Influencing Factors of Safety Management System Implementation on Traditional Shipping

Abstract

Traditional shipping, often referred to as ‘Pelra’, displays unique characteristics in contrast with conventional ships. This study hypothesizes that the implementation of a Safety Management System (SMS) in Pelra is influenced by both technical and non-technical factors. The primary objective of this research is to ascertain the degree of influence exerted by each factor on the implementation of SMSs in Pelra. Structural Equation Modeling was utilized to analyze expert opinions gathered for this study. The findings indicate that both technical and non-technical factors have direct and interrelated impacts on the safety management of Pelra. Notably, non-technical factors, particularly the authority and responsibility of the company and the ship’s crew, along with technical aspects such as the ship’s construction and stability, emerge as predominant influences. These factors act as control variables, guiding the prioritization of actions in SMS implementation. The insights gained from this research can inform policymakers in developing SMS regulations for traditional ships. A detailed examination of safety issues within the most influential factors is undertaken to reshape policies and program directives, aiming to enhance traditional shipping safety. The adoption of SMSs implies increased responsibilities and associated costs for ship owners and crew, necessitating consensus and commitment among all stakeholders, including the government, ship owners, and crew, for effective SMSs policy implementation in Pelra.

1. Introduction

Traditional shipping, commonly known as ‘Pelra’, has long been integral to trade and transport in Indonesia. Pelra, typically wooden ships, have transitioned from sail to engine propulsion (see Figure 1). These vessels play a crucial role in connecting growth centers to small and remote islands, remaining vital for the livelihood of smaller island communities. Despite their importance, Pelra ships are frequently involved in accidents.

In recent years, the Pelra ships has undergone significant changes, shifting from cargo transportation to tourism. This evolution marks a notable decline in its national significance compared to 20–30 years ago. Challenges such as reduced cargo demand and limited wood resources, the primary material for these ships, have led to safety concerns due to a higher incidence of accidents in traditional shipping [1].

Pelra ships, as non-conventional vessels, are mandated by government regulations to adhere to safety standards [2]. Despite this, the implementation of a Safety Management System (SMS) has been lacking. The risk of accidents and damage is particularly high in eastern Indonesian waters, which are known for their high waves. Data from 2018–2022 in

Table 1. Overview of maritime accident investigations conducted by the National Transportation Safety Committee (NTSC) in Indonesia.

No	Explanation of Causes	Year					Total
		2018	2019	2020	2021	2022
1	Sink	10	6	3	5	5	29
2	Burns/Explodes	12	6	2	6	5	31
3	Collision	3	9	2	4	1	19
4	Run aground	7	0	4	2	2	15
5	Others	7	4	1	2	0	14
Total		39	25	12	19	13	108

indicate that Pelra ships were involved in 10% of the 108 ship accidents investigated by the National Transportation Safety Committee (NTSC) [3]. Implementing an SMS is expected to mitigate these safety risks.

In SMS application, factors are generally categorized into technical and non-technical aspects. Human error, a non-technical factor, is often cited as a primary concern, which is usually due to crew members’ lack of safety awareness and training. This aligns with theories suggesting that most maritime accidents are attributable to human error whether intentional or not. Ensuring the crew’s health and physiological condition is therefore vital [4]. Consequently, SMS designs should aim to minimize human error [5,6].

Company responsibility also plays a crucial role. Companies must provide adequate resources for safety management and ensure crew members are well-versed in safety procedures [7]. This includes maintaining ships and equipment to safety standards, conducting regular maintenance, and implementing clear safety management policies. Effective risk analysis and incident reporting mechanisms are also essential [8], as is the use of early detection systems to mitigate environmental hazards [9].

From a technical standpoint, the construction and stability of ships are paramount factors in SMSs. Ships must meet regulatory standards and be built with appropriate materials and processes. Stability considerations, such as weight distribution and buoyancy, are critical for safety under various conditions [10]. The regular maintenance of construction and stability is essential, as Pelra ships often exhibit low stability due to non-standard construction, making them susceptible to high waves [11,12].

Domestic vessel SMS regulations cover a range of functions, from safety policies to company audits [13]. However, these standards cannot be directly applied to Pelra ships due to their unique characteristics. The goal of Pelra SMS should be to ensure operational safety by establishing standards and procedures to address potential risks to the ship, crew, and environment. This includes enhancing emergency preparedness for safety and environmental protection.

This paper builds on Wahid’s research [14], which identified factors and indicators affecting SMS implementation, by quantifying the influence of each factor. Previous studies, such as that of Malisan et al. [15], have focused on stability, strength, and human resources. This research aims to expand on these factors and quantify their influence. Additionally, it is the first study to develop a concept for Pelra SMS policy in Indonesia.

Understanding both technical and non-technical influencing factors is vital for developing Pelra SMS policy. Recognizing the interplay among these factors will aid in comprehensive risk analysis, effective policy formulation, resource optimization, stakeholder engagement, and adaptation to environmental changes. This approach will enable the creation of more contextual, adaptive, and effective safety management policies for future ship safety.

2. Research Methods

2.1. Population Type and Sample

This study is a cross-sectional analysis focusing on the elements of the Safety Management System (SMS) when applied to traditional maritime transportation. The primary method used is Structural Equation Modeling (SEM), which is a multivariate statistical technique that integrates regression, factor analysis, and path analysis. SEM’s application in transportation research is well documented, including in areas such as sea transportation network integration [16], accessibility and connectivity [17], public bus transportation [18], aviation service quality [19], and quality of goods distribution services [20].

The data for this study were collected through an online questionnaire distributed via Google Docs. The questionnaire targeted respondents with experience in marine transportation safety and included items related to research variables. Responses were measured using a Likert scale, with 5 indicating ‘strongly agree’, 4 indicating ‘agree’, 3 indicating ‘moderately agree’, 2 indicating ‘disagree’, and 1 indicating ‘strongly disagree’. Both reliability and validity tests were conducted on all variables.

A total of 265 respondents participated in this study, meeting the minimum SEM requirement of 200 participants [21]. The selection of informants was conducted using a purposive sampling technique. The demographic breakdown of participants was as follows: 90% male and 10% female, with educational backgrounds of 32.4% high school, 39.7% Bachelor’s or Diploma, and 27.9% postgraduate. Age distribution included 38.9% aged 30-40 years, 45.3% aged 41–50 years, and 15.8% over 51 years. Regarding work experience, 45.7% had 5–10 years, 31% had 11–15 years, and 23.3% had over 15 years in the field. Additional details on informant characteristics are provided in

Table 2. Distribution of respondents by occupation.

Characteristics	n	%
- Academics	30	11.3
- National Search and Rescue Agency	44	16.6
- Indonesian Classification Bureau (ICB)	7	2.6
- Shipping Directorate, Ministry of Transportation	6	2.3
- Shipping Court	12	4.5
- Marine Inspector	29	10.9
- Ship Owner/Ship Crew	78	29.4
- Harbormaster	59	22.3

.

2.2. Research Variables

This study involves the identification of both latent and manifest variables, which are critical in assessing the performance improvement of Pelra’s Safety Management System (SMS). These variables were identified through a comprehensive review of relevant literature and in-depth interviews conducted using the Delphi method. The detailed list and descriptions of these variables, including their sources and potential impacts on Pelra’s SMS performance, are presented in

Table 3. Research variables.

Variable Latent	Variable Manifest		References	Factor Type
Corporate responsibility and authority (X1)	X1.1	Establish rules and procedures for ship safety and environmental protection	[13,15,22,23,24]	Non-Technical
	X1.2	Regularly monitor crew compliance with vessel safety requirements
	X1.3	Ensure safety rules are implemented by all crew
	X1.4	Ensure the availability of crew resources in accordance with manning regulations
	X1.5	Prepare operation checklist for vessel operations related to safety and personnel
	X1.6	Consistent implementation of SMS regulations
	X1.7	Implementation of ongoing safety management training for crew members
	X1.8	Consistently conduct regular meetings to find solutions to safety management issues
	X1.9	Appoint crew members who understand the safety aspects of the vessel (skipper and shore personnel)
	X1.10	Program and internally evaluate safety activities
	X1.11	Evaluate the effectiveness of SMS and review in accordance with established procedures
Crew responsibility and authority (X2)	X2.1	Routinely check the completeness requirements of safety systems on board	[13,15,23,25,26,27,28,29,30,31,32]	Non-Technical
	X2.2	Understand the duties and responsibilities related to ship safety management system
	X2.3	Obtain precise, clear and easy clarity of instruction in the implementation of safety systems
	X2.4	The skipper motivates the crew to implement the safety policy
	X2.5	Routine strengthening of leadership to captains
	X2.6	Able to operate navigational equipment
Resources and personnel (X3)	X3.1	Receive regular vessel safety training (skipper and shore personnel)	[13,30,33]	Non-Technical
	X3.2	Psychological examination of crew members before sailing
	X3.3	Checking the physical condition of the crew before sailing
	X3.4	Crew health check before sailing
Emergency readiness (X4)	X4.1	Identify potential emergency situations on board	[13,22,23,27,28]	Non-Technical
	X4.2	Establish procedures for responding to emergency situations
	X4.3	The crew must be able to respond quickly when conditions occur that jeopardize safety
Administration and Documentation (X5)	X5.1	Establish procedures for controlling documents and data related to the safety management system	[13]	Non-Technical
	X5.2	Organizing document and data control procedures related to the safety management system
	X5.3	Establish and document authority, responsibility, and coordination patterns among crew members in the implementation of the safety management system
Ship maintenance (X6)	X6.1	The ship owner establishes regular ship maintenance procedures	[13,15,22]	Technical
	X6.2	The crew understands maintenance operation manuals and routine maintenance systems
	X6.3	The crew performs routine ship maintenance
Ship Construction (X7)	X7.1	Connection system	[14]	Technical
	X7.2	Ship body impermeability
	X7.3	Transverse watertight bulkhead
	X7.4	Reinforcement of machine foundation
	X7.5	Reinforcement of deck and deck house construction
	X7.6	Reinforcement of hatch area
Ship stability (X8)	X8.1	Cargo layout	[14]	Technical
	X8.2	Type of cargo transported
	X8.3	Ship shape and size
	X8.4	Wind, waves, currents and storms
Safety and navigation equipment (X9)	X9.1	Checklist the condition and quantity of safety and navigation equipment	[14]	Technical
	X9.2	Guidelines for the use of safety and navigation equipment
	X9.3	Placement of safety equipment in an easily accessible location
	X9.4	Crew skills using safety and navigation equipment

.

2.3. Conceptual Framework and Hypothesis

Drawing from the work of Ofori et al. [34], it is posited that a safety culture is established through knowledge and consistent safety training. However, these elements are only effective when there is adherence to safety regulations and standards. Additionally, the enhancement of safety performance is contingent upon the support of robust policies and institutional frameworks [35].

In this research, the conceptual framework encompasses safety knowledge, safety training, safety policy, and institutional support as key latent variables. These variables comprise nine factors and are further delineated into manifest variables, which include both technical/environmental and non-technical factors. The structural model (Y) is conceptualized as follows: Y = f (X₁, X₂, X₃, X₄, X₅, X₆, X₇, X₈, X₉), where the measurement model is X₁ = f (X_1.1, X_1.2, X_1.3, X_1.4, X_1.5, X_1.6, X_1.7, X_1.8, X_1.9, X_1.10, X_1.11); X₂ = f (X_2.1, X_2.2, X_2.3, X_2.4, X_2.5, X_2.6); X₃ = f (X_3.1, X_3.2, X_3.3, X_3.4); X₄ = f (X_4.1, X_4.2, X_4.3); X₅ = f (X_5.1, X_5.2, X_5.3); X₆ = f (X_6.1, X_6.2, X_6.3); X₇ = f (X_7.1, X_7.2, X_7.3, X_7.4, X_7.5, X_7.6); X₈ = f (X_8.1, X_8.2, X_8.3, X_8.4); and dan X₉ = f (X_9.1, X_9.2, X_9.3, X_9.4). The interplay between technical and non-technical variables is elucidated through both the structural and measurement models, as depicted in the conceptual framework presented in Figure 2. This framework forms the basis for understanding how these variables interact within the context of the study.

Based on the delineated relationships in the SEM conceptual framework, it is hypothesized that each variable (H₁ to H₉) exerts a direct and positive influence on the overall improvement of transportation safety performance, which is represented by hypothesis H₀. Therefore, the research hypotheses are formulated to reflect this direct positive relationship, which is denoted as [(H₁ to H₉) = (H₀)].

3. Results and Discussion

3.1. Statistical Test of Data and Measurement Model

The reliability of the questionnaire was assessed using Cronbach’s alpha (CA), yielding values between 0.73 and 0.86. This indicates a high level of reliability for the measurement tool. Validity testing on 44 indicators, with a 95% confidence level, resulted in a count value of r = 0.232, confirming the validity of all indicators. A normality test, using the critical ratio skewness value criterion of ±1.98 at a significance level of 0.03, showed that the data are normally distributed, with skewness and kurtosis values ranging between −2 and +2 [36]. The outliers test indicated a significant value of p < 0.001, suggesting an absence of outlier observations. Additionally, the multicollinearity test, with a determinant of the sample covariance matrix value of 4.438, indicated no issues of multicollinearity or singularity in the data.

Confirmatory Factor Analysis (CFA) was conducted to further ensure the reliability and validity of the model. The results, as shown in Table 3, demonstrate that the composite reliability (CR) for each construct exceeded the threshold of >0.7 [37], factor loadings were above >0.5 [38], and all average variance extracted (AVE) values surpassed the threshold of >0.5 [39], confirming the model’s high reliability. Discriminant validity was also established, as the square root of AVE values (Table 3) was greater than the correlation coefficients between the constructs [40], indicating a significant effect of all manifest variables on their respective latent variables. The CFA validation for the measurement model is detailed in Appendix A.

3.2. Structural Model Analysis

The model’s quality was evaluated using various goodness of fit (GOF) tests, including sig-probability, RMSEA, GFI, AGFI, CMIN/DF, TLI, and CFI, with the standardized criteria values listed in

Table 4. Full structural model test results.

Test Criteria	Standardized Cutoff Criteria	GOF Test Results	Description
Sig-Probability	≥0.05	0.083	Good fit
RMSEA	≤0.08	0.047	Good fit
GFI	≥0.90	1.484	Good fit
AGFI	≥0.90	0.915	Good fit
CMIN/DF	≤2.00	1.720	Good fit
TLI	≥0.95	0.945	Marginal fit
CFI	≥0.95	0.921	Marginal fit

Description: GOF = goodness of fit; RMSEA= root mean square error of approximation; GFI = goodness-of-fit index; AGFI = adjusted goodness-of-fit index; CMIN/DF = the minimum sample discrepancy function; TLI = Tucker–Lewis index, CFI = goodness-of-fit-index.

. The initial hypothesis posited that all latent variables (X1 to X9) significantly influence the application of SMSs in Pelra sea transportation, as depicted in Figure 3. However, the initial model test indicated that the GOF values did not meet the requisite criteria, necessitating modifications to the structural model. Subsequent tests and estimations involved altering paths between latent variables and manifest variables to optimize the model [16].

After several modifications, the optimal structural model was identified, as shown in Figure 3. These modifications involved adjusting the paths so that latent variables not only exert direct effects on the implementation of SMSs but also have indirect effects. Specifically, the safety and navigation equipment variable (X9) is influenced by both the company’s responsibility and authority variable (X1) and the crew’s responsibility and authority variable (X2), subsequently affecting SMS implementation. Variable X1 directly impacts variables X3, X7, and X9, while X2 directly influences variables X4, X5, X6, X8, and X9. Additionally, there is a reciprocal influence between variables X1 and X2. In this final model, the SMS is directly influenced by all latent variables from X1 to X9. The comprehensive structural model diagram is presented in Figure 3, and the goodness of fit (GOF) values of the model are detailed in Table 4.

The model fit test results, as shown in Table 4, indicate that the modified model (Figure 3) performs well since most test criteria, including sig-probability, RMSEA, GFI, AGFI, and CMIN/DF, meet the specified standards. Despite TLI and CFI values being slightly below the cutoff threshold, the model is considered acceptable due to the marginal fit acceptance and the proximity of these values to the cutoff limit [40]. Ghozali [36] suggests that if two or more test criteria indicate a good fit with the data, the model can be deemed satisfactory.

3.3. Model Interpretation

The analysis of the SEM model’s suitability demonstrates the significance of both direct and indirect relationships among variables, offering insights into the extent of their influence on the implementation of Pelra’s Safety Management System (SMS). The estimated values, as presented in

Table 5. SEM model fit test results for the implementation of SMS in Pelra shipping.

Variables	Estimate	S.E	C.R	Description	Influence Rating
X1 → X9	0.590	0.187	7.452	significant	-
X1 → X3	0.506	0.023	8.704	significant	-
X1 → X7	0.691	0.095	10.450	significant	-
X2 → X4	0.703	0.150	8.054	significant	-
X2 → X5	0.609	0.034	8.334	significant	-
X2 → X6	0.534	0.090	3.903	significant	-
X2 → X8	0.804	0.056	6.034	significant	-
X2 → X9	0.690	0.056	12.434	significant	-
X1 → SMS	0.810	0.138	10.034	significant	2
X2 → SMS	0.842	0.084	6.430	significant	1
X3 → SMS	0.585	0.041	4.899	significant	7
X4 → SMS	0.555	0.039	6.346	significant	8
X5 → SMS	0.503	0.040	13.441	significant	9
X6 → SMS	0.618	0.042	14.434	significant	6
X7 → SMS	0.803	0.023	8.233	significant	3
X8 → SMS	0.785	0.076	7.034	significant	4
X9 → SMS	0.704	0.092	8.438	significant	5

, provide a quantified understanding of the correlations between these variables.

Based on the estimated values in Table 5, which reflect the extent of positive and significant influences on SMS implementation in Pelra vessels, these values range from 50.3% to 84.2%. Accordingly, the hypothesis [(H₁ to H₉) = (H₀)]—based on the SEM model fit test—suggests that the latent variables (X1 to X9) have a direct impact on SMS application. Notably, the variable representing crew responsibility and authority (X2) emerges as the most influential, which is followed closely by the employer responsibility and authority variable (X1). Subsequent variables, including construction, ship stability, safety and navigation equipment, ship maintenance, administration and documentation, emergency preparedness, and resources and personnel, also exhibit significant effects. Specifically, variable X2, elucidated through manifest variables indicated by X2.1 to X2.6, shows an influence of 84.2%. Similarly, the administration and documentation variable (X5) demonstrates a positive and significant effect on SMS application in Pelra vessels (H5 = accepted), with its influence explained by indicators X5.1 to X5.3, accounting for 50.3%.

Furthermore, the estimated values indicate a hierarchy of influence among the factors affecting SMS application, ranked from the most to the least influential: crew responsibility and authority (X2), company responsibility and authority (X1), ship construction (X7), ship stability (X8), safety and navigation equipment (X9), ship maintenance (X6), resources and personnel (X3), emergency preparedness (X4), and finally, administration and documentation (X5). This order of influence is visually depicted in Figure 4.

Figure 4 illustrates the varying levels of influence exerted by each factor on the implementation of the Safety Management System (SMS) in Pelra shipping. It is crucial to note, however, that the lower influence of certain factors does not diminish their importance in maritime safety. In the context of safety, every factor, regardless of its perceived magnitude, has the potential to contribute to maritime accidents. Therefore, when implementing SMS on Pelra vessels, it is essential to conduct a thorough analysis of each factor. This comprehensive approach allows for the identification of potential issues and the development of targeted strategies and program plans, ensuring a robust and effective safety management system.

3.4. Implications of Non-Technical and Technical Aspects of SMS Implementation

The most influential non-technical aspect in the implementation of Pelra’s Safety Management System (SMS) is crew responsibility and authority (X2). This significance stems from the captain’s critical role in coordinating sea activities, ensuring the safety of both the crew and the ship, and making key technical decisions, especially in emergency situations [22]. The crew’s responsibilities extend to regular maintenance and understanding safety and navigation aspects, which are vital for keeping the ship seaworthy. Similarly, land-based personnel play a significant role in organizing and coordinating safety activities before departure. However, the initial competence of Pelra crews is currently deemed inadequate, lagging in technology mastery and correlating strongly with accident occurrences [15,23].

The second most influential variable is the responsibility and authority of the company or ship owner (X1). Traditional management structures in Pelra (kinship relations) place significant responsibilities on the ship owner for financing operations and ensuring crew safety [22]. However, challenges persist, including providing competent personnel and modern safety and navigation equipment as well as ensuring ship construction meets the standards of the Indonesian Classification Bureau (ICB).

Financial constraints are a major challenge for Pelra companies with a significant decrease in market share leading to imbalances between revenue and operational costs [41]. The responsibility for providing adequately educated and experienced personnel, despite the implication of higher salaries, falls on the company [23]. Furthermore, the cost of safety and navigation equipment, along with the necessity of regular maintenance for the predominantly old Pelra fleet, adds to the financial burden.

These findings underscore the high impact of human error (crew) and company management in implementing Pelra’s SMS. The alignment of crew ability with company safety management capabilities, particularly in fostering a safety culture, is crucial. Yet practical challenges, including the cost implications for companies amid competition with conventional ships, present significant hurdles in SMS implementation [42].

In terms of technical aspects, the SEM analysis indicates that ship construction (X7) directly and positively influences SMS implementation, also affecting ship stability (X8). Ship construction is crucial for SMS effectiveness, with dynamic stability being dependent on the ship’s design and dimensions. The longitudinal and transverse strength, which are critical for wave resistance, are significantly influenced by loading practices on Pelra ships. Research by Malisan et al. [15] confirms that most ships under 150 GT meet and often exceed regulatory strength standards. However, despite these strengths, traditional design elements in construction still pose safety risks, with vulnerabilities in areas such as the hull, bulkheads, engine foundation, and deck construction [22].

Malisan et al. [15] also found issues in watertight integrity, particularly in the bulkhead separating different ship sections. Inadequate engine foundation reinforcement, especially in ships under 150 GT, and insufficient deck house reinforcement for external loads are prevalent issues. Damage often occurs due to collisions, which is exacerbated by the lack of robust construction in these areas [22].

Stability studies on various Pelra cargo vessels indicate compliance with technical seaworthiness criteria based on stability measures like righting arm and metacenter height, as required by the International Maritime Organization (IMO). This includes stability in Beaufort wind conditions 4, 5, and 6 as well as for converted tourist pinisi ships in wave heights up to 3.12 m [43,44]. Despite this, challenges remain, especially in coastal and inter-island shipping in underdeveloped and border areas, which often face natural sea conditions ranging from Beaufort 4 to 6. Important factors include a reliable weather information system, careful cargo supervision, especially of hazardous materials, and adherence to safety protocols in cargo preparation. Crew understanding of ship hydrodynamics, proper loading arrangements, and avoiding overloading are essential to prevent loss of stability and potential sinking in adverse weather [15].

4. Conclusions

Traditional shipping, known as Pelra, plays a pivotal role in enhancing connectivity and the distribution of goods across inland regions in Indonesia. Pelra, being non-conventional ships, have yet to fully implement Safety Management Systems (SMSs), resulting in recurrent accidents each year, particularly in vessels under 150 GT. This study reveals that both technical and non-technical factors significantly impact the application of SMSs, thereby influencing Pelra ships’ seaworthiness and safety.

Our analysis identifies the most influential factors in SMS implementation as the authority and responsibility of the ship’s crew (84.2%) and the company (81.0%), which are followed by ship construction (80.3%). Other notable factors include ship stability (78.5%), safety and navigation equipment (70.4%), ship maintenance (61.8%), personnel resources (58.5%), emergency preparedness (55.5%), and administration and documentation (50.3%).

These findings imply an increased responsibility for ship owners, encompassing the provision of competent crew, ensuring technical feasibility (construction and stability), risk identification to minimize emergencies, compliance with international regulations for safety and navigation equipment, and regular ship maintenance. Future research should focus on developing comprehensive SMS policies that integrate all these influential variables.

However, this research has limitations. (1) The indicators used could be further refined to strengthen the factors and indicators applicable to Pelra’s safety management system. (2) The study’s focus on ships operating in Indonesia’s outermost, remote, and border areas may require validation for different regional characteristics. (3) The analysis’s reliance on respondent inputs introduces a degree of subjectivity, necessitating cautious interpretation of the results.