The marketing of slaughter livestock is characterised by a large number of geographically dispersed sellers and a smaller number of large purchasing abattoirs. The structure of physical livestock auctions in the UK was fixed in the late nineteenth century as government imposed regulation led to the establishment of locally owned auction marts at railheads. Animals are brought to the market by farmers and bid for in the ring by representatives of abattoirs and butchers. The successful bidder then transports the animals from the auction to the abattoir. In this system animals undergo two journeys; from the farm to the auction and from the auction to the abattoir, causing the animals to lose weight. If animals are not bought, they have to be transported back to the farm and the farmer incurs greater costs. For the last sixty years the meat industry has become increasingly concentrated, with the emergence of large meat processing corporations, the decline of small independent butchers and the closing of small abattoirs due to over-capacity and the imposition of tighter regulation. As the power of the large meat companies has grown there has been a growth in the purchasing of stock direct from farms, bypassing the physical market.
The traditional physical market is inefficient because of the costs of animal transportation, the time required for participation by both buyers and sellers and the distortion of price in local markets. It was recognised in the 1960s that using interorganisational systems to link buyers and sellers might improve the efficency of marketing agricultural products(Henderson, 1984). Electronic markets potentially reduce the cost of bringing buyers and sellers together and reduce the cost of completing the transaction (Graham et al, 1994). Stigler (1961) recognised that the clear dissemination of information in electronic markets would lead to an increasing reduction in search costs and as such would have an attributable effect on prices in the market. Stigler also suggested that electronic markets would have a tendency to become monopolies because "since the cost of collection of information is approximately independent of its use, although the cost of dissemination is not, there is a strong tendency toward monopoly in the provision of information."
Bakos (1991) lists five points that characterise an electronic market:
1. costs of obtaining information are reduced and users have access to a wide community of alternative buyers and sellers at no great additional cost. This reduces monopolistic/monopsonistic power relations in vertical markets and allows buyers to locate appropriate sellers.
2. the benefits realised by individual sellers increase as more organisations join the system. i.e. network externalities.
3. there may be significant costs for users in switching between competing systems.
4. electronic markets require large capital investments and provide economies of scale in terms of there being very few costs for additional transactions until capacity constraints are reached.
5. participants may be uncertain as to the actual benefits of joining which may result in individuals delaying entry until they can see the effects on their rivals.
The existence of net benefits from electronic markets does not necessarily lead to their formation. Bakos (1991) notes that the establishment of electronic markets requires a 'major fixed investment' in terms of the systems that are developed and implemented. The costs include the cost of installing the necessary hardware and software and the costs of marketing the system to potential buyers and sellers.
An electronic markets is inherently a network technology: without links to buyers and sellers the benefits cannot be realised, the size of the benefits depends upon the size of the trading community and the community must share a common technical solution (Economides, 1996). A key stage in the formation of electronic livestock auctions is therefore the construction of a community or network of institutions and expertise which can convert its technical feasibility into commercial reality.
Farrell and Saloner (1984) identify two problems in the adoption of network technologies: excess inertia and excess momentum. Excess inertia arises where the new technology would give net benefits to users if a group adopt it but no firm will take the first move. Excess momentum arises where a group of firms adopt the new technology for fear of being left behind, but there is no overall net benefit. Farrell and Saloner argue that excess inertia and momentum may be a problem where information and reaction time lags are long or where two conflicting standards are proposed, but they claim that excess inertia can be overcome by communication and agreement between users.
In building the electronic market there are choices which must be reconciled within the community, including its institutional structure, scope of activities, geographical coverage, contractual relationships and technical specification of the system. We should therefore expect electronic markets to be socially shaped by the aims and behaviour of the actors involved. To consider the processes shaping the formation of electronic markets we shall consider the use of electronic markets in finished livestock auctioning in the UK.
The electronic trading of livestock combines advantages of the physical auction and direct purchasing routes: the transport of animals is restricted to one journey, but there is potentially a greater number of buyers participating in the market. By removing the need for buyers and sellers to travel to the auction, it becomes efficient to establish a wider market., replacing a large number of local physical auctions with a large virtual auction. The history of electronic livestock auctions in the UK is particularly interesting because a dominant system has not yet emerged. Instead there are now five competing systems, which differ in their ownership patterns but are remarkably similar in their operation.
The structures of the five systems are similar. The meat company buyer can connect to the system to download a catalogue of the lots available. At the notified time the auction starts, with the lots auctioned sequentially. The lots are offered at an initial price. A buyer bids by pressing the return key on their terminal. If no buyer bids at the initial price, the price falls in increments until a bid is made. When a bid has been made other buyers have a set period of time to bid one price increment higher. The price rises in fixed increments per kilogram deadweight until a higher bid has not been received in the time interval. If the highest bid price exceeds the farmer's reserve price the lot is sold and the buyer has around five days in which to uplift the stock. The following Java applet simulates an auction. (Requires Java-capable Netscape Navigator or equivalent. Click left mouse on image to start and press return key to bid).
After a rapid growth following the introduction of electronic livestock auctions in 1991, the volumes of cattle and sheep traded have gradually fallen for the past two years. This paper will describe the history of electronic livestock auction systems in the UK and consider the implications for electronic markets in general.
The roots of electronic livestock auctioning in the UK can be traced to Canada. In 1978 the Ontario Livestock Exchange, a livestock auction in Ontario, Canada, investigated the possibility of setting up electronic livestock auctioning. They visited two electronic marketing projects funded by the US Department of Agriculture. In one, developed at Texas A & M University for marketing cattle, the system offered buyers several lots simultaneously, whereas the other system, developed at the National Electronic Marketing Association in Virginia to market lambs, offered the lots to buyers sequentially. OLEX wanted a system which simulated the existing auction mechanism as closely as possible, so considered using the Virginia system. OLEX joined with eight other Canadian auctioneers in 1981 to use the now renamed National Electronic Marketing (NEMI) service to auction slaughter cattle for five months. After six months OLEX and another auction took over the idea when the other participants withdrew. OLEX wrote a system based on the NEMI system and started electronic auctioning in 1982 based on a DEC computer using a dial-up data service.
In 1983 OLEX bought 33% of NEMI which was running electronic auctions as a bureau service for only three users in the US. The Canadians brought in an academic to rewrite the NEMI software so that it would run on a PC in exchange for 25% of NEMI. By 1985 OLEX had their own system running on a DEC computer and a 33% share of the NEMI system running on a PC.
Aberdeen Northern Marts, a farmers' co-operative based in the North East of Scotland were the first company to introduce electronic livestock auctioning to the UK. They own a large physical auction in Aberdeenshire, but had seen since the 1960's a move away from prime stock for slaughter being sold through the physical auctions towards direct purchasing by abattoirs off the farms. In 1988 ANM studied the two main electronic auction systems in use: OLEX in Ontario, Canada, and CALM (Clarke & Jenkins,1993). The OLEX system, based on the real-time auctioning of lots in sequence most closely simulated the physical auction system. In spring 1989 ANM bought the rights for the OLEX system in the UK and an option to buy the rights for the rest of Europe. ANM held their first electronic sale in autumn 1989. ANM contacted other auctioneers and negotiated a network of 9 franchises covering the UK mainland, for the system now named Electronic Auction Systems Europe (EASE). EASE franchisees entered into three year contracts which gave them monopoly rights to operate EASE auctions in defined geographical areas.
County Auctions of Wooler were initially contacted by York Auctions at the time of EASE's establishment because York and Carlisle did not believe they could cover Northumberland. Both York and Carlisle withdrew from becoming EASE franchises, so County Auctions took on an EASE franchise for Northumberland, Durham and Cumbria. They were second to ANM in using EASE to auction cattle. County, although based close to the Scottish border, was not allowed to procure stock in Southern Scotland, as this area was covered by the Lawrie and Symington franchise in Lanark. By agreement they started procuring stock in Scotland, but paid a fee to Lawrie and Symington. County developed a specialisation in sheep sales on EASE from Southern Scotland and their franchise area.
In 1992 County investigated the possibility of developing an equivalent system to EASE which would free them from the costs of their franchise, and paid a software firm to develop a system which would simulate the EASE system. ANM were given the required six months notice of County withdrawing from EASE, and the Direct system started operation in August 1993. In developing Direct, County maintained the auction mechanism, but incrementally improved the system, changing the screen layout so that the details of the lot on offer remained on the screen, rather than scrolling round.
Rugby Auction Marts were one of the auctioneers who were close to taking on an Ease franchise in 1990. Rugby objected to the terms of the franchise, particularly the restriction to a defined area. In 1992 Rugby joined with Montrose Auction company to purchase the UK rights to the NEMI system from Canada. Montrose covered Scotland and Rugby covered England, south of Manchester. The operator of the West Country EASE franchise defected and joined APEX. The APEX system runs on PCs and does not require the hardware support of EASE, but to users the system is essentially the same. APEX started trading electronically in June 1992.
The EASE network was therefore weakened by one franchisee dropping out to use APEX and another, County Auctions, withdrawing and setting up their own system. While the APEX system was internally different to the EASE system, their common ancestry meant that to users they were very similar. The establishment of APEX was particularly upsetting to ANM, who thought that they had bought exclusive rights to the Canadian system and within two years be faced by a competitor using a similar system. ANM's dominant position with EASE was also challenged by the emergence of two further electronic livestock auctioning systems.
Stephenson and Son are auctioneers based in York. At the time of EASE's foundation they were contacted by ANM to become EASE franchisees in Northern England. Stephenson were reticent about entering into an individual franchise relationship with ANM. Stephenson formed a group of interested auctioneers and proposed that they collectively take a franchise for EASE in England. This proposal was rejected by ANM and four of the group became EASE franchisees. The remaining members proceeded to explore developing their own system. After gaining a government grant to support specifying the software and approaching other large auction firms outside EASE, they formed Beacon Auctions Limited as a consortium of eight auction companies in late 1989. The basis of Beacon was that all members have an equal stake and a say in the development of the system. Each member can procure stock from anywhere, except within 10 miles of another member.
Beacon drew up a specification for the software and asked seven firms to quote for providing the hardware and software. The Beacon system incorporated improvements on the EASE system. One change was to introduce time fields into the messages so that the system can avoid accepting a bid delayed in transmission. The software was developed externally but is maintained by Beacon and run on a system at York, again using BT GNS. The Beacon system began trading in 1992. There are now 11 Beacon member firms, each of whom use the York computer to run their sales. A consolidated sales catalogue is circulated to buyers. The sales follow each other, starting with the sales of Northern members, then moving south. Each member uses their own auctioneer, who can log on from anywhere through GNS.
In parallel to the formation of Beacon, Lysis, a computer software developer, approached auction firms outside the EASE network, offering to develop an equivalent system. This was bought by a consortium of six firms who named themselves LEAN. The LEAN system was launched in 1992 with Lysis running the system on their computer and buyers and auctioneers connecting through GNS.
Midland Mart of Banbury were Beacon members, but left in 1995 to take over the CLASS satellite auction system. Later in the year ANM and Midland Marts of Banbury merged their electronic marketing operations, establishing AGVISION. This led to the withdrawal of further auctioneers from EASE, now Agvision's electronic auction. Four of the EASE franchisees left to join LEAN.
Figures 1 and 2 show the volumes of finished cattle and sheep traded on the five electronic auction systems, plus the cattle traded on the CLASS/SLAM satellite auction.
It is seen in figure 1 that the level of cattle sales has been falling for the past two years. There is also an apparent decline in sheep sales, but this is masked by the seasonality of lamb sales. Market share is very volatile, with EASE losing ground from their position of initial strength. This decline is explained by the defection of fieldsmen, managers and franchisees to competing systems. Beacon and Direct have both established strong positions in sheep sales. LEAN has continued operating with very low sales volumes, but this is now improving due to the recent defection of franchisees from EASE.
The economic logic for electronic livestock auctions over physical auctions is persuasive: they remove the need for heavy capital investment in physical facilities by the auctioneer, transport costs should be lower and the time expended on trading will be lower for both buyers and sellers. The main additional cost is the cost of assessing the stock. Compared to direct farm purchases, the benefit for sellers is the weakening of the oligopsonistic power of the major meat processors, whereas for the meat processors it removes the need to maintain a network of peripatetic buyers. The factors which explain its slow diffusion are less tangible. There is a social value for farmers in participating in physical auctions and, not surprisingly, meat processors are not keen to relinquish the influence they have in direct farm purchases. There is also uncertainty over the accuracy of fieldsmen's stock assessments, but this is offset in part by the use of compensations for misclassification. The same uncertainty exists for meat processors buying through physical auctions or direct purchases, the only difference being that in the other routes they select the assessor whereas in electronic auctions the auction company appoints the assessor.
For auctioneers their evaluation of electronic auctions was shaped by two conflicting pressures. The growth of direct sales was eroding their physical auction sales and electronic auctions offered the prospect of halting this trend. However, if their electronic auctions displaced throughput from their physical markets, the viability of the traditional auction would be further squeezed.
When launched by ANM, the impacts of electronic livestock auctioning were difficult for auctioneers to predict. Unsurprisingly auctioneers did not want to be left out of a technology with potential to radically affect the industry. For ANM the policy of franchising the EASE system to local monopolies had the advantage of maintaining their control over the network and reduced competition amongst members of the network by removing the fear of them finding another auctioneer using EASE in their area. However, this structure left the majority of auctioneers outside the network and encouraged auctioneers to join the network for defensive reasons, believing that by joining they were preventing anyone else developing electronic auctions in their area.
The structure of local area franchises caused friction within EASE because it constrained auctioneers who wanted to expand the area in which they traded electronically. Also the wave of interest generated by ANM as they sought to enrol other auctioneers in EASE left auctioneers outside the network who wanted to trade electronically without an EASE franchise. The first of these pressures led to the formation of Direct and the latter to the formation of APEX, Beacon and LEAN. Each of these competing systems adopted a structure shaped in reaction to the EASE: collaborative in the case of Beacon and Lean, or operating through a two auctioneers in the case of APEX, and a single auctioneer in the case of Direct. Whilst the ownership patterns were different to EASE, they all operate using the model of assessment by fieldsmen and bidding in real time on deadweight pioneered in the UK by EASE.
The existence of five systems imposes additional costs on electronic livestock auctioning because each system maintains its own network of fieldsmen. For both buyers and sellers the proliferation of systems reduces the size of each market and therefore its value. The low profits made by the systems and the less sanguine assessment of the prospects for electronic auctions inhibits investment in the updating and development of the systems.
The similarity between the systems has been a major cause of the instability of the competing communities; it makes it easy for auction firms, their key marketing staff or fieldsmen to switch between systems, taking a large proportion of their trade with them. These transfers between communities are the main cause of the volatile swings in market share between the systems.
Now, with the systems paid for and most of the fieldsmen working on commission, the systems can operate at low trading volumes without incurring heavy losses. This means the pressure to rationalise the industry by merging systems is low and it is unlikely that any system will withdraw.
Bakos, J. Y. "A Strategic Analysis of Electronic Marketplaces", MIS Quarterly, September 1991, pp 295 - 310.
Clarke, R. & Jenkins, M. (1993) The Strategic Intent of On-line Trading Systems: a Case Study in National Livestock Marketing, Journal of Strategic Information Systems, 2,1 pp 57 - 76.
Economides, Nicholas, (1996), "The Economics of Networks," plenary address, EARIE conference, September 1994, Discussion Paper no. EC-94-24, Stern School of Business, N.Y.U., forthcoming, International Journal of Industrial Organization. [HTML] [Abstract] [Acrobat ~1249Kb] [Postscript ~3409Kb] [Zipped ~606Kb]
Farrell, J & Saloner, G. (1984) "Standardization, Compatibility and Innovation" Rand Journal of Economics, 16: 70-83.
Graham, I., Lobet-Maris, C. & Charles, D. R. (1994) Social and Economic Impacts of EDI (Final Report TEDIS Project C9), Http://www.ed.ac.uk/~ehja36/tedis_c9.html
Henderson, D. R. (1984) Electronic Marketing in Principle and Practice, Am. J. of Agr. Econ. , 66,5, 848 - 853.
Stigler, G. "The Economics of Information", Journal of Political Economy, Vol. 69, No. 3, 1961.