful management of resources but also indicates pathways for resilience and long-term competitiveness, in line with broader sustainability objectives and the SDGs.

In

Table 3. Specific comparisons for microenterprises: efficient group vs. low-efficiency group.

Sales
Comparison	Mean Difference 1	Confidence Interval (95%)	Comparison	Mean Difference 1	Confidence Interval (95%)	Comparison	Mean Difference 1	Confidence Interval (95%)
3112 vs. 3111	534 *	[131, 923]	3252 vs. 3111	603 *	[156, 1036]	3254 vs. 3111	748 *	[349, 1134]
3112 vs. 3113	324 *	[215, 626]	3252 vs. 3113	393 *	[35, 752]	3254 vs. 3113	539 *	[240, 837]
3112 vs. 3118	540 *	[252, 828]	3252 vs. 3118	609 *	[263, 956]	3254 vs. 3118	755 *	[470, 1039]
3112 vs. 3119	522 *	[227, 816]	3252 vs. 3119	591 *	[239, 943]	3254 vs. 3119	736 *	[445, 1027]
3112 vs. 3255	532 *	[200, 805]	3252 vs. 3255	601 *	[213, 930]	3254 vs. 3255	746 *	[418, 1016]
3112 vs. 3256	504 *	[147, 736]	3252 vs. 3256	573 *	[158, 863]	3254 vs. 3256	719 *	[365, 947]
3112 vs. 3259	481 *	[138, 807]	3252 vs. 3259	551 *	[156, 928]	3254 vs. 3259	696 *	[356, 1018]
3112 vs. 3363	491 *	[234, 648]	3252 vs. 3363	561 *	[228, 794]	3254 vs. 3363	706 *	[454, 857]
3112 vs. 3369	542 *	[14, 973]	3252 vs. 3369	611 *	[139, 1084]	3254 vs. 3369	757 *	[328, 1186]
3112 vs. 3252	−69	[−404, 265]	3252 vs. 3254	−145	[−476, 186]	3254 vs. 3112	215	[−54, 484]
Training
3112 vs. 3111	92	[−107, 292]	3252 vs. 3111	297 *	[76, 519]	3254 vs. 3111	240 *	[42, 438]
3112 vs. 3113	−213 *	[−365, −61]	3252 vs. 3113	−9	[−189, 172]	3254 vs. 3113	−65	[−216, 85]
3112 vs. 3118	96	[−41, 234]	3252 vs. 3118	301 *	[133, 469]	3254 vs. 3118	244 *	[109, 380]
3112 vs. 3119	78	[−70, 226]	3252 vs. 3119	283 *	[106, 460]	3254 vs. 3119	226 *	[80, 372]
3112 vs. 3255	48	[−97, 207]	3252 vs. 3255	253 *	[79, 440]	3254 vs. 3255	196 *	[52, 353]
3112 vs. 3256	51	[−98, 199]	3252 vs. 3256	256 *	[78, 433]	3254 vs. 3256	199 *	[52, 345]
3112 vs. 3259	79	[−69, 245]	3252 vs. 3259	284 *	[109, 477]	3254 vs. 3259	227 *	[81, 391]
3112 vs. 3363	68	[−36, 172]	3252 vs. 3363	273 *	[131, 415]	3254 vs. 3363	216 *	[115, 318]
3112 vs. 3369	102	[−143, 346]	3252 vs. 3369	307 *	[44, 570]	3254 vs. 3369	250 *	[6, 493]
3112 vs. 3252	−205 *	[−283, −12]	3252 vs. 3254	57	[−110, 223]	3254 vs. 3112	148 *	[13, 283]
Wages
3112 vs. 3111	1833 *	[443, 3008]	3252 vs. 3111	327	[−1, 2]	3254 vs. 3111	1404 *	[23, 2570]
3112 vs. 3113	1061 *	[81, 2042]	3252 vs. 3113	−445	[−2, 98]	3254 vs. 3113	632	[−336, 1600]
3112 vs. 3118	1277 *	[342, 2211]	3252 vs. 3118	−229	[−1, 2]	3254 vs. 3118	847	[−75, 1770]
3112 vs. 3119	344	[−612, 1300]	3252 vs. 3119	−1162 *	[−2305, −19]	3254 vs. 3119	−85	[−1029, 858]
3112 vs. 3255	1692 *	[195, 2156]	3252 vs. 3255	186	[−1832]	3254 vs. 3255	1262	[−222, 1715]
3112 vs. 3256	1115 *	[160, 2071]	3252 vs. 3256	−391	[−2752]	3254 vs. 3256	686	[−257, 1630]
3112 vs. 3259	776	[−308, 1860]	3252 vs. 3259	−730	[−2521]	3254 vs. 3259	347	[−726, 1420]
3112 vs. 3363	652	[−18, 1323]	3252 vs. 3363	−854	[−1771, 63]	3254 vs. 3363	223	[−431, 876]
3112 vs. 3369	1480 *	[81, 2879]	3252 vs. 3369	−26	[−1559, 1506]	3254 vs. 3369	1051	[340, 2441]
3112 vs. 3252	1506 *	[422, 2590]	3252 vs. 3254	−1077 *	[−1559, −3, 88]	3254 vs. 3112	−429	[−1, 443]

* p < 0.05, ¹ [thousand MXN].

—micro-size enterprises,

Table 4. Specific comparisons for SMEs: efficient group vs. low-efficiency group.

Sales
Comparison	Mean Difference 1	Confidence Interval (95%)	Comparison	Mean Difference 1	Confidence Interval (95%)	Comparison	Mean Difference 1	Confidence Interval (95%)
3111 vs. 3113	126,243 *	[98,774, 153,713]	3252 vs. 3113	202,783 *	[173,081, 232,484]	3366 vs. 3113	82,306 *	[29,639, 134,972]
3111 vs. 3118	152,400 *	[125,628, 179,172]	3252 vs. 3118	228,939 *	[199,881, 232,484]	3366 vs. 3118	108,462 *	[56,156, 160,769]
3111 vs. 3363	93,956 *	[72,558, 115,354]	3252 vs. 3363	170,496 *	[146,299, 194,692]	3366 vs. 3363	50,018 *	[248, 99,789]
3111 vs. 3364	124,059 *	[87,149, 160,968]	3252 vs. 3364	200,598 *	[161,999, 239,198]	3366 vs. 3364	80,121 *	[21,970, 138,272]
3111 vs. 3252	−76,539 *	[−106,351, 46,728]	3252 vs. 3366	120,477 *	[66,551, 174,403]	3366 vs. 3111	−43,938	[−96,667, 8791]
Training
3111 vs. 3113	15	[−16,47]	3252 vs. 3113	90 *	[56, 124]	3366 vs. 3113	−9	[−69, 51]
3111 vs. 3118	29	[−1, 59]	3252 vs. 3118	104 *	[71, 137]	3366 vs. 3118	5	[−55, 65]
3111 vs. 3363	−27 *	[−51,−3,0]	3252 vs. 3363	48 *	[20, 75]	3366 vs. 3363	−52	[−108, 5]
3111 vs. 3364	−71 *	[−113, 51, −29, 17]	3252 vs. 3364	4	[−40, 48]	3366 vs. 3364	−96 *	[−162, −29]
3111 vs. 3252	−74 *	[−108, 75, −41, 04]	3252 vs. 3366	100 *	[38, 161]	3366 vs. 3111	−25	[−85, 35]
Wages
3111 vs. 3113	647	[−1657, 2952]	3252 vs. 3113	12,795 *	[10,300, 15,291]	3366 vs. 3113	−905	[−5, 3]
3111 vs. 3118	4053 *	[1806, 6299]	3252 vs. 3118	16,201 *	[13,760, 18,642]	3366 vs. 3118	2500	[−1893, 6894]
3111 vs. 3363	−5019 *	[−6813, −3225]	3252 vs. 3363	7129 *	[5096, 9161]	3366 vs. 3363	−6572 *	[−10,753, −2390]
3111 vs. 3364	−9591 *	[−12,690, −6492]	3252 vs. 3364	2556	[−687, 5799]	3366 vs. 3364	−11,144 *	[−16,029, −6258]
3111 vs. 3252	−12,147 *	[−14,650, −9645]	3252 vs. 3366	13,701 *	[9170, 18,231]	3366 vs. 3111	−1552	[−5981, 2876]

* p < 0.05, ¹ [thousand MXN].

—SMEs, and

Table 5. Specific comparisons for large enterprises: efficient group vs. low-efficiency group.

	Sales		Training		Wages
Comparison	Mean Difference 1	Confidence Interval (95%)	Mean Difference 2	Confidence Interval (95%)	Mean Difference 1	Confidence Interval (95%)
3361 vs. 3111	33,345 *	[31,940, 34,750]	2988 *	[2783, 3193]	624,467 *	[593,097, 655,837]
3361 vs. 3112	32,518 *	[31,212, 33,824]	2910 *	[2720, 3101]	617,795 *	[588,626, 646,965]
3361 vs. 3113	32,756 *	[31,621, 33,892]	2969 *	[2804, 3135]	574,459 *	[549,104, 599,814]
3361 vs. 3114	33,204 *	[31,986, 34,423]	2956 *	[2778, 3133]	621,395 *	[594,197, 648,593]
3361 vs. 3115	32,847 *	[31,526, 34,169]	2993 *	[2801, 3185]	599,542 *	[570,079, 629,005]
3361 vs. 3116	32,950 *	[31,706, 34,194]	2903 *	[2721, 3084]	624,675 *	[596,899, 652,451]
3361 vs. 3117	33,260 *	[31,515, 35,006]	3018 *	[2764, 3273]	634,452 *	[595,476, 673,428]
3361 vs. 3118	32,497 *	[31,310, 33,685]	2716 *	[2543, 2889]	577,700 *	[551,178, 604,222]
3361 vs. 3119	32,884 *	[31,638, 34,131]	2689 *	[2508, 2871]	603,649 *	[575,848, 631,450]
3361 vs. 3251	32,906 *	[31,599, 34,212]	2679 *	[2489, 2870]	558,015 *	[528,846, 587,184]
3361 vs. 3252	33,219 *	[31,655, 34,783]	2886 *	[2658, 3114]	607,596 *	[572,673, 642,518]
3361 vs. 3253	32,732 *	[30,812, 34,653]	2393 *	[2114, 2673]	579,648 *	[536,765, 622,531]
3361 vs. 3254	32,771 *	[31,600, 33,943]	2471 *	[2300, 2642]	549,019 *	[522,900, 575,138]
3361 vs. 3255	33,032 *	[31,545, 34,519]	2810 *	[2593, 3026]	595,614 *	[562,416, 628,812]
3361 vs. 3256	32,751 *	[31,498, 34,005]	2657 *	[2474, 2840]	601,469 *	[573,485, 629,452]
3361 vs. 3259	32,883 *	[31,410, 34,356]	2905 *	[2690, 3120]	579,004 *	[546,212, 61,796]
3361 vs. 3362	33,171 *	[31,762, 34,579]	2823 *	[2618, 3028]	584,440 *	[552,923, 615,956]
3361 vs. 3363	33,050 *	[32,018, 34,081]	2673 *	[2523, 2824]	548,422 *	[525,387, 571,456]
3361 vs. 3364	33,164 *	[31,753, 34,576]	2232 *	[2026, 2438]	508,646 *	[477,130, 540,163]
3361 vs. 3365	32,206 *	[30,159, 34,253]	1148 *	[850, 1447]	323,482 *	[277,779, 369,186]
3361 vs. 3366	32,681 *	[30,462, 34,899]	2807 *	[2483, 3130]	560,043 *	[510,505, 609,581]
3361 vs. 3369	32,789 *	[31,107, 34,472]	2708 *	[2463, 2953]	544,705 *	[507,146, 582,262]

* p < 0.05, ¹ [MXN mn], ² [thousand MXN].

—large-size enterprises, the results of the specific comparison analyses (Tukey) carried out between the EABs belonging to the efficient group and the low-efficiency group are shown in relation to their S, TI, W, and H. It is important to highlight that, when performing the comparisons of the variable H, in no case—regardless of the size of the organization—was a statistically significant difference identified. For this reason, it was not included in any of the respective tables, nor was it considered for the subsequent analyses of this research. This methodological decision ensures the robustness of the results while still allowing the findings to inform discussions not only about efficiency but also about sustainability, since the way HC is managed has direct implications for economic, social, and environmental outcomes.

• Microenterprises

Table 3 shows the results of the specific comparison analyses (Tukey) carried out between the EABs belonging to the efficient group and the low-efficiency group in relation to their S, TI, W and H. The average scale efficiency (SE) for the total sample of microenterprises was 0.5, a result that coincides with those reported by other authors [71,72,73].

Within microenterprises, three EABs are identified in the efficient group: grain and oilseed milling and production of oils and fats (3112), manufacture of resins, synthetic rubbers, and chemical fibers (3252), and manufacture of pharmaceutical products (3254). This suggests optimal use of their HC resources in the production process. Of these, two belong to the chemical industry subsector and one to the food industry subsector, thus confirming the first and second hypotheses of this research. In contrast, the EABs within the low-efficiency group are animal feed production (3111), sugar, chocolate, candy, and similar products (3113), bakery products and tortillas (3118), other food industries (3119), paints, coatings, and adhesives (3255), soaps, cleaners, and toilet preparations (3256), other chemical products (3259), motor vehicle parts (3363), and other transportation equipment (3369). This behavior may result from factors such as suboptimal resource management, inefficient processes, or an inadequate scale of operation.

The results of the variance analysis show significant differences among the averages of sales per year, annual average investments in training, and annual average wages reported by each EAB: average sales per year; F (17,265) = 15.86, p < 0.001, annual average investments in training; F (17,270) = 11.35, p < 0.001, and annual average wages; F (17,266) = 8.49, p < 0.001. These results coincide with those reported by [72].

When analyzing specific comparisons (Table 3), average sales per year of the efficient group (3112, 3252, 3254) versus the low-efficiency group (3111, 3113, 3118, 3119, 3255, 3256, 3259, 3363, 3369) show that, in all possible combinations, the average sales per year of the efficient group are significantly higher, thus confirming the third hypothesis of this research.

For the training variable, the fourth hypothesis of this paper is confirmed in the EABs 3252 and 3254, since their annual average investments in training are significantly higher compared to all EABs in the low-efficiency group. However, this is not the case for EAB 3112, since its average investment in training is not significantly higher compared to the low-efficiency EABs. Nevertheless, when observing the annual average wages in this EAB, in 77.78% of cases, the annual average wages of HC are significantly higher than in the low-efficiency group. In contrast, for the EABs 3252 and 3254, this condition is met in only 33.33% and 55.56% of the cases, respectively. Thus, the fifth hypothesis of this paper is partially confirmed. This information suggests that HC is a fundamental element within the organization, since investment in it—through wages and/or training—allows for a significant increase in efficiency.

Within the efficient group, it is observed that EAB 3254, whose average sales per year are higher, invests significantly more in training and/or wages compared to the other two EABs (3252 and 3112). This condition highlights the effect of HC investment on sales: greater investment in HC results in higher sales.

The variability in efficiency among EABs underlines the need to understand how certain microenterprises manage to maximize sales through investment in their HC, while others struggle with suboptimal productivity [74]. From a sustainability perspective, this variability also reflects uneven capacities to create quality employment and resilient business practices, which are central to long-term economic and social sustainability. In microenterprises, the identification of low-efficiency EABs represents an area of opportunity to improve operational and technological practices [75].

• SMEs

The average SE for the total sample of SMEs is 0.69, a result that matches those reported by other authors [13,57,61].

The efficient group comprises the following activity branches: manufacture of animal feed (3111), manufacture of resins, synthetic rubber, and artificial and synthetic fibers and filaments (3252), and ship building (3366). In contrast, the low-efficiency group consists of manufacture of sugars, chocolates, candies, and similar products (3113), manufacture of bakery and tortilla products (3118), manufacture of motor vehicle parts (3363), and manufacture of aerospace products and parts (3364).

It is observed that within the efficient group, the activity branches with 100% scale efficiency belong to three subsectors (food, chemicals, and transport equipment manufacturing). This makes it clear that the “chemical industry” subsector is not the one with the highest number of efficient activity branches, a situation that leads to the rejection of the first and second hypotheses.

As in microenterprises, significant differences are identified in the averages of sales per year (F (22,5139) = 91.85, p < 0.001), annual average investments in training (F (22,5357) = 33.81, p < 0.001), and annual average wages (F (22,5147) = 98.28, p < 0.001) reported by each EAB. These results are consistent with those found by other authors [18].

Table 4 shows the results of the specific comparisons between the averages of sales per year, annual average wages, and annual average investments in training of the efficient group versus the low-efficiency group. The third hypothesis of this research is confirmed, since in all cases, the average sales per year are significantly higher. Regarding the fourth and fifth hypotheses of this research, they are partially confirmed.

Importantly, within the efficient group, the chemical industry EAB (3252) shows the highest average sales per year and the highest annual average investment in training and wages for its HC compared to its efficient counterparts. As in microenterprises, this shows that investing in HC is essential: training enhances productivity while contributing to sustainable competitiveness by reinforcing economic performance and the capacity of firms to adapt to changing market and social demands [76].

• Large Enterprises

The average SE for the total sample of large-size enterprises is 0.24, a result that does not coincide with those reported by other authors [13,56,57,71]. It is important to consider that previous analyses were disaggregated at the subsector level, and no information has been reported at the EAB level, a condition that inevitably impacts efficiency results.

Among large-size enterprises, out of 23 EABs in the manufacturing subsector, 22 are in the low-efficiency group, with only automobile and truck manufacturing (3361) belonging to the efficient group. This situation results in the rejection of the first and second hypotheses of this paper.

The variance analysis confirms what is evident at first sight: there are significant differences in average sales per year (F (22,5054) = 624.31, p < 0.001), annual investments in training (F (22,5067) = 258.07, p < 0.001), and annual average wages (F (22,5060) = 446.30, p < 0.001) in at least one EAB. These results coincide with those found in [72,77,78,79].

The specific comparisons (Table 5) show that in all possible combinations, the automobile and truck manufacturing EAB invests significantly more in its HC [80]. This condition has a direct impact on its average sales per year—as in the case of micro and SMEs—since average sales per year are also significantly higher in all cases. These results confirm the third, fourth, and fifth hypotheses of this research.

The identification of efficiency based on HC across the three company sizes by EAB has profound implications for understanding the importance of HC, economic development, and business strategy. It makes it possible to allocate capital more effectively, ensuring that resources are directed where they can generate the greatest impact. In addition, it reveals operational gaps as well as opportunities for improvements—investment in technological infrastructure and/or broader implementation of training programs—for low-efficiency EABs, which can replicate the practices of high-efficiency EABs to increase productivity and improve organizational performance [81,82].

There is clear heterogeneity in efficiency among EABs regardless of company size [13]. While previous research identified the chemical industry subsector (311) as efficient across the three company sizes [4,22], when disaggregated by EAB, the same behavior is not observed. As mentioned above, in micro-size enterprises, SMEs, and large-size enterprises, only 2, 1, and 0 EABs, respectively, of this subsector are efficient. This reveals the variability that exists in HC investment.

Understanding efficiency at the EAB level of the three main manufacturing subsectors has great relevance for social welfare, GDP contribution, and the country’s economic trajectory. In sustainability terms, allocating HC resources more efficiently contributes not only to economic growth but also to fairer labor practices and more resilient industrial development. In a developing country like Mexico, the budget for investing in HC—i.e., training and wages—is often limited, which makes it crucial that investments are directed where they can generate the highest return. More informed allocation fosters greater impact and stronger economic growth, rather than inefficient use of resources.

The high competitiveness of today’s globalized market demands high productivity from companies. By recognizing the most efficient EABs, best practices—whether operational, technological, or managerial—can be identified, allowing for the maximization of HC performance and enabling the diffusion and replication of these successful models in less efficient EABs.

The efficient use of HC is fundamental for the resilience of an economy in the medium and long term. By continuously improving the way people contribute to productivity, a developing country can build a stronger manufacturing base that is less vulnerable to external shocks and better positioned for sustained growth. This, in turn, attracts greater national and foreign investment, since investors seek economies where resources—especially HC—are effectively managed. All of this contributes to the stability and prosperity necessary for a developing country to improve its economic position.

Across the three company sizes, clear differences emerge in the efficiency patterns of HC investment at the EAB level. Microenterprises show a relatively balanced presence of efficient and inefficient EABs, SMEs show cases of full efficiency across subsectors, and large-size enterprises are characterized by a strong concentration of inefficiency, with very few efficient EABs. These findings underline the heterogeneity of HC efficiency across company sizes and subsectors, highlighting the importance of tailoring HC investment strategies according to firm size and the specific dynamics of each EAB. These differentiated strategies are essential not only for improving firm-level efficiency but also for advancing sustainable development goals that depend on inclusive, resilient, and competitive industrial growth.

5. Discussion of Contributions

The results of this research provide important insights into the efficiency of HC investment at the EAB level within Mexico’s three main manufacturing subsectors—transport equipment manufacturing (336), the food industry (311), and the chemical industry (325). By applying DEA methodologies, this paper confirms the existence of substantial heterogeneity across company sizes and subsectors, which has direct implications not only for enterprise productivity and competitiveness but also for the sustainability of economic and social development.

For micro-sized enterprises, the identification of three efficient EABs—grain and oilseed milling and production of oils and fats (3112); manufacture of resins, synthetic rubbers, and chemical fibers (3252); and manufacture of pharmaceutical products (3254)—shows that, despite their limited scale, some companies can achieve optimal use of HC resources. These findings confirm the first and second hypotheses of this research for this group and align with previous evidence that highlights the role of HC in explaining differences in productivity among small firms. From a sustainability perspective, these efficient EABs show how HC allocation can generate higher sales, quality employment, and better working conditions, which are central to long-term resilience. In contrast, most micro-sized enterprises remain in the low-efficiency group, reflecting challenges in resource management, training investment, and wage allocation, which may limit their contribution to sustainable growth.

SMEs show a diversified efficiency landscape, with one efficient EAB in each subsector (3111, 3252, 3366). These results lead to the rejection of the first and second hypotheses of this research, since efficiency is not concentrated in the chemical industry, and the transport equipment manufacturing is not the least efficient. Nevertheless, the performance of efficient SMEs is consistent with their ability to combine structured organizational practices with targeted HC investments. Significant differences in average sales per year, annual average investments in training, and annual average wages across EABs confirm that different HC strategies influence efficiency outcomes. In sustainability terms, the results show that SMEs contribute to economic stability and workforce development as drivers of inclusive and competitive industrial growth.

For large-size enterprises, efficiency levels are notably lower, with only automobile and truck manufacturing (3361) reaching efficiency. This contrasts with previous research at the subsector level, suggesting that disaggregation to the EAB level uncovers new patterns. The rejection of the first and second hypotheses of this research for this group demonstrates that scale advantages alone do not ensure efficiency when HC allocation is uneven. From a sustainability viewpoint, this finding is important because it demonstrates that large-size enterprises—which are often central to industrial development—may not always maximize their HC resources, minimizing their potential to create sustainable value.

Across the three company sizes, the third hypothesis of this paper—that efficient EABs achieve higher sales—is consistently confirmed, showing the strong relation between HC efficiency and economic outcomes. The fourth and fifth hypotheses of this paper, which are related to annual investments in training and annual average wages, are partially confirmed, underscoring those different combinations of HC investment strategies can foster efficiency. The sixth hypothesis of this paper is supported, as no significant differences are found in average working hours per day, reiterating that efficiency depends more on the quality than the quantity of HC allocation. The results of this research underline that sustainable competitiveness is built on strategic investment in HC, not merely on expanding working hours per day.

By examining efficiency at the EAB level, this study highlights how targeted HC strategies can support sustainable industrial development. The results suggest that policies designed to strengthen competitiveness should also emphasize social sustainability by improving human capital training, ensuring fair wages, and promoting resilient work environments. These measures enhance the likelihood that companies will achieve efficiency while simultaneously contributing to broader economic and social sustainability goals.

6. Conclusions

This research applied DEA methodologies to analyze the efficiency of HC in relation to average sales per year at the EAB level of the three main manufacturing subsectors—transport equipment manufacturing (336), the food industry (311), and the chemical industry (325)—in micro-sized, SME, and large-sized enterprises in Mexico.

The literature on assessing the TE of the manufacturing industry in Mexico using DEA is quite limited. Therefore, this research represents a novel contribution by applying DEA to analyze the efficiency of HC in relation to sales across each EAB within the three main subsectors of micro, SME, and large manufacturing enterprises in Mexico.

This approach enables a robust differentiation between productive units by identifying performance patterns, best practices, and areas for improvement in HC management. It offers a comprehensive perspective on how different EABs leverage their intellectual resources to generate value, thereby providing a deeper understanding of operational efficiency.

Investment in HC not only fosters greater organizational development but also enhances individual well-being, which is reflected in higher productivity and, consequently, in the organization’s sales. All technical progress should yield a positive impact and be reflected in increased output; this impact can be measured through the TE indicator. If a company is efficient, it means it is increasing its output while using its resources optimally. Efficiency is a crucial element at both the microeconomic and macroeconomic levels and is one of the conditions of achieving widespread welfare.

The results reveal marked heterogeneity in efficiency across EABs within each of the three firm sizes, a condition that underscores the importance of researching HC and its contribution to these differences. The findings derived from comparing the efficient group (SE = 1) with the low-efficiency group (SE < 0.5) for each firm size provide a nuanced perspective on how investment in HC (i.e., wages and training) translates into increased sales within Mexican manufacturing landscape.

• Microenterprises

Efficient microenterprises are primarily found in the chemical industry subsector, while the least efficient are from the transport equipment manufacturing subsector. This may be because EABs of the chemical industry benefit from specialization in market niches, the production of goods with higher added value per unit volume, and processes that, although complex, can be managed efficiently on a small scale if supported by highly specialized HC. These sectors often rely more on knowledge and formulation than on large infrastructures or extensive supply chains. Conversely, the low efficiency of EABs in the transport equipment manufacturing subsector is due to the inherent complexity of the products, high quality and safety standards, economies of scale that favor large producers, and dependence on larger supply chains, all of which hinder their operational efficiency and competitiveness on a small scale.

On average, in 63% of cases for training and 56% for wages, investment levels are significantly higher in the efficient group compared to the low-efficiency group. This pattern results in sales, as in 100% of cases, the sales of the efficient group are significantly higher compared to those of the low-efficiency group.

• SMEs

In small and medium-sized enterprises, the efficiency landscape is more diversified, as one EAB from each of the three main economic subsectors is 100% efficient. Therefore, none of the three main subsectors is inherently more or less efficient. This suggests that SMEs have the capacity to identify and exploit market niches where they can be highly efficient, regardless of the subsector. These results indicate that efficiency at this company size is not determined by belonging to a specific subsector but rather by business strategy, market adaptation, and specific HC management. The ability of SMEs to be flexible and specialize allows them to break away from the efficiency trends observed in micro or large enterprises.

On average, in 50% and 67% of cases, investment in training and wages is significantly higher in the efficient group compared to the low-efficiency group, and in 100% of cases, sales are significantly higher compared to the low-efficiency group.

The reason why, in both microenterprises and SMEs, a significant increase in both training and wages is not observed in 100% of the cases in the efficient group compared to the low-efficiency group is that efficiency is a multifaceted phenomenon. That is, an increase in sales can be achieved not only through greater direct investment in intellectual capital but also through factors such as the quality of pre-existing HC, specialization in niches, innovative management, as well as random events and the macroeconomic environment. It is important to highlight that in micro, small, and medium-sized enterprises, the “effective” implementation of training remains a challenge, which can limit the direct and measurable impact of the investment.

• Large Enterprises

In large enterprises, the only efficient EAB is the manufacture of automobiles and trucks, belonging to the “Transport Equipment Manufacturing” subsector. This information is highly revealing, as, on a large scale, the automotive industry is not only not the least efficient but positions itself as the only efficient one. This is because large enterprises benefit enormously from economies of scale, advanced automation, integration into global supply chains, massive investment in research and development, and the optimization of production processes. These factors allow them to achieve massive efficiency that surpasses other subsectors, even those that are efficient on a smaller scale.

In 100% of cases, investment in training, wages, and sales are significantly higher in the efficient group compared to the low-efficiency group. This stems from the fact that, at this group size, the capacity to invest massively and structurally in training and wages becomes a direct and measurable driver of their efficiency and, consequently, of their sales.

A consistent finding across all three enterprise sizes is the absence of a significant difference in the number of hours worked per day between the efficient group and the low-efficiency group. This finding indicates that the length of the workday (within the analyzed parameters) is not a determining factor of efficiency; rather, efficiency resides in the intensity, quality, or strategic use of those hours.

The RTS results indicate that many micro-sized enterprises operate under IRS, suggesting that scaling up their operations and HC investments could further enhance efficiency. In contrast, large-sized enterprises are often classified under DRS, showing possible diseconomies of scale and the need for better organizational alignment. SMEs appear under CRS, indicating that they operate closer to their optimal scale. These results support our conclusion that efficiency is strongly influenced by company group size; micro-sized enterprises may gain efficiency through expansion, large-sized enterprises must manage the risks of over-expansion and complexity, and SMEs achieve a balanced operational scale.

This research demonstrates that HC efficiency and therefore sectoral performance are “multifaceted phenomena” highly conditioned by company size. There is no universal rule of efficiency for entire subsectors that applies uniformly across all business strata.

The variability in efficiency underscores the importance of considering sectoral factors and the specific capabilities of companies to foster robust economic development in micro, SME, and large manufacturing enterprises. Public policies and business strategies must be differentiated and adapted to the realities and capacities of each company size, recognizing that there is no single formula for success in HC investment. A “one-size-fits-all” approach to promoting efficiency or sectoral development would be ineffective. It is crucial to understand the specific drivers of efficiency for each company size and sector to implement interventions that genuinely enhance productivity, competitiveness, and social welfare within the diverse manufacturing landscape. In general terms, it can be concluded that, regardless of the company size, efficient companies invest significantly more in their HC, which generates a significant increase in their sales. From a sustainability perspective, this demonstrates that improving HC management can enhance economic productivity while simultaneously promoting social outcomes such as job quality and workforce resilience.

Finally, future research could expand our analysis by adding additional variables to our analysis, such as investment in R&D, technological adoption, or export orientation. A longitudinal analysis could determine whether efficiency levels are sustained over time or change in response to evolving economic and social conditions. This would clarify whether efficiency advantages persist or fluctuate with these transformations and providing insights for sustainable industrial development.

Another future research direction is the application of the meta-frontier DEA model and Technology Gap Ratio to benchmark efficiency across heterogenous enterprise groups. This approach would enable comparison between enterprise group sizes and provide deeper insights into efficiency gaps that are outside of the scope of this paper.