Based on “factor production theory”, the firm size reflects the full extent of available production factors such as labor force, means of labor, subject of labor, and other resources. The optimized allocation of production factors and resource deployments among businesses is the primary component of corporate strategy [6]. Diversification is one of the good examples of corporate-level strategy, which is driven by resources [7]. In this paper, business diversification refers to a firm simultaneously operates several different businesses, which requires allocating and utilizing accumulated resources and capabilities among business to the fullest extent possible based on the identification, development, cultivation, and promotion of heterogeneous resources [8].

In practice, some studies show that the sharing of both tangible and intangible resources accumulated in the scale expansion promotes diversification [9] [10]. Besides, the results of empirical research in agriculture, manufacturing, and service industries have shown that the firm size can have a significant impact on the degree of related and unrelated diversification [11] [12]. That is to say, due to the competitive advantage from resources, large enterprises are more diversified, and the survival status of it is better compared with the small enterprises. Hence, our research stems from the vision of better sustainable development in logistics enterprises, so we investigate the relationship between firm size and diversification.

The resources owned by logistics enterprises include concrete equipment, warehouses, vehicles, and other resources, as well as intangible organizations, technologies, and reputations. In terms of the created services, it can be divided into related essential services such as transportation and storage and unrelated value-added services such as sales and supply chain solutions for the logistics industry [13]. The firm size reflects the firm’s resources, which is the basis and guarantee of diversification, thus influencing the degree of diversification of the firm [14]. Also, the difference in resources between different business units reflects the degree of corporate diversification [9] [15]. Moreover, a firm’s investment and services capability is also limited by the firm size. The representative views about the relationship between logistics enterprise size and diversification include: 1) Firm size determines the business scope of logistics enterprises by affecting their investment in infrastructure and information technologies. For larger sized logistics enterprises, diversifying into related and unrelated areas may be a crucial way for development and competitiveness. Thus the degree of related and unrelated diversification is higher. Due to the financial limits, Smaller enterprises may concentrate on the local market, providing relevant services to meet customer needs to increase profitability as well as market value [16]; 2) Yusufoglu [14] argues that the size of an enterprise can affect its access to the market and limit its ability to serve. Logistics enterprises with weak service capabilities are increasingly keen to focus on niche markets to operate related diversification businesses. In contrast, for logistics enterprises with a higher service level, developing a diversified economy and ensure all-round development, such as “one-stop” logistics services, surpasses absorbing in the core business. In general, the current literature supports the correlation between size and the degree of diversification. This is also following Martha’s view [17]. Thus, firm size is an essential factor that is invaluable for a firm’s diversification.

Based on the above discussion, the following hypotheses are proposed:

H1a: Firm size will be positively associated with related diversification.

H1b: Firm size will be positively associated with unrelated diversification.

Different-sized enterprises have distinct capabilities that enable them to allocate and transfer different resources variously, forming different diversification evolutionary patterns and strategies [18] [19] [20]. So it is appropriate for assessing the various diversification measures of different-sized companies.

Large-sized companies have performed well and have maintained a competitive edge in terms of resources and capabilities in the market. Moreover, the resources such as skills and knowledge acquired in one business can be applied to solve problems and exploit opportunities in new areas by related diversification strategy, improving their market competitiveness [21]. At the same time, an unrelated diversification strategy has the advantage of absorbing investment risks and achieving economic value in the process of scale expansion. Based on the above analysis, we assume that:

H2a: There is a significant correlation between the size of large logistics enterprises and the degree of related and unrelated diversification.

According to the research of Andersson [22], logistics alliances can enable enterprises to integrate, share the capabilities, resources as well as risks with each other. So the members of alliances can quickly expand their business into related and unrelated areas to achieve economies of scale and scope [16] [23] [24]. This is a joint development path for small and medium-sized logistics companies [25] [26]. In China, since convenience for joining the logistics alliances provided by continuous development of the network information technology together with the solid customer base in a niche market, the resources and capabilities of managing a variety of business accumulated in the scale expansion can be utilized by related and unrelated diversification strategy which also help small- and medium-sized firms operate more effectively [27]. We believe that:

H2b: The size of medium-sized logistics enterprises will be positively associated with the degree of related and unrelated diversification.

H2c: The size of small-sized logistics enterprises will be positively associated with the degree of related and unrelated diversification.

Due to insufficient resources such as physical, human, or organizational capabilities, micro-sized logistics enterprises have weak strength in the 800 billion less-than-truckload (LTL) market in China [28]. For one thing, as the “escape hypothesis” suggests that firms that experience poor prior performance in current industries are likely to prefer unrelated diversification [29] [30]. For another, compared with tangible resources with higher transfer costs and limited scope economic effects, the cost of transferring or integrating intangible resources is lower, which is the growth path of micro-sized enterprises to avoid industry barriers and risk, obtain superior performance and market advantages [20] [31]. Lastly, with the continuous development of the platform economy integrating manufacturing and service industries is emerging [32]. Many companies rely on JD- and Alibaba-based platforms to transfer or acquire intangible resources at lower entry costs. Therefore, micro-sized enterprises with flexibility and a quick decision chain can leverage and invest intangible resources into unrelated areas [21]. We assume that:

H2d: There is no significant correlation between the size of micro-sized logistics enterprises and the degree of related diversification. There is a significant

positive correlation with the degree of unrelated diversification.

Based on the above assumptions, the research framework of this paper is shown in Figure 1.

This research derived from the National Enterprise Credit Information Publicity System (NECIPS), an official database that includes information of all business registered establishment in China, which is constructed by the State Administration for Industry and Commerce (SAIC) in recent years driven by the internet economy. This paper mainly selected “logistics” as the keyword and acquired data on 7012 logistics enterprises registered in Chongqing from 1982 to 2016. We deleted unusable or incomplete data; the final data set consisted of 5923 valid samples. The main attribute includes company name, registered address, business unit, registered capital, business type, and other information. These data are of authority, reliability, and accuracy, but they were not open to the public until 2014, and empirical research based on this did not be used until 2018 [33]. As the only municipality directly under the central government in the western region of China, the development of logistics enterprises is under tremendous pressure to transform. Of course, as a China-Europe logistics hub under the Belt and Road Initiative, Chongqing’s logistics market has expanded, and the development environment has improved significantly. Choosing Chongqing as the object of the study will play an exemplary role in representing the china logistics market.

Business diversification is usually measured by objective methods built on established classification systems in which each of a firm’s establishments is classified according to its primary classification or activity. The logistics industry, a typical composite industry with economic activities scattered in different industry codes, is not included as a separate category in the CIC. Therefore, the previous research is inconclusive and has led to a need for establishing a particular Logistics business classification system to extend and support the researches. Based on the logistics industry classification in the logistics cluster [34], the logistics enterprise’s businesses can be divided into two categories: related businesses and unrelated businesses. Then, this paper builds a logistics business classification system based on the Chinese Industrial Classification (CIC) of Economic Activities (GB/T4754-2017), as shown in Table 1.

According to China’s “Logistics Terms” (GB/T18354-2006), the related business of logistics enterprises is divided into transportation, warehousing, loading, and unloading, distribution processing, express delivery, and port services. In terms of the division of unrelated businesses, according to the Chinese State Council’s Medium and Long-term Plan for the Development of the Logistics Industry (2014-2020), the unrelated businesses of logistics enterprises are divided into infrastructure and operations, information consulting services, supply chain services, multimodal transportation, sales and recycling logistics. According to the nature of the related and unrelated business combined with the rules of CIC, the category can be divided into different sub-categories.

1) Business diversification

Objective methods are the most common measurement index of diversification in research on strategic management and industrial organization [35]. The objective methods have the advantages of standardization, objectivity, and replicability. Considering that research samples are mostly micro-, small- and medium-sized logistics enterprises whose information is not fully disclosed, based on the logistics business classification system, this paper uses the mean narrow-spectrum diversification (MNSD)-based business count methods to measure the degree of diversification which does not require data on sales or revenues of businesses but still provides insights into both the degree of diversification and its direction [36]. The calculation formula is:

d r = d d u = M N ( N ≠ 0 , M > N ) (1)

where d indicates the total degree of diversification of logistics enterprises, du indicates the degree of unrelated diversification, and dr indicates the degree of related diversification. M is defined as the number of four-digit industry a firm operates in. N indicates the number of two-digit CIC codes in which a firm concurrently participates. Using du, dr as two dimensions of a four-lattice cell-matrix, to divide the degree of corporate diversification into four groups, as shown in Figure 2.

The low-diversification group refers to logistics enterprises with low unrelated and related diversification. The high-diversification group refers to logistics enterprises with high related and unrelated diversification. The related diversified group refers to logistics enterprises with a high degree of related diversification but a low degree of unrelated diversification. The unrelated diversified group refers to logistics enterprises with a low degree of related diversification but a high degree of unrelated diversification.

2) Firm size

According to domestic and international studies, scholars mainly select EMPL (total number of employees), SALES (sales revenue), and ASSETS (total assets) as indicators of size [37] [38]. The logistics industry is labor-intensive and capital-intensive. There may be deviations in measuring the scale based only on the number of employees. Enterprise sales data fluctuate significantly due to occasional factors such as market demand and are only disclosed in listed enterprises. The firm size emphasized in this paper is the allocation level of the production factors. Total assets, as a kind of financial capital, can scientifically reflect and monitor the production factor resources existing in the enterprise and provide decision-makers with information about resource allocation [39]. Considering the data availability, this paper uses registered capital instead of the total asset to measure the firm size [40]. Based on China’s classification criteria for small and medium-sized enterprises, considering the quality of business registration data of logistics enterprises, the registered capital of 100,000 and below, 100,000 to 5 million, 5 million to 10 million, above 10 million, respectively represent micro-, small-, medium- and large-sized logistics enterprises.

1) Degree of diversification of all-sized logistics enterprises

Figure 3 shows descriptive statistics of the degree of enterprise diversification in Chongqing from 1982 to 2016. From 1982 to 1996, the changes in the level of diversification are not significant due to the small number of logistics enterprises. After Chongqing became the only municipality directly under the Central Government in West China in 1997, the logistics industry developed rapidly, with an average annual increase of 256.5 enterprises. Overall, the change in the average level of unrelated diversification and related diversification of all enterprises was stabilized, with the former increasing from 2.12 to 2.41 and the latter from 1.10 to 1.16. In terms of the changes in the four diversified groups, the percentage of enterprises in the low-diversified group declined at an average annual rate of 16.32%, the percentage of enterprises in the high-diversified group increased at an average annual rate of 14.23%, and the percentage of enterprises in the related-diversified group changed little and remained stable. As for the percentage of enterprises in the unrelated diversified group, it shows an inverted U-shaped trend, reaching a peak of 27.95% in 2009 and then declining.

2) Degree of diversification of different-sized logistics enterprises

Figure 4 shows the diversified distribution of micro-, small-, medium- and

large-sized logistics enterprises from 1982 to 2016.

In micro-sized enterprises, the average proportion of low diversification increased from 56.21 percent (1997-2004) to 71.13 percent in 2005 and then decreased to 55.1 percent (2009-2016), the average proportion of unrelated and high diversification grew slowly (average annual growth rates are 0.3 percent and 0.4 percent respectively). In small-sized enterprises, the proportion of low and unrelated diversification has declined slowly (the average annual decline rate is 1.3% and 0.4% respectively). As for medium-sized enterprises, from 1997 to 2008, the proportion of the low diversification and unrelated diversification has increased rapidly (overall growth rates were 20 percent and 17 percent respectively), while the proportion of the related diversification and high diversification has declined significantly (overall decline rate were 10 percent and 27 percent respectively). From 2009 to 2016, the proportion of related diversification and high diversification has grown slowly (average annual declines rate were 0.5 percent and 1.4 percent, respectively), while the proportion of low diversification and unrelated diversification has declined slowly (average annual declines rate were 1.5 percent and 0.9 percent respectively). For large-sized enterprises, from 1997 through 2016, the proportion of low diversification decreased by about 20 percent, the proportion of related diversification increased by approximately 16 percent. The proportion of unrelated diversification and high diversification fluctuated between 15% - 19% and 26% - 30%, respectively, and remained stable.

In general, all enterprises have been moving toward high diversification (average annual growth rates are 14.23%). Before 2008, the trends of diversified development in large and medium-sized enterprises were more diverse and did not show consistent characteristics. Subsequently, small and medium-sized enterprises developed toward relevant diversification and high diversification, micro-sized enterprises developed toward unrelated diversification and high diversification, and large enterprises were more stable in all four diversification directions.

1) The relationship between the firm size and the degree of enterprise diversification

To determine whether there is a correlation between the firm size and the degree of enterprise diversification, a regression analysis is performed on the sample data of 5923 enterprises registered. We used OLS to test hypotheses, which were developed by Legendre and Legendre [41]. The OLS is a widely used technique for estimating the parameters of a linear regression model. Chen and Ho [11] advised analyzing the relationship between diversification and firm size by OLS regression analysis. Based on the above literature and theoretical hypothesis, we constructed the following regression equations to test the relationships in the theoretical model:

diversifition i = α 0 + β ln_scale i + ε i (2)

This article adapts the statistical software, Stata 12.0, to carry out the data analysis, the descriptive statistics of the variables, the correlation analysis among the variables and the regression analysis. The detailed analytical procedure is as follows: Model 1-1 consider the impact on the firm size on the total diversification, and independent variable is d; Model 1-2 consider the impact on the firm size on the unrelated diversification, and independent variable is du; Model 1-3 consider the impact on the firm size on the related diversification, and independent variable is dr. Table 2 reports the estimated results of the relationship

***p < 0.01.

between firm size and total diversification, related diversification, and unrelated diversification. In Hypothesis H1a, H1b, the authors proposed that the firm size would positively affect the level of related diversification and unrelated diversification. Model 1-2 and Model 1-3 provide strong support for this hypothesis under the condition of p = 0.01. This implies that firm size positively affects a firm’s diversification operation, whether diversify into unrelated business or related business. The result is in line with previous research accomplished by previous studies [14] [18] [19] [23]. Thus, business owners or managers should take into account the implanting diversification strategy in expanding scale.

2) Degree of diversification of enterprises in different sizes

Even if there is a positive correlation between firm size and the degree of diversification, it remains to be seen whether this conclusion is still applied to the different-sized enterprises, especially for micro-and small-sized logistics enterprises. Therefore, regression analysis is performed on the relationship between the size and the degree of the total, unrelated and related diversification in micro-, small-, medium- and large-sized enterprises. The results are shown in Table 3 and Table 4, Table 5.

Model 2-1, Model 3-1, and Model 4-1 are used to explore the relationship between firm size and total diversification, the unrelated diversification, and the related diversification in large-sized logistics enterprises. Model estimation result shows that firm size is not related to the degree of total diversification, unrelated diversification, and related diversification. So, the H2a is not supported.

Model 2-2, Model 3-2, and Model 4-2 are used to examine the relationship between firm size and total diversification, the unrelated diversification, and the related diversification in medium-sized logistics enterprises. In three models, the results showed that firm size is not related to the degree of total diversification, unrelated diversification and related diversification. So, the H2b isn’t supported.

***p < 0.01.

***p < 0.01.

***p < 0.01.

Model 2-3, Model 3-3, and Model 4-3 are used to examine the relationship between firm size and total diversification, the unrelated diversification, and the related diversification in small-sized logistics enterprises. In three models, the results showed firm size is positively related to the degree of total, unrelated and related diversification. So, the H2c is supported.

Model 2-4, Model 3-4, and Model 4-4 are used to test the relationship between firm size and total diversification, the unrelated diversification, and the related diversification in micro-sized logistics enterprises. In three models, the results showed that firm size is positively related to the degree of total diversification and the unrelated diversification but does not correlate with related diversification. So, the H2d is supported.