Analysis of the Efficiency of Transport Infrastructure Connectivity and Trade

Abstract

Analyzing the efficiency of transport infrastructure connectivity and trade in the Regional Comprehensive Economic Partnership (RCEP) is very important for regional integration for international trade in the RCEP. This study aims to significantly measure the efficiency of the connectivity of infrastructure in the RCEP for improving the performance of infrastructure connection and suggest the way to improve the connection of infrastructure. Therefore, the input and output variables of infrastructure connectivity have been inserted to achieve this objective. The inputs are: the number of ports, rail range, and road networks, the number of land borders, the number of maritime borders, number of cross border points, railway linkage with other countries, number of ports connected with railways, and the number of ports connected with road base on the “intermodal and multimodal concept”. On the other hand, the output factors most related to trade and economics are GDP, transport, import, and export volume. The paper applied DEA (Data Envelopment Analysis) model by using DEAP software to analyze the data. The result reveals that the efficiency of infrastructures connectivity and international trade in 10 countries were efficient and 5 countries were inefficient. The research study presents ways of development to improve the connectivity by investing in the basic infrastructures, such as increasing the logistics connection points and driving forward for international trade in the RCEP.

1. Introduction

The Regional Comprehensive Economic Partnership (RCEP) is one of the new accesses for international trade with fifteen countries: Australia, Brunei Darussalam, Cambodia, China, Japan, Indonesia, the Lao People’s Democratic Republic, Malaysia, Myanmar, New Zealand, the Philippines, Singapore, the Republic of Korea, Thailand, and Vietnam [1]. The RCEP plays a significant role in the world of transport and economy, with a huge land area and a population of around 2.2 billion people, with a total GDP of 38 trillion USD. It takes 30% GDP of the world’s total global trade. The RCEP also aims to promote economic integration, facilitate trade and tariff concessions, and reduce border costs and customs procedures [2].

The infrastructures and connectivity in the RCEP are the main challenges to promoting economic and international trade [3]. The connectivity is important to link the community, regional integration, transportation, and direction of international trade import and export performance [4]. Also, connectivity has increased the interaction between productivity, operation, economic competition, and regional integration in the RCEP. However, the infrastructures in the RCEP have major problems, facing by the geographical location, poor quality of infrastructures, connecting of transport modes, and infrastructure networks [5,6].

Previous studies related to “connectivity” have been terminated with transport service, network analysis, and concerns to logistics performance with international trade flow. Moreover, some research related to the measurement of the performance using Data Envelopment Analysis (DEA) has been carried out to analyze the performance of complex infrastructures such as ports, airports, and roads quality containing private and public transport [7,8]. Nevertheless, studies have been carried out on DEA (Data Envelopment Analysis) to measure infrastructure connectivity with “intermodal and multimodal concept” versus trade to measure the efficiency of infrastructure connectivity in each country of the RCEP.

Also, the purpose of this study is to evaluate the efficiency of infrastructure connectivity combined with international trade in the RCEP. This paper can defend infrastructure connectivity using the “intermodal and multimodal transport concept,” which widely considers the various modes of transport, connections between modes, and transfer from one mode to another by linking infrastructures and transportation modes [9]. It can be applied by using road and rail networks, the number of borders and mode connection inserted to be an input to compare with the output, which is trade (GDP and volume of import and export) [7]. The DEA result can then create awareness of the inefficient infrastructure connectivity and proposes a way to improve efficiency. On the other hand, the research question in this study focuses on: (1), which countries have good and efficient infrastructure connectivity effects to trade in the RCEP? And (2), what are the suggestions and solutions to improving the efficiency of infrastructure connectivity and trade efficiencies? This study uses archival data for efficiency by applying the Data Envelopment Analysis (DEA) approach.

Generally, this research contributes to measuring and providing the means of improving the infrastructure’s connectivity; the connection between transport mode and infrastructures drives the efficiency of transshipment. It also matters to decrease the cost of trade and upgrade the economics on the regional integration to enrich the connectivity, and fulfilling the weak points of inefficient connectivity by taking an action plan and improving trade and transportation.

The main issue addressed in this chapter is: (1) measuring the efficiency of infrastructures and connectivity, and (2) improving inefficient infrastructure connectivity. Therefore, the second part gives the literature review of Data Envelopment Analysis (DEA) and related research. Thereafter, the methodology and data collection, and the main results were presented. The research also provides a way to improve the inefficiency of infrastructures and connectivity. In the last section, the conclusions to the main finding and implications are outlined.

2. Literature Reviews

Several studies on DEA on transport infrastructure have been established in various areas such as efficiency port performance, trade economics, and transport infrastructures such as road, railway, port, and airport. The previous research that analyzed cross-country evaluation of rail freight transport in the (European Union) EU during 2005–2009 using data envelopment analysis (DEA) in areas related to the infrastructures, connectivity, and trade shows that Poland is ranked first in rail transport efficiency for cross–countries [10]. Another study was carried out to measure the logistics performance index and trade supply chains in the EU by using geographical and income factors to consider efficiency. The study found that the developed countries had good performance in transport and logistics [11]. Wiegmans et al. used DEA and Stochastic frontier analysis (SFA) to analyze the efficiencies, and applied SFA to estimate a transformation and explain the inefficiencies between rail and road freight transport networks in Canada, the USA, and Europe by using infrastructure factors such as length, number of vehicles, number of enterprises, employment in the transportation sector, and freight and vehicle movement. The results showed that the infrastructures in those countries were efficient and suggested policy creation with an aim to increase efficiency [12]. Another paper aimed to establish the coordination between freight transportation and economic growth. The authors presented DEA and spatial-autocorrelation tests to analyze freight transport of four regions in China [13]. The study of the international efficiency of logistics using DEA to evaluate logistics performance index (LPI) were related to the quality of trade and transport infrastructures in European countries. The results found that the LPI in the EU was in the high range, but some were not efficient [14]. Kyriacou A et al., 2019, used DEA to analyze the efficiency of transport investment by measuring the infrastructure quantity given the investment volume and explaining the government’s role in the investment of infrastructures by crossing countries [7]. Another author (Caulfield B et al., 2013) applied DEA on transport to analyze the efficiency among airport routes to compare the value of money [15]. Similar to Barros’s research, the efficiency of airports in Italy was compared to the economy, and the least efficient airports were improved [16]. The author also applied DEA to analyze the performance of benchmarking in the Indian railway container business by considering the organization’s profit. Lastly, this research provides the potential and establishes improvement methods [17]. Similarly, Hilmola used the DEA method to study railway freight transportation in Europe, which considers the demand, market share, and productivity in railway transport. The results found that improving freight wagons, the staff, and transit traffic can increase efficiency [18].

Alternative approaches to infrastructure connectivity measurements have been presented by Decrute. He used the network analysis to analyze the connectivity of liner shipping networks and traffic to consider centrality, connectivity, and vulnerability [19]. On the other hand, the researchers studied the complex infrastructure’s performance while investigating Malaysia’s trade by applying the Gravity model. Then, the random effect model and fixed effect mode were used for robustness [20]. Another research by Bensassi S et al., 2015, also applied the gravity model to establish the relationship between transport infrastructure and trade, to reduce the transportation cost and increase the trade flow, with concerns on the border, distance, total income, trade agreement, logistics index, and infrastructure [21]. Likewise, the paper presented by A. Host uses the Gravity model to examine the international trade and Logistics Performance Index (LPI). The results found that the economy has a positive impact on trade and distance leads to negative trade flow [22]. The study conducted by S. Li et al. focuses on analyzing the geography factors and international trade in China, Myanmar, and Vietnam using the conditional logit model to analyze the relevant factor related to the geography. The results show that the scale of freight, duration, transportation charge, distance, quality of infrastructures, geographical location, and characteristics of the shippers and POEs influence and propose new policies [23]. While some authors (S. Hanaoka, 2021), propose to identify the factors for crossing borders between the land corridor and maritime border which influence trade by using the Tobit model to analyze the factors, the results of this study showed that distance and infrastructure quality are the most important factors [24]. On the other hand, the connectivity-related network and trade presented by L. Tavasszy et al., used a strategic model and scenario analysis to analyze transport demand and predict constrainer flow at port and hinterland based on the trade information between the Trans-Siberian rail line and arctic shipping route [25]. This research can link the quality of logistics indicated by the logistics Performance Index (LPI) and international trade as the quality of the infrastructure combined with economics by using the gravity model by Gani. The outcome showed that LPI and trade flow could increase international trade [26]. Also, other transport connectivity indexes such as rail network connectivity, road network connectivity, and shipping network connectivity indicated the logistics and transport services.

Several studies related to international trade measure the factor on the movement of goods, policy, investment, and government processes by using the gravity model to analyze the effect on trade facilitation [10]. In contrast, studies have been carried out to measure infrastructure performance such as port, seaport, container terminal efficiency, and others [27], although no research has been carried out to measure the connectivity. This aspect brings novelty to this study to analyze infrastructure connectivity efficiency in the RCEP using Data Envelopment Analysis (DEA). After reviewing many research papers related to the DEA, the measure of infrastructure connectivity performance by applying the “intermodal and multimodal concept” was used to insert input factors. The output factors carried out to trade such as GDP and GDP of transport (import export) in the RCEP area can bring novelty for these studies. However, this research uses DEAP software version 2.1 developed by Tim Coelli [28] to measure the data and identify the infrastructure connectivity and trade efficiency level. On the other hand, the outcome can point out the advantage of intermodal transport and international connectivity to support the transport policy, international trade, and implications for development infrastructures and logistics. Lastly, it provides an improved scale of efficiency for the increase and decrease in the input and output to make connectivity efficient. All these reasons make this research different from the previous studies.

3. Materials and Methods

DEA is the most popular method used to measure efficiency and performance in several areas such as operations, production, infrastructures and economics [29,30]. Data Envelopment Analysis (DEA) is one of mathematical programs to evaluate the efficiency of various inputs and outputs data by giving the results of efficiency called “Decision management units” (DMU). Thus, the results of this method can suggest the direction for developing the unit performance and more efficiency [31].

The unit of $DM{U}_s$ where each unit $DM{U}_j$ $DM{U}_j$ (j = 1..., $n$) by inserting $m$ to be an input factors (i = 1,...,$m$). Later, perform to insert output factors $s$ = ($r$= 1,...,$s$) to measure the performance of connectivity by $DM{U}_j$. The DEA model can express on the mathematical equation below.

$$Max{h}_j=\frac{{{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{rj}}{{{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{ij}}\mathrm{for}\mathrm{all}j=1,2,\dots \dots ,n$$

(1)

Subject to

$$\frac{{{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{rj}}{{{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{ij}}\le 1, j=1,2,\dots \dots ,n.$$

(2)

$$\mathrm{Where};v_i\ge 0,(i=1,2,3\dots ,m)$$

(3)

$$u_r\ge 0(r=1,2,3\dots ,s)$$

(4)

The equation is explained as; $h_j$ is a result the efficiency of DMU, $y_{rj}$ is amount of output factor and $x_{ij}$ is amount of input factors. Then $n$ shows the number of DMU involving; $s$ is number of output and $m$ the number of inputs. In addition, $u_r$ is the weight given to output $r$ of DMU unit and $v_i$ is the weight given to the input i of DMU unit. However, the number of $DM{U}_s$ modify where each $DM{U}_j$ (j = 1,..., n) by insert $m$ input on the variable $x_{ij}$ (i = 1,..., m), then utilized to the variable of output $y_{rj}$ (r = 1,..., $s$) to measure the efficiency performance by $DM{U}_j$. Moreover, the value of $h_j$; can express by applied DEA for evaluation. The results of $h_j$; should range between 0 and 1 to show the value of efficiency. When the results of DMU $h_j$ is equal to 1, it means that the DMU is efficient. In contrast, if the results of DMU $h_j$ is nearly but not equal to 1, it will be indicated inefficient [31,32].

However, the DEA model created by Charnes, Cooper and Rhodes (CCR) measured the efficiency the results were call constant returns to scale (CRS) which earn from insert the total input and total output solving by mathematic program for. Later, BCC model have been developed by Banker, Charnes and Cooper in 1984 to measures one of the technical efficiency index called “Variable Returns to Scale” (VRS) [29,33].

3.1. DEA-CCR Method

CCR (Charnes, Cooper and Rhodes) model delivers the results call CRS (Constant Return to Scale). The CCR model can generate the efficiency results by inserting the input and output data of each DMU.

$$Max h_0={{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{r0}$$

(5)

Subject to

$${{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{i0}=1$$

(6)

$${{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{rj}-{{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{ij} \le 0$$

(7)

$$j=1,2,\dots ,n.$$

(8)

$$u_r,v_i\ge \forall r,i$$

(9)

The equation above represents $h_0$ is the efficiency of DMU, A number of DMU represented by $n$, is the amount of input factors is $r$, the number of output values is $i$. Then, the volume of the input factors for the DMU present by $x_{ij}$ and the output factors of DMU present by $y_{rj}$. Whereas the weight of the output factor existing by $u_r$ and $v_i$ is existing for the weight of the input factor [32,34].

3.2. DEA-BCC Method

Banker Charnes and Cooper developed Constant Return to Scale (CRS) of CCR model to be a “BCC model”. So, the BCC model returns the efficiency scale which is called a Variables Return to Scale (VRS).

$$Max h_0={{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{r0}+u_0$$

(10)

Subject to:

$${{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{i0}=1$$

(11)

$${{\displaystyle \sum }}_{r=1}^s{u}_r{y}_{rj}-{{\displaystyle \sum }}_{i=1}^m{v}_i{x}_{ij}+u_0\le 0$$

(12)

$$j=1,2,\dots ,n$$

(13)

$$u_r,x{v}_i \ge \in , \forall r,i$$

(14)

A return to scale of decision-making unit’s (DMU) is presented by $u_0$. The definition of $u_0$ it is can explain that in the case of $u_0$ equal 0, it can indicate that the DMU will be operated under the utility of production. While, $u_0$ > 0 the DMU suggest the improved way to represent the decreased return to scale (drs). In contrast, $u_0$ < 0 DMU provides an increased return to scale (irs) to improve operation and performance for more efficiency [33,34].

Therefore, the value of DMU can indicate efficiency by the results of $h_0$ equal to 1 in the CCR model and BCC model. In contrast, the value of DMU indicates inefficiency in case $h_0$ is less than 1. However, the results of $h_0$ provide the percentage of enhancement by increased or decreased input and output of DMU [35].

3.3. Scale Efficiency

The efficiency scale refers to the technical efficiency, which conducts CRS and VRS in the same data. It is used to calculate the differences between BCCTE (Banker, Charnes and Cooper Technical efficiency) of the BCC model, and CCRTE (Charnes, Cooper and Rhodes Technical efficiency) of the CCR model [28]. The results of the technical efficiency score should be between 0 and 1. If the result is nearly zero, it implies that the unit is far from being efficient. Likewise, if the result is close to 1, it indicates that the unit is nearly efficient [36].

$$\mathrm{Scale}\mathrm{efficiency}=\frac{CCRTE}{BCCTE}$$

(15)

3.4. Data Collection

The DEA method was used for the assessment of the efficiency; the processing starts by selecting the appropriate inputs and outputs factors. The data chosen were related to the “Intermodal and multimodal transport concept,” with connection to various modes of transport, transit of freight, connections between modes, and transfer of cargo from one mode to another by linking infrastructures and transportation [9]. In addition, this research creates the study idea in the new problem, such as considering the efficiency of infrastructure connectivity instead of only the infrastructure performance and trade in the RCEP by using the DEA method. Therefore, the input factor is related to the infrastructures based on the “Intermodal and multimodal transport concept,” such as the number of ports, rail range, and road network in each nation. Also, the input factor related to connectivity is expressed in the number of land borders, maritime borders, cross border points, railway linkage with other countries, the number of ports connected by railway, and the number of ports connected by road. On the other hand, the output factors most related to trade and economics are GDP, transport, and import and export volume which the variable explanation have been presented in

Table 1. DEA input and output variable explanation.

DEA Input and Output Variable Explanation
Input	(A) Number of ports	Number of ports in each country
	(B) Land border	Number of land borders in each country
	(C) Maritime border	Number of maritime borders in each country
	(D) Number of cross border point	Number of cross borders for crossing in each country
	(E) Railway linkage with other countries	Railway linking
	(F) Rail range	Total length of rail
	(G) Road network	Total length of road
	(H) Number of port connection with rail	Number of port terminal have rail connect
	(I) Number of port connection with road	Number of ports with road connect
Output	(W) GDP	GDP transport in RCEP (US Billion)
	(X) GDP in Transport	GDP transport in RCEP (US Million)
	(Y) Import	Amount of import in (thousands of US million dollar)
	(Z) Export	Amount of export in (thousands of US million dollar)

. Therefore, all the data was collected from the official website, research paper, and annual report is showed in

Table 2. Infrastructures connectivity data.

Countries	A	B	C		D	E	F	G	H	I	W	X	Y	Z
		Input									Output
Australia	106	0	7		8	0	14,814	36,064	5	106	1396.57	122.30	221.48	226.37
Brunei-Darussalam	1	2	5		0	0	4400	3029	0	1	13.469	0.01	5.10	7.03
Cambodia	3	3	3		14	3	652	45,256	1	3	27.08	0.03	20.27	14.82
China	172	17	9		115	8	35,000	97,000	21	172	14,280.00	350.51	2068.95	2498.56
Indonesia	154	4	12		16	0	4553	532,837	2	4	1119.19	26.57	171.27	167.68
Japan	292	0	7		7	0	30,625	1,215,000	3	292	5081.77	208.83	720.89	705.63
Laos	0	5	0		24	3	854	21,716	0	0	18.17	0.19	5.80	5.79
Malaysia	25	4	8		25	2	2262	82,144	3	1	364.68	10.30	204.90	238.08
Myanmar	5	5	3		11	4	7944	150,000	1	5	76.08	4.00	18.61	18.10
New Zealand	25	0	2		8	0	4375	10,895	8	25	206.92	1.41	42.70	39.54
Philippines	68	0	8		7	0	77	217,317	4	68	377.00	2.14	117.24	70.92
Singapore	2	1	2		4	1	240	2056	1	1	372.06	4.51	358.97	390.33
South Korea	17	1	3		4	0	2790	4285	4	17	1647.00	12.52	503.26	542.17
Thailand	21	4	7		18	3	4507	390,000	1	21	544.00	4.61	216.80	233.67
Vietnam	32	3	8		30	3	1600	256,684	1	15	262.00	5.50	253.44	264.61

Data Input: A: number of ports, B: land border, C: maritime border, D: number of cross border point, E: Railway linkage with other countries, F: rail range, G: road network, H: number of port connection with rail, I: Number of port connect with road. Data Output: W: GDP (US Billion), X: GDP in Transport (US Million), Y: Import (thousands of US million dollar), Z: Export (thousands of US million dollar).

. Lastly, the DEAP software was used to analyze the efficiency of infrastructures connectivity to enhance the trade in the RCEP to meet the efficiency [30,37].

4. Discussion

The results of the efficiency of infrastructure connectivity and trade by using the DEA model in the RCEP shows the input data considered the number of ports, number of land borders, number of maritime borders, number of cross border points, railway linkage with other countries, rail range, road network, number of port connections with rail, and the number of ports connected with road. Also, the output data considered GDP (US Billion), GDP in transport (US Million), imports (thousands of US million dollar), and exports (thousands of US million dollar) by DEAP software. The results of CRS which calculate by DEA-CCR on the Equations (5)–(9), and the results of VRS receive from DEA-BCC model on the Equations (10)–(14) model to measure the performance of connectivity and scale efficiency from the evaluation of the efficiency of infrastructures connectivity and trade is shown in

Table 3. DEA results of input and output variable.

DMU	Countries	CRS	VRS	Scale Efficiency	Reason	Remark
DMU1	Australia	1.000	1.000	1.000		Efficient
DMU2	Brunei-Darussalam	1.000	1.000	1.000		Efficient
DMU3	Cambodia	0.064	0.7	0.091	IRS	Inefficient
DMU4	China	1.000	1.000	1.000		Efficient
DMU5	Indonesia	1.000	1.000	1.000		Efficient
DMU6	Japan	1.000	1.000	1.000		Efficient
DMU7	Laos	1.000	1.000	1.000		Efficient
DMU8	Malaysia	1.000	1.000	1.000		Efficient
DMU9	Myanmar	0.382	0.677	0.565	IRS	Inefficient
DMU10	New Zealand	0.851	1.000	0.851	IRS	Inefficient
DMU11	Philippines	1.000	1.000	1.000		Efficient
DMU12	Singapore	1.000	1.000	1.000		Efficient
DMU13	South Korea	1.000	1.000	1.000		Efficient
DMU14	Thailand	0.777	0.778	0.998	DRS	Inefficient
DMU15	Viet Nam	0.710	0.712	0.997	IRS	Inefficient

and

Table 4. The value of improving the efficiency.

DMU	Countries	Scale Efficiency	Input	Output	Actual	Target	Potential Improvement	Improving Percentage
DMU3	Cambodia	0.091	A		3	1.408	−1.592	−53.07
			B		3	1.475	−1.525	−50.83
			C		3	1.444	−1.556	−51.87
			D		14	9.794	−4.206	−30.04
			E		3	1.574	−191.426	−47.53
			F		652	456.125	−195.875	−30.04
			G		45,256	7795.404	−37,460.000	−82.77
			H		1	0.700	−0.300	−30.00
			I		3	0.709	−2.2291	−74.30
				W	27.080	265.702	238.622	881.17
				X	0.030	3.211	3.181	10,603.33
				Y	20.270	252.864	232.594	1147.48
				Z	14.820	274.817	259.997	1754.37
DMU9	Myanmar	0.565	A		5	2.819	−2.181	−43.60
			B		5	1.260	−3.74	−74.80
			C		3	2.030	−0.970	−32.30
			D		11	7.443	−3.557	−32.30
			E		4	1.258	−2.742	−68.60
			F		7944	1062.326	−6881.670	−86.60
			G		150,000	11,734.040	−138,265.954	−92.20
			H		1	0.677	−0.323	−32.30
			I		5	2.157	−2.843	−56.90
				W	76.080	275.620	199.540	262.30
				X	4	4.000	0.000	0.00
				Y	18.610	243.060	224.450	1206.10
				Z	18.100	264.027	245.927	1358.70
DMU14	Thailand	0.998	A		21	16.340	−4.660	−22.19
			B		4	2.496	−1.504	−37.60
			C		7	1.520	−5.480	−78.29
			D		18	14.006	−3.994	−22.19
			E		3	1.877	−1.123	−37.43
			F		4507	2442.939	−2508	−55.65
			G		390,000	69471.300	−320528	−82.19
			H		1	0.778	−0.222	−22.20
			I		21	15.891	−5.109	−24.33
				W	544	544.000	0.000	0.00
				X	4.610	15.142	10.532	228.46
				Y	216.8	216.8000	0.000	0.00
				Z	233.67	234.040	0.378	0.16
DMU15	Vietnam	0.997	A		32	4.688	−27.310	−85.34
			B		3	1.554	−1.447	−48.23
			C		8	1.436	−6.564	−82.05
			D		30	10.244	−19.756	−65.85
			E		3	1.610	−1.390	−46.33
			F		1600	777.351	−822.64	−51.42
			G		256,684	22,000.57	−234.684	−0.09
			H		1	0.712	−0.288	−28.80
			I		15	4.010	−10.990	−73.27
				W	262.00	315.916	53.916	20.58
				X	5.500	5.500	0	0.00
				Y	253.44	253.440	0	0.00
				Z	264.61	274.528	9.918	3.75

. The most interesting aspect of the tables is that the scores of CRS and VRS are similar and equal, which implies that each country can operate infrastructure and organize the connectivity to support trade efficiency. The results show that of 10 countries, (66%) were efficient; Australia, Brunei-Darussalam, China, Indonesia, Japan, Lao, Malaysia, Philippines, Singapore South Korea. These countries show a high trade volume appropriate to the infrastructure’s connectivity. Some of the countries, especially in island countries which are located separately from the mainland, such as Australia, Brunei-Darussalam, Indonesia, Japan, Philippines, and South Korea. It has a good efficiency from the high quality of roadway and railway linkage, trade volume is still high, and the domestic infrastructure network is good. On the other hand, China has a huge area and a high number of land borders connected to other countries. The results show efficiency because China can utilize the quality of infrastructure and benefit from transport connectivity with international trade. In addition, Laos, Malaysia, and Singapore, located in the regional interrogation such as Greater Mekong Subregion (GMS), ASEAN Economic Community (AEC) integration, were efficient. These countries utilize the land border to support international trade.

If the efficiency score of CRS and VRS is less than 1, this means the infrastructure connectivity is inefficient in regards to trade [31]. Although, based on the total efficiency score, 5 countries (Cambodia, Myanmar, Thailand, Viet Nam, and New Zealand) were considered inefficient under CRS because the results are not equal to 1, except New Zealand which shows CRS score equal to 1, but VRS is not equal to 1. This can imply inefficiency regardless of the cause, whether from inappropriate management between infrastructure connectivity, land border, transport network, or trade. Thus, the outcome indicates that these countries should improve their infrastructure and connectivity to support more international trade. Based on the efficiency scale, there are two ways to improve the efficiency; by increasing returns to scale (IRS) and decreasing returns to scale (DRS) [14]. The IRS means that an increase in the input factor for leading the outcome should be appropriate with the output factors. In contrast, DRS can imply that a decrease in input factors for leading the outcome should be appropriate with the output factors. The results of this research can reveal the improvement in the efficiency by increasing or decreasing the scale. This can add more transit points (transit mode), extension infrastructures, and connectivity to enhance the connection of transport and infrastructures of their own countries, to bring forward more efficient trade and infrastructures [5]. Lastly, the scale efficiencies being lower than 1 means that infrastructure connectivity is inefficient, because countries manage an excess of inputs related to outputs that affect the scale efficiency results [34].

Scale Efficiency Improvement

After using the DEA method to measure the efficiency of infrastructure connectivity and trades by using DEAP software, the result found that five countries; Cambodia, New Zealand, Myanmar, Thailand, and Vietnam, were inefficient. The details are presented in Table 4. This table also presents the percentage of increase or decrease in the input and output data to improve the infrastructure and connectivity for more efficiency. Some countries have good quality infrastructure and share borders with many countries, but the results show inefficiency, which suggests that connectivity affects trade and transport. Some countries have a high quantity of infrastructure in port, road, and rail networks, which have border connections with the neighboring countries, but the outcome of contribution between the infrastructure and connectivity is less than the beneficial results of trade. Hence, those countries can improve the number of logistics infrastructures and connectivity by decreasing the amount of infrastructure or reducing the connectivity points of infrastructure to consider the trade outcome on GDP, GDP of transport, and the number of imports and export. On the other hand, the results of performance found that each country can maintain the number of infrastructure and connectivity, but all of them need to increase the output of trade. This can then imply that all the infrastructure and connectivity of their countries are efficient for transport and that it can increase international trade in the RCEP.

The results of Cambodia, Myanmar, Thailand, and Vietnam show inefficiency. These countries need to improve on their input and operations to make the output more suitable. Firstly, Cambodia can improve the efficiency according to the results by reducing the road network by 82% and number of ports connected with roads by 72%, and also reduce the number of ports, land borders, maritime borders, and railway linkage with other countries by approximately 50% with the same outcome of trade. Alternatively, Cambodia can maintain the connectivity of infrastructures by increasing the international trade to increase all GDP and import-export volume. On the other hand, Myanmar can improve its efficiency by increasing the input level of its land border by 74.8%, rail range by 92%, road network by 86%, and an average of 30% on the number of ports, maritime border, number of cross border point to drive more international trade and good infrastructures connectivity. Furthermore, Thailand can enhance the road network by 82% and maritime border by 72%, and the other factors between 20–40% to drive more efficient infrastructures connectivity and trade. Lastly, Vietnam can increase the number of ports by 85%, its maritime border by 82%, and the number of ports connecting with road by 73% to make the infrastructure of this country more efficient.

Table 5. The overview of Result analysis in infrastructures connectivity and trade.

DMU	Result Analysis
	CRS	VRS	Scale Efficiency
Efficiency	10 (66.6%)	11 (73.33%)	CRS	10 (66.60%)
Inefficiency	5 (33.3%)	4 (26.66%)	DRS	1 (6.88%)
Total	15 (100.0%)	15 (100.00%)	IRS	4 (26.66%)

shows an overview of the analysis of infrastructure connectivity. The analysis found that in 10 countries, the number of infrastructures, crossing borders, and physical infrastructures connectivity has been efficient with trade value. This means those countries can manage the infrastructures to support transport and freight movement to increase the GDP, transport, import, and export. On the other hand, five countries were found to be inefficient (the results of CRS and VRS lower than 1) which the reason comes from the infrastructure connectivity improperly utilizing with trade, and this paper provides the score for improving the connectivity to be efficient.

In this research, the reason for inefficiency came from the infrastructure, and the connectivity in each of the countries, which have excess input on the number of infrastructure and connectivity, but is underutilized for transport and the effect on the GDP of trade and amount of import and exports. Hence, the countries need to perform by increasing utilization of infrastructures and connectivity. Alternatively, based on the DEA approach, each country should be reducing the number of infrastructures and lowering the connectivity for the same value of trade. However, the realistic basic infrastructures can not be removed, but can be improved for the most utilization. Furthermore, this research suggests that the inefficient countries have an advantage in the land (road and rail) border and connectivity. So they need to use the benefit of infrastructures and connectivity to drive the numbers of GDP of transport, including the numbers of imports and exports. It should also increase the operation, and support multimodal transport to improve connectivity and trade.

5. Conclusions

This study attempts to measure the performance of infrastructure connectivity in the RCEP, and provides suggestions for improving their efficiency. In the infrastructure, connectivity has been applied with the “intermodal and multimodal concept” to consider the connection of infrastructures and transport. The type of the concept was generated in infrastructures (number of ports, rail range, and road network), and connectivity (number of land borders, number of the maritime border, number of cross border points, railway linkage with other countries, number of port connect with railway and number of port connect with road). The output factors are most related to trade and economics as GDP, transport, import, and export volume. However, most of the research was related only to the performance of infrastructure efficiency such as port, airport and railway performance. In this study, the infrastructure connectivity applied the “intermodal and multimodal transport concept” with trade in the RCEP, which measures the overall efficiency. Furthermore, the study applied DEA to measure the efficiencies. Then, the results are generated and scale efficiencies to improve the inappropriate performance and inefficiency of connectivity by demonstrating VRS and DRS.

Based on the analysis results, the efficiencies of infrastructures connectivity showed that ten countries: Australia, Brunei-Darussalam, China, Indonesia, Japan, Laos, Malaysia, Philippines, Singapore, and South Korea, were efficient in terms of infrastructures connectivity given to trade, because they utilize their infrastructures and benefit of connectivity to support international trade more effectively. Whereas, Cambodia, Myanmar, New Zealand, Thailand, and Vietnam were inefficient. The study suggests these countries’ efficiency improve by reducing the number of infrastructures and connectivity points to receive the same output by closing some border or temporary closed some infrastructures such port or railway following with the suggestion of DEA approach. In realistically, the author suggest to retain the same quantity and quality of infrastructures and connectivity points. But each country should utilize an infrastructure and trade potential for more efficiency. In addition, the negative influence on infrastructures and connectivity performance will limit driving trade performance and impact on freight flow. Moreover, a chance to expand more networks, cross border points, and improve connectivity might increase the GDP and volume of import and export.