e-Commerce and Supply Chain Management in China

Theoretical background

Relevance of international trade models and transport cost theories

E-commerce is one of the prospering and competitive spheres of businesses in China. Jiang and Prater correctly estimated that e-commerce adaptation would take time from Chinese logistics system and infrastructure, however, once done, it would change significantly this country. The long-term expectations for this digital market remain high not only within one country, but across the global market, where a full potential is far from being realized, especially in the sphere of cross-border trade. Absolute e-commerce cross-border trade was estimated to be 6.66 trillion yuan (around one trillion USD) in 2016, and this mostly to export direction. Based on the statistics of China's Custom, absolute e-commerce cross-border trade is 8.06 Trillion yuan in 2017 (approximately 1.2 trillion USD). In the first half of 2018, absolute e-commerce cross-border trade reaches 4.5 trillion yuan (almost 0.67 trillion USD; 100ec.cn 2019). One of the driving factors of its growth is in the development of platform-based services, and particular, shopping platforms of Alibaba Group, such as Aliexpress, Pandao, Joom, Wish that are used in Russia. For some other foreign markets is used Lazada in Vietnam, Indonesia, Malaysia, Singapore and the Philippines, as well as widespread growth of mobile users contributing to m-commerce evolvement.

The current study looks into the invisible part of the online trade, such as logistics, purchasing and distribution practices, as well as initial manufacturing phases that represent the offline side of omnichannel shoppers and should be considered from an integral point of view. Before any goods (tangible or intangible) appear on the company's websites and purchased by consumers, they have to be produced, undergo the value-added operations and then delivered to the required locations. Forward-thinking companies may be aware of this need for simultaneous developments, but do not know where to start and make the right decisions on the online business development that drive margin and profitability growth in both channels.

One of the sources for gaining financial benefits is hidden in the economic concept, which dates back to Smith. In particular, the economies of scale theory can be applied to a variety of organizational and business situations, because when the economy of scale is realized, the economic growth can be achieved. According to the theory, when average costs starts falling as output increases, then economies of scale are occurring. For doing so, it is possible to purchase inputs at a lower per-unit cost when they are purchased in large quantities, or use other sources of economies. For example, in wholesale and retail distribution, the increase of the speed of operations, such as order fulfilment, can lower the overall costs of doing business. Continuing notions from the landmark book of A. Smith 'The Wealth of Nations', where transport supports the trade in goods on the national level, the article takes logistics and trade perspective under the focus.

Logistics is considered as indispensable part of studies on location of e-commerce enterprises. For example, the research on dual-channel pricing under deterministic demand discusses the various logistical determinants of price. Authors, in turn, apply location-based pricing of a food retail company taking into account spatial distribution of customers on logistics costs, customers' channel preference and service providers' pricing. Li et al. studied the phenomenon of returns in e-commerce through the location-inventory problem, which was solved by lagrangian relaxation combined with ant colony algorithm (LRCAC). XiaoYan et al. also developed reverse logistics network model, but under the assumption of uncertain demand and return. Other recent studies likewise focus on logistics-oriented decisions for online shopping, providing location orientation of E-shops and joint distribution center by a new hybrid fuzzy multiple-criteria decision-making methods.

Basic approaches for the analysis of international e-commerce

Apart from logistics and supply chain theories, the feasibility studies on the e-commerce enterprise location in China include other well-tested approaches and methods of scientific analysis. Due to the fact that omnichannel enterprises in cross-border trade are represented by large ecosystems processing of goods in international trade, the explanation of the arrangement of their multi-functional structure must be found additionally in the theories of trade and transport cost theories (Fig. 1). Specifically, the figure schematically shows the relationship of the basic approaches that were taken into consideration to justify the Chinese provinces advantageous for omnichannel shoppers, with the allowance for logistics and trade perspective of cross-border trade.

Fig. 1 Theories advocating for omnichannel shopper's development

Despite the fact that some of the theories could be called by the authors as approaches or simple models, they are still fundamental in the international business. For the development of trade theory, first and foremost, Smith made significant contribution. Adam Smith argued that countries should specialize in production of goods for which they have absolute advantage (e.g. French in production of wine). Depending on countries absolute advantages in producing specific types of products, it is possible to find out what products are profitable to export and what is useful to import. On the other hand, countries should import products, which can be made with lower cost in the partner country. Specifically, David Ricardo's theory of comparative advantage took one step further from A. Smith's theory, suggesting in his book 'Principles of Political Economy' that countries should specialize in the production of those goods they produce most efficiently and buy goods they produce less efficiently from other countries. These theories include many unrealistic assumptions and naturally have a lot of limitations that make them suitable mainly for explaining trade flows between the nations.

Unlike Ricardo's theory, the Heckscher-Ohlin argues that the pattern of international trade is determined by differences in factor endowments (like land, labour, and capital), rather than differences in productivity. Thus, Swedish economists Eli Heckscher (in 1919) and Bertin Ohlin (in 1933) put forward a different explanation of comparative advantage. Another famous Harvard strategy professor, Michael Porter, has also written extensively on international trade. Porter theorizes that four broad attributes (factor endownments; demand conditions; related and supporting industries; firm strategy, structure, and rivalry), constituting a diamond, bring the creation of competitive advantage. In the framework of Porter' work "National Demand of Competitive advantage", Chen and Ning evaluated China's progress towards building electronic commerce.

It should be noted that these international trade theories, at the same time, have important implications for business practice (location, first-mover, and government policy implications). Specifically, many economists believe and stress in their classical trade theories (Smith, Ricardo, and Heckscher-Ohlin) that unrestricted free trade between nations will raise the economic welfare of countries, which participate in a free trade system. In light of current governmental policy, e.g. tariffs, subsidiaries, import quotas, export restraints, local content requirements, administrative policies, and antidumping duties, which frequently intervene and influences the international trade, the attention should be drawn to benefits of the free trade explained by the theories (higher level of domestic consumption, more efficient utilization of resources that stimulate growth and creation of wealth of nations). Still, the critics, for example from Paul Samuelson, show that free trade has historically benefited rich countries, and, of course, the political reality of international trade is beyond from the books' models and approaches.

Moreover, in the works of authors on international trade the attention to transport costs and related barriers is scarce. Meanwhile, transport significantly determines the trade in goods and services, therefore, the location of e-commerce enterprise was considered with the accessibility to the transport infrastructure and with the allowance for transport costs. Economic analysis of the transport sector is elaborated in the works of Oliver E. Williamson and Peter Nijkamp, theoretical and empirical studies of transport costs were provided by Paul A. Samuelson, D. Hummels, P-P. Combes and M. Lafourcade, while Russian school in the sphere of transport and logistics is represented by K.V. Kholopov, L.B. Mirotin, E. L. Limonov, E.G. Efimova, N.A. Zhuravleva, and A.S. Balalaev.

Further on, going down from the national level to the regional, the omnichannel ecosystem was regarded as a network of various companies, referring to logistics and supply management theory. The question of the formation of the scientific base of the integrated supply chain management is reflected in many works of B. A. Anikin, V. V. Dybskaya, V. S. Lukinskiy, L. B. Mirotin, Yu. M. Nerush, O. D. Protsenko, Yu. I. Ryzhikov, V. I. Sergeev, S. A. Uvarov and others. The theories of logistics and supply chain management are correlated with the 40 theories of marketing, management and organization theory.

The scientific basis for the management of material, information, financial, service and personnel flows is, first of all, is grounded on models and methods of the theory of logistics, management and marketing: transport models; network models; deterministic and probabilistic models of dynamic programming; deterministic and probabilistic models of inventory management; Markov decision-making processes; Queuing systems; theory of games and decision-making, respectively, in terms of certainty, risk, uncertainty; simulation modeling; method of dynamics of averages; forecasting methods.

Location problem and multi-criteria decision-making methods

To define the advantageous location of the enterprise in the sphere of e-commerce cross-border trade, depending on the type or complexity of location problem (single facility location problem, multi-facility allocation or location-allocation problem) different methods can be applied (e.g. linear, non-linear programming, simplex algorithm, lagrangian relaxation, branch & cut methods, branch and bound, local beam search, tabu search, artificial neural network, expert systems, fuzzy control, generic algorithms, multi-agent systems, and so on). In the current study, the general choice algorithm is applied, which was for the first time proposed by Lukinskiy et al. and further used by Lukinskiy and Lukinskiy for the selection of intermediaries in the supply chain. In the earlier publication, Lukinskiy and Katkova provided the comparative evaluation of the analytic hierarchy process (AHP), point-rating assessment method (PRA) and proposed general intermediaries' choice algorithm (ICA).

According to the authors, the ICA method is more difficult than the PRA one, but its objectivity is higher. At the same time, the ICA method produces almost the same evaluations as the AHP method does. In spite of AHP method popularity, it is often criticized because of a series of drawbacks. It allows some small inconsistency in augments because human is not always consistent, therefore, there is an inability to adequately handle the inherent uncertainty with the AHP method. Moreover, the AHP method possesses an advantage only when there is no any quantitative (tangible) information except for the experts' evaluations. In the below analyses, the tangible information was collected about all relative factors for the choice of alternatives.

The choice was made among 31 Chinese provinces based on the 15 factors that were partly mentioned in the similar studies: (1) Number of cross-border e-commerce zones, (2) Access to the sea, (3) Number of postal offices, (4) Number of outlets for express services, (5) Length of postal routes, (6) Retail purchase value, (7) Retail sales value, (8) Retail stock value, (9) Number of corporate enterprises of retail trade, (10) Number of websites, (11) Number of Internet users, (12) Broadband subscribers port of Internet, (13) Labour productivity, (14) Wages level, and (15) Unemployment rate.

For multi-criteria decision-making, the general choice algorithm (GCA) was used to rank the most advantageous regions for the development of e-commerce enterprise. GCA, as AHP, is grounded on the pairwise comparison. In a pairwise comparison, two alternatives are compared according to criterion and one is preferred. The objective of AHP is to choose the best alternative. The decision maker selects the alternative that best meets the decision criteria. As a rule, the general mathematical process includes several steps. In the first step, it is required to mathematically determine preferences (e.g. for the site) with respect to each criterion. In the second step, mathematically identify preferences for criteria (rank order of importance). Then, combine these two sets of preferences to mathematically derive the composite score for each site. Finally, select the site with the highest score.

According to GCA, at the beginning, the criteria should be divided into three groups: quantitative, qualitative and relay (or 'killer-evaluation'). Then, the pairwise comparison is used to rank the criteria. Quantitative data processing is carried out by the qualimetry methods, and to obtain quality criterion values authors suggested to use Harrington desirability function. The calculation of the integral estimates is a sum of qualitative and quantitative criteria evaluation. In the current research, all data were of the quantitative nature, which again justifies the use of GCA, because AHP method 'possesses substantial advantages without having any quantitative (tangible) information except for the experts' evaluations'.

GCA and AHP both utilize pairwise comparison. In this regard, GCA is similar to AHP. Pairwise comparison was carried out based on the mathematical steps. Firstly, a pairwise comparison matrix for each decision alternative for each criterion was developed. Afterwards, the synthesization was done in several sub-steps. That is the row value of the pairwise comparison matrix was summed up in the last column. Next, the sum of the last column was found. Afterwards, each value in last column was divided by its column sum. Finally, the value in each row of normalized matrices was averaged, and the sum of the row averages in the normalized matrix equaled to 1. The row average values represented the preference vector. Preference vectors for other criteria are computed similarly, resulting in the preference matrix (that is step 1 in AHP).

As a result, the final criteria preference matrix was used for the overall ranking for the decision alternatives, assuming the equal importance of the criteria. So, with GCA, preferences for the site with respect to each criterion were determined mathematically. GCA did not utilize preference scale, which assigns numerical values (as a rule, from 1 to 9) to different levels of importance. GCA did not resort to preference scale for pairwise comparison because quantitative (tangible) information was available for each criterion. In conclusion, it is worth to mention that the research was partly done in the Henan Key Laboratory for Big Data Processing & Analytics of Electronic Commerce, which has an affiliation to Luoyang Normal University.