This article was downloaded by: [Chung Ang University]On: 18 June 2014, At: 00:10Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Maritime Policy & Management: Theflagship journal of internationalshipping and port researchPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tmpm20

Enforcement of law by the Port StateControl (PSC)Kevin X. Li a & Haisha Zheng aa Department of Logistics , The Hong Kong Polytechnic University ,Hung Hom, Kowloon, Hong Kong, PRCPublished online: 13 Feb 2008.

To cite this article: Kevin X. Li & Haisha Zheng (2008) Enforcement of law by the Port State Control(PSC), Maritime Policy & Management: The flagship journal of international shipping and portresearch, 35:1, 61-71, DOI: 10.1080/03088830701848912

To link to this article: http://dx.doi.org/10.1080/03088830701848912

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

MARIT. POL. MGMT., FEBRUARY 2008VOL. 35, NO. 1, 61–71

Enforcement of law by the Port State Control (PSC)

KEVIN X. LI* and HAISHA ZHENG

Department of Logistics, The Hong Kong Polytechnic University,Hung Hom, Kowloon, Hong Kong, PRC

Since the International Maritime Organization (IMO) introduced the Port StateControl (PSC) programme in 1982, it has been playing a vital role in theenforcement of safety law at sea worldwide. This paper addresses the effectivenessof PSC and the effectiveness of the methods for selecting ships to be inspectedadopted by regional PSC. Data on ship total loss (from 1973–2003) and on thePSC records (from 1994–2005) have been collected. The study reveals that theenforcement of PSC is effective in improving the safety level of maritimetransport. The methods adopted by regional PSC are compared on theireffectiveness, efficiency and stability in identifying substandard ships. Suggestionsare made on the improvement of the enforcement of PSC.

1. Introduction

The International Maritime Organization (IMO) introduced the first Port StateControl programme (PSC) in 1982. After that, PSC, which is organized by nineregional PSC, were under rapid development, covering most ports and coastlines [1].The main idea of PSC is that when a ship goes to a foreign port, the ship can beinspected by the port state to ensure that the ship meets the IMO standards. The portstate has the right to require defects to be put right, and detains ships for this purposeif necessary. Those ships which cannot comply with the IMO standards are calledsubstandard ships.

PSC is often referred as the ‘last safety net’ [4]. It is accepted that PSC is a measureto consolidate the former maritime safety net since the former net, constructed byflag states and classification societies, cannot work effectively [2, 3]. The main task ofport states is to identify substandard ships and prevent the occurrence of shippingaccidents in their water.

Methods for selecting ships to be inspected are a necessary part of theenforcement of PSC. The new trend of the methods, developed by each regionalPSC, is that the methods can detect potential substandard ships in advance toimprove inspection efficiency and save inspection cost. Port states, based on themethod, can focus their efforts on these identified potential substandard ships, soonboard inspections may not hinder the schedule of a quality ship.

In a few words, PSC has been regarded as the main measure to improve themaritime safety level and the methods for selecting ships to be inspected are seen aseffective. However, there is no evidence to support this point of view.

This paper aims to investigate the effectiveness of PSC and the effectiveness ofmethods for selecting ships to be inspected; the following questions are studied:(i) after the introduction of PSC, does PSC work effectively to improve the safety

*To whom correspondence should be addressed. e-mail: k.x.li@ polyu.edu.hk

Maritime Policy & Management ISSN 0308–8839 print/ISSN 1464–5254 online � 2008 Taylor & Francishttp://www.tandf.co.uk/journalsDOI: 10.1080/03088830701848912

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

level of maritime transport? (ii) Are the methods of selecting ships to be inspectedeffective? Does the effective level of each method indicate an increase, a decrease orproduce no change?

The paper is organized as follows: section 1 presents a brief overview of theliterature on the development of PSC to establish a context for the analysis. Section 2presents the methodology and scope of the study. Section 3 outlines more than 30years of safety records of the world fleet (from 1973–2003) and analyses theeffectiveness of PSC. In Section 4, the effectiveness of the methods of selecting shipsto be inspected (hereafter, represented by selection methods) is investigated.The three methods, adopted by the USCG, the Paris MOU and the Tokyo MOU,as the three representatives among all different regional methods, are analysed on a10-year data collection of the three regions basis. In section 5, the main findings aresummarized and some suggestions about improving selection methods are provided.

2. Method and scope

2.1. Method for maritime safety level analysisTo investigate the variation of the maritime safety level before and after theintroduction of PSC, two aspects are combined (figure 1). One aspect is related to theindicators that represent the whole maritime safety level, and the other aspect focuseson the variation of the indicators during the two periods.

Maritime accident analysis is an important part of maritime safety level analysis,which is one of the research directions of risk assessment in maritime transport [16].The standard approach to calculate the total risk can be simply stated by multiplyingseverity with the probability of occurrence of such severity [10, 14]. In practice,however, probability and severity are generally difficult to estimate in quantitativeterms [17]. Maritime accident statistics is one way to overcome these

Maritime safety levelafter PSC introduction

Maritime safety levelbefore PSC introduction

Comparison

Maritime accident statistics:

(1) Total loss number

(2) Total loss rate

Maritime accident statistics:

(1) Total loss number

(2) Total loss rate

Statistical analysis:

(1) Growth rate

(2) Mean comparison

(3) Hypothesis test

Figure 1. Framework for maritime safety level analysis.

62 K. X. Li and H. Zheng

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

difficulties [16, 17]. Studies based on maritime accident statistics can provide anoverall view of the maritime safety level.

Total loss, adopted in many research studies [5, 6, 8, 15], is to present the maritimesafety level in maritime accident statistics. Total loss is also adopted in this study. Bydefinition, total loss of a ship, as a direct result of being a marine casualty, meansthat the ship has ceased to exist ‘either by virtue of the fact that the ship isirrecoverable or has subsequently been broken up’ [7].

Total loss is measured by the involved ship number in this study, which is calledthe total loss number. The total loss can be measured by the involved number or theinvolved gross tonnage. Individual ships are the subject of maritime safetymanagement, so the number of ships involved in shipping accidents is sensible interms of social, economical and safety studies. The total loss number is calculated bycounting the ships involved in total loss. This indicator can express intuitiveinformation on the safety records. In terms of safety, it is better that the value of theindicator per year is as low as possible.

The total loss rate is a relative indicator to show the maritime safety level. Theinfluence of the number of world fleets is removed from this indicator, since thenumber of world fleets is a factor to influence the amount of the occurrence ofshipping accidents, which is measured by the total loss number in this study.Generally, the bigger the number of world fleets, the bigger the total loss number.The total loss rate can be calculated as follows:

Total loss rate ¼total loss number

number of world trading fleet

This indicator can be used to compare the safety level at different times. In termsof safety, the lower is the value of the indicator, the better is the maritime safety level.

To investigate the variation of the maritime safety level before and after theintroduction of PSC, the growth rate, mean comparison and hypothesis tests wereconducted on the two indicators, which are total loss number and total loss rate.Growth rate is to show the developmental trend of the maritime safety level. Meancomparison is to investigate the differences of the maritime safety level during thetwo periods of before and after the introduction of PSC. A hypothesis test is to verifythat the differences of the means have statistical significance.

The growth rates are calculated based on the methods proposed in the research ofLi and Wonham [5]. In this study the decrease rates are calculated, so a minus isadded in each calculation to indicate the improved maritime safety level.Accordingly, the equations of growth rate are stated as: the growth rate is equalto �ðMy2 �My1Þ=My2, in which My1 is the measurement value in the previous year,and My2 is the measurement value in the present year. An average growth rate is theaverage of the growth rates per year.

Means were calculated and compared before and after the introduction of PSC.A hypothesis is assumed that the maritime safety levels are different during thetwo periods. T-test was adopted to investigate the statement.

2.2. Method for assessment of the selection methodsThe three typical selection methods used by the USCG, the Paris MOU and theTokyo MOU are analysed in this study. The three regions are advanced in the PSCimplementation and have developed complete selection methods.

Enforcement of law by the PSC 63

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

Three concepts were used to assess the effectiveness of these selection methods

(figure 2). Effectiveness is the power of these selection methods to target substandard

ships in advance. Efficiency means that the inspected ships based on these selection

methods are highly likely to be substandard ships. Stability means that the efficiency

can remain at a fixed value at any time. An ideal selection method is an effective one

with high efficiency and high stability.Inspection number and detention number are the two main items to present the

operation of the selection methods (figure 2). The two indicators were used in former

studies [8, 9], and annual reports issued by the USCG, the Paris MOU and the

Tokyo MOU. Inspection number is the number of inspected ships selected on the

basis of these methods. Detention is a result of ship inspections. A detained ship is

regarded as a substandard ship. The detention number is the number of substandard

ships. In this study, inspection number and detention number were adopted.To investigate the effectiveness of these selection methods, inspection number and

detention number were observed. The effectiveness of the selection methods can be

judged directly, when the condition is satisfied that after checking the ships selected

according to these methods, port States can find substandard ships.To investigate the efficiency of these selection methods, a detention-inspection

rate (DIR) was used. The concept of efficiency is borrowed from economics, and

economic efficiency is a general term for the value assigned to a situation under

which a measure is designed to reduce the amount of waste. Economic efficiency is

achieved though dividing the produced output by the cost. In this study, these

selection methods were the measures for decreasing the amount of inspections

conducted on substandard ships. Inspection number can be regarded as cost and the

detention number is the output. A DIR per year can be calculated by using the

following formula:

DIR ¼ 100%�NodetentionNoinspection

ð1Þ

Selection methods

Effectiveness(observation)

Efficiency(DIR)

Stability(SD)

Assessmentoperation

Inspection Detention

Figure 2. Framework for assessing selection methods.

64 K. X. Li and H. Zheng

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

For example, the USCG in 1998 inspected 12 448 ships and detained 373 ships.Substituting these data in the above formula, a DIR of the USCG in 1998 can beobtained:

DIR ¼ 100%�373

12448� 3:00%:

The efficiency of selection methods means that the big value of DIR is better thanthe small value. The reason for this criterion is that the production of a unit of goodsor services is termed economically efficient when that unit of goods or services isproduced at the lowest possible cost.

Standard deviation (SD) was adopted for stability of selection methods. Since thevariation of DIR obtained from each selection method shows the method’s stability,the standard deviation of each selection method was used to measure the variationrange. A method with low stability has a greater standard deviation.

2.3. Study scope and source of dataThe study focuses on propelled sea-going merchant ships. Data are collected fromCasualty Return (annually, from 1973–1993) and World Fleet Statistics (annually,from 1994–2003) published by Lloyd’s Register of Shipping, Shipping StatisticsYearbook (annually, from 1973–2003) published by the Institute of ShippingEconomics and Logistics, and annual reports published by the USCG (from 1998–2005), the Paris MOU (from 1994–2005) and the Tokyo MOU (from 1994–2005).

3. Effectiveness of the PSC programs

3.1. Reduction of the total loss numberThe total loss number has a reduced trend during the period of 1973–2003, from 363in 1973 to 144 in 2003 (figure 3). The total loss number per year decreased by a rateof almost 5%, with an average total loss number of 270. In other words, the safetyrecord of the world fleet has improved as displayed by the total loss number.

Figure 3. Total loss number from 1973–2003.

Enforcement of law by the PSC 65

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

The extent of the shift differs during the two periods of 1973–1982 and 1983–2003.An interesting observation in figure 3 is that the four years with total loss numbermore than 400, happened during the period of 1973–1982. The average total lossnumber (218) during the period of 1982–2003 was lower than the value of the periodof 1973–1982 (378), with the degressive range of 160 and the decreasing rate of morethan 40%.

It is expected that there will be a significant difference in the means of the totalloss number during the two periods of 1973–1982 and 1983–2003. The t-test resultsare t¼ 7.071(0.000) under df¼ 29. These results reveal that the means of the totalloss number of the two periods have a significant difference. There was a significantdecrease in the period of 1983–2003.

3.2. Reduction of the total loss rateThe total loss rate steadily reduced during the period of 1973–2003, from 6.09ø in1973 to 1.61ø in 2003 (figure 4). The total loss rate per thousand ships per yeardecreased by a rate of 5.92ø, with an average total loss rate of 3.64ø. In otherwords, the safety record of world fleets has improved as reflected by the totalloss rate.

There are many differences in the safety records of the two periods of 1973–1982and 1983–2003. An interesting observation shown in figure 4 is that the minimumvalue of the total loss rate (4.86ø) during the period of 1973–1982 is bigger than themaximum value (4.30ø) during the period of 1983–2003. The mean total loss rate is5.55ø during the period of 1973–1982, while it is 2.73ø during the period of 1982–2003, a decreasing rate of more than 50%, which is lower than the value of the periodof 1973–1982.

A significant difference is expected in the means of the total loss rate during thetwo periods of 1973–1982 and 1983–2003. The safety record during the period of1983–2003 is expected to improve. The t-test results that compare the means of thetwo periods show that t¼ 9.292(0.000) under df¼ 29. The results reveal that themeans of the total loss rate of the two periods have a significant difference. It means

Figure 4. Total loss rate from 1973–2003.

66 K. X. Li and H. Zheng

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

that the mean of the total loss rate during the period of 1983–2003 shows asignificant decrease.

3.3. Improvement of safety recordThe improvement of the maritime safety level, shown by the manifestation of thetotal loss number and the total loss rate during the period of 1973–2003 can bemainly attributed to the enforcement of PSC to strengthen the safety net of maritimetransport.

After 1982, the major measure to strengthen the maritime safety net is to establishand spread PSC. The greater the extent of the enforcement of PSC, the more effectiveis the control of shipping accidents. The reduced occurrence of shipping accidentscannot be explained by the influence of the number of world fleets since it is oppositeto the common case that the total loss number increases when world fleets increase.

On the first question, it can be summed up that PSC works effectively toimprove the safety level of maritime transport. It was proven that there weresignificant differences in the safety record of world fleets between the periods beforeand after 1982.

4. Effectiveness of the selection methods

4.1. Effective selection methodsThe power of selection methods can be manifested by observing the inspectionnumber and the detention number (figure 5). Guided by these selection methods, in2005 the USCG inspected 10 430 ships and detained 127 ships, the Paris MOUinspected 21 302 ships and detained 994 ships, and the Tokyo MOU inspected 21 058ships and detained 1097 ships, shown in figure 5. The observation for inspection anddetention number shows that the selection methods have the power to help portstates identify substandard ships in advance.

4.2. Variant efficiency and stability of the selecting methodsDIR is calculated based on equation (1) and the results are shown in figure 6. Thethree lines, in figure 6, represent the DIR yearly changes for each selection methodadopted by the USCG, the Paris MOU and the Tokyo MOU, respectively.

The yearly DIR values of the USCG are lower than the values of either the ParisMOU or the Tokyo MOU, since the USCG’s line lies at the bottom of the figure 6.On the other hand, the USCG has the highest stability since it has the lowest SDvalue (SDUSCG¼ 0.0057). It indicates that no matter what ships enter into the UnitedStates’ water, there is a relatively fixed percentage to identify substandard ships.These results show that the selection method of the USCG is stable but the effectivelevel is low.

The selection methods adopted by the Paris MOU and the Tokyo MOU haverelatively high efficient level, since their lines lie in the upper part of the figure 6.However, the trends of the two lines are downward. The DIR of the Paris MOU wasfrom 9.5% in 1994 to 4.7% in 2005. The downward trend of the Tokyo MOU’s DIRis shown since 2003, from 8% in 2003 to 5.2% in 2005. There is a low stability ofthese two types of methods (SDParis¼ 0.0192; SDTokyo¼ 0.0123). Therefore, themethods of the Paris MOU and the Tokyo MOU have relatively high effective levels,with low stability shown in downwards trends.

Enforcement of law by the PSC 67

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

25000Inspection Number

Detention Number

20000

15000

10000

5000

1994 1995 1996 1997

USCG Paris MOU Tokyo MOU

USCG Paris MOU Tokyo

1998 1999 2000 2001 2002 2003 2004 2005

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

0

2000

18001600

1400

1200

1000

800600

400

200

0

Figure 5. Inspection number and detention number.

Figure 6. DIR of the three regional PSC.

68 K. X. Li and H. Zheng

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

Another interesting observation in figure 6 is that there is a convergence betweenthe methods of the Paris MOU and the Tokyo MOU. In 1994, for example, the DIRof the Paris MOU was 9.4%, the Tokyo MOU’s was 3.5%, and the differencebetween them was 5.9%. In 2005, the difference was reduced to 0.5% since the valuesof DIR of the Paris MOU and the Tokyo MOU were 4.7% and 5.2%, respectively.This indicates that there is some reason that causes the effectiveness of these twomethods possibly to have the same developmental trend.

4.3. Reason for variant efficiency and stabilityThese selection methods are effective with different efficient levels and stability.The method adopted by the USCG is of low efficient level and high stability.Comparatively, the methods adopted by the Paris MOU and the Tokyo MOU are ofa higher efficient level and low stability.

This may be caused by the differences among the selection methods. The primarydifference is the constructive principles behind these selection methods. The risk-controllers idea is the principle to construct the USCG’s method. The identifiedcharacteristics from shipping accident investigations construct the methods of theParis MOU and the Tokyo MOU.

The detailed manifestations are risk indicator systems (table 1) and risk leveldetermination approaches in these methods [11–13]. The second column in table 1lists the risk indicators in the three selection methods and the third, forth and fifthcolumns show the risk indicators of the three selection methods, respectively.In terms of quantity, the USCG has nine indicators, the Paris MOU has ten and theTokyo MOU has eight.

Table 1. Risk indicators comparison.

Risk dimensions Risk indicators USCG Paris MOU Tokyo MOU

Ship management (RD1) A ship owner/manager/chartererdetention ratio

� � �

Flag state (RD2) A flag state detention ratio � � �

Flag state has not ratified allconventions

� � �

Classificationsociety (RD3)

A classification societydetention ratio

� � �

Non-recognized classificationsociety

� � �

Vessel history (RD4) Detention number � � �

Other operation control number(i.e. Customs hold)

� � �

Casualty number � � �

Time since last initial inspection � � �

Deficiency � � �

Outstanding deficiencies � � �

Ship type and shipage (RD5)

Ship type and ship age � � �

Ship age � � �

Ship type � � �

Note � means that this factor is considered; � means that this factor is not considered.

Enforcement of law by the PSC 69

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

The Tokyo MOU has a relatively simple version of the risk indicators of the ParisMOU. The risk indicators of the Tokyo MOU can be included in those of the ParisMOU, except one indicator, ship type. Yet the indicator, ship type and ship age ofthe Paris MOU is a more complicated indicator than ship type. Meanwhile, the ParisMOU has more risk indicators than the Tokyo MOU.

The range of the collected information on a vessel is different between themethods adopted by the Paris MOU and the USCG. Four risk indicators includedonly in the Paris MOU are flag state that has not ratified all conventions; non-recognized classification society; outstanding deficiencies; and ship age. Three riskindicators included only in the USCG are a ship owner/manager/charterer detentionratio; other operation control number (i.e. Customs hold); and casualty number.

The USCG has a wide range of information collection. Its information involves aship owner/manager/charterer, and the Customs and historical casualty of a ship. Onthe other hand, the Paris MOU collects more detailed information than the USCG.Non-recognized classification society, for example, indicates a poor technical controlby a classification society on a ship and is often observed in the casualties.

In summary, from the aspect of the risk indicator system, the Tokyo MOU has asimpler version of the Paris MOU, and relatively detailed information is included inthe Paris MOU and a relatively wide range of information is collected by the USCG.

There are differences in the risk level determinations approach between themethod adopted by the USCG and the methods adopted by the Paris MOU andthe Tokyo MOU. Giving mark to every risk indicator and summing them up are theapproaches used by the Paris MOU and the Tokyo MOU.

The USCG’s risk level determination takes into account the whole risk level andthe individual risk dimension listed in the first column in table 1. The risk level isdetermined by equations: (1) VRDi

� C and 8i i ¼ 1, 2, 3, 4, 5; or (2) �5i¼1VRDi

� C,in which RD represents risk-controllers shown in table 1, C is the risk levelbenchmark and the value adopted by the USCG in 2007 was seven. The firstequation realizes to control the individual risk dimension. It means that if only oneof the risk dimensions involved in a ship is higher than the regulated risk level (C),the ship must be inspected. The second equation is used to control the whole risklevel under condition that each risk dimension has a relatively low risk mark. Thecase is that if the whole risk level is higher than the regulated risk level (C), the shipmust be inspected. Hence this method is sensitive to every risk dimension and thewhole risk level.

In summary, from the aspect of the risk level determination approach, the USCGhas a more complicated approach which is sensitive to every risk dimension and thewhole risk level, and the Paris MOU and the Tokyo MOU adopt a summing methodwhich takes into consideration the whole risk level.

Additionally, the convergence of the DIR lines between the Paris MOU’s and theTokyo MOU’s could be explained by the similarity of the two methods: (1) they havethe same constructive principle, i.e. identified characteristics from shipping accidentinvestigations; (2) risk indicators of the Tokyo MOU is a simpler version of the thoseof the Paris MOU; and (3) they have the same risk level determination approach, i.e.giving mark to every risk indicator and summing them up. The convergence provesan economic assumption that under the same context the similar system producessimilar efficiency after a long time running. It is sensible to predict that a methodcombining the methods of the USCG and the Paris MOU may produce higherefficiency with higher stability.

70 K. X. Li and H. Zheng

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014

5. Conclusion

The PSC programs are effective in raising the maritime safety level. The analysis of

safety records of world fleets yields evidence that the maritime safety level is

improved after the introduction of the PSC programs.The three selecting methods are effective and have quite variant efficient levels and

stability. The method adopted by the USCG has high stability and the methods used

by the Paris MOU and the Tokyo MOU have high efficient levels.These differences are resulted by the constructive principles behind these methods.

As a general rule, an ideal method should not only have a high level of efficiency but

also high stability. Therefore, a further requirement of selection methods is to possess

both high efficiency and high stability.

Acknowledgment

The study is supported in part by the research grants from the Hong KongPolytechnic University. Thanks go to Prof. John Liu, Dr Heather Leggate and Dr Jia

Yan for their valuable comments on the study.

References1. HOPPE, H., 2000, Port state control—an update on IMO’s work. IMO news, 1, 9–19.2. MIDDLETON, I., 1994, Holes in the system. Seatrade Review, January, 6–7.3. OZCAYIR, Z. O., 2004, Port State Control, 2nd edn (London: LLP).4. STREETER, B., 1999, No place to hide, Address given at Mare Forum 99, Amsterdam, the

Netherlands, 22 June.5. LI, K. X. and WONHAM, J., 1999, Who is safe and who is at risk: a study of 20-year-record

on accident total loss in different flags. Maritime Policy and Management, 26(2),137–144.

6. ALDERTON, T. and WINCHESTER, N., 2002, Flag states and safety: 1997–1999. MaritimePolicy and Management, 29(2), 151–162.

7. WORLD CASUALTY STATISTICS, Lloyd’s Register of Shipping 1996, Preface 4.8. LI, K. X., 1999, The safety and quality of open registers and a new approach for

classifying risky ships. Transportation Research Part E: Logistics and TransportationReview, 35(2), 135–143.

9. CHEN, Y. H. and WENG, J. C., 2001, Establishment Taiwan’s ship safety inspection andmanagement system (in Chinese), http://www.safetysea.org/TAMSS/document/safety/PSC/CHINESE/reference/2001-from_psc_to_us_safety/CGA—seminar011214.htm#_ftn33.

10. LI, K. X. and CULLINANE, K., 2003, An economic approach to maritime riskmanagement and safety regulation. Maritime Economics and Logistics, 5(3), 268–284.

11. USCG MARINE SAFETY MANUAL, Vol. II, Section D, Chapter 4, A.1, USCG.12. Paris MOU, Target factor, http://www.parismou.org/upload/pdf/tf.pdf13. TOKYO MOU, 2004, Asia-Pacific Port State Control Manual, section 4–3.14. KRISTIANSEN, S., 2005, Maritime Transportation Safety Management and Risk Analysis

(Oxford, Burlington, MA: Elsevier Butterworth-Heinemann).15. PRONCE, P. V., 1990, An analysis of marine total losses worldwide and for selected flags.

Marine Technology, 27(2), 114–116.16. SOARES, C. G. and TEIXEIRA, A. P., 2001, Risk assessment in maritime transportation.

Reliability Engineering and System Safety, 74, 299–309.17. GAARDER, S., RONGNSTAD, K. and OLOFSSON, M., 1997, Impact of human element in

marine risk management. In: Advances in Safety and Reliability, edited by C. GuedesSoares (Oxford: Pergamon).

Enforcement of law by the PSC 71

Dow

nloa

ded

by [

Chu

ng A

ng U

nive

rsity

] at

00:

10 1

8 Ju

ne 2

014