It is much easier to dematerialise non-negotiable shipping documents, such as waybills, than it is to dematerialise the negotiable bill of lading, since it is necessary to send only information by a computerised system, rather than proof of title. Indeed, various such attempts have been made. A three month trial scheme was operated as a result of the D.I.S.H. project (Data Interchange for Shipping), set up in 1985 by P. & O. Containers Ltd. and a number of other large carriers and exporters. P. & O. say that: ``As well as reducing paper work, other benefits of the E.D.I. system include a speed up in communications, improved overall efficiency and reduced administration costs. It also eliminates double handling of information and all risk of documentary transcription error.'' P. & O. also say that developments are continuing under a newly-formed association. Another example is the Atlantic Container Line's Cargo Key Receipt Scheme, where a computer print-out in effect takes the place of a conventional waybill, called a DataLading, and also under the Cargo Key Receipt Scheme, where under a NODISP clause the consignee has the power to dispose of the cargo to one other person, the necessary information as to the identity of the carrier is sent to the carrier electronically.
There is no doubt that dematerialisation of waybills allows all the advantages of paperless electronic trading:
``It is widely expected that the impact of computerization will be as great as that of the industrial revolution. Computers are already providing all sorts of services at rising speed and diminishing costs. International trade data communication, however, seems to be the missing link. Yet the need is great. Not only do paper documentation and procedures represent as much as 10 per cent of goods value; they are slow, insecure, complicated and growing. The possibilities of cost reduction are in the order of 50 per cent, to the benefit of not only the main parties, but everyone involved, not least the authorities.''
``Any company that can reduce the time and money spent on costly trade documentation will have an immediate competitive advantage. The cost saving goes well beyond more efficient processing of orders and faster invoicing, to strategic priority areas and setting up just-in-time delivery systems.''
The problem is, of course, that replacement of bills of lading with non-negotiable documentation is not always either appropriate or possible, for example with carriage of oil where cargoes are typically sold and re-sold many times on the voyage. Yet it is also clear that at the moment the bill of lading has lost almost the whole of its security role in oil transactions. Bills of lading are not in practice used to obtain delivery of cargo, shipowners usually acting on the instructions of time charterers and relying on indemnities should delivery be made to the wrong person. Banks advancing funds on documentary credits obtain no real security, and indeed the documentary credit seems to have been largely replaced by the stand-by letter of credit in this type of transaction. Yet if electronic documentation could be made to work here, then these problems could all be resolved. I would suggest, then, that while replacing the bill of lading with the waybill and dematerialising that seems an obvious and simple solution, in reality not only does it fail to address many of the problems of present-day international trade, but it also fails to develop anything like the full potential of paperless systems.
It is of course possible partially to dematerialise and to keep paper documentation for some purposes. Yet this seems pointless once the reality is grasped that there is nothing that paper can do that computerised documentation cannot. An attempt partially to computerise documentation in the bulk oil trade was actually made in the Seadocs project (Seaborne Trade Documentation System), instigated in 1983 by Intertanko (the International Association of Independent Tanker Owners) and the Chase Manhattan Bank, which failed to get beyond an initial testing stage. Seadocs was envisaged as a compromise between traditional paper documentation and a fully electronic system. The essential idea of Seadocs was to avoid delay by lodging the bill of lading with a central registry, from which it did not move, and notifying change of ownership electronically. Delays would in theory be significantly reduced, the documents would be available for inspection, and the shipowner could always be certain that he was delivering to the correct party.
Leaving aside the numerous detailed problems with the scheme, among the reasons for the failure of the project was the monopoly position of the registry, which was perceived to be acting in its own interests rather than those of the trading parties. This may well be a problem with any closed system, although it would be less so if there were a number of registries competing for business. Another problem was the lack of enthusiasm from the P. & I. Clubs, and hence ultimately the shipowners. It is clear, however, that major problems were also caused by the retention of the paper bill of lading, albeit that it was centrally lodged. Kathy Love writes:
``A more serious problem, from the point of view of the banks, was that without physical presentation of the bill a bank would not be able to scrutinise the bill (and the documents associated with it, which could also be lodged in the Registry) for conformity with a letter of credit. In order to meet this objection the Registry, somewhat grudgingly, undertook to examine a bill, if requested to do so, to see whether it complied with particular requirements stated by the requesting party. It was hoped that this facility would also go some way towards satisfying the concerns of cargo buyers who, particularly in view of the frequency of fraudulent bills, often had careful in-house programmes for scrutinising documents.
However, this offer by the Registry was perceived to be seriously flawed ...''
The schemes described in this paper avoid these problems, since once the need for the physical document is removed, there is no reason why the proposed holder cannot examine its electronic replacement at his own place of business and at his leisure. In short, one of the major problems of Seadocs was the retention of the paper bill of lading.
Apart from a pilot project using shadow documents, Seadocs was never operated and therefore was not examined by the courts. Had it been I suggest that it would have been perceived to be fatally flawed in other respects as well, since no provision was made for the transfer of contractual rights and liabilities to holders of the bill, apart from the original shipper. The Bills of Lading Act 1855, section 1, which was then the operative legislation, would certainly not have applied in the absence of physical indorsement and transfer, and it is by no means certain that the common law would have recognised the static bill as a document of title. It is by no means clear, for example, who if anyone could have sued the shipowner had he delivered to the wrong person, or what protection a bank as holder under a documentary credit might have had in the event of the bankruptcy of the buyer. In any system of electronic documentation the parties themselves must make provision for these eventualities, and cannot rely on legal mechanisms already being in place once they move away from traditional documentation.
By comparison with the non-negotiable waybill, it is much more difficult to replace the bill of lading by electronic documentation. Whereas waybills contain only information, the possession of a bill of lading also connotes proof of entitlement to the goods. A bill of lading can after all be a document of title to goods, and it is essential that any electronic replacement should also offer the advantages of a document of title. The bill of lading also transfers contractual rights and liabilities under the Carriage of Goods by Sea Act 1992, and again, any electronic equivalent must be capable of doing the same.
The law might seem immediately to pose insurmountable difficulties in this regard. It seems most unlikely that the common law would recognise as a document of title something which has no physical presence (except insofar its presence and content is recorded electro-magnetically). The common law also requires a document of title to be signed, and although the matter has not been tested, it seems unlikely that the courts would accept electronic authentication as equivalent to signature, although it is almost certainly more difficult to forge an electronic authentication than it is to forge a handwritten signature. A further difficulty is that both bills of lading and waybills come within the Carriage of Goods by Sea Act 1992, and although provision has also been made (in section 1(5)) for the inclusion by regulation of electronic documentation, it would probably be a mistake to assume that appropriate regulations will necessarily be forthcoming.
These problems are not insurmountable, but on the assumption that the law will not help, any electronic replacement for the bill of lading should provide its own means for the transfer of contractual rights and liabilities along with the documentation itself. Any systems that are devised will also need to provide security equivalent to the document of title to goods, albeit (or so it may be assumed) without any assistance from the law.
In spite of these arguments, however, and the conservative practice of the transport industry to date, there is no reason why electronic documentation could not also be used in place of a negotiable document, where for example re-sales of the cargo at sea are envisaged. It is also necessary that it should do so. The waybill may well be an adequate substitute on container routes, where the carrier is reputable and the cargo rarely re-sold on the voyage, and (for example, with creative application of NODISP clauses) it is even possible to provide banks with some form of security where waybills are used, but it is simply not possible to use waybills in oil carriage, for example, where the cargo is re-sold many times. Yet problems caused by the failure of bills of lading to arrive in time are every bit as acute for the oil transport trade as for the container trade. There the only solution is the electronic replacement of the bill of lading itself.
We need first to identify the minimum requirements for any electronic replacement for a bill of lading:
1. The most important requirement is that the carrier needs to be informed of the identity of the ultimate receiver of the cargo, to whom (and to whom alone) he would be under an obligation to deliver, without there being any requirement for a paper document to be presented, or even to exist at all.
2. Rights and liabilities under the carriage contract must be transferred to successive owners of the cargo.
3. The electronic documentation must perform the same evidential functions as existing paper documentation.
4. The system must be secure against fraud. This is not an absolute requirement, since the present system is by no means proof against fraud, the evidence suggests that commercial men are prepared to accept a certain level of fraud in order to effect speedier and less costly transactions. The system needs only to be at least as good as the present system in this respect, but in reality, a far greater degree of security is possible.
If electronic documentation is to be truly equivalent to the present-day bill of lading, then it should also offer certain other features:
1. It should be capable of transferring property in the goods.
2. It should be open to anyone to use.
3. If it is intended to speed up the documentary process, it ought to be possible not only to transfer but also to check the documentation electronically. Using present-day levels of technology, this would require a considerable degree of standardisation of documentation.
Present-day technology allows all of these to be achieved, although some of the above objectives tend to conflict. For example, the more open a system the more difficult it is to make it secure. It will also certainly be necessary for commercial men to adjust to the technology, and to a considerable extent to standardise their procedures. Whether these sacrifices are worthwhile probably depends on the extent to which the existing documentation has become deficient in particular trades.
In 1990 the C.M.I. published its Rules for Electronic Bills of Lading, which represent an ingenious method of overcoming the problems of proving title to goods by electronic means. The essence of the system is as follows. Article 4 provides for an electronic document containing information similar to that on a paper bill of lading to be sent by the carrier to an electronic address specified by the shipper. In addition a private key is sent to the shipper to be used in subsequent transactions. The private key is known only by the shipper and the carrier. The shipper (and any subsequent holder) can transfer what the C.M.I. calls the ``Right of Control and Transfer'' to a subsequent holder (Clause 7(b)):
``(i) by notification of the current Holder to the carrier of its intention to transfer its Right of Control and Transfer to a proposed new Holder, and (ii) confirmation by the carrier of such notification message, whereupon (iii) the carrier shall transmit the information as referred to in article 4 (except for the Private Key) to the proposed new Holder, whereafter (iv) the proposed new Holder shall advise the carrier of its acceptance of the Right of Control and Transfer, whereupon (v) the carrier shall cancel the current Private Key and issue a new Private Key to the new Holder.''
Article 4(c) allows the proposed new holder to advise the carrier that he refuses to accept the transfer, and requires the carrier to assume this unless the new holder accepts within a reasonable time. In this event the carrier advises the current holder, and the current private key retains its validity. Under Article 7(a), only the holder can claim delivery of the goods from the carrier, nominate a consignee or substitute for a consignee already nominated, or transfer the right of control and transfer to anybody else.
At an initial glance the C.M.I. system looks very promising. It satisfies the most important requirement stated in the previous section, that the carrier needs to be informed of the identity of the ultimate receiver of the cargo, to whom (and to whom alone) he would be under an obligation to deliver. Because of the requirements of Article 4, the electronic message performs evidential functions similar to the traditional bill of lading, stating the name of the shipper, the description of the goods, representations and reservations as with a paper bill of lading, the date and place of receipt and/or shipment of the goods, and a reference to the terms of the carriage contract. No doubt the parties could vary these requirements if additional evidential functions were required. Furthermore, the proposed new holder gets the opportunity to inspect the electronic documentation before accepting it, and if he does not accept he does not obtain any right of control and transfer over the goods, those rights remaining in the seller just as is the case if a paper bill of lading is rejected.
The system is also (unlike all existing electronic systems) theoretically open to anyone to use. There is no need for the parties to be members of a group, and any carrier with the necessary technology can operate the system. In fact, the system requires no more than that all parties are in telephone or radio communication with the carrier, and have a computer and modem (or other access to the telecommunication network, such as a PAD).
So what, if any, are the disadvantages? First, the C.M.I. appears to make no provision for contractual rights and liabilities to be transferred along with the documentation. If the carrier refused to deliver to the eventual holder, he would certainly be in breach of contract, but only the original shipper could sue him. It would not be difficult to make appropriate provision, however, since (unlike the conventional position) each holder is inevitably in direct communication with the carrier. It would not be difficult for the carrier to offer a new contract to each holder, and for each holder to accept when the right to control and transfer is accepted. There is no need for legislation to enable this to operate smoothly.
The second problem is that it is not clear what happens if a holder who has accepted the right of control and transfer does not pay for the goods. The previous holder must not be allowed to retain any right, as against the carrier, in the goods after a transfer, since otherwise a subsequent holder who does pay would be obtaining no security of any value. This is not a problem that is unique to electronic documentation, since section 19(3) of the Sale of Goods Act was enacted specifically to deal with a similar problem with a conventional bill of lading. There can be no equivalent of section 19(3) operating in the present situation, but there is no reason why a seller should not make equivalent provision by an express term in the contract of sale, requiring the buyer to re-transfer the right of control and transfer to him in the event of non-payment, and reserving a right of disposal in the goods until paid. Just as with section 19(3), this would protect the seller in the event of the buyer's bankruptcy, but would not necessarily protect against a fraud, where for example, the defaulting buyer re-sold the goods to a third party who bought in good faith. The best method of protection for the seller in this case would be to require payment by an electronic equivalent of a documentary credit, and only to transfer the right of control and transfer to a reputable bank against payment.
A third point, which is related to the second, is that the C.M.I. rules make no provision for the passing of property in the goods. But transfer of even a conventional paper document of title will not necessarily pass property in the goods. It would be necessary for the parties to make appropriate provision in the sale contract, and in the other contracts making up the credit. This would obviously require new contracts to be drafted, but this is unavoidable in any case, since special forms of agreement will be required whenever electronic forms of transmission are used. It is of course to be hoped that industry-standard contracts can be developed, in order to avoid the necessity to engage lawyers to draft each individual transaction. One of the advantages of, for example, the c.i.f. contract, is that its meaning is well-known anywhere in the world, and a c.i.f. purchaser can re-sell c.i.f. without the need for complex negotiations and re-drafting.
I would suggest, however, that the greatest difficulty with the C.M.I. proposals is that they seem to be relatively insecure against fraud. It is undoubtedly possible to modify the system so as to make it as secure as is wished, and far more secure than any paper system, but it is difficult to see how this can be achieved without imposing additional levels of bureaucracy, possibly to the extent of militating against a truly open system. It may well be impossible to provide a system which is both completely open and completely secure, although either one or the other ought to be possible. No doubt different compromises will be reached between these conflicting objectives in differing trades.
The essential problem with the C.M.I. model is that it assumes transmission of secret codes between ship and shore. The C.M.I. calls the secret code the private key, but I prefer to use ``secret code'' for reasons which will become apparent in a moment. It is difficult to see how this transmission can be other than by microwave or other form of radio communication, to which anybody can in principle listen in. Furthermore, any fraudster who by listening obtains the secret code thereby obtains directly the key to the goods. He can now pretend to be the genuine holder and re-sell or pledge the goods to an innocent third party. No doubt the fraud would be discovered when the ship discharged, but by then the fraudster would presumably have banked the money and departed the scene. If the C.M.I. model is to be secure, it is essential that the secret code is encrypted.
There are no problems about transmission of the secret code to the original shipper, or of receiving the secret code from the ultimate receiver. There is also no problem about the first stage of the first transaction, since only the shipper knows the secret code, and even if it is intercepted during transmission to the carrier, this will be of no use to a fraudster, since the code is used only once. The problems arise with transmissions from the carrier to subsequent holders. If the fraudster intercepts a ship-to-shore transmission of the code, he can in principle use this information to impersonate the holder and steal the goods.
The problem with encrypting the code before transmission is that the trader or carrier (but not the fraudster) needs to be able to decrypt it. If all the trading parties were known in advance this would not be a problem, since the encryption algorithm could be agreed between them privately, and nobody else would be able to encrypt or decrypt the secret code. However, the C.M.I. model assumes that the goods can be sold by anybody to anybody. In an open system the general encryption algorithm would have to be public, but to avoid anybody breaking in, an encryption key would be used which would be unique to the transaction. The idea of using a key is that only someone who receives both the encrypted message and the key can decrypt the message.
The problem with any key system of this type is that the key has to be notified to each recipient, and since this must itself be done over an insecure channel, this lets the fraudster in, since with the key he can obtain the secret code, and hence deal with the goods. A reasonable degree of security may be possible by notifying encryption keys sufficiently in advance of transmission of the secret codes, but this introduces a delay, in a system whose main purpose is after all to minimise delay, and fraudsters can get to work on decryption as soon as the code is sent out by the carrier. They do not need to await transmission by the trader. Alternatively the key could itself be encrypted, but then the key for this would need to be transmitted, and eventually a key would have to be transmitted in the clear.
The best solution to the problem is to use a public key/private key encryption system, which can in principle be kept entirely secure. We have to make four further assumptions in order to operate the C.M.I. model using public key/private key encryption:
1. Each party has a unique public key which is known to everyone.
2. Each party has a unique private key which is kept secret, and never disclosed or transmitted.
3. The encryption method (or algorithm) is publicly available, and universally adopted in the trade. The system can only work if everybody uses the same encryption algorithm.
4. Any message sent from A to B encrypted with A's private key and B's public key can only be decrypted using A's public key and B's private key. This is not reversible. Thus, only B should be able to decrypt the message, and B can also tell that the message must have come from A.
The system is modified on the basis of these assumptions to work as follows:
1. Each trader knows everybody else's public key, because it is public.
2. The carrier gives the shipper the secret code as before. There is no need for this communication to be encrypted, since there is no need to use an insecure channel of communication.
3. The shipper performs a transaction by returning the secret code (and the identity of the transferee) to the carrier, the secret code being encrypted using his private key and the carrier's public key. The carrier, and only the carrier, can decrypt this using his private key and the shipper's public key.
4. The carrier sends the electronic bill of lading and a new secret code to the transferee, the secret code being encrypted using his private key and the transferee's public key. The transferee, and only the transferee, can decrypt this using his private key and the carrier's public key, and he also knows that the transmission must have come from the carrier.
5. Further transactions are performed in the same way.
6. The ultimate receiver uses the secret code as before to obtain the goods. No encryption is necessary at this stage.
It is essential to this system that the private keys remain secret and never need to be transmitted, so that the fraudster never gets the opportunity to discover any private key, and therefore cannot encrypt or decrypt any secret codes. Obviously, for encryption and decryption to work between the legitimate parties, there must be a relationship between the public and private keys. However, whereas it should be easy to calculate the public key from the private key, the reverse process should be infeasible (e.g., a computer could do it, but it would take billions of years).
The C.M.I. system can therefore be made secure, since neither the secret code nor any encryption key needs to be sent over insecure channels in plaintext. Nor do private keys ever need to be disclosed or transmitted. Only if the fraudster is able to obtain a private key does it break down. However, the very idea of a public key suggests a central authority at a minimum to keep a directory of keys. For maximum security, each trader's public and private keys need to be changed regularly, since if the same keys are used over and over again the fraudster (who we must assume will have a bank of very fast computers) has an increased chance of discovering a private key by cryptanalysis or some other method (e.g., bribing an employee). If new keys are frequently to be generated and (in the case of public keys) circulated, then the central authority has a far greater role to play. The more secure the system, the greater the role of the central authority, until we no longer have a truly open system. Of course, if a certain level of fraud can be tolerated by the trade (and it is by no means entirely absent in the present system), then the role of the central authority can be reduced, and more open systems become possible.
Any type of electronic communication poses problems that are not posed by ordinary paper transactions, and Article 3 of the C.M.I. rules provide rules of procedure to deal with this. It is particularly important to adopt rigorous procedures in case of dispute, since it may be intended to use the electronic record as evidence. One problem, for example, is the fleeting nature of an electronic record:
``Because of its physical characteristics, the traditional paper document is accepted as evidence. It is durable, and changes or additions will normally be clearly visible. The electronic document is quite different. It takes the form of a magnetic medium whose data content can be changed at any time. Changes or additions will not appear as such.''
``Computer records are subject to accidental corruption, a phenomenon almost unknown to paper-based records, and are intrinsically more subject to deliberate alteration than are paper-based records.''
Even for them to be acceptable as evidence between the parties to the transaction, therefore, some method must be found on ensuring that all changes appear as such, and can be authenticated. Standards need to be agreed between the parties to ensure that the information sent is accurate, that it reaches the recipient, and that any changes are recorded as changes.
Article 3(a) of the C.M.I. rules provide for U.N.C.I.D. to govern the conduct between the parties. U.N.C.I.D. is intended to provide the procedures necessary for the accurate transmission of electronic data. For example, Article 10 makes provision for the keeping of a ``Trade Data Log'', which is intended to ensure that all changes are properly recorded.
What is actually transmitted is a string of 0s and 1s, which are meaningless unless the parties agree on the meaning of the string. Article 3(a) of the C.M.I. rules accordingly provide for communications to conform with the relevant UN/EDIFACT standards, unless the parties choose some other protocol acceptable to all the users. A number of different protocols exist, and it is obviously important that the parties agree in advance as to which they intend to operate.
There is a further difference in principle between the C.M.I. electronic bill of lading and a conventional bill of lading. A conventional bill of lading passes from trader to trader, retaining its identity as a single document, not returning to the carrier until the goods are discharged. By contrast, the C.M.I. electronic bill of lading returns to the carrier every time it is negotiated, and effectively each successive trader is issued with a new document transmitted from the ship. This point is of interest only in that section 1(5) of the Carriage of Goods by Sea Act 1992 clearly pre-supposes the conventional, rather than the C.M.I. model.
The C.M.I. model assumes reliable communication between ship and shore, over channels that are essentially open and insecure. In principle, the C.M.I. model is an open system, but this is less clearly the case if it is to be operated securely.
If one were prepared to operate a closed system, to which the parties had to register, it would be possible to cut out altogether communication between ship and shore, and to confine transmission to channels which are more secure. It is, after all, unnecessary for the carrier to know the identity of every trader in a chain, so long as he knows to whom he is supposed eventually to discharge the goods. So long as this were known at the port of discharge, there would be no need to communicate with the ship at all. For this system to operate effectively, we need to make the following assumptions:
1. All trading parties are registered users of a mainframe computer, maintained by a central registry, which they can access from their places of business. The carrier is also a registered user, and can access the mainframe from a terminal at the port of discharge.
2. All transactions are recorded centrally on the mainframe computer. The computer can also store all the details that are typically entered on to bills of lading, including (if desired) the terms of the carriage contract.
3. The carrier can access the identity of ultimate receiver at port of discharge, but not necessarily in contact during voyage.
In principle, transactions of this type can be made very secure, since all we have in effect is a single multi-user computer network. Transactions are recorded within the mainframe computer itself, so the opportunities for a fraudster to listen in are considerably curtailed. Each carrier and trader have a unique user_id (user identity), and a secret password (which they can and should change frequently).
The carrier would inform the system of the user_id of the shipper. The shipper would be able to effect a sale or pledge by using his user_id and password to access the system, and notifying the system of the user_id of the buyer or bank in whose favour the sale or pledge is to be made. Subsequent transactions would be performed in a similar manner, the current holder alone having the right to deal with the goods on the computer. The mainframe would in effect be maintaining a database containing the details of each consignment and the current ``owner''. It would be programmed also to control access to data. There is no reason why access to documents could not be given to buyer or bank for purposes of inspection well in advance of actual transfer, nor would the data necessarily have to be stored in any particular format, as long as they could be displayed or printed by the computer.
As there are no longer any open channels of communication, the only serious risk of fraud would be if a fraudster discovered a trader's password, in which case he could imitate a trader and effect unauthorised transactions. A great deal has been written on effective security of passwords, and it would obviously be important to follow best practice. Passwords should also frequently be changed, in case a fraudster was able to discover one by (e.g.) bribing an employee, or intercepting the communication channels into the system from outside.
Obviously, there are advantages in terms of security over the C.M.I. model, and it is easier to provide inspection facilities for documentation. However, the system is not open, but depends on membership by all parties of a group. It is therefore less flexible, and there is the additional problem that a central registry may take advantage of its monopoly power. It is also worth noting that the carrier is no longer directly involved with every transaction, so that the method of transfer of contractual rights and liabilities suggested for the C.M.I. model would not work here. Since membership of a group is essential, however, it would be possible to draft the rules of membership in such a way as to effect the appropriate transfers.
As with the C.M.I. model, there is no transfer of an electronic document from trader to trader in the manner of the traditional bill of lading. Indeed, there is not really a transfer of any documentation at all: once the details are entered on to the mainframe computer that is where they remain. Again, this is of interest in the context of the models assumed by the Carriage of Goods by Sea Act 1992.
I would conclude that it is possible, on the basis of existing technology and under the existing legal framework, to replace bills of lading by electronic documents, which can in principle afford to the parties security at least as great as existing paper documents. Critics would no doubt object that making the models described in the previous sections workable depends heavily on the contractual provision made by the parties. New contracts of sale and carriage will have to be devised in any event, if only to agree on rules of conduct and protocols, without which any electronic data interchange is impossible. Given that new forms of contract will have to be devised in any event, it seems appropriate to ensure that they also make appropriate provision for the transfer of property, constructive possession and contractual rights and liabilities. Of course, for the system to be as open as the present system these contracts will need to be standardised, which may be considerably more difficult.
Nonetheless, it must also be observed that the existing legal framework is not ideal for electronic communication. Under existing documentary credits, payment is often effected by the negotiation of bills of exchange, and there seems little point in replacing bills of lading but not bills of exchange with an electronic equivalent. Bills of exchange are defined in section 3 of the Bills of Exchange Act 1882, which begins:
``(1) A bill of exchange is an unconditional order in writing, addressed by one person to another, signed by the person giving it, requiring the person to whom it is addressed to pay on demand or at a fixed or determinable future time a sum certain in money to or to the order of a specified person, or to bearer.
(2) An instrument which does not comply with these conditions, or which orders any act to be done in addition to the payment of money, is not a bill of exchange.
...''
It may be possible to argue that an electronic order authenticated electronically is an unconditional order in writing, signed by the person giving it, but it is by no means certain that the courts would accept that argument. It is also possible for the parties to make provision for payment without using bills of exchange, but the ideal solution would be for section 3 to be amended or clarified as necessary to allow for the possibility of electronic bills of exchange.
If rules of conduct such as U.N.C.I.D. are adopted then there is no reason why electronic documentation cannot be used as evidence in court. The Civil Evidence Act 1968 has long been regarded as poorly drafted for these purposes, however, and again an ideal solution would involve a change in the law.
I have assumed throughout that the Carriage of Goods by Sea Act 1992 will not be extended to apply to electronic documentation, and that the common law would not regard an electronic bill of lading as a document of title. Yet there is provision within the 1992 Act itself for extension to deal with electronic documentation by regulation, and it would not be difficult to amend the property sections of the Sale of Goods Act 1979, and in particular section 19, as appropriate. Of course, an electronic document could become a document of title with proof of custom, but customs take a long time to establish: legislation may well be the best solution to that particular problem.
Article III(3) of the Hague-Visby Rules allows the shipper to demand a shipped bill of lading. Obviously, if he did so then the entire system described here could break down. As with the other issues discussed in this section, this could be resolved by agreement, or alternatively the carrier could charge a prohibitive sum for a paper bill of lading, but again, a better solution would be the amendment of the Hague-Visby Rules to allow the carrier to use equivalent electronic documentation instead.
Finally, before electronic data interchange can be used worldwide, changes may be required in some states to the rules on personal data protection, admissibility of computer-generated documents in court, and the transmission of encrypted data across national boundaries.
Top of page, Back to legal home page, Back to home page, Back to international trade page.
This page was last updated on 08 May 97.
SLAPNT@cf.ac.uk. See, in general, International carriage of goods: some legal problems and possible solutions (Centre for Commercial Law Studies, 1988), Chapter 2, ``The Paperless Transfer of Transport Information and Legal Functions'', by Professor Kurt Grönfors.
. Indeed, the difficulties of replacing bills of lading with an electronic equivalent have led many to argue that bills of lading should be replaced by waybills wherever possible. An good example is the passage, set out in Thomsen and Wheble, ``Trading with EDI - The Legal Issues'', I.B.C. Financial Books (1989), at p.26, which is taken from the WP4 study, leading to Recommendation No.12 1979 on ``Measures to Facilitate Maritime Transport Documents Procedures''. WP4 is the United Nations Economic Commission for Europe (E.C.E.), Working Party No. 4, set up in 1960 for what was later named ``Trade Facilitation'': ibid., at p.11.
. The Merchants Guide (4th ed., 1987), p.11. Even P. & O., who are very much in the forefront of E.D.I. development, accept that the bill of lading is not susceptible to this approach, because it exists as a tangible document requiring surrender to obtain delivery: The Merchants Guide, 4th ed. (1987), p.32.
. For an example of such a clause, see Thomsen and Wheble, ``Trading with EDI - The Legal Issues'', I.B.C. Financial Books (1989), p.182:
``By acceptance of this waybill the shipper irrevocably renounces any right to vary the identity of the consignee of these goods during transit.''
. For a description of this scheme, see ``Electronic Banking: The Legal Implications'', edited by Professor R.M. Goode, joint publication of the Institute of Bankers and Centre for Commercial Law Studies (Queen Mary College, London, 1985), p.114.
. The passages are respectively U.N.C.I.D. Uniform Rules of Conduct for Interchange of Trade Data by Tele-transmission, I.C.C. Publication No.452 (1988), p.7, and ``Cutting costs by more-efficient shipping procedures'', Ideas in Action (Business International), 14 June 1986., p.10. Although these passages place the greatest emphasis on cost savings, the speed advantages also should not be forgotten.
. Indeed, tanker charterers sometimes demand a clause to be inserted into the charterparty requiring the shipowner to deliver without production of a bill of lading, and against only a personal indemnity. The authorities were reviewed in Kuwait Petroleum Corp. v. I. & D. Oil Carriers Ltd., The Houda [1993] 1 Lloyd's Rep. 333, where Phillips J. interpreted the employment and indemnity clause in a time charterparty, which provided that the master was to be under the orders and direction of the charterers as regards employment of the vessel, as encompassing an order to deliver without production of bills of lading. In his view, the shipowners would also have been indemnified under the same clause.
. Kathy Love, ``Seadocs: The Lessons Learned'', [1992] 2 Oil and Gas Law and Taxation Review 53, at p.55.
. The 967 pages of documentation in Bankers Trust Co. v. State Bank of India [1991] 2 Lloyd's Rep. 443 could probably not be processed any faster electronically than they were by hand.
UCP 500 & 400 Compared, I.C.C. Publication No. 511 (1993), notes at p.40 that ``a considerable degree of standardisation of formatting and message content practice has been attained by SWIFT, and more will be obtained once a uniform EDIFACT document `syntax' becomes operative''. The comments in I.C.C. Publication No. 511 (1993) are directed towards the development of internationally standardised banking practice, and something like this would obviously be necessary for the speed potential of EDI fully to be recognised.
Even without electronic checking, in chain sales where the identity of successive holders is known by the time of shipment, there is no reason why electronic copies could not be circulated for inspection purposes well in advance of the actual transfers. So long as the transfer itself is rigourously controlled, and so long as each electronic copy is properly authenticated by the carrier, perhaps using public and private key system similar to that described below, there is no obvious reason why as many copies as one wished could not be circulating simultaneously. This would effect a significant speed advantage. Obviously, the copy could not be used as proof of title, although it could (subject to the proper authentication) be used as evidence that the shipper had shipped goods of the contract description.
. This is essentially the solution suggested by Jan Ramberg in his contribution to Thomsen and Wheble, ``Trading with EDI - The Legal Issues'', I.B.C. Financial Books (1989),, pp. 193 ff.
. It does not seem to be envisaged that the terms are themselves transmitted electronically, but that they are readily available. Presumably there is no reason why they should not be transmitted electronically, at any rate in principle, although the cost of repeatedly doing so by what could well be satellite communication may rule this out as a practical proposition, at any rate for the moment.
. The mechanism of payment would probably have to be constrained, since it is likely that the law would not recognise the concept of an electronic bill of exchange. It is clear that new forms of contract will need to be drafted for electronic documentation, and when they are property issues should also be considered.
. At any rate by comparison with what is theoretically possible. It may well be that even as it stands, it is more secure than present-day paper systems. After all, signatures, and even entire bills of lading for non-existent goods, are not difficult to forge under the present system. There are those who argue that so long as we can produce a system which is as secure as the present we should be content. That attitude seems to me to be the very negation of progress. The present paper system is insecure because it is difficult to make a paper system secure. Electronic systems can be made much more secure, and that should therefore be the aim. The fact that the railway locomotive could replace the horse was not a good reason for limiting it to that role.
. For a short and straightforward description of encryption using keys, see Gordon C. Everest, ``Database Management Objectives, System Functions, & Administration'', McGraw-Hill International Editions (1986), Chapter 14.
. Only the secret code needs to be encrypted. The contents of the bill of lading should be sent in the clear, as should most certainly (if they are sent at all) the terms of the carriage contract. This is because if the fraudster can discover the plaintext by other means (and carriage contracts are usually easy to obtain) and the encrypted text it will be easier for him (knowing, like everyone else, the public keys) to discover the private keys, and hence to discover the secret code.
. It is not possible to devise one-way algorithms which are truly irreversible, but it is quite easy to devise algorithms which are quick to operate one way but not the other. A commonly-used public key/private key technique is to derive the private key from two or more very large prime numbers, and the public key from the product of those numbers. Obtaining the public key from the private key is easy (just multiply the numbers), but if the numbers are sufficiently large (e.g., numbers of 200 digits), the reverse process (finding prime factors of very large numbers) can be computationally infeasible. Exhaustive search (division of the public key by all known prime numbers up to its square root) is the only known method for very large numbers, and so long as the numbers are sufficiently large this can take even a fast computer billions of years, by which time one would hope the cargo has been legitimately discharged. Of course, mathematicians may one day discover a quick method of factorising very large numbers, in which case a different system will have to be devised. Any system will do, so long as it is not feasible to calculate the private key from the public.
. On the problems of maintaining and circulating public keys, and generating new keys in a secure fashion, see Everest, op. cit.
. The passages are respectively U.N.C.I.D. Uniform Rules of Conduct for Interchange of Trade Data by Tele-transmission, I.C.C. Publication No.452 (1988), p.8, and Eric Bergsten, ``Paperless Systems: The Legal Issues'', (1988) 3 Computer Law and Security Report 25.
. The Uniform Rules of Conduct for Interchange of Trade Data by Teletransmission, as adopted by the I.C.C. Executive Board at its 51st Session (Paris, 22 September 1987). On U.N.C.I.D. see Savage and Walden's contribution to ``EDI and the Law'', edited by Ian Walden, Blenheim Online (1989), Ch. 5.
. United Nations Rules for Electronic Data Interchange for Administration, Commerce and Transport. For a description of the standard, see John Berge's contribution to ``EDI technology'', edited by Mike Gifkins, Blenheim Online (1990), Ch. 5. UN/EDIFACT is an example of what is referred to in U.N.C.I.D. as a trade data interchange application protocol (TDI-AP).
. Indeed, buyers and banks some distance ahead in the chain could have access for inspection purposes, thereby effecting substantial savings in time.
. E.g., Everest, op. cit. On Unix systems, see Bruce H. Hunter and Karen Bradford Hunter, ``Unix Systems: Advanced Administration and Management Handbook'', Macmillan (1991), Ch. 6.
. The arguments are rehearsed in ``EDI and the Law'', op. cit., Ch.2 (by Rod Bradgate).
. Ibid.
. See generally Ian Walden and Nigel Savage: ``The Legal Problems of Paperless Transactions'' [1989] J.B.L. 102, 106-107.
. Article III(8) would not apply to an agreement of this type. Clause 10(a) of the C.M.I. rules allows the holder to demand a paper bill of lading, which seems regressive.
. See generally Ian Walden and Nigel Savage: ``The Legal Problems of Paperless Transactions'' [1989] J.B.L. 102, 108-109.
Top of page, Back to legal home page, Back to home page, Back to international trade page.
This page was last updated on 08 May 97.
SLAPNT@cf.ac.uk