Scholarly article on topic 'Ship Safety Policy Recommendations for Korea: Application of System Dynamics'

Ship Safety Policy Recommendations for Korea: Application of System Dynamics Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"System Dynamics" / "Ship Accident" / "Scenario Analysis" / "Policy Recommendation"}

Abstract of research paper on Economics and business, author of scientific article — Jun Woo Jeon, Ying Wang, Gi Tae Yeo

ABSTRACT This study aimed to demonstrate and quantify the factors that influence ship safety and ship accidents and suggest policy recommendations for both government and the marine industry in Korea based on the application of system dynamics (SD). Korea was selected as the target country because a number of recent ship accidents have focused attention on ship safety in the region. Three factors that can influence ship safety were considered: economic factors, ship-handling and management factors, and government budget allocation of ship industry. SD was then applied to model the factors involved in ship accidents. Data on ship accidents in Korea from 2009–2014 were included in the model simulation. To measure the simulation accuracy, mean absolute percentage error (MAPE) analysis was employed. Following the simulation, a sensitivity analysis was conducted to determine the relationships among the factors involved in ship accidents. Finally, a simulation incorporating ship accident data and government budget allocated to the ship industry was performed. The results of this study can be used in government policy recommendations on ship safety to prevent and reduce ship accidents.

Academic research paper on topic "Ship Safety Policy Recommendations for Korea: Application of System Dynamics"

The Asian Journal of Shipping and Logistics 32(2) (2016) 073-079

¡11? c Aisiait í mi roa I itf O gljlpplitfl un il ïnyi ultra

Contents lists available at ScienceDirect

The Asian Journal of Shipping and Logistics

Journal homepage: www.elsevier.com/locate/ajsl

Ship Safety Policy Recommendations for Korea: Application of System Dynamics

Jun Woo JEONa , Ying WANGb , Gi Tae YEOc

a Ph.D. Candidate, Incheon National University, Korea, E-mail: j-wjeon0329@gmail.com (First Author) bPh.D. Candidate, Incheon National University, Korea, E-mail: yingmickey@163.com c Professor, Incheon National University, Korea, E-mail:ktyeo@inu.ac.kr(CorrespondingAuthor)

A R T I C L E I N F O

A B S T R A C T

Article history: Received 30 January 2015 Received in revised form 15 May 2016 Accepted 31 May 2016

Keywords: System Dynamics Ship Accident Scenario Analysis Policy Recommendation

This study aimed to demonstrate and quantify the factors that influence ship safety and ship accidents and suggest policy recommendations for both government and the marine industry in Korea based on the application of system dynamics (SD). Korea was selected as the target country because a number of recent ship accidents have focused attention on ship safety in the region. Three factors that can influence ship safety were considered: economic factors, ship-handling and management factors, and government budget allocation of ship industry. SD was then applied to model the factors involved in ship accidents. Data on ship accidents in Korea from 2009-2014 were included in the model simulation. To measure the simulation accuracy, mean absolute percentage error (MAPE) analysis was employed. Following the simulation, a sensitivity analysis was conducted to determine the relationships among the factors involved in ship accidents. Finally, a simulation incorporating ship accident data and government budget allocated to the ship industry was performed. The results of this study can be used in government policy recommendations on ship safety to prevent and reduce ship accidents.

Copyright © 2016 The Korean Association of Shipping and Logistics, Inc. Production and hosting by Elsevier B.V. All rights reserved. Peer review under responsibility of the Korean Association of Shipping and Logistics, Inc.

1. Introduction

Ship accidents generally result in serious damage, death, loss, injury, or pollution and have major political, economic, and environmental consequences. They also affect several entities in the marine industry: the ship's company and owners, flag states, freight carriers, other sailing ships, coastal states, the shipbuilding industry, and the ship insurer (Poyraz, 1998). Improvements in marine technology and the implementation of

safety-related regulations in the International Maritime Organization (IMO) and shipping nations have succeeded in reducing the number of ship accidents (Celik, et al., 2010). However, as ship accidents can cause significant damage to the environment and result in substantial human losses, they remain a major concern for global maritime interests (Celik and Cebi, 2009). Today, almost all vessels are equipped with modern

http://dx.doi.Org/10.1016/j.ajsl.2016.06.003

2092-5212/© 2015 The Korean Association of Shipping and Logistics, Inc. Production and hosting by Elsevier B.V. All rights reserved.

navigation devices in accordance with regulation requirements. However, it is difficult to eliminate ship accidents because of the complex and high-risk environment at sea, where various dynamic agents, including geography, water, environment, and human-related factors, play a role in collisions. These collisions can result in fires, grounding, hull damage, or sinking. As the causative factors also change over time, preventing accidents can be difficult (Shu, 2014).

Most previous studies of ship accidents have focused on the role of human error (Hetherington et al., 2006; Celik and Cebi, 2009; Mullai and Paulsson, 2011; Ventikos and Giannopoulos, 2013; Akyuz and Celik; 2014) or the economic environment (Toffoli et al., 2005; Ugurlu et al., 2015a; Lu and Tsai, 2008; Knudsen and Hassler, 2011) and utilized simple data analysis or expert advisors. Although the investigation of a marine ship accident is complex and requires a professional and fair judgment, at the same time, it depends on accurate and comprehensive statistical data to detect the role of dynamic changes in ship accident causes.

In the present paper, system dynamics (SD) was employed to represent the complex causal relationships of ship accidents, with an SD model used to simulate the factors that influence ship accidents and maritime safety.

The contribution of this research is twofold. First, this research establishes a ship accident causation model to analyze the various factors responsible for ship accidents and to determine whether government budget allocated to maritime safety are sufficient. Second, the results of this research can be utilized by policy makers when allocating resources to ship safety and proposing strategies to reduce ship accidents.

In Section II, the factors influencing ship accidents are introduced. The SD model for ship accidents is established and validated in Section III. A sensitivity analysis based on various scenarios is described in Section IV to determine whether current budget allocated by the Korean government to ship safety are effectively used. Finally, the conclusions are presented in Section V.

2. Factors Influencing Ship Accidents

Ship accidents continue to occur, despite ongoing prevention efforts. The main purpose of most ship accident studies is to identify causative factors and provide valid and reliable information for decision makers (Psarros, et al., 2010; Nikolaos, et al. 2013). They then utilize this information to make informed decisions, which are aimed at avoiding injuries to seafarers and damage to property and the environment. According to ship accident analysis reports, the most common causes of ship accidents are human errors, technical and mechanical failures, legislative shortcomings, and environmental factors (Hassel, 2011; Weng and Yang, 2015; Sahin and Senol, 2015). Toffoli et al. (2005) proposed that there were two main sources of ship accidents, operational and maintenance, with about 60% due to operational causes (e.g., fire, collision, machinery damage) and 40% due to design and maintenance issues (e.g., water ingress, hull breaking in two, and capsizing). A grounded theory model based on a large amount of empirical data was previously used to analyze marine accidents caused by macroscopic and microscopic factors (Mullai and Paulsson, 2011). The macroscopic factors included the number of people on board, type of cargo, and environmental conditions. The microscopic factors consisted of (a) the construction of the ship; (b) equipment-related technical faults; (c) the operation, management, and design of the equipment; (d) communication, organization, and procedures; (e) human factors; and (f) noncompliance with regulations. Fuzzy theory has also been utilized as a

tool in an effort to prevent ship collisions (Shu et al., 2014). A review of ships involved in collisions and grounding based on historical data and expert opinion concluded that structural crashworthiness, oil outflow, and the residual strength of the damaged ship were the main factors governing ship accident causing (Wang et al., 2002). A framework for marine risk assessments of ship accidents was developed that considered the two main consequences of ship accidents: human losses and pollution of the environment (oil leakage from cargo or fuel) (Ventikos and Giannopoulos, 2013). Ugurlu et al. (2015a) proposed that human factors, heavy weather conditions, equipment failure, and unknown factors were critical causes of ship accidents. Blanc and Rucks (1996) analyzed data on 936 ship accidents that occurred from 1979-1987 for various reasons, including sea traffic levels, system utilization, accident type, weather, location, and other variables. They reported that a lack of knowledge about the position of other vessels and inadequate ship-to-ship communication played a pivotal role in ship accidents. Ugurlu et al. (2015b) found that human error, technical and mechanical failures, lack of communication, regulation violations, and environmental factors (bad weather or voyage conditions) were common factors underling ship accidents. Human error, which includes technical and management failures, as well as operational errors, is the primary factor that contributes to marine accidents. Celik and Cebi (2009) identified the role of human error in ship accidents by the Fuzzy-AHP method and quantified the human contributions to ship accidents. Their results showed that skill-based errors were the primary cause of ship accidents, followed by lack of coordination, communication, and operation planning. Akyuz and Celik (2014) proposed a ship accident analysis and prevention model basing on human error factors. They found that unsafe preconditions played a major role in ship accidents, and proposed preventative steps to reduce human errors leading to ship accidents. Hetherington et al. (2006) reported that various human factors, including poor communication, decision making, situational awareness, and teamwork, in addition to fatigue, stress, health, automation, and lack of a safety culture, were involved in ship accidents.

Maritime regulations have evolved over time in response to historical experience, mainly because of serious ship accidents. Knudsen and Hassler (2011) investigated the failure of the IMO to implement appropriate ship safety legislation and regulations to reduce the ship accident rate. They concluded that the core problem was the absence of a strong link between the IMO and national maritime administrations, with new rules sometimes negatively affecting the functioning of existing regulations or language difficulties impeding their implementation. As a result, some ship accidents occurred due to violations of regulations caused by faulty implementation or nonimplementation of the complex marine regulations and policies of each nation.

Economic growth stimulates increases in international business, including international marine transportation, leading to greater maritime traffic. Due to advantages of scale, large-sized ships are preferred. Increases in maritime traffic, cargo, and passenger numbers augment the risk of ship accidents.

Lu and Tsai (2008) evaluated the influence of climate on ship accidents in the container shipping industry. They assessed six dimensions of climate-related effects: safety attitudes, safety training, management safety practices, supervisor safety practices, job safety, and coworkers' safety practices.

According to the aforementioned literature and statistical data from the Ministry of Oceans and Fisheries in Korea (2014), the factors responsible for ship accidents can be divided into three areas: economic factors, ship-handling and management factors, and government budget allocation

factors shown in Table 1. The macro and micro economic factors are number of registered ships, number of abandoned ships, marine cargo volumes and the volume of travelers. The ship-handling and management factors include faulty ship building, watchman negligence, inappropriate response to weather conditions, sailing regulation violations, equipment-based errors and violations of onboard safety instructions. The government-related budget factors are seafarer qualifications, ship safety, maritime safety management system, marine transportation environment and marine safety global cooperation.

Table 1

Economic factors influencing ship accidents

Economic factors Definition Previous studies

Number of registered ships New ships registered in Korea Ventikos and Giannopoulos(2013) Akyuz and Celik (2014)

Number of abandoned ships Old ships disposed of in Korea Ventikos and Giannopoulos(2013) Akyuz and Celik (2014)

Marine cargo volumes Cargo volume in Korea, including imports and exports Mullai and Paulsson (2011) Akyuz and Celik (2014) Shu et al. (2014)

Volume of travelers Passenger volume in Korea by ship Mullai and Paulsson (2011) Akyuz and Celik (2014) Shu et al. (2014)

Table 2 Ship-handling and management factors influencing ship accidents

Ship-handling and management factors Definition Previous studies

Faulty ship building Ship accidents caused by poor ship building Mullai and Paulsson (2011) Akyuz and Celik (2014)

Watchman negligence Human error causing ship accidents due to negligence in guarding the ship Blanc and Rucks (1996) Celik and Cebi (2009) Knudsen and Hassler (2011)

Inappropriate response to weather conditions Unsuitable actions taken in response to inclement weather Lu and Tsai (2008) Mullai and Paulsson (2011) Ugurlu et al. (2015a) Ugurlu et al. (2015b)

Sailing regulation violations Violation of national or IMO legislation and ship safety regulations Hetherington et al. (2006) Knudsen and Hassler (2011) Ugurlu et al. (2015b)

Equipment-based errors Errors caused by lack of correct skills Celik and Cebi (2009) Hetherington et al. (2006) Mullai and Paulsson (2011) Ugurlu et al. (2015a)

Violations of onboard safety instructions Human error causing ship accidents due to violation of onboard safety instructions Mullai and Paulsson (2011) Akyuz and Celik (2014)

Table 3

Government budget allocation factors influencing ship accidents

Government budget allocation factors for ship accidents Definition

Seafarer qualifications Lack of resources for professional education in ship accident prevention

Ship safety Dispose of old ships and examine ship safety regularly to reduce ship

accidents

Maritime safety management system Draw up safety management regulations on ship safety

Marine transportation environment Improve sailing infrastructure and install E-navigation in ships for better knowledge of the ship's condition

Marine safety global cooperation Have a clear understanding of and communicate the IMO marine safety regulations and those of calling countries

3. SD Modeling of Ship Accidents

To date, most research on ship accidents has focused mainly on the contributory factors established by investigations. However, ship accidents are complex, with a variety of causes in most cases. To obtain a clearer picture of the causes of ship accidents, this paper utilized SD to create a dynamic model to simulate the causes of ship accidents arising from various factors, including government budgetary allocations. SD modeling, combined with cognitive feedback, offers a realistic way to represent complex causal relationships by modeling the impacts of critical success factors. As such, SD is an appropriate method to simulate the factors influencing ship accidents and provide suggestions for maritime safety.

3.1. Stock-flow Diagram

Data on ship accidents that occurred in Korea from 2009 to 2014 were included in the simulation model. Details on the variables employed in the stock-flow diagram are provided in Fig.1.-Fig.3. The aim of the model was to determine how the government-related budget factors affected ship accidents in Korea by applying the stock-flow diagram. To determine the potential role of budgetary factors in ship accidents, the following equation was used:

Ship accident = Ship accidenti * * (jfy * (-^J * *

(£L)as * * (a.)"7.....* * f^/11 * (^f112 * (-"T3 *

VCii/ \C2iJ \C3iJ \C6iJ wij/ \G2lJ \G3iJ

(IT* (IT d)

In Equation 1, accidenti is the initial value of a ship accident (i.e., the number of ship accidents that occurred in 2009), and R1, R2, R3, and R4 represent number of registered ships, number of abandoned ships, marine cargo volumes and the volume of travelers, respectively (economic factors). C1, C2, C3, C4, C5, and C6 represent various ship-handling and management factors, such as the faulty ship building, watchman negligence, inappropriate response to weather conditions, sailing regulation violations, equipment- based errors and violations of onboard safety instructions. Additionally, G1, G2, G3, G4, and G5 denote the mean government budget allocated to ship accident prevention (i.e., seafarer qualifications, ship safety, maritime safety management system, marine transportation environment and marine safety global cooperation, respectively). R1, R2, R3, and R4 are based on actual data collected by the Korean government. The items C1-C6 represent actual ship accidents and the factors responsible for these accidents. The items G1-G4 are based on the actual budget allocation of the Korean government in Korean Won. ax means the elasticity of the factors. Fig. 1.-Fig.3. presents the results of SD modeling of ship accidents using Equation (1).

DMnf pNiiMitpW

pirate if

Mmiáps C Urate if DNm]W

M hàdàp ltahsláp!

Amie mi olí „1

Clinofta*! 0 N< ¿ * ► ïihmfto*! ,_________i_____ VoSBMof

iTolBcgfMii tovdtiîNim

mlmsNiim

Fig. 1. SD modeling of ship accidents

Fig. 2. Ship-handling and management factors

Fig. 3. Government budget for ship accident prevention 3.2. Results and Validation of the Model

The elasticity coefficient of each variable was optimized by applying the calibration function of the SD software Vensim. The elasticity of the optimization value was incorporated in Equation 1, and the results are shown in Table 4. With regard to the economic factors, "number of registered ship" was the most influential factor in ship accidents. In the presence of limited voyage space, increasing numbers of registered ships at sea may lead to traffic congestion or even collisions. The variable "volume of travelers" was the least important because most international sea traffic transports cargo rather than passengers. For long-distance travel, passengers prefer airborne rather than sea transport.

With regard to the ship-handling and management factors, "watchman negligence" seemed to be the most critical variable. Given the various causes of ship accidents, vigilance by seafarers is of critical importance in preventing accidents. Most ship accidents occur at a moment by various causes, therefore guard against is an important approach for seafarers to avoid ship accidents. However, during long voyages, vigilance by seafarers generally decreases, increasing the likelihood of accidents. The variable "equipment-based errors" was not strongly correlated with ship accidents. Most of the seafarers were well educated and trained before working onboard. Therefore, they were well able to handle the equipment, with very few errors.

All the government-related budgetary variables had minus elasticity coefficient values. The budgetary items included in the simulation covered ship accident prevention. Thus, increases in the budget decreased the possibility of ship accidents. Among these items, the "marine transportation environment" was the most important. The budget for the "marine transportation environment" aimed to improve sailing infrastructure and install E-navigation in ships to provide more information on a ship's location. Investments in the marine transportation environment can aid seafarers to better guard the condition of the ship and surroundings under advanced technology support. The factor "marine safety global cooperation" was the least important of the five government

ship safety budget items. The budget for global cooperation in marine safety is aimed at ensuring that seafarers have a clear understanding of the IMO's marine safety regulations and those of calling countries. It provides for those in the shipping industry to attend conferences or meet with relevant officials. Although this kind of budget can help to reduce ship accidents caused by regulatory issues, most ship accidents are caused by human error.

Table 4

Elasticity coefficient of each variable

Factor Elasticity coefficient

Number of registered ship 0.0960941

Number of abandoned ship 0.0954212

Marine cargo volume 0.091697

Volume of travelers 0.32949

Faulty ship building 0.385929

Watchman negligence 0.668434

Inappropriate response to weather conditions 0.539409

Sailing regulation violations 0.329334

Equipment- based errors 0.23454

Violations of onboard safety instructions 0.3421

Seafarer qualifications -0.97898

Ship safety -0.551806

Maritime safety management system -0.938187

Marine transportation environment -0.98884

Marine safety global cooperation -0.121624

After obtaining the elasticity coefficient of each variable, the optimization was conducted by inputting the elasticity coefficient of each variable into the model. The results are shown in Fig. 4

2010 2010.60 2011.20 2011.KO 2012.40 2013 2013.60 Time ÇYsar)

O "Number of ship accident : Caj t-i-i---1--î--i--ï-L:-"-i-

Ship accident : Cal -H?--a--^-^-1*--£-?----3--fr-3

Fig. 4. Comparison of actual data with ship accident simulation values

In Fig.4, the line marked 1 denotes actual data on ship accidents in Korea, and the line marked 2 represents the simulation results. As shown in the figure, the simulation results are relatively well matched to the real data. Table 5 presents a comparison of the actual data with the ship accident simulation values. To verify the accuracy of the simulation model, a reliability analysis was conducted by applying a mean absolute percentage error (MAPE) analysis, as shown in Equation (2) below. Based on the reliability analysis, the MAPE value was 1.04%, with a range of 0%<=MAPE<10%. As reported earlier, this sort of simulation process is very accurate (Lewis, 1982).

MAPE—2—*1001

Table 5

Comparison of the actual data with the ship accident simulation values

Year Real data Simulation value

2010 1,942 1,942

2011 2,139 2,159

2012 1,854 1,876

2013 1,306 1,324

2014 1,565 1,593

4. Sensitivity Analysis under Various Scenarios

A sensitivity analysis was performed to examine the influence of the government budgetary factors on ship accidents. The following ship accident prevention items of government-related budget factors were included in the scenario analysis: seafarer qualifications, ship safety, maritime safety management system, marine transportation environment and marine safety global cooperation. Pessimistic scenarios and optimistic scenarios were examined by applying a variation of ± 10% of the variables studies. Five pessimistic scenarios based on a decrease of 10% in government spending were named cases 1 to 5, and five optimistic scenarios based on an increase of 10% in government spending were named cases 6 to 10, as shown in the example below.

Scenario 1 (pessimistic): A decrease of 10% in government budget For example, in Case 1: Ship accident (a decrease of 10% in investment in the seafarer qualifications)=

Ship accidentí * (ff * gf * gf * £T * (if * (|f *

(££\a?.....* (££]"" * ( Ci-Q-i'Ci) )a" t t (Mais * ( Ma" *

VC31/ \RelJ \Gll-(0.1tGlt)J \G2lJ VC31/ \GltJ

ma" (3)

2010.60 2013.20

Ship accident ; SesSíer qualiicatiens Gnmth rate -10% Ship accident : Ship safety GrowA tats -10% -

2011.80 2012.40

Time (Year) —Y-í--

2013.60

Ship accident : Maritime safety manageoKnt system Growth rate -lOT-o -

Ship accident : Marias tfsr.jpwtatior. ercrimmnait Growth rate -10% -

Ship accident : Marine safety global cooperation Growth rate -10% —5— Ship accident : Cal —G-G-G-G-

Fig. 5. Sensitivity analysis: pessimistic scenarios

0%<=MAPE<10%: Very accurate prediction; 10%<=MAPE<15%: relatively accurate prediction; 20%<MAPE<50%: very reasonable prediction; 50%>MAPE: inaccurate prediction.

Fig. 5. presents the results of the sensitivity analysis of pessimistic scenarios. They show that the "marine transportation environment" ranked first as the most influential factor in ship accidents, followed by "seafarer qualifications" and "maritime safety management system.

Scenario 2 (optimistic): An increase of 10% in government budget For example, in Case 6:

Ship accident (an increase of 10% in investment in the seafarer qualifications)=

Ship accidenti * (ff * (|f * gf * gf * * (ff *

(a.)"7.....* f^12 * ( Ci+CM'Ci) * ( Ma" „ ( * ( Ma" *

\C3tJ \RaJ Vcn+(0.1 »CliV \02tJ \G3tJ \GitJ

ma" (4)

2010 2010.60 2011.20 2011.80 2012.40 2013 2013.60 Time (Year)

Slap midan : Síiip séty toit, rafe -M--r-i--J--i--+—

Stipíiódan: SafaerjfiiliiatiHH tottratí-lC:(. —f--£---1-;L-

Stip Midan : Mantra safety oumjaiait avilar. Gurnt rae --¡I-%-3-1

Ship it&jol : tüiir.a crar_5porta[Lor. wùlnoài, toi: rati +10H -4-4-4-

Stip ao:idan : Mira afcty slotal caspaaiior. tetli rata -IOS —5-5-5-5-

Stip «tai —■--0-f--E--$-fi-F--

Fig. 6. Sensitivity analysis: optimistic scenarios

The results of the sensitivity analysis using optimistic scenarios are shown in Fig.6. They indicate that the government budget for "maritime safety management system" is being used effectively to avoid ship accidents but that the budget for "seafarer qualifications" is not.

The ranking of each factor changed in accordance with a decrease and increase of 10% in the government's budgetary allocation. With an increase of 10% in the government budget allocation, "maritime safety management system" changed from 3 to 1, indicating that budget on this item was very important in ship accident prevention and avoidance. Following a decrease of 10% in the government budget, the "marine transportation environment" ranked first. Therefore, the budget for the "marine transportation environment" is an important part of ship accident prevention. With an increase of 10% in the budget, the ranking increased from 1 to 2, indicating that the budget for the "marine transportation environment" is not used efficiently. The budget for "seafarer qualifications" was the same as that for "marine transportation environment." Both items were important in ship accident prevention, but the budget was not used efficiently. The budget for the "marine safety global cooperation" and "ship safety" factors was the same, and these two factors were ranked the same in the model. This indicates that these two

factors do not seem to have a substantial impact on ship accident prevention and that the amount allocated to these items in the budget can be gradually reduced.

In most studies of ship accidents, the most important cause was human error. To reduce human errors, the government should try to allocate more of the ship accident prevention budget to investing in seafarer education and training. The results of the present study showed that a decrease of 10% in the budget for "maritime safety management systems" and the "marine transportation environment" but not in the budget for "seafarer qualifications" had a greater impact on reducing the numbers of ship accidents. In the presence of a limited government budget, to enhance the efficiency of the budget and prevent or avoid ship accidents, the majority of the budget should be allocated to "maritime safety management systems" and the "marine transportation environment" rather than "seafarer qualifications."

Table 6

Government budget for the prevention of ship accidents

Factor Decrease of 10% in each government budgetary factor Increase of 10% in each government budgetary factor

Ranking Change ranking

Seafarer qualifications 2 2^3

Ship safety 4 4^4

Maritime safety management systems 3 3^1

Marine transportation environment 1 1^2

Marine safety global cooperation 5 5^5

5. Conclusion

With the increasing demand for marine safety and protection of property and the environment, the ability to forecast an accident, assess its causes, and ultimately minimize the damage an accident causes to ships, life, property, and the environment are especially significant (Wang et al., 2002). Most previous studies of the factors influencing ship accidents have mainly focused on the roles of human error or the economic environment. To fill the gap in the literature, this paper focused mainly on the impact of political factors (i.e., government budgets) on ship accident prevention in Korea.

This paper aimed to demonstrate and quantify the factors (mainly political) that influence ship safety and ship accidents and suggest policy recommendations for both government and the marine industry in Korea based on the result of SD simulations. A ship accident stock-flow diagram was developed that incorporated economic factors, ship accident causative factors, and government budgetary factors, with data on ship accidents that occurred from 2009 to 2014. A reliability analysis of the model revealed a MAPE value of 1.04%, with a range of 0%<=MAPE<10%, suggesting that the results of the simulation process were very accurate. Following the sensitivity analysis, the government's budget for ship accidents was examined. The results of the sensitivity analysis indicated that the government's budget for "maritime safety management system" was being used effectively to avoid ship accidents but that its budget for "seafarer qualifications" was not.

The implications of this paper can be summarized as follows (1) "number of registered ship," "watchman negligence," and the "marine transportation environment" are the most influential factors that affect ship accidents; (2) in the presence of limited budgetary expenditure, to

increase the efficiency of the budget and prevent or avoid ship accidents, the budget should focus more on "maritime safety management systems" and "marine transportation environment and less on "seafarer qualifications."

With regard to the government budget for ship safety, the present study employed data from only 2009 to 2014. This is too short a period upon which to base policy recommendations or propose efficient strategies to reduce ship accidents.

In future research, first, other factors that influence ship accidents should be added to the existing model to enhance the validity of the simulation. Second, additional data on the government's budget for ship safety should be collected and input in the model for future forecasting and strategy making.

References

AKYUZ, E. and CELIK, M. (2014). "Utilisation of cognitive map in modelling human error in marine accident analysis and prevention," Safety Science, Vol. 70, pp. 19-28.

BLANC, L.A.L. and RUCKS, C.T. (1996). "A multiple discriminant analysis of vessel accidents," Accident Analysis and Prevention, Vol. 28, No. 4, pp. 501-510.

CELIK, M., and CEBI, S. (2009). "Analytical HFACS for investigating human errors in shipping accidents, " Accident Analysis and Prevention, Vol. 41, pp. 66-75.

CELIK, M, LAVASANI, S.M., and WANG, J. (2010). "A risk-based modelling approach to enhance shipping accident investigation," Safety Science, Vol. 48, pp. 18-27.

HASSEL, M., ASBJORNSLETT, B.E., and HOLE, L.P. (2011). "Underreporting of maritime accidents to vessel accident databases," Accident Analysis and Prevention, Vol. 43, pp. 2053-2063.

HETHERINGTON, C., FLIN, R., and MEARNS, K. (2006), "Safety in shipping: The human element," Journal of Safety Research, Vol. 37, pp. 401411.

KNUDSEN, O.F. and HASSLER, B. (2011). "IMO legislation and its implementation: Accident risk, vessel deficiencies and national administrative practices,"Marine Policy, Vol. 35, pp. 201-207.

LEWIS, C. D. (1982). "Industrial and business forecasting methods: A practical guide to exponential smoothing and curve fitting," London: Butterworth Scientific.

LU, C.S. and TSAI, C.L. (2008). "The effects of safety climate on vessel accidents in the container shipping context," Accident Analysis and Prevention, Vol. 40, pp. 594-601.

Ministry of Oceans and Fisheries in Korea, (2014), "Report of maritime security investment plans."

MULLAI, A., and PAULSSON, U. (2011). "A grounded theory model for analysis of marine accidents," Accident Analysis and Prevention, Vol. 43, pp. 1590-1603.

NIKOLAOS P.V. and I. F. GIANNOPOULOS (2013). "Assessing the consequences from marine accidents: Introduction to a risk acceptance criterion for Greece," Human and Ecological Risk Assessment: An International Journal, Vol. 19, No.3, pp. 699-722.

POYRAZ, O. (1998). "Gemi Kazalarindan Dogan Krizlerin Kiyisal Yonetimi ve Turk Bogazlari Bolgesine Uygulanmasi," Istanbul Universitesi.

PSARROS, G., SKJONG, R., and EIDE, M.S (2010). "Under-reporting of maritime accidents," Accident Analysis and Prevention, Vol. 42, pp. 619-625.

SAHIN, B. and SENOL, Y.E. (2015). "A novel process model for marine accident analysis by using generic fuzzy-AHP algorithm," The Journal of Navigation, Vol. 68, pp.162-183.

SHU, C., AHMAD, R., LEE, B.G., and KIM, D.H. (2014). "Composition ship collision risk based on fuzzy theory," Journal of Central South University, Vol. 21, pp. 4296-4302.

TOFFOLI, A., LEF'evre, J.M., BITNER-GREGERSEN, E., and MONBALIU, J. (2005). "Towards the identification of warning criteria: Analysis of a ship accident database," Applied Ocean Research, Vol. 27, pp.281-291.

UGURLU, O., EROL, S., and BAÇAR, E. (2015a). "The analysis of life safety and economic loss in marine accidents occurring in the Turkish Straits," Maritime Policy & Management, DOI:10.1080/03088839.2014.1000992.

UGURLU, O., KOSE, E., YILDIRIM, U., and YUKSEKYILDIZ, E. (2015b). "Marine accident analysis for collision and grounding in oil tanker using FTA method," Maritime Policy & Management, Vol. 42, No. 2, pp. 163185.

VENTIKOS, N.P. and GIANNOPOULOS, I.F. (2013), "Assessing the Consequences from Marine Accidents: Introduction to a Risk Acceptance Criterion for Greece," Human and Ecological Risk Assessment: An International Journal, Vol. 19, No. 3, pp. 699-722.

WANG, G., SPENCER, J., and CHEN, Y. (2002). "Assessment of a ship's performance in accidents,"Marine Structures, Vol. 15, pp. 313-333.

WENG, J. and YANG, D. (2015). "Investigation of shipping accident injury severity and mortality," Accident Analysis and Prevention, Vol. 76, pp. 92-101.