The viability and revenue of the fisheries industry are substantially impacted by fluctuations in seafood prices (Dahl and Oglend, 2014). Changes in demand and supply are a primary cause of seafood price swings, which impede the fisheries sector’s performance (Samy-Kamal, 2021; Surathkal et al., 2017). Consumer behavior coupled with market structure account for the main drivers of demand shifts that affect stakeholder income and the sustainability of the fishing sector (Little et al., 2018). The erratic nature of demand poses challenges in formulating comprehensive marketing strategies, leading to less than optimal performance in the fisheries sector (Prosperi et al., 2019; Salas and Gaertner, 2004). Every fisheries-related business, regardless of its size, faces a range of challenges, including limited funding, insufficient funding, or inadequate levels of financial management literacy (Parappurathu et al., 2019; Muddassir et al., 2019). Growth in Pakistan’s fisheries industry is being hampered by higher tariffs (Mohsin et al., 2024).

Hypothesis 1 (H1). Economic risks affect performance of the fisheries sector in Pakistan.

Climate change, rising temperatures, fish migrations, alterations in current patterns, and variations in salinity induce a cascading effect on fisheries. Typhoons, cyclones, and floods harm coastal aquaculture, causing market instability and economic losses (Khan et al., 2016; Yadav et al., 2024). Reductions in capture productivity resulting from overexploitation has adverse economic implications (Colloca et al., 2017). Furthermore, the overexploitation of fishery resources is linked to environmental degradation, especially when big fishing vessels, such as trawlers, are employed (Pipitone and Colloca, 2018; Hiddink et al., 2011). Changes in salinity levels disrupt aquatic habitats, making them unsuitable for marine life (Röthig et al., 2023). Pakistan’s fisheries sector faces significant challenges, including destructive fishing techniques, exceeding capture limits, and excessive bycatch (Noman et al., 2022). Polluted water makes it difficult for fish to spawn, which diminishes biomass output in subsequent generations, lowering catch numbers and economic gain (Hamaguchi, 2024; Hall et al., 2000; Hamilton et al., 2016). Environmental disruption reduces steady fish supply to the market and raises market uncertainty (Fleming et al., 2014).

Hypothesis 2 (H2). Environmental risks affect performance of the fisheries sector in Pakistan.

Poor policy implementation and insufficient monitoring systems are the biggest obstacles faced by the fisheries sector in Pakistan (Noman et al., 2022). Inconsistent regulations and a lack of departmental cooperation are limiting the development of the fishing sector (Ullah et al., 2021). Stakeholder participation is critical for effective fisheries policy decisions (Pita et al., 2010; Mackinson et al., 2011; Mohsin et al., 2022). Additionally, in order to revitalize the fisheries industry in Pakistan, it is imperative that trade policies be brought up to date and that new trade agreements be negotiated (Yeo and Deng, 2019).

Hypothesis 3 (H3). Policies and regulations risks affect performance of the fisheries sector in Pakistan.

Non-compliance issues and consumer dissatisfaction are impacting consumption as well as fisheries trade (Masakure et al., 2009; Mohsin et al., 2024). Consumers are unaware of the full benefits of consuming fish, as well as the products made from it (Carlucci et al., 2015; Ueland et al., 2012). Moreover, the cultural and dietary preferences of the consumers play a major role in abruptly changing the market structure (Saidi et al., 2023; Mitra et al., 2021). Decreased fish consumption in the low-income class due to price fluctuation affects stakeholders’ income. Poor product quality and consumer dissatisfaction make the fisheries sector vulnerable to degradation (Carlucci et al., 2015; Rebezov et al., 2021). In Pakistan, fish market dynamics are heavily affected by the seasonal cycle of fish availability, which peaks in the winter season and falls in the summer season (Paudel et al., 2020). Thus, domestic and international supply chains continuously face inconsistency issues.

Hypothesis 4 (H4). Consumption risks affect the performance of the fisheries sector in Pakistan.

A significant portion of the fisheries products get spoiled on the way due to the lack of a swift transport system and cold storage facilities (Jan et al., 2014). This results in decreased market demand and increased waste. In the presence of an inefficient transport system, meeting international market standards becomes very difficult (Luqman et al., 2024; Mohsin et al., 2024). Quality and safety issues, limited market access, and low export competitiveness are the biggest challenges confronted by the fishing industry (Rehman et al., 2019; Masakure et al., 2011). The operational efficiency of the production plants in Pakistan is low, resulting in decreased productivity and financial losses (Noman et al., 2022; Mehak et al., 2023). The fisheries sector struggles to prosper due to market uncertainty, as well as infrastructure and technology problems (Bradley et al., 2019). Poor marketing strategies, in conjunction with traceability issues, affect the market competitiveness of the fisheries products of Pakistan (Hobbs et al., 2023; Khan and Khan, 2011). In Pakistan’s fisheries industry supply chain risks are made worse by seasonal changes in fish availability. This makes it hard to meet market needs and disrupts both local and international trade. Poor infrastructure and slow logistics also cause delays and reduce the quality and profits of the industry (Paudel et al., 2020; Mehak et al., 2023).

Hypothesis 5 (H5). Infrastructure and logistics risks affect performance of the fisheries sector in Pakistan.

We conducted this study in an orderly manner ( Figure 1 ). During the first phase, a comprehensive bibliographical review was done. It helped to identify research gaps and formulate research objectives and hypotheses. During the second phase, risks were classified into two tiers according to the explanations and guidelines of Gray et al. (2010) as well as Tingley et al. (2010). Several stakeholders were consulted and this list was modified to ensure risk prevelence. In this list, the first tier consisted of five main risks, followed by 23 sub-risks in the second tier ( Figure 2 ). Study variables’ operational definitions are given in Appendix 1. During the third phase, a questionnaire was prepared for the purpose of obtaining statistics according to the proposed hypotheses (Appendix 2). This questionnaire was discussed and reviewed by five professors working in the fisheries management field and by stakeholders. It was ensured that the questionnaire contained relevant questions and was designed appropriately for the research. A pretesting of questionnaire was carried out by involving 26 respondents. During the pretesting, reliability was assessed by calculating Cronbach’s alpha (CA) for all constructs, ensuring values above 0.6. Moreover, construct validity was tested through exploratory factor analysis (EFA). It was ensured that the questions were grouped appropriately under their respective constructs.

During the fourth phase, data was collected from two coastal provinces of Pakistan, i.e., Sindh and Balochistan. The snowball sampling technique was employed to seek potential and reliable respondents. It is important to mention that snowball sampling can introduce selection bias because it relies on participants referring others within their own networks. This limits the generalizability of the results as the sample may not represent the broader population of fisheries stakeholders. Additionally snowball sampling can overrepresent certain subgroups such as those with stronger ties to the industry while underrepresenting marginalized groups. As a result the findings may be less applicable to a wider more diverse population. However, these limitations are inherent in any statistical method and do not imply that the technique should be abandoned. In fact, snowball sampling is a widely used and effective method for collecting survey data (Noy, 2008; Emerson, 2015; Tubbs and Berggren, 2024). During the fifth phase, reliable statistical approaches such as multi-variate analysis were applied to the data to get dependable results. During the sixth phase, results were interpreted, and conclusions were drawn.

The present research examined how risk factors affect fisheries performance in Pakistan. Additionally, it evaluated the implications of risk perception within the risk control framework. A questionnaire survey was conducted to collect data from May 2024 to October 2024. From Sindh, three coastal districts were selected, viz., Sujawal, Thatta, and Karachi, whereas from Balochistan, two coastal districts, viz., Lasbela and Gwadar, were chosen for data collection ( Figure 3 ). Coastal districts of Sindh and Balochistan were chosen for data collection where the fisheries industry is predominantly concentrated. As a result, they provide a comprehensive representation of the Pakistani fisheries sector. As aforementioned, a wide range of stakeholders were surveyed to obtain questionnaire feedback. An aggregate of 982 participants filled out the questionnaires and were therefore deemed worthy of statistical assessment. First, stakeholder groups were contacted to determine their availability. Afterwards, personal interviews were conducted to gather data. It was attempted to collect data after stakeholder meetings or gatherings. Before the distribution of printed copies of survey forms, the purpose of the questionnaire survey was explained. Every survey respondent was urged to provide accurate answers to every question. Data was collected by telephone in some cases as well. It is important to note that respondent characteristics, such as gender, can lead to biased data. However, discussing this bias is beyond the scope of this study.

Two statistical programs, AMOS 18.0 and SPSS 18.0, were used to analyze collected data and verify the hypotheses established statistically. Inferential and descriptive statistics are the two main statistical techniques used to analyze data. The SPSS 18.0 program was employed to perform a frequency analysis, reliability check, EFA, and random sample t-test, whereas CFA was executed employing the AMOS 18.0 program. This model’s appropriateness and hypothesis verification were assessed using SEM estimates. This model not only estimates independent and dependent variable relationships, but this can also estimate the causal relationship between various dependent variables. SEM is a multivariate analysis consisting of confirmatory factor assessment and regression evaluation.

Further, analytical analyses can be performed with correlations between exogenous variables as well as correlations of errors between endogenous variables. A study that should be carried out individually can be performed simultaneously using one research model (Mirrasooli et al., 2019). SEM has the advantage of showing either direct effects or indirect effects across variables with causal relationships. It can grasp a more meaningful causal relationship by incorporating direct, indirect, and total effects at once (Gallagher et al., 2008). AMOS proposes several indicators to check whether data fits the model. The number of fit measures is around 20 (Amadu et al., 2021; Narayanan, 2012; Byrne, 2001; Zhang et al., 2022). This study used three matrix models to represent SEM. The first matrix is mathematically defined as follows (Equation 1):

In the above matrix, ξ represents exogenous and η denotes endogenous variables used in the study. All of the main risks, including economic risks, policy and regulation risks, environmental risks, consumption risks, and infrastructure and logistics risks, are exogenous, while their corresponding sub-risks are considered endogenous. Γ and B represent the coefficients of variables. The second matrix model or the measurement model can be expressed as below (Equation 2):

In this equality, the measurement of all the endogenous variables is represented by Y. In addition, Coefficients of correlation between endogenous variables and their corresponding measured variables are denoted by Λy. ϵ stands for all those errors that occur during this method. The third matrix model can be represented as follows (Equation 3):

In the above mathematical expression, the calculation of all exogenous variables is given by X. Here, Λx denotes the correlation coefficient between exogenous variables as well as corresponding measured variables. δ represents all errors during this algorithm (Narayanan, 2012; Shek and Yu, 2014).

The essential features of the survey respondents are covered in detail in Table 1 . Regarding the state of their relationships, 211 respondents (21.5%) were single, while 771 (78.5%) were married together. On the other hand, 899 respondents (91.5%) were male, and 83 respondents (8.5%) were female. The age distribution of the respondents was as follows: 208 (21.2%) were between the ages of 25 and 34, 630 (64.1%) were between the ages of 35 and 54, and 144 (14.7%) were between the ages of 55 and 65. Of the people who answered, 149 (15.2%) had only gone to elementary school, 636 (64.7%) had gone from secondary school to master’s degree, and 197 (20.1%) had PhDs. Of the total respondents, 491 (50%) were from Sindh and 491 (50%) were from Balochistan. Out of the total number of responders, 182 (18.5%) had 5~9 years of experience, 581 (59.2%) had 10~14 years, and 219 (22.3%) had 15 years or more. Concerning the stakeholder coalition among the respondents, 303 (30.8%) were fishermen, 177 (18%) were employees of fishing firms, 174 (17.7%) were members of public or private organizations, 138 (14.1%) were researchers, and 190 (19.4%) were customers who were adequately informed.

A reliability test was conducted to confirm the dependability of the data. The findings of this test are detailed in Table 2 . 4 questions were included in economics risks, and the estimated value of CA was 0.945. Besides, policies and regulations risk comprised five questions and a computed value of CA 0.962. In addition, environmental risk contained four questions, and the calculated value of CA was 0.914. Moreover, consumption risk included five questions; the CA accessed value was 0.937. Likewise, infrastructure and logistics risk consisted of 5 questions, and the estimated value of CA was 0.943. Furthermore, the performance comprised four questions, and the calculated value of CA was 0.872. There were 27 questions, and their computed value of CA was 0.981. It should be noted that CA estimates for all constructs were well above 0.6, validating the data’s reliability (El-Sheikh et al., 2017).

Estimates of CFA are given in Table 3 . Many indices representing the goodness of model fit were calculated. These indices included ‘Chi-square’ (x2) (984.547), ‘Normed Fit Index’ (NFI) (0.873), ‘Goodness of Fit Index’ (GFI) (0.831), ‘Root Mean Square Residual’ (RMR) (0.031), ‘Incremental Fit Index’ (IFI) (0.952), ‘Adjusted Goodness of Fit Index’ (AGFI) (0.814), ‘Comparative Fit Index’ (CFI) (0.933), and ‘Tucker-Lewis index’ (TLI) (0.916). These indices confirmed that the data fit the model well and are acceptable for further analysis. It is essential to mention that computed values of ‘Construct Reliability’ (CR) were higher than the standard acceptable value of 0.70. Moreover, the ‘Average Variance Extracted’ AVE estimates were also higher than 0.05, the standard value. Thus, all CFA estimates were acceptable and representative (El-Sheikh et al., 2017).

Table 4 lists the correlation results between all the constructs used in this analysis. Negative signs indicate a negative relationship between constructs, whereas their values indicate correlation strength. The acceptable discriminant accuracy of all the variables is shown by the fact that all of the calculated AVE values are greater than squared correlations.

Statistical indices such as RMR, CFI, CMIN/DF signpost structure fitting of equation modeling approach (El-Sheikh et al., 2017). In some circumstances, NFI is capable of producing more accurate estimations, particularly when complex equation frameworks are applied. In such cases, considering the CFI index is more suitable. Based on standard criteria, all of the index values were found to be valid. This means that the model structure can be trusted ( Table 5 ).

Using the structure equation modeling technique, two models, free and constraint, were evaluated. These models verified the mediating effect of risk perception on fisheries sector performance in Sindh and Balochistan. For the free model, x2 and DF were estimated at 2017.634 and 971, whereas these indices for the constrained model were calculated as 2033.472 and 978, in that order. The estimate of x2 in the free model was lower than the constrained model by 15.838. Thus, the free model performed well as a lower value of x2 can offset DF. Moreover, it was found that risk perception profoundly impacts performance ( Table 6 ).

Table 7 illustrates the direct effects of all constructs on performance and determines the acceptance of the proposed hypothesis. H1 was accepted as an economic risk (estimate = −0.425 and p = 0.000) and showed a noteworthy negative impact on the performance of the fisheries sector. Likewise, environmental risk (estimate = −0.251 and p = 0.007) and consumption risk (estimate = −0.265 and p = 0.000) also significantly affected performance. Considering these results, H3 and H4 were accepted. Conversely, policies and regulations risk (estimate = −0.113 and p = 0.121) and infrastructure and logistics risk (estimate = −0.073 and p = 0.411) were insignificant. In the light of these estimates, H2 and H5 were rejected.

A comparison of estimates for Sindh and Balochistan is presented in Table 8 . Estimates differ between both provinces. This discrepancy suggests that risk perception acts as a mediating factor, influencing the relationship between various risks and performance outcomes in both provinces. In Sindh, the economic risk has a significant negative impact on performance (Estimate = −0.363, CR = −2.399**), while in Balochistan, it positively influences performance (Estimate = 0.274, CR = 2.732**). Policies and regulations risk was found to have no significant effect on performance in Sindh (Estimate = 0.019, CR = 0.182), but in Balochistan, it negatively affects performance (Estimate = −0.359, CR = −3.384**). Environmental risk significantly impacts performance in Sindh (Estimate = 0.374, CR = 3.107***) but has no significant effect in Balochistan (Estimate = −0.146, CR = −1.345*). Both consumption risk and infrastructure and logistics risk do not significantly influence performance in either region, except for infrastructure and logistics risk in both Sindh (Estimate = 0.386, CR = 2.919, ***) and Balochistan (Estimate = 0.511, CR = 3.412***) where positive significant impacts on performance are observed ( Figures 4 , 5 ). Summary of all results for quick reference is given in Table 9 .

This study has several limitations. Such as various stakeholder groups within the fisheries sector have been surveyed to examine risk perception. This type of analysis is often limited because it only provides a snapshot of how different stakeholder groups view their fisheries. If this is the case, the study may only prove helpful for a limited period. Moreover, there is possibility of biasness in the data due to the use of snowball sampling technique, where initial respondents can influence the selection of subsequent participants, potentially limiting the diversity of perspectives. Additionally, the reliance on self-reported data may introduce response bias, as participants may feel inclined to provide socially desirable answers rather than their true perceptions. The use of a single methodological approach, namely the questionnaire survey, could limit the study’s ability to capture a broader range of insights, as it might not account for other qualitative factors that could emerge from in-depth interviews or ethnographic methods. Thus, the data may not truly represent entire fisheries sector. In addition, geographical scope of this study may limit the applicability of the findings of this study to the other regions of the country. Since, the data was collected during specific time period therefore temporal variations in the perceptions of the stakeholders may occur during other times in a year. Furthermore, the complexity of SEM may limit its ability to fully capture multifaceted nature of the risks. Moreover, there may a cultural and social biasness which can effect risk perception leading to less significant conclusions.

This study presents multiple practical implications that could guide fisheries policy decisions in Pakistan and other similar developing nations. A better understanding of risks and risk perceptions can be used to develop risk control and communication strategies. By combining various risk theories with the existing risks, further targeted efforts could provide more insight into people’s definition of fisheries risk. As a major food supply, improving fisheries performance would help achieve SDG 2. Furthermore, the development of resilient and sustainable fisheries plays a crucial role in the attainment of SDG 14 (Life Below Water). By defining risks and determining their impact on fisheries performance, better management policies may be developed, leading to objective-based sustainable growth. Additionally, the results of this study reveal important information that may enhance fisheries’ socioeconomic features by reducing risks. Some of the risks investigated, including environmental risk, are crucial for developing proactive management strategies and improving fisheries performance because they address environmental concerns and the impact of climate change. Moreover, greater risk awareness might result in the conservation of fishery resources, safeguarding their sustainable utilization.

Based on the findings of this study several policy recommendations can be offered. First, the study suggests that capacity-building and incentive-based fisheries management techniques are key to improving fisheries performance. This has been clearly demonstrated by major fish producers like China, Indonesia, and India, where these methods have successfully enhanced their fisheries (Tang and Tang, 2006; Bailey et al., 2016; Sun et al., 2017; Huang and He, 2019). Second, there is a dire need to enhance fisheries risk perception which can be achieved through tailored training programs. Third, stakeholders can be motivated toward adopting sustainable practicing by implementing incentive-based management system. Fourth, fisheries law enforcement is very important to control illegal practices and ensure compliance with the regulations. Fifth, in order to achieve evidence-based management a comprehensive and reliable data collection system should be developed. Sixth, stakeholders must be encouraged to participate in the decision-making process. Last but not least, a strict evaluation system must be employed to evaluate interventions and navigate mangement strategies.