Estimating Energy Efficiency and Energy Saving Potential in the Republic of Korea’s Offshore Fisheries

Abstract

The Republic of Korea’s government has established a carbon negativity policy to mitigate climate change in the fisheries sector. To achieve this objective, the government proposed enhancing energy efficiency in vessel fisheries, known for high carbon emissions. However, it was difficult to find research that investigated the energy consumption status of vessel fisheries. Thus, this study aims to calculate the offshore fisheries’ energy efficiency (EE) and to estimate the energy saving potential (ESP) needed in order to achieve efficient energy consumption. For this purpose, annual fisheries management surveys and data on the tax-free petroleum supply are employed. This study measures the EE and the ESP of offshore fisheries by year and fishing gear by employing the stochastic frontier analysis (SFA), which considers exogenous determinants of energy inefficiency. The analysis results show a decline in the EE over time and an increasing trend in the ESP. Notably, the trawl and fleet fisheries tend to have lower energy efficiency. Furthermore, the trawl and fleet fisheries were identified as having the highest ESP. Therefore, to utilize energy efficiently and reduce energy consumption in offshore fisheries, this study suggests scaling down fleet fisheries, developing energy saving fishing nets and eco-friendly fishing vessels, expanding modernization projects for fishing vessels, and revising the related acts.

1. Introduction

At the 2019 UN Climate Action Summit, countries around the world declared carbon neutrality, which means counterbalancing carbon emissions with carbon offsets as one solution to the climate change challenges [1]. Subsequently, this approach was recognized as a new global paradigm for mitigating climate change. In alignment with the international strategy towards climate change, the Republic of Korea formally presented its carbon neutrality framework to the UN in 2020. To participate in the international community’s strategy for addressing climate change, the Republic of Korea’s government submitted a carbon neutrality strategy to the UN in 2020. As part of its endeavor to achieve carbon neutrality, the Korean government devised a carbon neutrality strategy by sector [2]. Accordingly, the fisheries industry in the Republic of Korea set a goal to become carbon negative, which means the net carbon emissions below zero by 2050 and established an ambitious carbon reduction roadmap [3].

In 2018, the greenhouse gas (GHG) emissions from the fisheries industry of the Republic of Korea, including vessel fisheries, were reported to be 3.0 $\mathrm{M}\mathrm{t}{\mathrm{C}\mathrm{O}}_2\mathrm{e}\mathrm{q}$ [3]. In 2020, GHG emissions from vessel fisheries were recorded at 2.79 $\mathrm{M}\mathrm{t}{\mathrm{C}\mathrm{O}}_2\mathrm{e}\mathrm{q}$ [4]. Of these, offshore fisheries accounted for approximately 60% (1.68 $\mathrm{M}\mathrm{t}{\mathrm{C}\mathrm{O}}_2\mathrm{e}\mathrm{q}$) of the GHG emissions [4]. As a result, the Republic of Korea’s government has established a policy to enhance the energy efficiency of offshore fisheries to achieve carbon negativity in the fishery industry [3]. Moreover, improving the energy efficiency of offshore fisheries is an urgent task from the perspective of sustainable fishery business management. From 2012 to 2020, the proportion of fuel costs in offshore fisheries accounted for about 22% of the fishery costs [5], making it the second highest expense following labor costs (33%). In order for the business management of offshore fisheries to be sustainable in the face of the continuous decline of fish stocks [6,7], it is necessary to reduce fishery costs of the offshore fisheries enhancing the energy efficiency. In addition, ocean acidification and rising sea temperatures caused by carbon emissions have a negative impact on marine ecosystems [8], leading to the deterioration of fishery stocks [9]. Although the carbon emissions from the global fisheries sector account for only 0.5% of the total [10], a reduction in carbon emissions through enhanced energy efficiency in this sector is crucial for the sustainable use of fishery stocks.

To achieve the goals of carbon negativity and sustainable management in the fisheries industry through the improvement of energy efficiency in offshore fisheries, it is important to prioritize the investigation of offshore fisheries’ current energy efficiency utilizing the econometric method. Concurrently, it is essential to calculate the energy saving potential for enhancing the energy efficiency of offshore fisheries. Thus, this study aims to estimate the energy efficiency and energy saving potential of the Republic of Korea’s offshore fisheries, employing the stochastic frontier analysis (SFA), and to examine the determinants influencing energy inefficiency. Additionally, based on the results of this analysis, we suggest policy improvements that can aid in enhancing the energy efficiency of the offshore fisheries. The analytical process of this study is as follows. First, we estimate both the energy distance function of the offshore fisheries and the exogenous determinants of energy inefficiency, employing the SFA method. Based on these estimation results, we derive the energy efficiency of offshore fisheries by year and fishing gear. Subsequently, using this energy efficiency, we calculate the energy saving potential of offshore fisheries by year and fishing gear.

The rest of this paper is structured as follows. Section 2 summarizes the previous studies related to this research. Section 3 describes the data, variables, and the analysis method. Section 4 presents the estimated results, and the final section summarizes the main findings and provides concluding remarks.

2. Literature Review

Stochastic frontier analysis (SFA) has been extensively employed to appraise energy efficiency across diverse domains. Notably, several studies have examined energy efficiency in diverse sectors, such as the manufacturing industry in the United States [11,12,13], industries with high energy consumption in China [14], the steel and paper industries in India [15,16], the manufacturing industry in Sweden [17] and Kenya [18], the industrial sectors in Japan [19], and the major industries in Vietnam [20]. Additionally, some studies estimated regional energy efficiency in China [21,22,23,24,25,26], Japan [27,28], the United States [29], Portugal [30], and Cameroon [31]. Furthermore, many studies have estimated the energy efficiency of countries belonging to the OECD [32,33,34], the EU [35,36], the Baltic Sea [37], Africa [38,39], and emerging countries [40].

While numerous studies have focused on evaluating energy efficiency across various sectors, regions, and countries, research specifically addressing the estimation of energy saving potential remains comparatively limited. Research measuring the energy saving potential based on energy efficiency estimated through SFA has mainly been conducted in regions or industries of China (

Table 1. Summary of major previous studies.

Author(s)	Object	Main Contents
Lin and Yang [41]	Thermal power industry (China)	Energy efficiency, Energy saving potential, Energy inefficiency determinants
Lin and Wang [42]	Iron and steel industry (China)
Lin and Long [43]	Chemical industry (China)
Xie et al. [44]	Transport industry (China)
Xu et al. [45]	285 cities (China)
Liu et al. [47]	141 countries
Liu et al. [48]	Agricultural sector (27 emerging countries)
Lin and Du [46]	30 provinces (China)	Energy efficiency, Energy saving potential
Khraiche et al. [49]	44 European countries

). For instance, Lin and Yang [41] estimated the average energy efficiency of the thermal power industry across 28 provinces in China to be 0.85. The energy saving potential of this industry was calculated to be approximately 551.04 million tons of coal equivalent (Mtce). This study also found that as the degree of market openness increased, the energy efficiency in this industry improved. Lin and Wang [42] analyzed the average energy efficiency of the iron and steel industry in 26 Chinese cities to be 0.699, with an energy saving potential of 723.44 Mtce. Additionally, the iron and steel industry in Chinese cities was found to be more energy efficient at larger scales but less efficient as the degree of nationalization increased. Lin and Long [43] evaluated the energy efficiency of the chemical industry in 30 provinces of China, with values ranging from 0.3866 to 0.9743. The study determined an energy savings potential of 89.42 Mtce for this sector. In China’s chemical industry, larger entities or those facing higher energy prices tend to be more energy efficient. Conversely, as the degree of nationalization increases, the energy efficiency of this industry diminishes. Xie et al. [44] measured the average energy efficiency of the transport industry in 29 Chinese provinces to be 0.67 and calculated the energy saving potential to be 783.77 Mtce. The transport industry in China exhibits higher energy efficiency with increased government support and public transport utilization. Moreover, road condition improvements positively influenced this sector’s energy efficiency. Xu et al. [45] found the average energy efficiency across 285 Chinese cities to be 0.35, with an electricity conservation potential of 309 billion kilowatts. The energy efficiency in Chinese cities increased up to a certain threshold of economic agglomeration but decreased beyond that point. Additionally, Lin and Du [46] estimated the average energy efficiency of 30 Chinese provinces at 0.632 and calculated the energy saving potential of 1000 Mtce. However, this study did not examine the determinants of energy inefficiency.

We found studies estimating energy efficiency and energy saving potential in countries other than China. First, Liu et al. [47] measured the average energy efficiency of 141 countries at 0.77 and estimated that global energy consumption should be reduced by 23%. Notably, this study found that economic globalization positively impacted energy efficiency. Liu et al. [48] estimated the average energy efficiency of the agricultural sector in 27 emerging countries to be 0.74 and calculated the energy saving potential at 542.80 MTOE. Using a Tobit model, this study identified determinants of energy efficiency. Economic structure, urbanization, and GDP per capita were observed to negatively affect the energy efficiency of the agricultural sector, while the energy mix and pesticide use exhibited a positive impact. Khraiche et al. [49] computed the average energy efficiency of 44 European countries at 0.9416, with the energy saving potential calculated to be 130.15 MTOE.

On the other hand, it has been observed that the vessel fisheries have yet to utilize SFA to estimate energy efficiency and energy saving potential. Some studies have calculated energy intensity (energy or fuel consumption per catch), CPUF (catch per fuel consumption), and VPUF (fishing income per fuel consumption) using only the fuel usage, catch amounts, and fishing income of fishing vessels. These indicators were then used to assess the energy efficiency of vessel fisheries in the EU, Norway, and Denmark [50,51,52,53]. However, these studies have limitations due to evaluating the energy efficiency of vessel fisheries without considering essential input factors such as crew, vessel size, and the engine’s power.

To overcome the limitations of previous studies in the fisheries sector, we adopted the SFA model that considers exogenous inefficiency determinants from several articles [41,42,44,45]. Furthermore, using the energy efficiency estimated through this model, we utilized a method to derive the energy saving potential. This research signifies a progression from prior studies in the fisheries sector, which assessed energy efficiency solely based on energy consumption and fisheries revenue. In our approach, we included multiple inputs into the SFA model to gauge the energy efficiency of the Republic of Korea’s offshore fisheries. Specifically, we evaluated the energy efficiency of offshore fisheries by considering not only the fisheries revenue and energy consumption of fishing vessels but also crew, vessel tonnage, and horsepower. Additionally, a limitation was identified in previous studies in the fisheries sector, where determinants influencing energy inefficiency were not presented. Conversely, this study investigated the determinants of energy inefficiency in offshore fisheries. Lastly, unlike previous studies in the fisheries sector, this research distinguishes itself by measuring the energy saving potential that enables offshore fisheries to optimize energy consumption.

3. Materials and Methods

3.1. Data and Variables

The subjects of this study are 14 fishing gears of offshore fisheries in the Republic of Korea with a gross tonnage (GT) of fishing vessels exceeding 10 tons. These fishing gears include the offshore stow net (OSN), offshore long line (OLL), offshore gill net (OGN), offshore angling (OA), offshore trap (OT), anchovy drag net (ADN), large purse seine (LPS), large otter trawl (LOT), east sea Danish seine (EDS), east sea trawl (EST), medium Danish seine (MDS), large pair trawl (LPT), large Danish seine (LDS), and Diver. Among these gears, ADN, LPS, and LPT operate with a fleet composed of 5, 6, and 2 fishing vessels, respectively [54,55]. The seven fishing gears, including ADN, LOT, EDS, EST, MDS, LPT, and LDS, operate using trawling that involves dragging a fishing net [54,55].

We analyzed data derived from the annual fisheries management surveys and tax-free petroleum supply reports which are divided into fishing gear [5,56]. Due to limited tax-free petroleum data access, only the data from 2012 to 2020 were available for analysis. The data utilized for the analysis, such as fisheries revenue, the number of crews, vessel’s horsepower, vessel’s ton, vessel’s age, and depreciation cost, were obtained through annual Fishery Management Surveys from 2012 to 2020 [5]. Among the data, the fisheries revenue and the depreciation cost were converted into real values by applying the Fisheries Products Producer Price Index (2015 = 100) [5,57]. The energy variable (E) was calculated by multiplying the tax-free petroleum supply (PS) data, which was categorized by year, fishing gear, and fuel type, with the gross calorific value (GCV) per liter classified by fuel type [56,58]. This process can be described as follows:

$$E_{itk}={PS}_{ikt}\times {GCV}_k,$$

(1)

where the symbols i and t denote fishing gear and year, respectively. Moreover, k signifies the fuel type such as diesel, gasoline, and heavy oil. Moreover, the liters per GCV for diesel were applied at 0.0009 TOE (tons of oil equivalent), for gasoline at 0.00078 TOE, and for heavy oil at 0.00097 TOE [58]. To estimate the energy efficiency and the energy saving potential of offshore fisheries, we summed the energy data calculated by fuel type according to Equation (1). After that, we categorized it by fishing gear and year for the analysis. This study utilized the data per fishing permit for fishing gear. The fleet fisheries such as ADN, LPS, and LPT include 5, 6, and 2 vessels under the fishing permit. Additionally, the basic statistics of the variables used in the analysis are summarized in

Table 2. Basic statistics of analysis data.

Variable	Observation	Mean	Standard Deviation	Minimum	Maximum
Fisheries revenue (Thousand KRW)	126	2,098,527	2,845,626	154,428	16,439,039
Crew (Person)	126	17	18	3	76
Vessel HP (Horsepower)	126	1553	2010	379	8478
Vessel Ton (Gross tonnage)	126	154	252	4	1083
Energy (TOE)	126	574	825	15	3548
Vessel age (Year)	126	22	9	7	41
Depreciation cost (Thousand KRW)	126	60,786	135,450	972	765,603

Note: Parentheses inside the table mean the unit of variables.

.

3.2. Energy Distance Function

Debreu [59] and Farrell [60] conducted the first studies on production efficiency using different approaches. Debreu [59] measured technical efficiency based on the output direction, while Farrell [60] proposed a method of measuring efficiency based on the input direction. In this study, we employ the Shephard energy distance function based on input direction [33]. In line with the production structure, we consider a production process in which the offshore fisheries inputs energy (E), crew (L), vessel’s gross tonnage (GT), and horsepower (HP) to create fisheries revenue (R). Therefore, the production technology can be represented as follows:

$$T=\{\left(E,L,GT,HP,R\right):\left(E,L,GT,HP\right)canproduce\left(R\right)\},$$

(2)

where the set of all input and output vectors of T is assumed to satisfy the axioms of production technology [61]. Specifically, it is impossible to produce fisheries revenue without inputs. Thus, inactivity is always an option; only finite inputs can produce finite output [38]. Moreover, it is generally assumed that these inputs and outputs are disposable [47]. Furthermore, T is bounded and convex [47]. Based on the Shephard energy distance function proposed by Zhou et al. [33], we can express the Shephard energy distance function as follows:

$$D_E\left(E,L,GT,HP,R\right)=sup\{\beta :\left(\frac{E}{\beta },L,GT,HP,R\right)\in T\}.$$

(3)

Equation (3) is formulated to compute the most potential decrease in E while preserving the input and output vectors defined inside the restrictions of the production process. Consequently, $D_E\left(E,L,HP,Ton,R\right)$ signifies a vessel’s hypothetically optimal energy use. According to previous studies [45,46,62], energy efficiency (EE) is calculated by adapting the formula:

$$EE=\frac{1}{D_E\left(E,L,GT,HP,R\right)}.$$

(4)

The EE assesses the disparity between a vessel’s actual energy consumption and the optimal energy level necessary for output. Its definition restricts the range to between 0 and 1. When the EE is 1, the vessel utilizes energy optimally, and its energy efficiency is situated on the frontier curve. Conversely, an EE score below 1 denotes energy inefficiency in the production process, indicating that production remains within the frontier curve.

3.3. Estimating Energy Distance Function through SFA

Stochastic frontier analysis (SFA) is a parametric approach introduced by Aigner et al. [63] and Meeusen and van Den Broeck [64]. The SFA methodology, based on the distance function, is a frontier procedure and is oriented toward economic optimization [47]. The SFA enables the separation of residuals into random noise and inefficiency effect [65]. For that reason, considering statistical noise while calculating efficiency is a primary advantage of the SFA. Nevertheless, it is necessary to determine the function’s specific form and to predefine the probability distribution of the inefficiency error term before applying the SFA.

To analyze the energy distance function employing the SFA method, we have to determine a specific functional form first. In this study, we employ the translog function form following several research [38,42,45,47]. The translog function is more flexible than the Cobb–Douglas function because it does not impose any prior limitations on the production technology [66]. Therefore, we can represent the energy distance function in the translog form as follows:

$$\begin{array}{c}ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right)={\beta }_0+{\beta }_{E}ln{E}_{it}+{\beta }_{L}ln{L}_{it}+{\beta }_{GT}ln{GT}_{it}+{\beta }_{HP}ln{HP}_{it}+{\beta }_{R}ln{R}_{it}+\\ {\beta }_{E·L}\left(ln{E}_{it}\right)\left(ln{L}_{it}\right)+{\beta }_{E·GT}\left(ln{E}_{it}\right)\left(ln{GT}_{it}\right)+{\beta }_{E·HP}\left(ln{E}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{E·R}\left(ln{E}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{L·GT}\left(ln{L}_{it}\right)\left(ln{GT}_{it}\right)+\\ {\beta }_{L·HP}\left(ln{L}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{L·R}\left(ln{L}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{GT·HP}\left(ln{GT}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{GT·R}\left(ln{GT}_{it}\right)\left(ln{R}_{it}\right)+\\ {\beta }_{HP·R}\left(ln{HP}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{E·E}{\left(ln{E}_{it}\right)}^2+{\beta }_{L·L}{\left(ln{L}_{it}\right)}^2+{\beta }_{GT·GT}{\left(ln{GT}_{it}\right)}^2+{\beta }_{HP·HP}{\left(ln{HP}_{it}\right)}^2+{\beta }_{R·R}{\left(ln{R}_{it}\right)}^2+v_{it},\end{array}$$

(5)

where $i$ and $t$ mean fishing gear and year respectively. In addition, $v_{it}$ is the symmetric disturbance term and is assumed to follow a standard normal distribution with $v_{it}~i.i.d.(0,{\sigma }_v^2)$. Equation (5) is linearly homogeneous concerning energy input [33,67]; then, we can transform $ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right)$ to the following Equation (6):

$$ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right)=ln{E}_{it}+ln{D}_E\left(1,L_{it},{GT}_{it},{HP}_{it},R_{it}\right).$$

(6)

To estimate Equation (6) by applying the econometric method, we can expand the formula as below:

$$\begin{array}{c}ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right)=ln{E}_{it}+{\beta }_0+{\beta }_{L}ln{L}_{it}+{\beta }_{GT}ln{GT}_{it}+{\beta }_{HP}ln{HP}_{it}+{\beta }_{R}ln{R}_{it}+\\ {\beta }_{L·GT}\left(ln{L}_{it}\right)\left(ln{GT}_{it}\right)+{\beta }_{L·HP}\left(ln{L}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{L·R}\left(ln{L}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{GT·HP}\left(ln{GT}_{it}\right)\left(ln{HP}_{it}\right)+\\ {\beta }_{GT·R}\left(ln{GT}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{HP·R}\left(ln{HP}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{L·L}{\left(ln{L}_{it}\right)}^2+{\beta }_{GT·GT}{\left(ln{GT}_{it}\right)}^2+{\beta }_{HP·HP}{\left(ln{HP}_{it}\right)}^2+\\ {\beta }_{R·R}{\left(ln{R}_{it}\right)}^2+v_{it}.\end{array}$$

(7)

Making some rearrangement to Equation (7), we obtain:

$$\begin{array}{c}-ln{E}_{it}={\beta }_0+{\beta }_{L}ln{L}_{it}+{\beta }_{GT}ln{GT}_{it}+{\beta }_{HP}ln{HP}_{it}+{\beta }_{R}ln{R}_{it}+{\beta }_{L·GT}\left(ln{L}_{it}\right)\left(ln{GT}_{it}\right)+{\beta }_{L·HP}\left(ln{L}_{it}\right)\left(ln{HP}_{it}\right)+\\ {\beta }_{L·R}\left(ln{L}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{GT·HP}\left(ln{GT}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{GT·R}\left(ln{GT}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{HP·R}\left(ln{HP}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{L·L}{\left(ln{L}_{it}\right)}^2+\\ {\beta }_{GT·GT}{\left(ln{GT}_{it}\right)}^2+{\beta }_{HP·HP}{\left(ln{HP}_{it}\right)}^2+{\beta }_{R·R}{\left(ln{R}_{it}\right)}^2+v_{it}-ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right).\end{array}$$

(8)

Then, $ln{D}_E\left(E_{it},L_{it},{GT}_{it},{HP}_{it},R_{it}\right)$ in Equation (7) sets $u_{it}$ following Battese and Coelli [68]. We obtain:

$$\begin{array}{c}-ln{E}_{it}={\beta }_0+{\beta }_{L}ln{L}_{it}+{\beta }_{GT}ln{GT}_{it}+{\beta }_{HP}ln{HP}_{it}+{\beta }_{R}ln{R}_{it}+{\beta }_{L·GT}\left(ln{L}_{it}\right)\left(ln{GT}_{it}\right)+{\beta }_{L·HP}\left(ln{L}_{it}\right)\left(ln{HP}_{it}\right)+\\ {\beta }_{L·R}\left(ln{L}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{GT·HP}\left(ln{GT}_{it}\right)\left(ln{HP}_{it}\right)+{\beta }_{GT·R}\left(ln{GT}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{HP·R}\left(ln{HP}_{it}\right)\left(ln{R}_{it}\right)+{\beta }_{L·L}{\left(ln{L}_{it}\right)}^2+\\ {\beta }_{GT·GT}{\left(ln{GT}_{it}\right)}^2+{\beta }_{HP·HP}{\left(ln{HP}_{it}\right)}^2+{\beta }_{R·R}{\left(ln{R}_{it}\right)}^2+v_{it}-u_{it},\end{array}$$

(9)

where $u_{it}$ is assumed to be a non-negative and one-sided disturbance term with $u_{it}~N^{+}(\mu ,{\sigma }_u^2)$. This term represents the energy inefficiency of fishing gear $i$ over the $t$ year. Concurrently, $u_{it}$ and $v_{it}$ are assumed to be independent and identically distributed in all observations. The determinants of energy inefficiency have been articulated by Battese and Coelli [69]:

$$u_{it}={\alpha }_0+{\alpha }_{FLEET}{FD}_{it}+{\alpha }_{TRAWL}{TD}_{it}+{\alpha }_{DEPRECIATION}{DC}_{it}+{\alpha }_{AGE}{AGE}_{it}+{\epsilon }_{it},$$

(10)

where the ${FD}_{it}$, ${TD}_{it}$, ${DC}_{it}$, and ${AGE}_{it}$ are exogenous variables that affect the energy inefficiency of fisheries. In addition, ${\epsilon }_{it}$ is the truncated normal distribution with zero mean and variance ${\sigma }^2$ [69]. ${FD}_{it}$ denotes the dummy variable, which equals 1 when the fishing operation is conducted with more than two vessels, and otherwise 0. ${TD}_{it}$ is the dummy variable, which is set to 1 when the vessel operates using the trawl method, and otherwise 0. ${DC}_{it}$ denotes the depreciation cost, and ${AGE}_{it}$ represents the vessel’s age. If the coefficients of the variables in Equation (10) are positive, it can be interpreted that energy inefficiency increases. However, if the sign of coefficients is negative, it signifies a decrease in energy inefficiency.

There are single-stage or two-stage approaches to estimating Equations (9) and (10). Most previous studies have favored the single-stage approach to analyze energy efficiency and its determinants simultaneously because the two-stage approach tends to provide considerably biased results [70]. Utilizing this approach is also likely to solve the issue of conditional heteroscedasticity in energy efficiency [38]. Hence, we apply the single-stage approach to examine the influencing factors of the energy efficiency estimation based on the SFA method.

According to Battese and Coelli [69], the energy efficiency (EE) in fishing gear $i$ and year $t$ can be measured via the following:

$${EE}_{it}=\mathrm{e}\mathrm{x}\mathrm{p}(-u_{it}),$$

(11)

where the value of ${EE}_{it}$ is between 0 and 1, where 0 signifies the most energy inefficiency and 1 indicates the most energy efficiency. Following Lin and Wang [42] and Lin and Yang [41], energy saving potential (ESP) in fishing gear $i$ and year $t$ can be computed by the following:

$${ESP}_{it}=(1-{EE}_{it})\times E_{it}.$$

(12)

4. Results

4.1. Energy Distance Function of Offshore Fisheries

Table 3. Results of estimating energy distance function through SFA.

Coefficient	Estimate	Standard Error	z-Statistic
Dependent variable: ${-\mathrm{l}\mathrm{n}E}_{it}$
${\beta }_0$	11.9805 **	5.6938	2.1000
${\beta }_L$	1.4697	1.1285	1.3000
${\beta }_{GT}$	8.6202 ***	1.2068	7.1400
${\beta }_{HP}$	−6.9171 ***	1.1980	−5.7700
${\beta }_R$	−0.2489	0.6291	−0.4000
${\beta }_{L·GT}$	−0.3112 *	0.1602	−1.9400
${\beta }_{L·HP}$	−0.7760 ***	0.1235	−6.2800
${\beta }_{L·R}$	0.2591 ***	0.0732	3.5400
${\beta }_{GT·HP}$	−0.3928 ***	0.1136	−3.4600
${\beta }_{GT·R}$	−0.6749 ***	0.0744	−9.0700
${\beta }_{HP·R}$	0.1636 **	0.0663	2.4700
${\beta }_{L·L}$	1.2159 ***	0.1707	7.1200
${\beta }_{GT·GT}$	−0.2289	0.2439	−0.9400
${\beta }_{HP·HP}$	1.4154 ***	0.1421	9.9600
${\beta }_{R·R}$	−0.0259	0.0404	−0.6400
Dependent variable: $u_{it}$
${\alpha }_0$	−1.0647	0.7301	−1.4600
${\alpha }_{FLEET}$	0.8621 ***	0.1348	6.3900
${\alpha }_{TRAWL}$	0.2326 *	0.1301	1.7900
${\alpha }_{DEPRECIATION}$	0.0506	0.0587	0.8600
${\alpha }_{AGE}$	0.0116 *	0.0061	1.9000
$\lambda ({\sigma }_u/{\sigma }_v)$	3.0700 ***	0.0259	118.5400

Note: * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the results of estimating the energy distance function of the Republic of Korea’s offshore fisheries and the exogenous determinants of inefficiency employing the SFA method. These results were estimated with the ‘sfpanel’ package in Stata17. The Lambda ($\lambda ({\sigma }_u/{\sigma }_v)$) in Table 3 is the ratio of the standard deviation of $u_{it}$ to the standard deviation of $v_{it}$ [65]. Therefore, if lambda is not statistically significant even at the 10% significance level, the energy distance function of the offshore fisheries can be interpreted as not having any energy inefficiency [71]. As a result, the null hypothesis that lambda is zero was rejected at the 1% significance level, indicating the presence of energy inefficiency. It was also confirmed that $u_{it}$ contributed relatively more than $v_{it}$ in explaining the composed error terms [16,32]. Thus, to separately estimate $u_{it}$ and $v_{it}$ in the energy distance function of the offshore fisheries, it was more appropriate to adopt the maximum likelihood estimation (MLE) approach rather than the ordinary least squares (OLS) approach [65]. Additionally, we have prevented the estimation of biased results by using a single-stage approach that simultaneously estimates the energy distance function of the offshore fisheries and the exogenous determinants of energy inefficiency.

As exogenous determinants of energy inefficiency in offshore fisheries, we considered the two dummy variables: depreciation cost and vessel age. Among these, all variables except for the depreciation expense were statistically significant at a level of significance of 10% or less. First, ${\alpha }_{FLEET}$ was statistically significant even at a 1% significance level, and the fishing gears that conduct the fleet operation were analyzed to be more energy inefficient. This is because fleet fisheries deploy up to five additional vessels for operations compared to fishing gear that operate with a single vessel [54,55]. Next, ${\alpha }_{TRAWL}$ was statistically significant even at a 10% significance level, and it was found that energy inefficiency increases as it becomes a trawl fishery. This is because these fishing gears operate by dragging the fishing net and receive more underwater resistance [54,55]. Finally, ${\alpha }_{AGE}$ was statistically significant at a 10% significance level. The vessel age and energy inefficiency showed a positive (+) relationship, and it was demonstrated that energy inefficiency also escalates as the vessel age increases. Korean fishing vessels that have exceeded useful life (25 years) were generally either built by referencing the design of Japanese vessels or were used vessels imported from Japan [72,73]. However, due to the two countries’ distinct marine and fishing environments, these vessels could not be optimized for the fishing conditions in the Republic of Korea [72]. This issue ultimately led to an increase in the vessels’ propulsion resistance, causing a decrease in energy efficiency.

4.2. Energy Efficiency and Energy Saving Potential of Offshore Fisheries

Based on the estimated results of the energy distance function of the offshore fisheries in Table 3, the energy efficiency of the offshore fisheries for each year is shown in

Table 4. Results of estimating energy efficiency (EE) by year.

Year	EE	Ranking
2012	0.8594	1
2013	0.8457	2
2014	0.8384	3
2015	0.8289	4
2016	0.8202	6
2017	0.7949	7
2018	0.8269	5
2019	0.7730	9
2020	0.7806	8
Average	0.8187	-

and Figure 1. Throughout the analysis period, the annual energy efficiency of the offshore fisheries showed a decreasing trend over time. Taking a closer look at the annual changes in energy efficiency, the energy efficiency of offshore fisheries was highest in 2012 at 0.8594 and lowest in 2019 at 0.7730. Although the annual energy efficiency of offshore fisheries was on a downward trend, it increased to 0.8269 in 2018. The suspected reason for this phenomenon is that the international oil price based in Dubai increased by about 31% compared to the previous year [74], leading to an increased fuel cost burden for offshore fisheries. This economic pressure potentially resulted in the diminished energy consumption observed within the offshore fisheries.

After observing a decline in the annual energy efficiency of offshore fisheries, the National Institute of Fisheries Science (NIFS) developed techniques with which to enhance the propeller efficiency of offshore fisheries vessels and to improve LED fishing lights for OA vessels in 2022 [4]. Despite these efforts, the energy consumption of the offshore fisheries increased by about 16% from 2012 to 2020 [56]. However, the volume of fishery resources in the Republic of Korea dramatically decreased from 8.6 million tons in 2013 to 3.13 million tons in 2018, resulting in a reduction of approximately 44% in fisheries revenue during the analysis period [5,6,7,57]. Such an increase in energy consumption and a decrease in fisheries’ revenue are presumed to be significant factors contributing to the annual decline in the energy efficiency of offshore fisheries.

Table 5. Results of calculating energy saving potential (ESP) by year.

Year	ESP (TOE)	Ranking
2012	2763.43	9
2013	2916.16	8
2014	3099.73	7
2015	3533.08	4
2016	3748.80	2
2017	3690.21	3
2018	3187.31	6
2019	3362.22	5
2020	4040.44	1
Average	3371.27	-

Note: Parentheses inside the table mean the unit of ESP.

and Figure 2 present the computed ESP values for offshore fisheries for each respective year, determined via Equation (12). The data suggest that due to variable annual energy consumption, there is not a straightforward inverse correlation between the annual EE rankings and ESP values in offshore fisheries. Nevertheless, when observing the trend in ESP of the offshore fisheries throughout the analysis period, it is evident that as the annual EE of the offshore fisheries decreases, the ESP seems to increase year by year. Specifically, for the most efficient energy utilization in offshore fisheries while operating, the total ESP per fishing permit for the 14 fishing gears was 2763.43 TOE in 2012. This figure rose by 32% to 4040.43 TOE in 2020.

The energy efficiency of each fishing gear, based on the energy distance function of the offshore fisheries, is summarized in

Table 6. Results of estimating energy efficiency (EE) by fishing gear.

Fishing Gear	EE	Ranking
Offshore stow net (OSN)	0.9527	4
Offshore long line (OLL)	0.9616	2
Offshore gill net (OGN)	0.9566	3
Offshore angling (OA)	0.9480	5
Offshore trap (OT)	0.9372	7
Anchovy drag net (ADN)	0.4173	14
Large purse seine (LPS)	0.4411	13
Large otter trawl (LOT)	0.8198	11
East sea Danish seine (EDS)	0.9077	8
East sea trawl (EST)	0.9042	9
Medium Danish seine (MDS)	0.9388	6
Large pair trawl (LPT)	0.4522	12
Large Danish seine (LDS)	0.8624	10
Diver	0.9617	1
Average	0.8187	-

Note: Parentheses inside the table mean the abbreviations of the fishing gear.

and Figure 3. Among these fishing gears, Diver (0.9617), OLL (0.9616), and OGN (0.9566) were identified as the most energy efficient. These fishing gears were neither included in the trawl fisheries nor in the fleet fisheries, which led to them having lower energy inefficiency compared to other fishing gears. Furthermore, the average age of vessels (10–12 years) in these three fishing gears was lower than that in other fishing gears, which is speculated to have influenced the reduction in energy inefficiency.

The average energy efficiency (0.7575) of the trawl fisheries (ADN, LOT, EDS, EST, MDS, LPT, LDS) was analyzed to be lower compared to the non-trawl fisheries (0.8798). Trawl fisheries use an active fishing method involving the dragging of nets, which results in higher energy consumption compared to non-trawl fisheries. When calculating the energy consumption per fisheries revenue based on the analysis data, trawl fisheries (0.000287 TOE/Thousand KRW) were found to consume 9.84% more energy than non-trawl fisheries (0.000261 TOE/Thousand KRW) to generate the same amount of revenue [5,56,57]. In terms of energy consumption per vessel’s gross tonnage, trawl fisheries (4.38 TOE/GT) were 34.04% higher than non-trawl fisheries (3.27 TOE/GT) [5,56]. Moreover, the energy consumption per vessel’s horsepower was greater in trawl fisheries (0.42 TOE/HP) compared to non-trawl fisheries (0.33 TOE/HP) [5,56]. Considering that trawl fisheries had higher energy consumption across all three indicators compared to non-trawl fisheries, it is speculated that this influenced the lower energy efficiency of trawl fisheries. The age of the vessels in trawl fisheries (27 years) is also about ten years older than those in non-trawl fisheries (17 years) [5]. This suggests that vessels in trawl fisheries might face greater propulsion resistance during operation [72]. The aging of the vessels could be a contributing factor to the lower energy efficiency observed in trawl fisheries. Furthermore, the sharp decline in fishery stocks from 8.6 million tons in 2013 to 3.13 million tons in 2018 might have significantly impacted the energy efficiency of trawl fisheries [6,7,51].

The government and academia, which were aware of these results, consistently developed an energy saving fishing net that reduces underwater resistance in various studies to improve the energy inefficiency of trawl fisheries [75,76,77,78]. Despite these developments, the reason for the lower energy efficiency of trawl fisheries compared to non-trawl fisheries is supposed to be that energy saving fishing net has not been sufficiently distributed in the fisheries field.

The fleet fisheries (ADN, LPS, LPT) involving two or more vessels operating simultaneously have shown an average energy efficiency of 0.4369, which is lower than the non-fleet fisheries (0.9228). According to the data from this study, the energy consumption per fisheries revenue for fleet fisheries (0.000319 TOE/Thousand KRW) was 53.85% higher than non-fleet fisheries (0.000207 TOE/Thousand KRW) [5,56,57]. This implies that fleet fisheries expend more energy to generate the same fisheries revenue compared to non-fleet fisheries. While the energy consumption per ton was 14.20% lower for fleet fisheries (3.57 TOE/GT) compared to non-fleet fisheries (4.16 TOE/GT) [5,56], the energy consumption per horsepower showed fleet fisheries (0.41 TOE/HP) consumed 32.06% more than non-fleet fisheries (0.31 TOE/HP) [5,56]. Since fleet fisheries had higher energy consumption across two of three indicators compared to non-fleet fisheries, it is inferred that this impacted the lower energy efficiency of fleet fisheries. The age of vessels in fleet fisheries (25 years) is about four years older than those in non-fleet fisheries (21 years) [5], implying that vessels in fleet fisheries might experience greater propulsion resistance during operation [72]. This can subsequently affect the increased energy consumption of fleet fisheries. Moreover, despite a 46.64% decrease in fisheries revenue of fleet fisheries over the analyzed period [5,57], these fisheries consumed 9.70% more energy [56]. This also likely contributed to the deterioration of energy efficiency in fleet fisheries.

To address this concern, the Ministry of Oceans and Fisheries (MOF) proposed a measure to reduce the number of vessels in the ADN fleet, where five vessels operate together, and the LPS fleet, where six vessels operate at the same time [79,80]. From 2014 to 2018, the MOF conducted an empirical study of the standard vessels for ADN and developed a standardized fleet model by reducing the number of vessels in the ADN fleet to four. However, this model was not widely disseminated in the ADN fisheries [79]. Furthermore, the MOF attempted to save energy consumption by reducing the number of vessels from six to four in LPS fisheries. However, this initiative was prematurely terminated, primarily due to insufficient engagement from LPS fisheries [80].

The ESP for each fishing gear calculated using Equation (12) is summarized in

Table 7. Results of calculating energy saving potential (ESP) by fishing gear.

Fishing Gear	ESP (TOE)	Ranking
Offshore stow net (OSN)	9.17	10
Offshore long line (OLL)	5.74	12
Offshore gill net (OGN)	4.32	13
Offshore angling (OA)	8.46	11
Offshore trap (OT)	21.13	6
Anchovy Drag net (ADN)	435.65	3
Large purse seine (LPS)	1733.06	1
Large otter trawl (LOT)	137.41	4
East sea Danish seine (EDS)	13.83	9
East sea trawl (EST)	18.25	7
Medium Danish seine (MDS)	14.35	8
Large pair trawl (LPT)	940.07	2
Large Danish seine (LDS)	29.10	5
Diver	0.74	14
Average	240.80	-

Note: Parentheses inside the table mean the abbreviations of the fishing gear or the unit of ESP.

and Figure 4. Due to differences in energy consumption based on the vessel’s scale of each fishing gear, the rankings of EE and ESP do not have a straightforward inverse correlation. However, LPS (1733.06 TOE), LPT (940.07 TOE), and ADN (435.65 TOE), which had the lowest EE and the highest energy consumption, showed higher ESP per fishing permit compared to other fishing gears. Additionally, the trawl fisheries with lower EE generally showed a higher ESP per fishing permit than non-trawl fisheries. Among the trawl fisheries, LPT (940.07 TOE), ADN (435.65 TOE), LOT (137.41 TOE), LDS (29.10 TOE), EST (18.25 TOE), MDS (14.35 TOE), and EDS (13.83 TOE) had the highest ESP per fishing permit in that order. Conversely, Diver (0.74 TOE), OGN (4.32 TOE), and OLL (5.74 TOE), which had the highest EE, were cataloged as fishing gear with the lowest ESP per fishing permit.

5. Discussion and Conclusions

The Republic of Korea’s government proposed improving energy consumption efficiency in the fisheries sector as a means to achieve carbon neutrality and negative. Following this, this study assessed the energy efficiency of the Republic of Korea’s offshore fisheries and calculated the potential energy reduction required for these fisheries to become the most energy efficient. Employing the stochastic frontier analysis (SFA) method, we estimated the energy distance function of the offshore fisheries and the exogenous determinants of energy inefficiency. Based on these results, we computed the offshore fisheries’ energy efficiency (EE) and energy saving potential (ESP) by year and fishing gear. At the same time, this paper also proposed policies to enhance the energy consumption efficiency of offshore fisheries and reduce energy consumption.

The primary findings of this study can be summarized as follows: First, from the analysis results on the determinants of energy inefficiency in offshore fisheries, it was found that energy inefficiency increased for fishing gears that belong to fleet or trawl fisheries. Furthermore, it was observed that the energy inefficiency rises as the vessel’s age increases. Second, the annual EE of the offshore fisheries was highest in 2012 at 0.8594 and lowest in 2019 at 0.7730, showing a declining trend over the analysis period. This was due to the increased energy consumption in the offshore fisheries, while fisheries revenue decreased due to the deterioration of the fisheries stock’s status. Based on the annual EE of the offshore fisheries, the total ESP per fishing permit of the 14 fishing gears was 2763.43 TOE in 2012 and increased by 32% to 4040.43 TOE in 2020, showing an upward trend throughout the analysis period. Third, the EE by fishing gear in the offshore fisheries was identified as highest in Diver (0.9617), OLL (0.9616), and OGN (0.9566). These fishing gears were not included in the fleet or trawl fisheries, and such gears had a lower average age of vessels than other fishing gears. On the other hand, ADN (0.4173), LPS (0.4411), and LPT (0.4522), which belong to fleet or trawl fisheries, had the lowest EE. These fishing gears had the lowest EE and higher energy consumption, resulting in the highest ESP per permit for LPS (1733.06 TOE), followed by LPT (940.07 TOE) and ADN (435.65 TOE).

The specific policy proposals for improving energy efficiency and energy saving potential in Korea’s offshore fisheries can be summarized as follows. First, it is necessary to expand the targets of the modernization project for fishing vessels, which anticipates improving energy consumption efficiency by replacing the hulls of the offshore fishing vessels that have exceeded useful life (25 years). To encourage more fishermen to participate in this project, adjustments to loan interest rates are essential. Second, for LPS (six vessels) and ADN (five vessels), it is vital to explore measures to reduce the number of vessels composing the fleet without adversely affecting fishing operations. Third, the continuous development and dissemination of energy-saving fishing nets that reduce underwater resistance are crucial for trawl fisheries. If the public direct payment in the fisheries sector in the Republic of Korea is utilized to support fishermen using gear that is efficient in energy consumption, it is expected that the energy efficiency of trawl fisheries will improve. Fourth, the government should revise the Fishing Vessel Act and its enforcement degree and rule to develop and distribute fishing vessels powered by electricity, LPG, hydrogen, and hybrid technologies to reduce energy consumption. Moreover, it is necessary for the government and fishermen to jointly establish a fund to ensure sufficient resources are available to promote eco-friendly vessels. Fifth, to motivate offshore fisheries to use tax-free petroleum efficiently, the National Federation of Fisheries Cooperatives (NFFC) should revise the Petroleum Business Supply Guidelines to ensure that vessels do not exceed allocated tax-free petroleum supply limitations.

The significance of this study lies in its groundbreaking effort to estimate the energy efficiency and energy saving potential of the fisheries sector employing SFA for the first time globally. Additionally, it holds importance in examining the energy consumption status of the Republic of Korea’s offshore fisheries, which had yet to be introduced internationally. Furthermore, beyond the energy intensity, CPUF, and VPUF, which were previously utilized as indicators for assessing the energy efficiency of the fisheries sector in preliminary research, this study also proposes an alternative approach to evaluate the energy efficiency of the fisheries sector.

This study can serve as foundational data for the Republic of Korea’s government when establishing specific policies to enhance the energy efficiency of offshore fisheries. Especially when the government is promoting policies for offshore fisheries, the findings of this study can be used to determine policy application priorities by fishing gear. In addition, each fishing gear within the offshore fisheries can identify levels of energy efficiency and energy saving potential through this study. Finally, the 14 fishing gears of offshore fisheries are expected to refer to the policy improvement measures mentioned above to enhance energy efficiency and increase energy saving potential.

While this study made efforts to reflect the recent trends in energy consumption in offshore fisheries, it could not estimate the energy efficiency and energy saving potential in 2021 and 2022 due to limitations in our data access. Additionally, a limitation of this study is that it estimated energy efficiency without considering the energy saving technologies of offshore fisheries. Future research should explore variables that can represent the energy saving technologies of the offshore fisheries and include them in the energy distance function to analyze energy efficiency and energy saving potential. Finally, if models like TFE (true fixed effects) or TRE (true random effects) included in the SFA are utilized in the analysis, it is expected that energy efficiency and energy saving potential can be computed considering the individual characteristics of each fishing gear in the offshore fisheries.