Factors Affecting the Efficiency of Tuna Fishing Vessels: Implications on Tuna Production and Improvement

1

Factors Affecting the Efficiency of Tuna Fishing Vessels:

Implications on Tuna Production

Pedro A. Alviola IV*

Associate Professor, School of Management

University of the Philippines Mindanao

Mintal, Davao City, Philippines

Adjunct Faculty, University of Arkansas

[email protected]

Jon Marx P. Sarmiento

Assistant Professor, School of Management



[email protected]

Larry N. Digal

Professor, School of Management



[email protected]

Sherleen M. Comidoy

School of Management



[email protected]

_____________________ *Corresponding Author: Pedro A. Alviola IV, PhD, School of Management University of the

Philippines Mindanao, Mintal, Davao City, Philippines, 8022, Telefax: +6382 295 2750

[email protected]

2

Factors Affecting the Efficiency of Tuna Fishing Vessels:

Implications on Tuna Production

ABSTRACT

With tuna production declining and importations increasing every year, there is a need to

improve the efficiency of tuna vessels in the Philippines. In this research, we aimed to identify

and evaluate the factors affecting tuna fishing vessels’ efficiency in General Santos City,

Philippines. We used the stochastic frontier analysis and utilized inputs such as gross register

tonnage, effort days, fuel and water consumption. In measuring technical efficiency, results

reveal that a Cobb-Douglas production function with inefficiency model following a half normal

distribution specification be utilized. Our results indicate that fishing vessels’ efficiency can still

be improved by 21%. We also find that increasing the berthing days and enhancing the

operations of purse seiner and ring netter vessels may improve technical efficiency. Also our

findings indicate that handline fishing vessels are more technically efficient compared to purse

seiners and ring netters and policies that limit vessel’s effort days may improve efficiency.

Keywords: Production Frontier Model, Stochastic Frontier Analysis, Technical Efficiency, Tuna

3

INTRODUCTION

In 2013, the Philippines incurred a trade deficit of approximately 16,229 t of tuna.

According to the Bureau of Agricultural Statistics (2013), tuna imports (34,126 t) almost

doubled relative to the level of exports (17,897 t). For example, the skipjack tuna species which

is an important input in canning companies has been declining overtime especially those caught

in Philippine waters (Yamashita, 2008; Aprieto, 1995). As a result, input usage and fishing

vessel depreciation has increased because fishermen opted to expand their fishing grounds

outside Philippine waters (Vera and Hipolito, 2006). Also, Aprieto (1995) reported that most of

the commercial fleet in the Philippines consisted of second-hand vessels which were either

bought or rented from countries such as the United States, Taiwan and Japan. Since the ‘70s, the

Philippines has been a major producer of tuna (Vera and Hipolito, 2006) and in 2010 became the

seventh largest producer of tuna and tuna-like species in the world with approximately 120,553 t

of tuna stocks (FAOSTAT, 2013). The major tuna fishing grounds in the Philippines include the

Celebes, Sulu and South China seas. In a report prepared by the Tuna Sampling Program, in as

early as 1992 (Aprieto, 1995), tuna fishers expanded their operations to include Indonesian

waters, Papua New Guinea and Solomon Islands because of the continued decline of local stocks

(Vera and Hipolito, 2006; Yamashita, 2008).

The capacity to reach a fishing vessel’s catch target is influenced by factors such as fleet

age, proximity to fishing grounds and management schemes that regulate fishing output and

inputs. And subsquently, this affects the vessel’s efficiency levels (New, 2012). Thus, it becomes

apparent that due to the random nature of fishing, the usage of the stochastic frontier analysis to

estimate technical efficieny becomes especially important. Several studies have used this

approach to examine a wide variety of fishery interventions such as permits, quotas and measure

4

vessel efficiency. In Australia for example, New (2012) cites that management schemes that

govern the eastern tuna and billfish fishery include interventions that regulate longline permits

and total allowable effort, and the shift from an effort based system to the quota system as the

newer approach to managing the above mentioned fishery resources. In this case, New (2012)

utilized the technical efficiency approach in measuring the different management schemes

affecting vessel efficiency. In other countries such as Iran, the technical efficiency approach was

used by estimating a stochastic frontier production estimation method in measuring vessel

efficiency in major fishing ports in Southern Iran (Jamnia et al., 2015; Zibaei, 2012). Also, Solis

et al. (2014) used the stochastic distance frontier analysis in evaluating the effect of introducing

fishing quotas on catching red snappers in the Gulf of Mexico. Their findings indicate that the

fishing capacity declined as a number of operators left the red snapper fishing industry and in

turn concentrated on catching vermilion snappers and red groupers.

In using the technical efficiency approach, the output variable can be expressed in

several ways. For example, Herrero (2005) used the total value of the targeted species (revenue)

because multiple species tend to have different prices. This approach is similar to the output

variable used by Sharma and Leung (1999). However, New (2012), used an output index that

utilized the Elteto-Koves-Szulc (EKS) extension of the Fisher index because the fishery

characteristics of the Eastern Tuna and Billfish Fishery are multi-species in nature. Also, New

(2012) reported that despite the fishery’s multi-species characterisation, the quantity of fish

landings (in tonnnes) by vessel was used because the information pertaining to the species catch

breakdown was deficient. Also, del Hoyo et al. (2004) used daily catches as output variable and

were aggregated to construct the annual catch data that is similar to the approach followed by

Esmaeili (2006).

5

In terms of input requirements, Reid et al. (2005) identified the following fixed inputs

that are needed in measuring fishing capacity. These include gross register tonnage (GRT),

engine horsepower and per day freezing capacity. They also accounted for other inputs such as

fishing/searching days, number of hooks set, and fuel consumption. Likewise, New (2012)

considered the number of shots and average hooks per shot to substitute for the total effort in the

fishery and average gear size. Other inputs such as trends in stock availability and productivity

over time were included in New’s (2012) specification. Also, New (2012) identified the

determinants of technical inefficiency which includes fleet size, hooks, the dummy interaction

variable of hooks and vessel length. Likewise, Herrero (2005) used GRT, engine power, number

of crew and effort trips. However, instead of applying GRT, Sharma and Leung (1999) used

vessel size, where they constructed indicator variables to represent medium, large and the target

fish per vessel. The fish species included swordfish, tuna and other fishes. Also, indicator

variables representing ownership and boat operator’s education were considered. Moreover, the

authors also considered the age of the vessel and the fishing experience of boat operator. Other

key inputs include trip days, crew size and other operating costs which include fuel price, bait,

ice and other assorted items.

With the decreasing trend of tuna stocks around the country’s fishing grounds, the

relatively rapid depreciation of fishing fleets and the reported declining tuna projections, we

aimed to identify the drivers that affect the efficiency of the tuna fishing vessels in General

Santos, Philippines. In determining the vessels’ technical efficiency, we identify gross register

tonnage, effort days, fuel and water consumption as factors affecting catch volume. Other drivers

such as berthing days, harbors and markets, seasonality and fishing grounds were also included

6

as possible sources of inefficiency. Our results indicate that efficiency levels of fishing vessels

in General Santos City can still be improved by approximately 21%.

In the next sections of the paper, we give a brief discussion regarding the current trends

in the Philippine tuna industry and discuss the methodology of the stochastic frontier analysis

(SFA) approach. Afterwards, we discuss the model results and present our conclusions and

recommendations.

CURRENT TRENDS IN THE PHILIPPINE TUNA INDUSTRY

From 2002 to 2012 the average annual catch growth rate for the major tuna species was

2% (Bureau of Agricultural Statistics (BAS), 2013). The major tuna commercial species caught

in the Philippines include yellowfin tuna (Thunnus albacares), skipjack tuna (Katsuwonus

pelamis), and bigeye tuna (Thunnus obesus). This annual tuna catch includes all unloaded tuna in

Philippine ports regardless on where the tuna species were caught. The data does not separate

catches from foreign waters or those caught by foreign-flagged vessels (BFAR-NFRDI, 2012).

The 2012 Western and Central Pacific Ocean (WCPO) stock assessment of these species reveal

that overfishing is not present in yellowfin and skipjack stocks, but was confirmed in bigeye,

because of the increasing fishing mortality ratio (International Seafood Sustainability

Foundation (ISSF), 2012). According to the report, reducing the catches of immature bigeye tuna

will increase the overall catch levels. The report also showed that skipjack tuna stocks are

moderately exploited while yellowfin tunas are fully exploited with the least likelihood of

increasing catches that can sustain the yellowfin stocks. Based on the computed ratio of

spawning biomass, these tuna species were not overfished and are consistent with 2013 first

quarter results (ISSF, 2013). Throughout the year both skipjack and yellowfin were found to be

7

abundant in all Philippine waters (BFAR-NFRDI, 2012; Aprieto, 1995). From 2002 to 2012

skipjack, yellowfin and eastern little tuna landings had an annual growth rate of 6% while

negative growth rates were observed for frigate and bigeye tuna (BAS, 2013). Also, both

commercial and municipal tuna producing sectors posted increasing trends for all tuna species

with average annual growth rates of 1% and 3% respectively. Other tuna species such as eastern

little tuna (Euthynnus affinis), frigate tuna (Auxis thazard) and bullet tuna (Auxis rochei) relative

to the mentioned tuna species are minor but are fairly distributed in local markets.

At the regional level, SOCCSKSARGEN garnered the largest tuna production share of

approximately 46% of the total tuna landings in the Philippines (BAS, 2013). The region’s

dominant share is likely the result of improvements in infrastructure facilities (such as the

General Santos City Fishport Complex or GSCFC) for tuna commercial marketing in General

Santos City (Aprieto, 1995). Likewise, the country has a total of seven tuna canneries where six

are located in General Santos City and the remaining cannery is in Zamboanga City (BFAR-

NFRDI, 2012). Nevertheless, in 2012, the tuna species comprised 84% of the total marine fish

landings in Region 12 (BAS, 2013).

DATA

In this article, we use GRT, total effort days, fuel and water consumption (Table 1) and

employ other possible sources of inefficiency such as berthing days, market types (1, 2 and 3),

quarterly fishing period indicator variables and fishing ground location (Moro Gulf, Sulawesi,

Sarangani, Pacific Ocean and Kalamansig). The data on effort in days was derived using the

starting and unloading dates where the days spent for searching and fishing were added and the

number of berthing or mooring days excluded. Also if indicated in the database, the calculation

8

of the total berthing days included the extension requested by each fishing vessel. The market

type variables describe the correspondence between the type of market and gear used. Market 1

corresponds to handline fishermen who unload their catches while markets 2 and 3 are

interlinked where small and large purse seiners and rignetter unload their haul. For the fishing

ground and period variables, a value of 0 or 1 was imputed depending on the share of output

level since there were fishing vessels operating at different fishing time periods in multiple

fishing grounds. The cross-sectional data used in this study is the 2012 vessel-specific daily fish

landings recorded by the Philippine Fisheries Development Authority (PFDA) located in General

Santos Fish Port Complex (GSFPC). Finally, since the input records were relatively complete,

we considered 41 mixed fishing vessels.

METHODOLOGY

In this research, we utilize the stochastic frontier analysis (SFA) approach to measure the

technical efficiency of tuna fishing vessels operating in General Santos City, Philippines. We

used SFA because it can incorporate both a technical inefficiency term and a random error

component that accounts for other sources of statistical noise (Coelli et al., 2005). Moreover, the

SFA becomes the appropriate approach for this study because the method of harvesting/catching

marine resources is inherently random (Sharma and Leung, 1999). An advantage of using the

stochastic frontier analysis is that it can accommodate functional forms such as Cobb-Douglas

and Translog in estimating production functions1.

Following Coelli (1996) and Coelli et al. (2005), we specify both a Cobb-Douglas and

Translog stochastic production functions (Aigner, Lovell and Smidt, 1977; Meeusen and van den

1 A shortcoming of the the stochastic frontier analysis is that it cannot control for zero input, zero output values and

multiple output conditions (FAO, 2011).

9

Broeck, 1977) and denote these in equations 1 and 2. On the other hand, equation 3 specifies the

sources of inefficiency.

�� = + �� + �� + �� + �� + �� − �� (1)

�� = + �� + �� + �� + �� + �� +

��+ ��

�+ ��

�+ �� +!�� +

�� + �� + �� !+!�� + �� −

�� (2)

�� = " + "�#��ℎ%�&� + "�'(�(� + "�)��*�+%� + "�,�-%�%-� +!"�.��/��+%&� +

"�0&'1� + "�0&'2� + " 0&'3� +!"�51� +!"��52� +!"��53� +!"��54� +*� (3)

The Cobb-Douglas model in equation 1 specifies the input drivers of catch output2 such as gross

registered tonnage (GRT), effort in days (Eff), fuel and water consumption (Fuel and Water)

while the Translog specification (equation 2) in addition includes the square and interaction

terms of the inputs3. The terms �� !and ��! represent the random error and error associated with

technical inefficiency while *�!is the unobservable random variable in the technical inefficiency

effects model. The sources of technical inefficiency ��! in equation 3 include berthing days,

harbor and market type, fishing period and fishing grounds (Table 1). The technical inefficiency

term is measured through the error terms, and can be decomposed into statistical noise and

factors associated with inefficiency (Bakhsh, 2007).

2 We follow del Hoyo et al. (2004) and Esmaeili (2006) approach in measuring catch output, where the former

measures total volume in tonnes while the latter in kilograms. 3 While multicollinearity is expected among the independent drivers of catch output, a variance inflation factor

(VIF) benchmark value that is less than 10 imply that there is no degrading multicollinearity present. In this study,

the VIF among the explanatory variables were less than 10, (GRT :1.33, Effort :1.80, Fuel :2.61 and Water :2.14),

thus no degrading multicollinearity was observed.

10

In this paper, we estimated the stochastic production function and technical inefficiency

equations using maximum likelihood estimation with the one-stage approach (Battese and Coelli,

1995). We used the FRONTIER (v.4.1) program developed by Coelli (1996) to estimate

equations 1-3 and performed the likelihood-ratio test of different null hypotheses in order to

identify the model that best fits the data. In addition, we obtained the critical values from the

mixed chi-squared distribution (Kodde and Palm, 1986) and calculated the test statistic using the

formula found in the paper of New (2012).

7� = !−2{ln [<=>?@

<!=>A@]} = !−2{ln[7=0@] − ln![7=0�@]} (4)

RESULTS AND DISCUSSION

Model Inference

The first hypothesis test verifies the assumption of exhibiting a half-normal distribution

(Coelli, 1996) or equivalently determining the absence of the constant term in the inefficiency

model (δ0=0). From Table 2, both the Cobb-Douglas and Translog specifications have likelihood

ratio (LR) statistics less than the critical values (0.21,-0.62 < 5.41). Thus, the assumption of

half-normal distribution was not rejected. Likewise, the two specifications were subjected to the

second hypothesis test of determining whether ɣ = 0. This involved examining the restriction on

variance parameter gamma, or equivalently testing the the absence of the technical inefficiency

effects by comparing the Ordinary Least Squares (OLS) and Maximum Likelihood Estimator

(MLE) estimates. In the Cobb-Douglas specification, the LR statistic was greater than the critical

value (30.58 > 27.03), thereby rejecting the null hypothesis whereas in the Translog model, we

failed to reject the null hypothesis (16.67 < 27.03). Therefore in this model, the OLS estimates

were preferred over those generated by the MLE procedure.

11

Finally, we assumed the absence of squared and interaction terms and carried the

comparative test between Translog and Cobb-Douglas models. With the null hypothesis

validating the choice of the Cobb-Douglas function (-9.11 < 10.50), the final model resulted in a

Cobb-Douglas specification that contained the sources of inefficiency but lacked the constant

term in the inefficiency model. Also, we utilized the stochastic frontier production’s variance

parameters, gamma and sigma squared to validate the functional form of the model. The gamma

parameter represents the proportion of the composite error term characterised by the technical

inefficiency (New, 2012). From table 3, following Taru et al. (2011) paper, 74% of the variation

in the fishing vessels’ output levels was due to technical inefficiency. The model has a sigma-

squared value of 0.39 and is statistically significant at 1 percent level (Table 3).

Also from table 3, the estimated coefficients for each input in the Cobb-Douglas

specification were considered in the constant elasticity of the input. The model results suggest

that increasing the vessel size by 10 percent in terms of GRT, increases output by approximately

5.2 percent. Likewise, a 10 percent increase in fuel consumption and effort days increases output

by 3.9 percent and 2.5 percent, respectively. Also, if water consumption increases by 10 percent,

then output will increase approximately by 0.8 percent. Also, we observe the diminishing returns

to scale in the Cobb-Douglas function. Finally, inputs that significanly drive output responsivess

are vessel size, fuel consumption, effort days and water consumption in decreasing order.

Technical Efficiency Performance

The technical efficiency score for the sample vessels ranged from 0.12 to 0.95 with 0.79

as the mean value (Figure 1). The majority of the vessels (41.5%) had technical efficiency

indices greater than or equal to 0.9, while 29% belong to 0.8-0.9 indices and 12% into the indices

12

equal to 0.7-0.8. This suggests that in 2012, almost half of the tuna fishing vessels in General

Santos City, operated close to the efficient frontier and only 10% performed below the 0.50

index level. In terms of GRT classification, vessels with less than 50 GRT had an average

technical efficiency score of 0.82, while vessels with GRT of 50 to 100 garnered an efficiency

score of 0.73. However, for vessels with above 100 GRT, the average score efficiency score was

lowest at 0.71. From the sample, the results indicate that larger vessels are less technically

efficient compared to smaller vessels.The presence of technical inefficiency is verified because

the gamma term is statistically significant. From the model, variables such as berthing days,

harbors and markets 1 and 3, and 1st quarter fishing period were significant at α=0.05 level of

significance (Table 5). In this case, a positive coefficient indicates diminishing technical

efficiency while negative values represent effciency improvements.

Berthing Days

The average tuna fishing vessel’s unloading period (days) showed a significant negative

relationship with technical inefficiency (positive relationship to technical efficiency). This

implies that increasing the berthing period by one day increases technical efficiency (decreases

technical inefficiency) by approximately 3%. This is possible because the longer the vessel’s

berthing days implies longer time to unload catches because of higher volume harvested.

Furthermore, longer berthing days would translate to relatively lower operational cost on a per

vessel basis. Tuna fishing vessels with berthing periods exceeding 10 days had an average

technical score of 0.93 while those with 5 to 10 days and less than 5 berthing days had lower

technical efficiency scores equal to 0.84 and 0.78 respectively. This is because when berthing

days are prolonged, input consumption of water and fuel become minimal. Furthermore, vessels

with more than 10 berthing days had the lowest input consumption per tonnage for water and

13

fuel (0.28 and 13.91) but incurred the highest level of output per trip for the year 2012 (14 t)

(Table 4).

Harbors and Markets

The fishing port complex of General Santos City contains four harbors and two wharves.

Each harbor corresponds to the type of fishing vessel and tuna markets. Harbors and markets 1 to

3 and wharf 1 cater domestic fishing vessels while harbor and market 4 and wharf 2 accomodate

foreign operators. In this study, we consider the first three harbors and markets to be relevant.

We found using field observations and key informant interviews that harbor 1 deals with

handline tuna fishermen while harbors 2 and 3 transact with purse-seiners and mini purse seiners

or ring netters. Furthermore, harbor 1 is used for market transactions that involve export and

domestic market of Grade A whole, frozen and chilled yellowfin tuna and by products such as

blue marlin (market 1). On the other hand, market 2 deals with assorted fish types for the

domestic market while market 3 process tunas that are usually purchased by canning companies

such as skipjack.

The Cobb-Douglas technical inefficiency model suggests that fishing vessels that unload

in harbor and market 1 display a positive relationship with technical efficiency. This means that

handliners are more technically efficient compared to other tuna fishing vessels that discharge in

harbors and markets 2 and 3. In 2012, the average handline fishing vessel holds a gross tonnage

size of 15 GRT and usually unloads majority (84%) of its catch in harbor and market 1 (Table 5).

Also, the data indicate that the average handline vessel’s fuel consumption was 784 liters for 16

days of operation and unloaded its catch for two mooring days. In terms of gross tonnage, the

fuel consumption of handliners appear to be higher, but the daily fuel consumption was lower by

14

49 liters per day compared to 77 liters and 82 liters for vessels unloading in harbors and markets

2 and 3 respectively. The berthing days were also shorter for handline tuna fishing vessels

because of frequent yearly trips. Also, the technical efficiency of tuna handliners is driven in

part by the usage of fish aggregating devices (FAD)4 or payaos, because handliners share the

technology with tuna purse seiners which target small pelagic tuna species such as skipjack. For

example, Aprieto (1995) found that this technology sharing enabled handline fishermen monitor

fish concentrations and guard fish clusters from intruders. In this manner, handliners can easily

catch tuna species such as yellowfin and bigeye because these are usually found at the bottom of

skipjack schools. Also, this arrangement enables handliners to generate capital savings from

purchasing additional aggregating devices.

The inefficiency model further suggests that tuna fishing vessels that unload in harbor

and market 3 are relatively technically inefficient (0.74 technical efficiency score) compared to

handline fishing vessels. An average purse seine vessel’s capacity is 88.5 GRT and ranges from

35 to 142 GRT. The purse seine vessels had longer berthing days and the daily fuel consumption

is higher. The less efficient purse seine tuna fishing vessels are usually cast-off vessels that were

bought or chartered from countries such as the United States, Taiwan and Japan (Aprieto, 1995).

Fishing Periods

In general, fishing operations are restricted during monsoon seasons. The season usually

starts from May to October (Southwest monsoon (SWM) or Habagat) and also from November

4 The usage of FADs has generated a number of studies that recommend either limiting, regulating or banning FADs

because of its effects on tuna sustainability. For example, Cabral, Aliño, and Lim (2014) propose to transform FADs

into fish enhancing devices (FEDs) because putting FADs into an already overfished area accelerates the decline of

tuna stocks. However, Davies, Mees, and Milner-Gulland (2014) recommends that FADs be actively managed

because small scale fishermen derive their livelihoods from operating these devices. Finally, Baske et al. (2012)

notes that assesment methods for FADs need to be improved in order to accurately measure the number of stocks so

that superior strategies in managing FADs can be introduced.

15

to February (Northeast monsoon (NEM) or Amihan). The fishing months in Mindanao usually

coincides with monsoon months because it starts from February to June and from August to

October. Likewise, Aprieto (1995) revealed strong correlations between recorded peaks in catch

rates and SWM and the inter-monsoon periods. However, our technical inefficiency model

suggests that vessels operating in the first quarter of 2012 have an inverse relationship with

technical efficiency. The technical efficiency score for the first quarter is 0.79 and is the lowest

among the quarter periods (Table 6). From the total output sample, 27% (21 t) of the catch was

unloaded in the first quarter while the bulk were distributed among the remaining quarters. The

average fuel consumption and number of effort days did not vary across quarters. Similar trend

can be observed in the output levels of South Cotabato province. The total volume of tuna

landings in the province have been declining during the 1st quarter. There was a 9% decline from

the 4th quarter of 2011 to the 1st quarter of 2012 and another 11% decrease from the 1st quarters

of both years. This may provide evidence that during this period, tuna catch levels were low.

Fishing Grounds

The major tuna fishing grounds in Mindanao are found in Moro Gulf and those extending

in the North of Celebes sea (Aprieto, 1995). Because local stocks continue to decline

significantly, Philippine tuna fishing vessels started expanding fishing operations into Indonesia

particularly in the North Sulawesi waters (Yamashita, 2008). In this paper, we considered five

fishing grounds namely Moro Gulf, Sulawesi (Celebes Sea), Sarangani, Pacific Ocean and

Kalamansig. The bulk (95%) of the sampled vessels’ total annual catch were harvested in the

Moro Gulf while 1.7% were caught in Kalamansig, an adjacent fishing ground from Moro Gulf.

However, no statistically significant relationship existed between fishing grounds and the sample

fishing vessels’ technical efficiency.

16

CONCLUSIONS AND RECOMMENDATIONS

The annual tuna catches from the Philippines have been declining while tuna imports

continue to increase. With this trend, there is a need to improve the efficiency of tuna vessels in

the Philippines. In this study, we utilized the technical efficiency approach by applying the

Stochastic Frontier Analysis. From the hypotheses tests and parameter values, we used a Cobb-

Douglas production function in specifying the technical inefficiency model. With an average

technical efficiency score of 0.79, our findings suggests that the efficiency levels of fishing

vessels operating in General Santos City can still be improved by approximately 21%. The likely

source of improvement may emanate from increasing the berthing days because this will shorten

effort days at sea. This implies that improvements in efficiency maybe gained if the effort days

of the vessels were restricted. In terms of vessel type, the focus of the improvement should target

the operations of purse seiners and ring netters which unload in harbor and market 3. Also, we

found that handline fishing vessels were more technically efficient than the other two types of

fishing vessels. Moreover, vessels operating in the first quarter of 2012 exhibited a negative

relationship with technical efficiency.

Also, we recommend that all other primary inputs that affect the technical efficiency of

tuna fishing vessels be validated through the use of primary data in surveying specific vessel

types (i.e. handline, ringnet or purse seine). This approach may provide more specific solutions

for each line of tuna fishery. Furthermore, we recommend that an ecological efficiency analysis

that considers the undesirable output of juvenile and undersize tuna catch as the dependent

variable be conducted in order to address the technical and ecological efficiency considerations

of the vessels. This may provide more policy options which the tuna industry can use in order to

sustain its catching operations.

17

ACKNOWLEDGEMENT

This research was funded under the Higher Education Regional Research Center

(HERRC) - XI hosted by the University of the Philippines Mindanao and supported by the

Philippine Commission on Higher Education (CHED).

18

REFERENCES

Aigner, D.J., C.A.K. Lovell and P. Schmidt. 1977. “Formulation and Estimation of Stochastic

Frontier Production Models.” Journal of Econometrics. 6:21- 37.

Aprieto, V. L. 1995. “Philippine Tuna Fisheries: Yellowfin and Skipjack”. Quezon City: The

University of the Philippines Press.

Bakhsh, K. 2007. “An Analysis of Technical Efficiency and Profitability of Growing Potato,

Carrot, Radish and Bitter Gourd: A Case Study of Pakistani Punjab.” University of

Agriculture, Department of Environmental and Resources Economics, Faisalabas.

http://prr.hec.gov.pk

Baske, A., J. Gibbon, J. Benn, and A. Nickson. 2012. “Estimating the use of fish aggregation

devices (FADs) around the globe.” Discussion Paper. The Pew Environment Group.

Washington, D.C.

Battese, G, E and T.,Coelli. 1995. “A model for technical inefficiency effects in a stochastic

frontier function for panel data.” Empirical Economics, 20: 325 – 332.

Bureau of Agricultural Statistics. 2013.” Commercial and Marine Municipal Fisheries: Volume of

Production by Species, by Region and by Province, 2002-2013; Fishery: Supply Utilization

Accounts, 1990-2012.” Available at <www.bas.gov.ph>

Bureau of Fisheries and Aquatic Resources – National Fisheries Research and Development

Institute. 2012. Available at <www.nfrdi.da.gov.ph>

Cabral, R.B., P.M. Aliño, and M. Lim. 2014. “Modelling the impacts of fish aggregating devices

(FADs) and fish enhancing devices (FEDs) and their implications for managing small-scale

fishery.” ICES Journal of Marine Science. doi: 10.1093/icesjms/fst229.

19

Coelli T.J. 1996. “A Guide to FRONTIER Version 4.1: A computer program for Stochastic

Frontier Production and Cost Function Estimation.” Center for Efficiency and Productivity

(CEPA) working papers, University of New England, Australia.

Coelli,T, Rao, D, O’Donnell, C & Battese, G.2005. An Introduction to Efficiency and

Productivity Analysis, Springer,USA.

Davies, T.K., C.C. Mees, and E. Milner-Gulland. 2014. “The past, present and future use of

drifting fish aggregating devices (FADs) in the Indian Ocean.” Marine Policy 45:163–170.

Esmaeili, A. 2006. “Technical efficiency analysis for the Iranian fishery in the Persian Gulf.”

ICES Journal of Marine Science, 63: 1759 – 1764.

Food and Agriculture Organization of the United Nations. 2011. “Measuring and assessing

capacity in Fisheries.” FAO Fisheries technical paper. www.fao.org

Food and Agriculture Organization of the United Nations, FAOSTAT database. 2013. Available

at http://faostat.fao.org/site/362/DesktopDefault.aspx?PageID=362.

Herrero, I. 2005. “Different Approaches to Efficiency Analysis: An Application to the Spanish

Trawl Fleet Operating in Morrocan Waters.” European Journal of Operations Research,

167(1): 257 – 271.

del Hoyo, J. J. G., Espino, D., and Jiménez Toribio, R. 2004. “Determination of technical

efficiency of fisheries by stochastic frontier models: a case on the Gulf of Cádiz (Spain).”

ICES Journal of Marine Science, 61: 416 – 421.

International Seafood Sustainability Foundation. 2012. “ISSF Stock Status Ratings – 2012: Status

of the World Fisheries for Tuna.” ISSF Technical Report 2012 – 04A. International Seafood

Sustainability Foundation, Washington, DC, USA.

20

International Seafood Sustainability Foundation. 2013. “ISSF Stock Status Ratings – 2013: Status

of the World Fisheries for Tuna.” ISSF Technical Report 2013 – 04. International Seafood

Sustainability Foundation, Washington, DC, USA.

Jamnia, A.K., S.M. Mazloumzadeh, and A. Keikha. 2015. “Estimate the Technical Efficiency of

Fishing Vessels Operating in Chabahar region, Southern Iran.” Journal of the Saudi Society of

Agricultural Sciences 14(1): 26–32.

Kodde, D. A. and Palm, F. C. 1986. “Wald Criteria for Jointly Testing Equality and Inequality

Restrictions.” Econometrica, volume 54 (5): 1243-48.

Meusen, W. and J. van den Broeck. 1977. “Efficiency Estimation from Cobb-Douglas Production

Functions With Composed Error.” International Economic Review. 18:435-444.

New, R. 2012. “Management changes and vessel-level technical efficiency in the Eastern Tuna

and Billfish Fishery: a stochastic frontier analysis.” Australian Government: Department of

Agriculture, Fisheries and Forestry. ABARES.

Reid, C., Kirkley, J.E., Squires, D. and Ye, J. 2005. “An analysis of the fishing capacity of the

global tuna purse seine fleet,” in W.H. Bayliff, J.I, de Leiva Moreno, and J. Majkowski (eds).

Management of tuna fishing capacity: conservation and socio-economics. Proceedings of the

Second Meeting of the Technical Advisory Committee of the FAO Project. Madrid (Spain),

15-18 March 2004, Fisheries Proceedings No. 2. Rome.

Sharma, K. R. and Leung, P. 1999. “Technical Efficiency of the Longline Fishery in Hawaii: An

Application of a Stochastic Production Frontier.” Marine Resource Economics, 13: 259 –

274.

Solis, D., J., del Corral, L., Perruso and J., Agar. 2014. “Individual Fishing Quotas and Fishing

Capacity in US Gulf of Mexico red snapper fishery.” Agricultural and Resource Economics,

58: 1 – 20.

21

Taru, V.B., Lawal, H. and Tizhe, L. 2011. “Technical efficiency of sloe cowpea production in

Adamawa State, Nigeria: A Cobb-Douglas stochastic frontier function.” Journal of

Economics and International Finance, 3(8): 504-507.

Vera, C., and Z. Hipolito. 2006. “The Philippines tuna industry: A profile.” SAMUNDRA

Monograph. International Collective in Support of Fishworkers.

Yamashita, H. 2008. “The Value Chain for Philippine Tuna Commodity: Recent Developments

and Future Directions.” IIFET 2008 Vietnam Proceedings.

Zibaei, M. 2012. Technical Efficiency Analysis of Fisheries: Towards an Optimal Fleet Capacity.

Sustainable Agriculture Research, 1(1): 96–102.

22

Table 1. Description of output, inputs and technical efficiency variables

Variable

Name Description

Mean Std. Dev. Minimum Maximum

Output

Output Total volume in t 82.04 89.72 1.20 370.00

Inputs

GRT Gross register tonnage (t) 39.18 43.75 5.00 141.48

Eff Effort in days (days) 160.44 91.01 7.00 335.00

Fuel Fuel consumption (L) 10700.40 10328.80 150.00 50600.00

Water Water consumption (m3) 122.85 130.97 5.00 486.00

Determinants of technical inefficiency

Berthing Berthing days 36.27 42.97 1.00 177.00

Moro Fishing ground: Moro Gulf 0.93 0.18 0.00 1.00

Sulawesi Fishing Ground: Sulawesi 0.05 0.18 0.00 1.00

Sarangani Fishing Ground: Sarangani 0.01 0.04 0.00 0.27

Pacific Fishing Ground: Pacific Ocean 0.01 0.05 0.00 0.22

Kalamansig Fishing Ground: Kalamansig 0.00 0.01 0.00 0.04

H&M1 Harbor and Market 1 0.53 0.46 0.00 1.00



Q1 Fishing period: 1st quarter of 2012 0.22 0.20 0.00 1.00

Q2 Fishing period: 2nd quarter of 2012 0.25 0.16 0.00 0.63

Q3 Fishing period: 3rd quarter of 2012 0.20 0.17 0.00 0.61

Q4 Fishing period: 4th quarter of 2012 0.34 0.28 0.00 1.00

23

Table 2. Test of Hypotheses

Log Likelihood

Hypothesis Null

(H0)

Alternative

(H1)

Likelihood

Ratio

statistic

Degrees of

freedom

(alpha)

Critical

Value Decision

H0: δ0 = 0

Cobb-Douglas -18.97 -18.86 0.21 1 (0.01) 5.41 Fail to reject Ho

Translog -15.19 -15.50 -0.62 1 (0.01) 5.41 Fail to reject Ho

H0: ɣ = 0

Cobb-Douglas -34.26 -18.97 30.58 13 (0.01) 27.03 Ho rejected

Translog -23.52 -15.19 16.67 13 (0.01) 27.03 Fail to reject Ho

H0: β5 = ... = β14 = 0

Cobb-Douglas vs. Translog -18.97 -23.52 -9.11 3 (0.01) 10.50 Fail to reject Ho

24

Table 3. Parameter Estimates of Stochastic Production Frontier and Technical Inefficiency Models

coefficient t-ratio

Stochastic production frontier

Constant -2.47 -4.86 ***

GRT 0.52 6.80 ***

Effort 0.25 1.68

Fuel 0.39 3.36 ***

Water 0.08 0.82

Technical inefficiency model

Berthing days -0.03 -2.46 **

Fishing Ground: Moro Gulf 0.27 0.39

Fishing Ground: Sulawesi 2.03 1.39

Fishing Ground: Sarangani -0.13 -0.13

Fishing Ground: Pacific Ocean -2.38 -1.36

Fishing Gound: Kalamansig -0.10 -0.10

Harbour and Market 1 -1.96 -2.48 **

Harbour and Market 2 0.37 0.52

Harbour and Market 3 1.55 2.16 **

Fishing period: 1st quarter 2.18 2.24 **

Fishing period: 2nd quarter -0.58 -0.60

Fishing period: 3rd quarter -1.00 -0.92

Fishing period: 4th quarter -0.70 -0.76

Variance parameter

sigma-squared (σ2) 0.39 2.53 **

gamma (ɣ) 0.74 7.56 ***

Ln (likelihood) -18.97

Notes: t-test (2-tail, df=28); @10% 1.70*, @5% 2.05**, @1% 2.76***

25

Table 4. Berthing days according to GRT, technical efficiency and input levels

Berthing Days Output per

trip (t) GRT

Efficiency

score

Water Fuel

Average Per tonnage Average Per tonnage

Less than 5 days 5.14 28.16 0.78 9 0.32 895 31.78

5 to 10 days 13.59 82.8 0.84 24 0.29 1,369 16.53

More than 10 days 14.05 90.12 0.93 25 0.28 1,254 13.91

26

Table 5. Efficiency score, input and output levels according to harbour

Harbour 1 Harbour 2 Harbour 3

Range Average Range Average Range Average

Efficiency score 0.57 to 0.95 0.87 0.12 to 0.95 0.81 0.17 to 0.95 0.74

Gross register tonnage (GRT) 5 to 77.29 14.17 5 to 141.5 42.2 35 to 141.5 88.5

% of catch unloaded 2 to 100 84 4 to 100 44 48 to 100 82

Annual catch unloaded (t) 2 to 202 37.5 1 to 106 33.63 23 to 370 140.25

Fuel/trip

150 to

1,827 784 240 to 1,583 870 1,000 to 2,625 1,468.00

Effort days/trip 5 to 33 16 5 to 32 13 9 to 32 18

Berthing days/trip 1 to 6 2 1 to 16 4 1 to 16 6.4

27

Table 6. Technical efficiency score, input and output levels according to fishing period

1st Quarter 2nd Quarter 3rd Quarter 4th Quarter

Efficiency score 0.79 0.81 0.83 0.83

Gross register tonnage (GRT) 41 39 42.4 41

% of catch unloaded 27 29 27 38

Catch unloaded (t) 21 27.7 27.7 24

Fuel/trip (liter) 998 955 1,043 1,008

Effort days/trip 16.98 16.92 17.24 17.53

28

Fig. 1 Frequency distribution of tehcnical efficiency scores for the tuna fishing vessels in General Santos City

9.8

4.9 2.4

12.2

29.3

41.5

-

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

<0.5 0.5 - 0.6 0.6 - 0.7 0.7 - 0.8 0.8-0.9 >0.9

Fre

quen

cy (

%)

Range of Technical Efficiency Scores

Factors Affecting the Efficiency of Tuna Fishing Vessels: Implications on Tuna Production and Improvement

Documents