Competition and Single Desk Selling of Wheat From Australia: An Economic Analysis

Competition and Single Desk Selling of Wheat
from Australia: An Economic Analysis
by
Joshua S. Gans
University of Melbourne
6th July, 2000
This report was commissioned and funded by AWB (International) Limited.

However, the views expressed do not necessarily reflect those of AWB
Limited, AWB (International) Limited, or associated persons and corporations.
Responsibility for all views expressed and errors lies with the author.
Executive Summary
The Wheat Marketing Act 1989 (WMA) governs the trading regime for Australia wheat
in international markets. Its most salient feature is allow AWB (International)
Limited (hereafter AWB) to export wheat while providing that organisation with
veto power over wheat exports by other traders. However, coupled with these rights,
is an obligation on AWB to purchase all wheat offered to it that meets its receival
standard and to reward growers through pooling arrangements.
The present competition review asks whether it is possible to meet the goals of the
WMA (namely, maximising the net return to growers of Australian wheat) by
placing less restrictive conditions on its trade? From an economics viewpoint,
because AWB is wholly owned by growers, the answer to this question is intimately
tied with whether, in the absence of such restrictions, AWB would perform its
marketing and logistics functions as effectively and also whether affording growers a
choice of international trade options would improve their net returns on average.
In terms of its effect on AWB performance, the removal of single desk restrictions
would likely diminish its marketing function (i.e., its ability to negotiate high prices
internationally) and may result in a loss of economies of scale in its logistics function
that would not be restored by cost reductions from competitive pressure.
• Marketing: economic theory suggests that the prices for Australian wheat will be
high when it is sold to customers located relatively closer to Australia and also
when the wheat produced is of a higher quality. In this environment, single desk
selling can potentially result in premiums because of the stronger bargaining
position of that single seller. In addition, the single desk can assist in maintaining
brand image and assuring that contracts are honoured. However, given its
statutory purchase obligation, the single desk cannot increase prices by the
exercise of monopoly power (i.e., restricting supply). Our empirical analysis has
demonstrated that AWB has obtained price premia beyond those that can simply
be explained by timing, distance and quality issues. This supports the hypothesis
that the removal of the single desk would, on average, reduce the price received
for Australian wheat.
• Logistics: in taking wheat to market, the AWB performs many operations that
could involve the realisation of economies of scale. To the extent that the removal
of single desk restrictions would invite other traders to perform its logistics
functions, there may be a consequent loss in these scale economies and an increase
in the average o l gistics costs across the Australian wheat industry. This cost
increase could, in principle, be offset by an improvement in performance afforded
by competitive pressure. This would require, however, more efficient traders to
enter the Australian wheat industry (most likely multi-national corporations) so
that competitive pressure is real. Of course, to the extent that the Wheat Export
Authority and grower-owners can effectively monitor AWB’s performance, the
additional push of competitive pressure may be small relative to the potential loss
in scale economies.
Thus, it appears that there is economic support for the notion that the removal of
single desk restrictions would reduce rather than increase the net return to growers
who continue to trade through AWB.
On the other hand, it is often argued that giving growers more choice in their trading
options could not hurt but only expand grower returns on average. This argument,
while important, does, in its simple form, neglect the important externalities that
exist between grower decisions precisely because of the bargaining power afforded
single desk selling as well as the potential realisation of economies of scale. So while
some growers may benefit from the ability to trade outside pooling arrangements,
this will impose a negative impact on other growers. Indeed, that impact may be
sufficiently large as to prevent those growers from effectively accessing the
international market altogether. This highlights the fact that single desk restrictions –
in particular, the pooling and purchase obligations – serve not only to increase net
returns on average but also to spread those returns across growers. Hence, removal
of the single desk is likely to have distributional consequences.
Finally, the single desk restrictions do impact on other participants in the Australian
wheat industry. While the domestic market is deregulated, there is an important
linkage between domestic and international wheat prices. Indeed, a reform that
reduces international wheat prices is also likely to reduce domestic prices; to the
benefit of domestic purchasers. In addition, other potential export traders are clearly
harmed by the single desk restrictions. Finally, it is often argued that international
purchasers are harmed by the high prices afforded by single desk selling. While true,
these high prices are likely to represent distributional benefits accruing to Australian
growers rather than restrictions on international trade from the exercise of monopoly
power. That is, AWB is limited in its ability to exercise such power because of its
statutory purchase obligation.
This report concludes by suggesting that the evaluation of the single desk involves
consideration of the appropriate counterfactual; that is, what will happen in the
absence of a single desk? While there are theoretical arguments suggesting that the
net returns to growers on average may be reduced by the removal of a single desk,
what is more certain is that such a reform would alter the distribution of returns
among growers. Indeed, the environment where the removal of single desk
restrictions could improve grower returns on average also potentially means that
some growers will lose access to those international markets.
Contents Page
1 Background.............................................................................3
2 Evaluating the Single Desk..................................................7

2.1 Classifying AWB’s Functions ........................................ 8
2.2 Marketing Advantages of a Single Desk ....................... 9
2.3 Logistics Advantages of a Single Desk.........................20
2.4 Overall Effect of Non-exclusivity on AWB...................24
3 Effect on Growers ................................................................25

3.1 Allowing Grower Choice..............................................26
3.2 Summary.......................................................................28
4 Effect on Others....................................................................29
4.1 Domestic Market...........................................................29
4.2 International Purchasers...............................................30
4.3 Other Export Traders....................................................32
4.4 Overall Impact ..............................................................33
5 Conclusion.............................................................................35
6 Technical Appendix A: Theory.........................................39

6.1 Bargaining with a Single Desk......................................40
6.2 Contestable Trading......................................................41
6.3 Few Traders..................................................................42
7 Technical Appendix B: Empirical Analysis (with Joseph

Hirschberg)............................................................................44
July 2000 i
Contents Page
7.1 Data...............................................................................44
7.2 The Model .....................................................................51
7.3 Results...........................................................................54
Attachment A: Recent Research....................................................63
Attachment B: Consultant Profiles...............................................65
July 2000 ii
1 Background
In 2000, the Federal government announced a National

Competition Policy (NCP) review of the Wheat Marketing Act 1989 (WMA).
This Act establishes the regulatory framework for the wheat industry and
also specifies operating conditions and rights for the AWB Limited; the
privatised entity of the original statutory authority, the Australian Wheat
Board. For the purpose of aiding in its submission to the NCP review,
AWB Limited asked me to provide an economic analysis of competition
aspects associated with single desk selling and associated regulations.
This paper reports the findings of that economic analysis.
The WMA contains several sections that impose restrictions on the

degree of competition in various areas of the Australian wheat industry.
The most important of these are:
• Export Controls: There are several ways in which the WMA

places restriction on the free export of wheat from Australia.
First, the export of the wheat from Australia requires the consent
of the Wheat Export Authority (WEA). The WEA is a statutory
authority that overseas the operation of the WMA. AWB
International Limited (a subsiduary of AWB Ltd) has a general
exemption from requiring the consent of the WEA in order to
export wheat. Second, AWB International has a veto over any
consent that might be given to a person or firm that may wish to
export wheat in ‘bulk’ quantities.
• Receival Standards: the WMA gives AWB Ltd the right to

establish standards for wheat it receives.
The first set of provisions effectively imposes single desk selling of

Australian wheat overseas through AWB International. Note that this
does not restrict trade in the domestic market nor does it apply to non-
bulk trade through bag or by container. Nonetheless, these provisions give
AWB International power to exclude others from exporting Australian
wheat; potentially in direct competition with it. The second set of
provisions may give AWB the ability to restrict competition by raising
Section 1 Background
receival standards ‘too high.’ In each case, the potential exclusionary

nature of the provisions does not, in of itself, give the AWB an incentive to
exercise its rights in all circumstances.
It should also be noted that the WMA contains some pro-

competitive elements. The most significant of these are:
• AWB Purchase Obligation: Under s84 of the WMA, AWB

International is obligated to purchase all wheat offered to it that
meet its receival standards and to operate a pooling arrangement
for the calculating the price of that wheat. This purchase
obligation would not remain in force if AWB International lost
its exemption from applying for a license to export.
• WEA Regulatory Powers: AWB’s performance is subject to

regulatory review by the WEA.
As will be discussed later in the paper, the AWB’s purchase obligation is

pro-competitive because it prevents AWB from exercising monopsony
power when purchasing from Australian growers. The regulatory
oversight is also pro-competitive for, to the extent that it is effective, it will
improve AWB’s performance and raise the productive efficiency of the
Australian wheat industry.
Also of relevance for the economic analysis is the corporate

structure of AWB. First, the AWB is effectively a growers’ cooperative. All
operating growers in Australia are entitled to A-class shares that enable
them to elect the majority of directors to the Board of AWB Limited.
Growers were assigned B-class shares on the basis of their equity in the
Wheat Industry Fund (WIF).1 Holders of B-class shares can receive
dividend payments from AWB. In the future, these shares may be listed
on the Australian Stock Exchange (ASX). These shares are tradable.
Essentially, this means AWB is owned and controlled by growers.2
1A levy on grower farm gate production over a 9-year period enabled B-class shareholders to convert their
WIF units to shares on a one for one basis. The net asset backing of the WIF units at the time of conversion
was $2.51.
2 A-class shares are also related to production levels but the incremental power as a function of quantity
produced diminishes quickly.
July 2000 4
Second, and related to the first, is that the corporate objectives of

the AWB are tied to the interests of growers, rather than the pure
maximisation of shareholder value that arises for firms that are not
cooperatives. In particular, the AWB operates to maximise the net returns
to growers selling pool return wheat to the company and to provide
services with a view to realising that objective.3 Finally, AWB
International holds the single desk and this operates separately from the
other functions of AWB.4
A final important feature of AWB for the purpose of economic

analysis is its use of pools for determining disbursements to growers. In
the absence of pools, growers would receive a price that reflected the
particular price paid for their grain by final consumers. A pool means that
growers receive a price that reflects the average price paid by final
consumers for grain. The pool, therefore, provides an insurance
mechanism for growers against short-term price fluctuations and also
against heterogeneity among consumers of grain in terms of the value
they place on that grain and the competitive options open to them. The
AWB operates pools in an overlapping manner with different pools
opening and closing at different times of the year. Therefore, due to
harvest cycles there is a regional element to pooling. Moreover, AWB
distinguish between different qualities of grain supplied to the pool; so
that growers are rewarded from grain quality. Essentially, by selling grain
to a pool, the grower is buying shares in a fund the return of which is the
overall revenue generated by all of the wheat in that pool less common
costs. In this respect, marketing costs are shared among growers.
This report will focus on the ‘single desk’ aspects of the above
legislation. However, as will be elaborated upon below, these cannot be
reviewed without consideration of the restrictions imposed by the other
key provisions of the WMA. Hence, this paper will also demonstrate how
each of these impacts on the benefits and costs of the single desk
arrangements.
3 Constitution of AWB Limited, Article 2.
4 In what follows, for convenience, I will use AWB and AWB International interchangeably unless, of
course, the distinction is critical for a particular issue.
July 2000 5
The paper proceeds as follows. First, the single desk itself is

evaluated. In particular, the functions of AWB Limited are identified and
classified as marketing and logistics categories respectively. From that
point the economic theory regarding the impact of single desk selling on
AWB’s performance of these functions is described and evaluated. That
section also provides a discussion of our economic analysis of actual price
premiums paid to AWB. Section 3 then turns to consider the overall
impact on growers that would arise if single desk selling were removed. It
is argued there that – aside from a potential loss in marketing and logistics
benefits that arise from single desk selling – there will likely be a
distributional impact on growers from removing single desk restrictions.
In particular, some growers may find themselves without cost-effective
access to international markets. Finally, the impact of the removal of
single desk selling on other key participants in the wheat industry is
considered. This includes the impact on the domestic market,
international purchasers and other export traders. It is found there that
other export traders (in particular, multi-national corporations) would
benefit from the removal of single desk selling, as would international
purchasers (although this would be a distributional rather than an
efficiency effect). The effect on domestic wheat prices would depend upon
whether the net returns to growers fell or rose as a result of the change. If
they were to fall, then domestic wheat prices would also likely fall;
otherwise domestic wheat prices could rise.
July 2000 6
Section 2 Evaluating the Single Desk
2 Evaluating the Single Desk
The objective of the WMA is to “maximise the net returns to

Australian wheat growers” who are providing wheat for export. An NCP
review asks whether such legislative objectives can be achieved without
any specified restrictions on competition that might exist in the legislation.
It is with this in mind that this economic analysis takes place.
Nonetheless, it is worth emphasising that broader issues of

economic welfare and efficiency can also be of relevance. In particular,
restrictions on competition, while of benefit to a particular group (e.g.,
Australian wheat growers), may not be in the interests of all people and
businesses, in both Australia and internationally. Hence, it will also be of
relevance, throughout this paper, to identify those persons that might
potentially be harmed by the competitive restrictions in the WMA.
However, throughout this report, I will frame the analysis in terms of how
a particular legislative provision impacts upon the returns to Australian
wheat growers.
The key question for analysis, therefore, is: if other traders were
permitted to export bulk Australian wheat, what would happen to the net returns
Australian wheat growers received? The answer to this question depends
upon:
• What determines the net returns to growers in the current

environment?
• What will change if the single desk provisions were removed?
Answering these questions involves describing the AWB’s current role, in

particular, what services it provides to growers and, also, how that role
would change if its exclusive export right were removed. It is the latter set
of questions that are most complex because of the potentially large
number of alternative scenarios that might evolve. For example, the
removal of a single desk could lead to a large number of export traders
selling Australian wheat overseas or, alternatively, the removal of the
provisions may have a limited effect on the AWB’s current role as it has
July 2000 7
intrinsic advantages in marketing Australian wheat that cannot easily be

replicated by others.
Given this, in what follows, I will consider first the impact of the
removal on the single desk on the functions performed by AWB and then
consider how these changes might flow on to growers and others with an
interest in the industry.
2.1 Classifying AWB’s Functions
An individual grower considering selling wheat internationally

will earn a greater return if two things occur:
1. The price paid by international buyers for the wheat (P) is high; and
2. The price paid by export traders (p) is high.
In actuality, a grower’s main concern is the price they receive, p; however,

these two prices are intimately linked. The domestic purchase price, p,
cannot be higher than the international purchase price, P.
AWB operates as an intermediary primarily focussed on export

trading. Their operations and practices impact on both P and p. In
particular, we can identify two key functions:
• Marketing: this is AWB’s role as a negotiator and seller of Australian

wheat overseas. In its marketing role, AWB’s goal is to ensure that P is
as high as possible. Its ability to maximise P depends on its negotiating
power, demand management and branding activities.
• Logistics: this is AWB’s role as a coordinator of the international

transportation and handling of Australian wheat. By performing this
function well, AWB can reduce costs of getting Australian wheat from
the farm to international purchasers. If its costs can be kept low, AWB
can ensure that the differential, P – p, is as small as possible.
What we wish to focus upon is how the removal of the AWB’s exclusive
export right impacts upon its ability to perform each of these functions.
Consequently, each will be examined in turn.
July 2000 8
2.2 Marketing Advantages of a Single Desk
A selling price, such as P (the international price paid for

Australian wheat), will be higher when (1) buyers face fewer substitutes
for the product being sold; and (2) when sellers have more potential
purchasers of that product. In evaluating the single desk, therefore, we
need to consider how having an exclusive right to sell Australian wheat
internationally impacts upon each of these dimensions of price
determination.
2.2.1 Substitutes Available to Buyers
An exclusive right to sell, by definition, reduces the substitutes

available to international buyers of Australian wheat. In negotiations, if
those buyers fail to reach an agreement with the AWB, they are forced to
negotiate with sellers from other countries for wheat. This will be difficult
for those buyers if:
• Location: other countries are located further from those buyers.
• Quality: the wheat from other countries does not have the quality of
Australian wheat.
• Differentiation: Australian wheat supply is more reliable or better suited

to the seasonal demand patterns of those buyers (e.g., they may have a
preference for a southern hemisphere harvest schedule).
If any of these dimensions is salient for a particular buyer, then that buyer
will have imperfect substitution possibilities for Australian wheat.
Consequently, in negotiations with that buyer, the AWB will be able to
command a relatively high price for wheat.
Notice that these dimensions are not under the complete control of
AWB. For example, location is a geographic constraint and Australian
wheat will always command a locational advantage for certain buyers.
Similarly, quality and differentiation is a function of weather and
geographical characteristics and for some buyers this will always be of
value. Hence, whether trading is through a single desk or not will not
change the fact that Australian wheat may command a higher price for
any of the above reasons. However, as demonstrated in the Technical
July 2000 9
Appendix, the single desk can ensure that Australian growers capture a
relatively higher proportion of the value created by those advantages than
would be the case with more competitive export trading of Australian
wheat. Hence, it is the existence of those advantages that gives a single
desk some value. Moreover, to the extent that a single desk allows quality
or differentiation to be higher than they might otherwise be, this also
gives the single desk marketing value for growers.
To see this, consider what might happen if another trader could sell
Australian wheat to that buyer. In that case, the negotiating position of the
AWB would be diminished because that other trader would be able to
supply a product that from the perspective of buyers is a close substitute
to that provided by the AWB. Indeed, the competition may be quite
intense because if bargaining with the AWB broke down, the buyer would
still potentially have access to the entire crop in Australia through the
alternative trader.5 The end result is that the price, P, obtained by either
the AWB or other traders would lower; reducing potential returns to
growers.
Of course, in this circumstance that AWB and other traders would

have an incentive to find ways by which their products could be
differentiated on the basis of quality from one another. But the development
and maintenance of a brand and reputation for quality is a difficult exercise.
From buyers’ perspectives, a brand will be associated with a quality only
if historically quality has been high. Keeping quality high requires
monitoring and control by the trader. So, if other traders are able to sell
Australian wheat, then any single trader attempting to establish a brand
will need to control quality from the source (farm) right to the customer.
To do otherwise would be to allow for free riding on that brand name by
others and its eventual dilution.
Two broad outcomes might arise from this effect. On the one hand,
the additional costs in terms of establishing and maintaining a brand for
Australian wheat may be so high that it proves prohibitive. Hence,
Australian growers will lose any quality premiums they might have
otherwise commanded. On the other hand, a few traders might establish
5 That is, AWB would not have a credible ability to create a shortage of Australian wheat from the
perspective of any buyer.
July 2000 10
an effective control system. These traders, including AWB, would contract

in advance with certain growers and limit the supply of new varieties to
them. While that option would be open to other traders, it is possible that
each would act to limit the supply of new varieties overall; with some
growers benefiting from being able to produce higher quality grain and
others being excluded from that possibility.6
In contrast, AWB as the sole trader does not have to concern itself
with differentiating its product from other traders with the same
geographic supply of grain. Instead it is interested in differentiating all
Australian wheat from that grown elsewhere. By so doing, it can ensure
that international buyers see wheat from other countries as a less perfect
substitute for the Australian product. This gives it an incentive (and
obligation) to ensure that opportunities from growing quality varieties are
more freely available than might otherwise be the case. More importantly,
it makes a quality premium easier to maintain and less costly to achieve;
thereby, raising the value of Australian wheat.
It is also important to emphasise the role of the single desk in

countering the vagaries of international trade. When it negotiates a sale
with a particular buyer, there are difficulties with enforcing that particular
contract. In some cases, AWB will potentially be subject to what
economists term the hold-up problem.7 This occurs when a buyer tries to
renegotiate a sale price after the seller has incurred some key sunk costs.
For AWB, hold-up may occur just prior to grain being delivered when
transport and marketing costs have already been incurred. At that point, a
buyer might be able to renegotiate price terms with AWB that ignores
costs that have already been incurred.
To counter this, AWB requires a credible threat that it either has

alternative buyers (see 2.2.2) or alternatively that it will refuse to trade
with that buyer in the future. The latter threat is salient for that buyer
when Australian wheat is an imperfect substitute with other wheat and
the buyer has no alternative source for that wheat. If another trader
6Jean Tirole, “A Theory of Collective Reputations,” Review of Economic Studies, Vol. 63(1), January 1996,
pp.1-22.
7The classic reference on such hold-up problems is Oliver E. Williamson, The Economic Institutions of
Capitalism, Free Press: New York, 1985.
July 2000 11
existed, then the buyer could hold-up AWB with the knowledge that they
could still have access to Australian wheat in the future through another
trader. All traders of Australian wheat would have to coordinate a boycott
in order to make international contracting more reliable. However, such
boycotts are themselves vulnerable to free riding and ‘running.’8 Thus, an
advantage that comes with a single desk is that it gives the sole trader
more power to maintain contracts and counter potential vagaries involved
in international marketing.
In summary, a single desk has greater bargaining power compared

to a situation where there are many sellers of Australian wheat. While
Australian wheat will always likely command a locational and quality
premium, the single desk increases prices (through bargaining power and
branding ability) and also the certainty of successful trade; thereby,
raising net returns to growers.
2.2.2 Options Available to the Single Desk
AWB is also able to negotiate a higher price, P, if it can potentially

play buyers off against one another. That is, if AWB has more options
available to it in selling wheat, it can more readily insist on a higher price
in negotiations with any single buyer.
One way AWB does this is through the use of logistics and
planning. By timing sales and strategically managing the quality of
Australian wheat sold over the course of a year, AWB can ensure that it is
able to flexibly make stronger marketing options available at all times. It
can also ensure that the correct type of wheat is available at the port
closest to a particular customer. In this way, if a buyer were to be difficult
in negotiations, AWB would face fewer costs in supplying that wheat to
other buyers.
In the absence of a single desk, AWB would have more difficulty in

matching logistics to buyer negotiations. Instead, it would not be as sure
where a particular grower would be selling wheat and, indeed, whether it
8The ability of organisations to coordinate boycotts has been demonstrated to be critical in resolving hold-up
problems in medieval trade. See Avner Greif, Paul Milgrom and Barry R. Weingast, “Coordination,
Commitment and Enforcement,” Journal of Political Economy, 102 (4), August 1994, pp.745-776.
July 2000 12
would trade through AWB at all. Thus, its flexibility would be

undermined and this would diminish AWB’s negotiating position with
buyers. Moreover, these logistical advantages would potentially be lost to
all traders (except maybe for multi-national corporations) to the long-term
detriment of Australian growers.
Another way AWB could reap benefits from buyer competition

would be to limit the supply of Australian wheat. This is the textbook way
a monopolist exercises its market power; restricting supply in order to
create shortages and increase price.
However, there is a complication to this textbook case of cartel use

of market power for the case of the AWB. The statutory obligation on the
AWB to purchase all wheat (of acceptable quality) limits its ability to exercise any
market power it might have. To see this, consider how a firm may exercise
any market power it possessed. That firm would wish to restrict output in
order to push up prices and maximise the total profits earned. However,
for the AWB, the extent to which it can limit output is restricted by the
behaviour of growers. Suppose that in one particular year, wheat prices
were to rise to monopoly levels. This would induce some growers to
expand their harvest and maybe other farms to plant wheat.9 The AWB
could not continue to maintain the output levels that led to high prices.
Instead, it would have to market any wheat produced in that year and if
necessary accept a lower price in order to sell its stocks.10,11
The end result is that while the AWB may have market power, its
ability to exercise it by restricting output is limited by its statutory
obligation to accept all wheat delivered. The amount of wheat produced is
a decision of individual growers and not the AWB. This effect makes it
unlikely, as a matter of economics, that the AWB will be able to extract
9 As those growers are relatively small (compared with the volume of Australian wheat), they ignore the
effects their expansion of output has on the prices earned by themselves and by other growers.
10 It could conceivably only release a limited quantity onto world markets but this could not be done
indefinitely as its stockpiles would rise.
11 Mueller argued that the inability of cooperatives to control supply makes the customary index of market
power, a seller’s share of the market, virtually meaningless when applied to agricultural cooperatives.
Mueller points directly at the lack of control over production as a severe limitation on the ability to enhance
prices (W.F. Mueller, The National Antitrust Commission: Implications of Cooperatives. In Economics
Issues. Department of Agricultural Economics, University of Wisconsin – Madison, 1979).
July 2000 13
maximal price premiums in its current form.12 Only by being able to

control the output of all Australian wheat could the AWB perhaps fully
exercise market power and increase wheat prices. However, in order to do
this it would have to ensure that individual growers did not exceed
specified quotas or restrain total production by limiting grower access to
export markets.13,14
In summary, while market power is often extolled as a benefit of

exclusive marketing arrangements, the particular institutional structure of
the AWB makes it difficult for such market power to be exercised. This is
because the output decisions continue to reside with growers who can
respond to supra-competitive price premiums by increasing output.15 In
order for the AWB to be able to exercise market power, it would have to
be able to impose production maxima (quotas) on growers or to restrict
access to AWB services. This, however, would have the potential
implications of either placing disproportionate harm on small growers
who may be unable to produce enough to be viable or, alternatively, may
be denied access to export markets entirely.16
12The AWB’s ability to price discriminate is not necessarily a function of its market power. Many firms in
competitive markets practice price discrimination. For example, cinemas offer discounts for screenings on
certain days, or to particular classes of customers (e.g. pensioners). The ability to price discriminate is more
an indicator of good performance in negotiations than the exercise of market power per se.
13 Setting production quotas is a common practice of cartels seeking to raise world prices (e.g., OPEC).
14 Cotterill notes that closed membership arrangements produce different results. “… farmer members of the
closed membership cooperative capture.. a higher price for their farm product” (R. Cotterill, “The
performance of agricultural marketing cooperatives in differentiated production product market”. Strategy
and policy in the Food System: Emerging Issues, Proceedings of NE-165 Conference June 20-21, 1996, Wash
DC.
15 Perhaps this explains why it has been difficult to establish the existence of persistent price premiums
earned by single desks around the world. The difficulties of evaluating single desk selling is outlined by R.R.
Piggott, “Some old truths revisited” Australian Journal of Agricultural Economics 36, 1992, pp.117-40.
Piggott argues the problem lies in establishing the correct counterfactual. However, to this we would add that
the supply response of growers erodes whatever premia are available.
16Cotterill ibid. cites the case of Welch’s-National Grape and Ocean Spray system wherein farms that have
marketing rights through the cooperative are much more valuable than those that do not. The closed
cooperative price premium is capitalised into the value of the farm.
July 2000 14
2.2.3 Price Discrimination
An argument often advanced for the retention of a single desk is

that a single desk has an ability to practice price discrimination. Price
discrimination occurs where a seller is able to charge different buyers
different prices based on differences in buyers’ respective demand.17 In
order to do this, a seller must be able to charge a premium over its
marginal cost of production; something usually associated with existence
of market power.
AWB is currently able to practice price discrimination with the

effect of increasing the net returns to growers. This price discrimination
comes principally from the lack of good substitutes that some buyers of
AWB have.
From this perspective, if the single desk were not retained, AWB
would lose some of its negotiating power and hence, buyers would have
more substitutes available. Principally, these substitutes may have
comparable quality and locational advantages compared with AWB. As
argued earlier, the end result may be a reduction in premiums that AWB
is able to negotiate.
It should be emphasised, therefore, that it is not the ability to price

discriminate that allows a single desk to earn higher returns but rather the
overall strength of its bargaining position that arises from the lack of
buyer options to it. This will translate into price differentials among
customers as AWB gathers information about a particular customer’s
options and any international competition AWB may face. Hence, price
discrimination is a profitable manifestation of the greater marketing
opportunities open to AWB as a sole trader of Australian wheat; rather
than an option only available to a monopolist.18
17 Strictly, demand elasticities.
18See Michael Levine, “Price Discrimination without Market Power,” Discussion Paper, Harvard Law
School, 2000; for an elaboration of this argument.
July 2000 15
2.2.4 Evidence on Price Premiums
Associate Professor Joseph Hirschberg of the University of

Melbourne has conducted an extensive empirical investigation of the
determinants of any price premiums achieved by AWB over the past few
years. His analysis is provided in the technical appendix below and his
results are summarised here.
The difficulty associated with examining price premiums is that

identified above: namely, when we observe that Australian wheat is
commanding a premium over Pacific Northwest Wheat (US) at a
particular time, while some of that premium may be the result of AWB
marketing efforts (and the single desk) it also could be explained by the
particular mix of customers they were trading with in a particular month
(be it varied in distance and overall demand) and also the mix of wheat
grade traded at that time. Thus, to obtain a clearer picture of the price
premium that could be attributed to AWB choice rather than the pure
characteristics of Australian wheat requires the use of sophisticated
statistical techniques.
The methodology employed was to look at the variation of

percentage price premiums in AWB contracts from 1997 to 1999 and to see
how much that variation was explained by factors such as distance to
customer and the particular grade of wheat sold; as well as variables that
may be under the control of AWB. Figure 1 depicts the results of this
analysis of variation. Note that most of the variation in the percentage
price premium is explained by the distance advantage accruing to
Australia and the precise grade of wheat being offered. Also, note that
Japan as a customer was a significant factor; that is, Australian wheat
commanded a premium in Japan above that explained by distance and
grade. About 40 percent of the variation, however, was unexplained by
these factors. This amount could in principle be attributed to the activities
of the AWB and potentially the single desk increments to bargaining
power. However, it is also possible there could be other unexplained
factors in the international wheat market that could account for such
variation.
July 2000 16
Figure 1: Explanation of Variation in Percentage

Price Premium
Date
7%
Distance
Advantage
18%
Other
40%
Japan
Grade of 26%
Wheat
9%
This analysis allows one to obtain a prediction of the price

premium that would arise at a particular time given the mix of customer
demands and wheat quality. This was done for the data available and the
results are plotted in Figure 2. Notice there that while average actual price
premiums have been increasing over time, predicted price premiums are
falling. Indeed, the two series are converging. What this means is that in,
say 1997, when actual premiums appeared to be low, this was in fact
accounted for by perhaps a poor customer mix/wheat quality (or other
related poor trading conditions in world markets) and, in fact, when these
are taken into account the marketing performance of AWB appeared to be
relatively high. In contrast, in 1999, actual and predicted premiums are
closer in value indicating that actual premiums are relatively reflective of
actual AWB performance.
July 2000 17
Figure 2: Actual and Predicted Percentage Price Premiums
14
12
10
8
% Premium
2 % Premium
0 Estimated
-2 Average
SEP 97
SEP 98
SEP 99
JAN 97
JAN 98
JAN 99
JUL 97
JUL 98
JUL 99
MAY 97
NOV 97
MAY 98
NOV 98
MAY 99
MAR 97
MAR 98
MAR 99
Month
It is possible to provide a similar comparison, not based on time,

but in terms of grade of wheat. Table 1 summarises the results regarding
the actual and predicted average price premium for each grade of wheat
for the data at hand. Notice that for some grades of wheat, the actual
average premium (per contract) understates the true marketing conditions
as predicted by our model. This is true for Australian General Purpose,
Australian Hard, Australian Prime White, Durum and Soft. While for
Australian Prime Hard, Australian Standard White, Australian Standard
White Noodle and Feed actual premia overstate true marketing
conditions. This table gives us a clearer picture of the true premia
achieved by AWB over the period taking into account customer mix,
distance and timing effects as well as interactions among different grades
of wheat.
July 2000 18
Table 1: Actual versus Predicted Percentage Price Premiums per Grade

of Wheat
Grade Total Tonnes Average Percentage Price Premium

Sold
Actual Predicted
Australian 4,741,487 0.68 1.24

General Purpose
Australian Hard 5,067,144 1.06 4.49
Australian 3,220,714 6.90 5.29

Prime Hard
Australian 9,370,155 1.64 3.87

Prime White
Australian 22,580,362 4.98 4.26

Standard White
Australian 74,978 19.96 8.23

Standard White
Noodle
Durum 372,968 24.12 28.55
Feed 100,934 4.44 0.48
Soft 757,967 4.82 5.82
Significantly, these results indicate that Australian wheat has

appeared to command real premia beyond which that would be explained
by its distance advantages and particular mix of grades.19 Our theoretical
19This is significant because previous studies have often failed to find the single desk sellers achieve price
premia; see Colin Carter, Donald MacLaren and Alper Yilmaz, “How Competitive is the World Wheat
July 2000 19
model suggests that some of this premia can in fact be attributed to single
desk effects and indicates that realised prices might fall if such restrictions
are removed. Nonetheless, one should be cautious in interpreting such
results as other factors – not accounted for in our data (e.g., variation in
weather patterns etc.) – may also account for the premia achieved.
2.2.5 Summary
Due to its purchase obligations, economic theory suggests that

single desk sellers are able to command price premia based on bargaining
rather than market power. Moreover, the price premia can be further
enhanced by the control afforded single desk sellers in developing brand
image and also in terms of minimising potential hold-up problems.
When examining the actual price premia achieved by AWB, we

found that by taking into account distance advantages and also the mix of
wheat grades sold, we could uncover a more accurate picture of AWB’s
performance in this regard. This performance is consistent with the notion
that AWB commands price premia as a result of bargaining power and,
consequently, that such premia might be eroded somewhat if single desk
restrictions were removed.
2.3 Logistics Advantages of a Single Desk
The other function performed by AWB is that of logistics. They

coordinate the supply of all wheat for export in Australia and then
negotiate favourable terms for freight overseas. The issue, however, is
how AWB’s performance of this function would change if other traders
were permitted to seller wheat internationally. Recall that, given AWB’s
corporate structure, when its logistics function is performed well, P – p is
minimised as those cost savings are passed directly onto growers. As
such, here we will focus on the magnitude of P – p in evaluating the merits
of single desk selling.
Market?” mimeo., University of California, Davis, 1999. Of course, the data made available for the present
study is significantly more detailed than previous studies.
July 2000 20
2.3.1 Production Economies of Scale and Scope
The usual justification for monopoly rights across all industries is

the ability to achieve economies of scale. The monopoly protection
afforded to inventors by patents is an example of this as well as the
statutory protection to organisations such as Australia Post.20 For each, it
is argued that costs would not be minimised if those organisations were
not guaranteed certain volumes of production.
It has been argued that there exist similar economies of scale in

wheat trading.21 The ‘system’ element to wheat handling, storage,
distribution and transport costs as well as the savings on the costs of
transacting with many international customers all have qualities that the
average cost of each falls when a greater quantity of wheat is serviced.22
But the economies of scale can go beyond these functions to the
profitability of research and development and the dissemination of
information to growers. In addition, the ability of the AWB to offer
insurance options to growers depends on the diversity and numbers of
growers across Australia as well as its ability to store wheat from year to
year.23 These systemic elements to wheat marketing all potentially serve to
lower logistics costs for all wheat as more wheat is traded through AWB.24
20In the case of Australia Post, one of the arguments for retention of the exclusive arrangements has been the
preservation of the ‘universal service’ obligations, particularly service to remote and rural areas.
21For example, Cargill’s acquisition of Continental’s grain merchandising business was considered to be
motivated by the desire to achieve economies of scale and scope in grain handling (M. Hayenga and R.
Wisner, “Cargill’s acquisition of Continental Grain’s grain merchandising business,” Staff Paper 312,
Department of Economics, Iowa State University, 1999.
22Such pressures for consolidation are occurring throughout the agricultural industry. See, for example, Mark
Drabenstott, “Consolidation in US Agriculture: The New Rural Landscape and Public Policy,” testimony to
the Senate Committee on Agriculture, Nutrition and Forestry, January 1999.
23 Richard Caves found that both information and insurance were important reasons why scale economies
appeared to be present in agriculture. On the information-side, economies came from information about
demand (in terms of quantity, timing and quality) and the expectations of international trade flows. This
information could be used to assist in coordination and other logistical activities. On the insurance side, risks
in grain trading can more easily be managed through pooling arrangements. See Richard E. Caves,
“Organisation, Scale and Performance of the Grain Trade” Food Research Institute Studies, Vol. 16 Stanford
University, 1977-78; and Richard Caves, Multinational Enterprise and Economic Analysis, Cambridge
University Press: Cambridge, 1982.
24The issue of economies of scale and scope raises the empirical question as to the optimal size of the AWB,
and how this compares to the Australian market.
July 2000 21
If the single desk were removed then presumably other traders

would handle some wheat. If scale economies are present, this means that
the average cost of wheat handled by AWB would rise and hence, all other
things being equal, for growers trading through AWB, returns would fall.
2.3.2 Marketing Scale Economies
A subtler form of scale economy surrounds the issue of ‘branding.’

As argued earlier, AWB’s methods of assuring wheat quality and
improving wheat variety have enabled it to develop a reputation for
Australian wheat; thereby, improving returns per unit sold. If some
Australian wheat were marketed without these quality assurances, it may
lead to some deterioration in wheat quality and may harm the brand-
image of Australian wheat. That is, the reputation the AWB has developed
may be undermined.25 It may be possible to remove the statutory
protection of the AWB without erosion of its brand so long as strict
controls on the alternative marketing of Australian wheat were possible.
However, these very controls may be argued to be an impediment to the
competition any removal of protection may be designed to achieve.
2.3.3 Monopsony Power
While economies of scale are a potential reason for the retention of

single desk selling, an argument often given against it is one of
monopsony power. That is, AWB is currently the sole purchaser of wheat
for export from Australia; by virtue of it being the sole trader. Given that
80 percent of wheat grown in Australia is exported, despite the
deregulation of the domestic grain market, this gives AWB substantial
monopsony power as a purchase of grain. The removal of single desk
selling would, therefore, presumably erode the monopsony power of
AWB.
What needs to be evaluated is whether AWB’s monopsony position

acts to the detriment of growers. Note that the textbook monopsony story
25Jean Tirole, “A Theory of Collective Reputations,” Review of Economic Studies, Vol. 63(1), January 1996,
pp.1-22.
July 2000 22
(like the textbook monopoly story)26 does not apply to AWB in the face of
its obligation to purchase all wheat offered for export. A firm exercises
monopsony power by limiting its purchases from suppliers, thereby
lowering the price it has to pay them for their product. The AWB,
however, cannot exercise power in this way because it cannot restrict its
purchases. It must accept all wheat offered and also must compensate
growers through a pooling arrangement. Hence, it cannot use
discriminatory pricing terms to effectively limit its purchases of wheat.27
On the other hand, for AWB, its monopsony position and

consequently its monopoly position in terms of logistics to the wheat
export trade means that it is not subject to competitive pressure in the
performance of this function. Competitive pressure often forces firms to
keep costs down and ensure a reliable service, etc. The absence of such
pressure may mean that AWB’s costs of undertaking its logistics role are
higher than they might otherwise be.
Traditionally, however, the policy-maker’s response to this lack of

competitive pressure is to substitute performance regulation for
competition. This occurs in many network industries – rail, energy
distribution, roads, water etc. – that also have substantial economies of
scale in their productive arrangements. For AWB, this regulation exists
through the WEA. In principle, effective regulation can be a perfect
substitute for competition in assuring that a firm performs its functions in
an efficient manner. For the purposes of this review, it needs to be
considered whether the regulation and also the grower-ownership of
AWB means that, in a practical sense, AWB’s costs are the same or lower
than they would be under a purely competitive constraint.
2.3.4 Summary
Single desk selling potentially enables the realisation of economies

of scale in the Australian wheat industry. Eliminating the single desk, if it
26 For
this textbook story see, for example, Hal Varian, Intermediate Microeconomics: A Modern Approach,
rd
3 Edition, Norton: New York, Chapter 25.
27Monopsony power was a primary concern of the US Department of Justice in requiring divestitures of
several river and port elevators when approving the Cargill-Continental merger. This type of monopsony
power cannot be exercised by AWB because of its statutory purchase obligations.
July 2000 23
serves to reduce the average scale of wheat exporters could mean that the
average costs of logistics functions could rise. Nonetheless, single desk
selling does give AWB a monopsony over the logistics function for the
export of Australian wheat. While this is unlikely to result in lower prices
paid to growers, it does reduce competitive pressure on AWB to perform
this function. Nonetheless, effective regulation could act as a substitute for
this loss of competitive pressure.
2.4 Overall Effect of Non-exclusivity on AWB
The above discussion has generated several conclusions regarding

the likely effect of non-exclusivity on AWB.
• Diminished Bargaining Power in the International Market: the price,

P, that AWB receives for wheat it markets is likely to be lower if
other traders are able to trade wheat in bulk to international
clients.
• Cost Changes: the differential, P – p, may rise or fall depending

upon whether potential losses in economies of scale are
outweighed by improved performance brought about by
competitive pressure or not.
Consequently, the price received by growers selling to AWB, p, will only

rise if the effect on bargaining power is small, economies of scale are not
lost, and competitive pressure is high as a result of removing the single
desk requirements.
This, however, is only part of the analysis of the single desk. It

focuses on the return to growers who continue to market grain through
the AWB even after its exclusive right is removed. There are also growers
who will operate independently of the AWB. Thus, in the next section, we
consider the implications of removing the single desk for all growers and
not simply those who might remain with the AWB.
July 2000 24
Section 3 Effect on Growers
3 Effect on Growers
The previous examined the effect of removing AWB’s exclusive

export right and we noted that this would likely reduce P, the price AWB
received for wheat, and may increase or decrease P – p, depending upon
whether the loss of scale economies outweighed the effect of competitive
pressure or not. For growers, however, the main issue is the effect on p;
that is a sum of the two factors above.
Growers as a whole would benefit from the removal of the single

desk only if p were to rise. This would occur if the expected reduction in P
were not too large and also if the industry logistics costs were to fall. Note
that this is not simply a function of whether AWB’s costs would fall
following the removal of a single desk but also whether other traders
could operate at a lower cost than AWB. If they cannot, then competitive
pressure will not be sufficient to reduce costs.
This highlights that the critical issue for growers is whether the industry
costs of handling grain would fall by a substantial amount if there were other bulk
exporters. If economies of scale are important in keeping such costs low,
then having other exporters is unlikely to generate cost reductions. Of
course, some of the other traders may be multi-national corporations who
might be able to generate cross-border economies. So, even in the presence
of scale economies, costs could be reduced if Australian grain were to be
marketed by a global, multi-national trader.
The proponents of the acquisition by Cargill of Continental’s grain

merchandising business believed that economies of scale – not only
justified the merger – but made it socially desirable as well. Cargill argued
that the merger would realise economies of scale and that these would be
passed through to farmers in terms of higher grain returns. The US
Department of Justice was concerned about the new merged entity’s
monopsony power in some regions – especially given its control of key
handling facilities.28 However, as we noted earlier, AWB’s purchase
28 See M. Hayenga and R. Wisner, “Cargill’s acquisition of Continental Grain’s grain merchandising
business,” Staff Paper 312, Department of Economics, Iowa State University, 1999; and William Wilson and
July 2000 25
obligations virtually eliminate its ability to exercise monopsony power of

kind that concerned US anti-trust authorities. The very fact that the
imposed restrictions on the Cargill-Continental merger limited its
monopsony power, yet preserved sufficient cost savings to make the
merger profitable indicates the potential validity for the story that a break-
up may lead to losses in such economies.29
3.1 Allowing Grower Choice
One argument often offered for the removal of the single desk is
that it could not hurt and only benefit growers as they would have more
choice. After all, the argument goes, if a single desk is performing well in
maximising grower value, growers will continue to trade through it and
there would be not substantial difference to the current situation. On the
other hand, if the single desk does not maximise grower value, the choice
will reveal that.
While there is, of course, some merit to the benefits of choice – this
is the main route by which competitive pressure is brought to bear – the
argument is simplistic in that it neglects the very externalities among
grower decisions that lie at the heart of both the logistics and marketing
benefits that might be generated by having a single desk.
Consider first the scale economies argument. This argument

suggests that if scale economies where present, then having grower choice
will not change the fact that there will be a single exporter of Australian
wheat but it might change who that single exporter was. That is, it could
either by a trader other than AWB or it could be an AWB that performs
better because of the potential choice growers have.
Bruce Dahl, “Transnational Grain Firms: Evolution and Strategies in North America,” Agricultural
Economics Report, No.412, Department of Agricultural Economics, North Dakota State University, 1999, for
a discussion of economies of scale in handling brought about by mergers and joint ventures.
29For a discussion of the signals sent by merger restrictions see Teresa Fels, Joshua Gans and Stephen King,
“The Role of Undertakings in Regulatory Decision-Making,” Australian Economic Review, 33(1), 2000,
pp.3-16.
July 2000 26
However, the underlying assumption here – that a scale economy

implies a sole trader – is not necessarily true. Scale economies can be
present yet more than one firm performs a function. The end-result is
unnecessary duplication of costs and higher industry costs overall.
In addition, recall that AWB is essentially a cooperative. That

means that it, more or less, pools costs across growers. While it does
attempt to attribute costs, when there are scale economies, this cannot be
done perfectly. Consequently, there will likely be growers at the margin
who might be able bring grain to market in a way that is cheaper to them
if they did not have to share costs with others. These growers will be the
targets of alternative traders. The end result, however, is that there will be
a loss of scale economies at an industry level and hence, higher costs for
the average grower.
On the marketing side, there are strong reasons why some growers
might wish to by-pass AWB and trade elsewhere. AWB’s use of price
discrimination among buyers, translates into pooled returns to growers.
This means that some buyers could offer traders a lower price than they
pay AWB but this will translate into a higher price for grain for some
growers. Hence, the pooling mechanism itself generates incentives for
some growers to trade outside AWB.30
The end result of grower choice would be lower prices overall but
also a change in the distribution of returns among growers. In particular,
growers selling the same quality of wheat may receive different returns
depending upon whom they have export trading relationships with.
However, another implication is that a private firm – especially one
handling a large share of Australian wheat – would have an incentive to
actually exercise any market power it may have. To do this it would only
selectively purchase grain and perhaps restrict access to other marketing
services to growers. Also, varieties would be developed to suit the needs
30 Note that this is a different reason for by-pass than is usually associated with a cartel. For the cartel
monopoly that restricts output, the temptation for individual members to break the cartel is to sell high
quantities even if at a slightly reduced price. For AWB, however, this incentive is not present as it does not
restrict output overall.
July 2000 27
of the marginal grower (likely to be one of the largest growers) rather than
the average grower.31
Consequently, some growers may find themselves without efficient

and cost effective access to international markets.32 These are the growers
whose output may be constrained by a private firm to such an extent that
they can only find it worthwhile to sell domestically and perhaps not at
all. If AWB were to become unviable then one of its chief functions –
ensuring that all growers have access to international markets – would not
be performed. Ironically, the attempt to give growers more choice by
removing the AWB’s exclusivity may result in some growers having no
choice at all.
3.2 Summary
If there are benefits derived from single desk selling that arise from
marketing and logistics these are invariably a function of AWB’s ability to
internalise externalities among growers. Therefore, the simple argument
that the removal of single desk selling gives growers more options in
wheat trading and thereby must be in their collective interest is incorrect.
While some growers may benefit from the ability to trade internationally
outside of the AWB, as a whole -- to the extent there externalities among
growers – the removal of AWB’s exclusive export right would likely
reduce net returns to growers in total.
31Oliver Hart and John Moore, “Cooperatives versus Outside Ownership,” Working Paper, No.6421, NBER,
1998.
32Even if AWB retains an obligation to purchase, its costs may have become so high that some growers do
not find it worthwhile to supply wheat for export.
July 2000 28
Section 4 Effect on Others
4 Effect on Others
The stated objective of the WMA is to maximise the net returns to

growers of Australian wheat for export. Hence, in evaluating the
importance of restrictions on competition in achieving that goal, the
analysis thus far has focussed on growers. It is useful, however, to review
the impact of single desk selling on other interested parties. In particular,
here we review its impact on the domestic market, international
purchasers and other potential grain traders.
4.1 Domestic Market
Growers have a choice of whether to sell their grain domestically or

for export through the single desk. As grain is sold both domestically and
for export, this means that at the margin, grain is earning the same net
return for growers in each market. Of course, actual returns may vary
slightly because of differences in buyers’ value placed on quality and
other considerations, but in general this principle will hold.
This means that there is a clear relationship between the

performance of the single desk in maximising grower returns and the
prices traded for wheat domestically. In particular, the better is the
performance of AWB internationally, either in its marketing (increasing P)
or logistics (reducing P – p) functions, the effect will be to raise the price of
domestic wheat needed to keep wheat trading domestically.
If the single desk were removed, then the impact of this on

domestic wheat prices would depend upon its effect on the net return to
growers in the export market. If those returns were to fall, then domestic
wheat prices would also fall. However, if those returns were to rise, then
domestic wheat prices would also rise. Thus, in a sense, the interests of
domestic purchasers of wheat is diametrically opposed to the ability of a
single desk to maximise net returns to growers on international markets.
For the higher that ability, the more costly is domestic wheat likely to be.
July 2000 29
Of course, if economies of scale were important, then it may be that

domestic purchasers of wheat may have interest aligned with growers
selling for export. In particular, if eliminating the single desk meant the
loss of scale economies, this could also increase handling costs in the
domestic market – pushing up costs and prices there.
What should be emphasised, however, is that AWB’s inability to

exercise monopsony power by restricting the quantity of wheat for export
also constrains its ability to use this to raise domestic prices; potentially
for the benefit of its domestic trading operations. Hence, the usual concern
regarding dominance on one particular market potentially spilling over
into related markets is unlikely to be of primary concern here.
It is, therefore, hard to predict conclusively what the impact of

removing the single desk would be on domestic wheat prices. In general,
one expects that if this serves to push export prices received by growers
down, this will also lower domestic wheat prices. However, if it otherwise
raises costs in the domestic industry, the reverse effect might occur.
4.2 International Purchasers
The primary international concerns about single desk selling, such

as that by AWB, have come from the United States and Europe.33 They are
concerned about the ability of such producers to exercise market power
and hence, distort grain prices upward. Interestingly, many of the
distortions in grain trade come from EU and US subsidy and support
schemes. Here, however, we concentrate our discussion on the issue of
market power.
There are often legitimate concerns about the behaviour of

monopolists in reducing economic welfare. Specifically, the so-called
dead-weight losses described in economics texts arise from the loss in
both consumer and producer surplus when monopolists sell a lower
33 USGAO (United States General Accounting Office), “Report to Congressional Requesters: Canada,
Australia and New Zealand, Potential Ability of Agricultural STEs to Distort Trade,” GAO/NSIAD-96-94,
June 1996; and Sir Leon Brittan (vice president of the European Commission), “Competition Policy and the
trading system: towards international rules in the WTO,” speech to Institute for International Economics,
Washington DC, November 1997.
July 2000 30
quantity than would arise under more competitive conditions. However,

the exercise of market power that leads to such losses is not a given. Often,
monopolists – while able to command a favourable negotiating position –
are not able to easily reduce their overall output. The difference is
important as a reduction in output lowers overall value created by the
industry while the improved bargaining position merely serves to enable
to monopolist to appropriate a greater share of value created.
This distinction is well recognised in economics. Basically, not all

monopolists (i.e., single sellers) are ‘bad.’ A monopolist is considered
‘bad’ if it exercises its market power to reduce value created in the
economy. However, there are two reasons why monopolists – even if they
technically have market power – may not exercise it.
First, a monopolist may not be able to credibly reduce output. The

exercise of market power necessarily involves output contractions.
However, some monopolists may face markets and technology that make
it difficult for them to persistently contract output. As mentioned earlier,
the threat of potential competition and entry can constrain a monopolist
(the so-called contestable monopoly outcome). But monopolists can also be
constrained by their own future behaviour. A monopolist that restricts
output by building up a stockpile will eventually face pressure to sell
down those stocks. If customers realise this they might hold off their
purchases to wait for future price declines. This is especially a problem
that faces monopolists that sell products that are durable and can be
stored by customers. For AWB, their statutory obligations to accept
delivered grain, combined with stockpile constraints, also mean that it
may not be able to restrict output.
Second, a monopolist may not have an incentive to reduce its

output. This arises when the monopolist is not strictly maximising profit
per se but the returns to its own suppliers.34 For AWB, it has an obligation
to sell all wheat produced. Therefore, it has limited discretion to restrict
wheat output to customers in any given year.
34 Consider the recent case of the Australian Performing Rights Association (APRA) -- the copyright
collective recently authorised (by the Australian Competition Tribunal) to fix prices for composers
(Australasian Performing Right Association Limited (1998) ATPR (Com) 50-256.).34 APRA was held to have
market power but not exercise in the sense of restricting the supply of compositions. This is because it had an
incentive to place compositions in the hands of all customers and negotiate tailored pricing for each.
July 2000 31
These two elements mean that while AWB is a monopolist in the

strict sense of the word it is not necessarily a bad one – from the
perspective of competition policy. The AWB cannot restrict access by
growers to its services, nor control the production of grain. In part this
comes from its statutory purchase obligation and in part this comes from
its ownership by those growers.
This perspective helps explains the concerns of international

purchasers about AWB’s single desk. They realise that without the single
desk, prices for Australian wheat would likely be lower and that
Australian growers would appropriate a lower share of total value
created. It also explains, however, that their overall concerns about a loss
of efficiency are potentially unfounded. That is, the high prices may reflect
a superior bargaining position afforded by AWB’s sole right to export
rather than acts design to create wheat shortages. The total amount of
wheat available for export is not a choice of AWB and hence, it cannot
choose to reduce that quantity in an attempt in exercise monopoly power
in the international wheat market.
4.3 Other Export Traders
The single desk clearly deprives other potential exporters of

Australian wheat from an opportunity to do just that. Some of those
exporters may also be domestic traders and may be able to realise scale
economies in both markets that would enhance overall efficiency in the
wheat industry (subject of course to losses in scale economies from AWB).
In addition, other export traders without a statutory purchase obligation
may be able to exercise market power to the potential benefit of some
classes of growers.
Veeman, Fulton and Larue35 argue that the main beneficiaries of the
removal of single desk provisions would be multi-national enterprises
(MNE). This is because MNEs are primarily private operations that could
potentially utilise global scale economies to build up market share as
35M. Veeman, M. Fulton and B. Larue, “International Trade in Agricultural and Food Products: The Role of
State Trading Enterprises,” report for Agriculture and Agri-Food Canada, April 1999.
July 2000 32
exporters from many countries. This would assist their ability exercise
market power and also to exercise oligopsonistic power over the purchase
of grain from growers. Thus, they would stand to benefit from higher
profit margins in the marketing of grain throughout the world. However,
Veeman et.al. do not necessarily see that these benefits would flow back to
growers or necessarily result in a reduction in domestic grain prices. In
effect, one organisation with market power is replaced by another with
similar power; meaning little for the performance of the wheat industry.
4.4 Overall Impact
Given all of the potential interested parties in a single desk,

analysing the overall impact of its removal is difficult. Veeman, Fulton
and Larue have attempted this exercise – concentrating on the impact of
an export state trading enterprise (STE) on total economic surplus
generated throughout the world economy. They found that:36
1. STE are always less efficient than private traders if private traders are
perfectly competitive.
2. If the private traders are not perfectly competitive, the use of a price pooling
export STE can increase the total economic surplus available. This result can
hold for both the small country and the large country case.
3. The use of a price pooling export STE is more likely to increase the total
economic surplus available when:
(a) export sales are a relatively large percentage of total production
(b) the market power of the STE in the domestic processing market
is limited.
Notice that for Australia, the wheat industry is focussed on export while
AWB does have a large market share in the domestic market. This makes
it difficult to evaluate, in terms of their model, the overall impact of single
desk selling of Australian wheat. Nonetheless, they did conclude that the
36 ibid., p.85.
July 2000 33
AWB was unlikely to have a negative impact on economic surplus

generated in international wheat markets.37
There are, however, two assumptions in their model that do not

apply to the Australian case. First, they assumed that the STE pooled
prices across domestic and world markets. In Australia, AWB does not do
this and operates domestic and export trading through separate corporate
entities. Second, they assumed that the STE did not have a statutory
obligation to purchase wheat for export. Once again, AWB does have this
obligation. If these two assumptions are taken into account then their
model overstates both the potential harm an STE may place on surplus
generated by international trade and the potential harm to domestic wheat
prices. This suggests that the case for a positive overall impact of single
desk selling of Australian wheat for export is likely to be reasonably
strong.
37 ibid., p.96.
July 2000 34
Section 5 Conclusion
5 Conclusion
Analysing the desirability of single desk selling of Australian wheat

is a difficult task. In particular, it involves establishing what the likely
outcomes would be if single desk restrictions were removed and other
traders were able to sell Australian wheat. But how to predict that
counterfactual is not simple; a dilemma noted by others:
The impact of an STE (State Trading Enterprise) operating in the
international market can be determined by answering the
following question: What is the outcome (e.g., in terms of prices,
output, exports, imports, economic surplus) if an STE is assumed
to operate in the international market, and what is the effect on
this outcome if the STE is removed and replaced by privately-
owned traders?
The assumptions used in answering this question are very

important. Previous models of STE typically assume that STE
removal will result in the formation of a competitive trading
sector in that country or for that commodity. This assumption is
simplistic and can have important consequences for the results of
analyses that employ it. In reality, although the removal of an STE
could potentially result in a competitive trading sector, in many
industries the more likely impact of STE removal is the
replacement of one oligopolistic structure with another. This is
particularly likely for many export STE. … the replacements to an
export STE are often likely to be multinational enterprises (MNE);
these MNE are large, may be vertically integrated, and often are
privately owned firms that provide very little public
information.38
In the Australian context, the analysis is complicated further because the

domestic market is unregulated and because AWB has conditions that
require it to purchase all wheat put forward by growers for export, subject
38 Veeman et.al., op.cit., p.33.
July 2000 35
to minimal quality standards. This means that AWB can potentially derive
bargaining power but not market power in world markets.
On a theoretical level, we concluded that the impact of removal of

single desk selling may have an adverse impact on growers as a whole
and also a distributional impact that may affect some growers more than
others. In particular, it is likely that AWB can command price premia that
would be eroded in the absence of its exclusive trading rights. Moreover,
by examining actual pricing data, it was found that this data was
consistent with this theoretical hypothesis. Even entry of additional trader
could erode price premia significantly.
Similar losses could also arise when considering AWB’s logistic

function. In particular, the critical issue is whether new traders would
erode economies of scale and thereby increase average exporting costs
across the Australian industry. Moreover, it is important to consider
whether those traders would place sufficient competitive pressure on
AWB’s costs to outweigh any losses in economies of scale; specifically,
competitive pressure that could not be achieved by appropriate regulation
by the Wheat Export Authority. This is a key issue that needs to be judged
in any review of AWB’s exclusive export rights.
In terms of the potential gains to agents other than growers, we

were unable to predict definitively the impact on the domestic trading
market although we could conclude that international purchasers and
other traders would potentially benefit from the removal of single desk
selling of Australian wheat. Nonetheless, any gains here seemed to be
distributional rather than welfare enhancing. This was a similar
conclusion to other studies:
We believe that concerns over export STE are probably
exaggerated. The evidence available indicates that most export
STE receive little assistance from their government. The most
effective way of ensuring that SET do not subsidise exports is to
ensure that these are self-financing institutions that are insulated
from government. However, the most effective means of ensuring
that export subsidies do not distort world markets for agricultural
July 2000 36
products would be the complete prohibition of all agricultural

export subsidies. 39
From this perspective, this report has argued that the negative impact of
single desk selling on international trade is overstated and is in all
likelihood non-existent given the statutory purchase obligations of AWB.
In the end, it appears most likely that the impact of the removal of
single desk selling would be distributional. That is, some growers may
benefit while others might be disadvantaged. Such distributional concerns
have been common throughout the history of the Australian wheat
industry.
The standard work on the AWB (Whitwell and Sydenham, 1991)
is entitled A Shared Harvest , a title reflecting the essence of the
pooling principle that price risks and marketing costs are shared
among producers. The way farmers manage production,
marketing and financial risks is at the core of their individual
business strategies and competition for resources amongst
farmers. Pooling is designed to reduce this competition. Whatever
the economic effects of pooling and statutory marketing, grower
equality was paramount for its supporters. Their other concern
was the behaviour of middlemen. Whitwell and Sydenham (ibid.,
p.286) described growers’ ambitions for wheat marketing as
involving ‘three main principles, namely that the pool be
compulsory, that the marketing organisation be granted
monopoly powers and that it be grower-dominated.’40
Pooling is an example of practice that protects some growers. It means

that grower returns are average across wheat grade and other factors. In
practice, this means that there are potential opportunities for some
growers to receive greater returns from dealing with customers who have
high demand for certain grades of wheat. However, to allow this would
be to reduce the returns for growers unable to sell wheat to those
customers. The erosion in bargaining power and the loss of economies of
scale that might arise from the removal of AWB’s exclusive right – as well
39 Veeman et.al., op.cit., p.32.
40 A.S. Watson, “Grain Marketing and National Competition Policy: Reform or Reaction?” Australian
Journal of Agricultural and Resource Economics, 43 (4), 1999, p.436.
July 2000 37
as pressure to remove its statutory purchase obligation – could well result

in some growers being no longer able to access international markets. It is
in this light that the single desk policy needs to be evaluated. That is, is the
principle of equalising access to international customers worth sacrificing
efficiency losses (if any) that result from single desk selling?
July 2000 38
Section 6 Technical Appendix A: Theory
6 Technical Appendix A: Theory
Here we develop a model that demonstrates the impact of a single

desk of the price of Australian wheat. It should be noted that the model is
specified so as to be very simple in nature and that the majority of
assumptions embedded below can easily be generalised. Hence, the
conclusions reached are quite strong and robust.
We consider AWB’s negotiations with a single buyer. That buyer,

places a benefit v(q) on a quantity of wheat, q, that may come from
Australia (A) or the rest of the world (R). If qA is the quantity supplied
from Australia and qR the quantity from the rest of the world, then the
buyer’s total value on the wheat is v(qA + qR). This assumes that wheat
from any source is the same in the eyes of the buyer. Below we will amend
the model and demonstrate the effect of a quality differential between
wheat from Australia and the rest of the world.
On the supply-side, suppose that the (industry) short-run marginal

cost of supplying qA units of wheat from Australia is S(qA) and that of
supply qR from the rest of the world is S(qR).41 We assume that S(.) is an
increasing function in each case; that is, industry marginal costs rise as
output rises – hence, supply curves slope upwards. It is assumed,
however, that Australia has a locational advantage with respect to the
buyer. We model this by assuming that it costs an additional t per unit
shipped from R for it to supply the buyer.
Total value created from selling wheat to the buyer is maximised

by choosing qA and qR that maximises:
v (q A + qR ) − S ( qA ) q A − S ( qR )q R − tqR (1)
Let q *A and q *R be the quantities that maximise this function and V* its
maximised value. In negotiations, it is these quantities that will actually be
41 This assumes equivalent supply conditions for the two sources of wheat. It can easily be amended to
include differences in supply conditions.
July 2000 39
supplied to the buyer. Note that q *A > qR* when transportation costs, t, are
positive.
6.1 Bargaining with a Single Desk
A standard model of negotiations in economics is one based on core

theory. In its business applications, this model has been associated with the
concept of added value.42 Added value is the value a particular agent brings
to a negotiating situation. A given agent cannot expect to get more than
their added value when negotiating with other rational agents. In general,
an agent will receive from fraction, α, of their added value. α is therefore a
measure of an agent’s bargaining ability. This makes added value the
appropriate concept for determining the outcomes in negotiations.
For a single desk such as AWB, if it were not to supply a given

buyer, total value created would fall to some level, say, V − . Thus, AWB’s
added value is V * − V − ; that is, the difference between total value created if
it supplies the buyer and total value created if it does not. The price of
Australian wheat, P, would, therefore, become:
 V * −V −   V * −V − 
P =α  *  + S (q A ) = α 
*
*  + S (qR ) + t
*
(2)
 qA   qA 
where the last term comes from the fact that S ( q*A ) = S ( qR* ) + t .
When there is a single desk, if negotiations were to break down and

AWB were to withdraw from selling to the buyer, the buyer would have
to adjust its purchases accordingly. In particular, it would be forced to
purchase all of its wheat from the rest of the world. In so doing, it will
have to choose qR to maximise
v ( qR ) − S ( qR ) qR − tqR (3)
42The notion of added value was popularised by Adam Brandenburger and Barry Nalebuff, Coopetition,
Doubleday: New York, 1996.
July 2000 40
Let q −R be the quantity that maximised this function. Of course, the

maximised value of the function is V − . Notice that
V * − V− = v144
( q*A +42444 −
q R* ) − v( q3 S ( qR− ) qR− − S ( q*A ) qA* − S (qR* )3
R ) + 144444244444 qR* + t1 −
( q4
R2−4*
q3
R) (4)
Fall in buyer benefits Increase in supply costs Increase in transport costs
because q *A + q*R > qR− and q −R > q*R . So a high price, P, is driven by key
supply and demand characteristics as well as the bargaining ability, α, of
the single desk.
Note that when the buyer can obtain all of the wheat supplied by
AWB from a similar location and for the same quantity, AWB’s added
value would be zero. In this case, price would be equal marginal cost;
which at the value maximising quantity is equal to rest of world marginal
cost.
6.2 Contestable Trading
Suppose now that instead of bargaining with a single desk, export

trading from Australia is perfectly contestable so that the buyer can
potentially negotiate with a number of traders. Suppose that the buyer can
frictionlessly switch between any trader and still have access to all
Australian wheat. Then it is no longer the case that, in negotiations with
any one trader, the buyer would be denied supply of Australian wheat if
negotiations broke down. Instead, with contestable trading, the buyer
would be able to procure their entire desired supply with no additional
cost.
This means that the added value of any one trade would fall to
zero. In this case,
P = S ( q*A ) = S (q R* ) + t (5)
Thus, price is driven the marginal cost.
Comparing this with the single desk outcome, we can calculate the
theoretical single desk premium:
July 2000 41
 V * −V −   v( q*A + q*R ) − v ( q−R) + S ( q R− ) − S ( q*A ) − S (q *R ) + t ( qR− − q*R ) 

α *  = α   (6)
 qA   q*A 
If the absence of a single desk led to a contestable trading market, then it is

this that growers would lose in terms of a price premium.
6.3 Few Traders
Of course, the existence of economies of scale in trading makes it

unlikely that trading would become perfectly contestable if single desk
requirements were removed. In this case, the fall in P following the
removal of a single desk is likely to be smaller. Hence, the single desk
premium is likely to be smaller and perhaps take a different form than
that above.
To see this suppose there were two traders. We will assume for
simplicity that each cannot exercise market power per se.43 In that case, if
negotiations with any single trader broke down, the buyer would be faced
with the price calculated in the previous section. Suppose that α = ½, then
the new price would be determined by the following (Nash bargaining)
equation:44
( )
v (q *A ) − Pq *A − v( q*A ) − 12 (V * − V − ) − ( S ( qR* ) + t )q A* = P − ( S ( qR* ) + t )qA*
(7)
⇒P= 1
4 q *A
(V *
− V − ) + S ( qR* ) + t
where V * − V − is as defined in (4). If there are N traders, then this price

becomes:
P= 1
2 N q*A
(V *
− V − ) + S (q *R ) + t (8)
43 For AWB, this is reasonable so long as its statutory purchase obligations remain in force.
44 This solution is based on the model of Lars Stole and Jeff Zwiebel and has been extended to the monopoly
situation by Joshua Gans and Catherine de Fontenay. See Lars Stole and Jeffrey Zwiebel, “Organizational
Design and Technology Choice under Intrafirm Bargaining,” American Economic Review, 86 (1), 1996,
pp.195-222; and Joshua S. Gans and Catherine de Fontenay, “Extending Market Power through Vertical
Integration,” Working Paper, Melbourne Business School, February 1999.
July 2000 42
In this case, the single desk premium becomes: ( 1

2
− 21N ) (V
1
q *A
*
− V − ) . This is
a smaller amount than (6) calculated for the contestable trading case. Note
that if only a single new trader entered following the removal the single
desk, the premium would fall by 50 percent.
July 2000 43
Section 7 Technical Appendix B: Empirical Analysis (with Joseph Hirschberg)
7 Technical Appendix B: Empirical Analysis

(with Joseph Hirschberg)
In this study we attempt to construct a model to explain the

variation in the price premium for Australian wheat. The premium is
defined as the differential between the FOB price in the contract for the
shipment and the price on the market for Pacific Northwest Wheat (US)
on the day of the contract. The purpose of this exercise is to account for as
much of the variation in the price premium that is due to conditions over
which the AWB has no control and then to determine what proportion
they are able to influence via the institution of the single desk.
This appendix proceeds as follows. First, we describe the data used

in the analysis and what assumptions have been made in its construction.
Second, we define the model used in the analysis. Third we report on the
results of the models run and draw conclusions from the estimates that we
have made.
7.1 Data
The data used were contracts for wheat engaged by the AWB from
late 1996 to late 1999. A total of 2,660 records were made available each
representing a shipment. In some cases a single contract would result in
multiple shipments and these values were then combined to form 2,317
observations on a contract level. Each observation identified the agent
(which may be a national agent for an entire country), the country, the
customer (which may in some cases be the country), the FOB price agreed
to in the contract, the price premium defined as the difference between the
FOB price and the Pacific Northwest Wheat price in the US, the date of the
contract (which if missing was estimated as 60 days prior to the shipment
date), the total amount shipped, the total value of the shipment, the total
value of the premium, the % of the premium over the US price, the US
FOB price used for the comparison, and the number of shipments made
for the contract. Table 1 summarizes these quantities.
July 2000 44
Table 1. Summary of the Contract Data
Label N Mean Std Dev Minimum Maximum
Total Tonnes Shipped 2317 20,002 62,364 0 1,414,538
Total Value Of Premium 2308 49,352 467,369 -6,253,832 13,763,464
Total Value Shipped 2317 3,027,854 9,048,796 0 201571803
Fob Price 2317 156.36 28.34 73 325
% Premium 2308 4.35 8.32 -25.90 115.4301872
Premium To PNW Base 2308 6.17 12.41 -33.28 171.46
Base US FOB Price 2308 150.24 26.73 90.76 280.71
Number Of Shipments 2317 1.14 1.13 1 33
Due to missing data only 2,308 contracts had sufficient information

to use them in the analysis. In addition to this information we also had the
grade of the wheat sold. Table 2 lists the total tonnes of wheat shipped
under the contracts for each grade as well as the number of contracts and
their average % premium. Unfortunately, the comparison of the %
premiums from this table does not provide a very clear picture of the
value of these premiums because of all the other reasons for the %
premium to vary.
In addition to the data provided on contracts we also used data

concerning the share of the total wheat sales imported by each country
that originated in Australia. These values were converted into shares of
the total market that made up the Australian contribution. Table 3 lists the
market share values based on the market share in from 1994 to 1996.
July 2000 45
Table 2. Wheat Sales by Grade
Grade Total Tonnes % Total Number % Of All % Premium

Tonnes Contracts Contracts
Average St Dev Max Min
A Gen Pur 4,741,487 10.24 107 4.64 0.68 5.65 20.20 -18.35
A Hard 5,067,144 10.95 430 18.63 1.06 3.72 34.33 -9.94
A Prime Hard 3,220,714 6.96 294 12.74 6.90 7.36 20.04 -9.69
A P Wht 9,370,155 20.24 380 16.46 1.64 8.76 115.43 -17.96
A S Wht 22,580,362 48.78 721 31.24 4.98 6.70 23.09 -12.95
A S Wht 74,978 0.16 64 2.77 19.96 7.33 38.90 7.22

Noodle
Durum 372,968 0.81 29 1.26 24.12 19.81 58.47 -7.37
Feed 100,934 0.22 13 0.56 4.44 10.00 14.38 -25.90
Soft 757,967 1.64 270 11.7 4.82 8.41 42.88 -16.38
The other information needed for each contract was the relative
location of the market to Australia, the US, and Europe (using France for
reference). Using the Great Circle distance between capital cities45 we
derived the set of distances for the countries given in Table 4. Note that for
Cambodia and Vietnam we used a bordering country for reference.
45 These data are provided by Jon Haveman at his international trade data web site
http://www.eiit.org/Trade.Resources/TradeData.html.
July 2000 46
Table 3. The Average Proportion of Imports of Wheat Originating in

Australia (1994-1996)
Country Market Share Country Market Share
BAHRAIN 0.7105 NEW CALEDONIA 1.0000
BANGLADESH 0.0946 NEW ZEALAND 0.8650
CAMBODIA 0.4286 OMAN 1.0216
CHINA 0.1781 PAKISTAN 0.0595
CIS & BALTICS 0.0550 PHILIPPINES 0.0477
COLUMBIA 0.0843 PNG 0.9715
EGYPT 0.1184 QATAR 0.9091
ETHIOPIA 0.0496 SINGAPORE 0.6574
FIJI 0.8800 SOMALIA 0.3164
GREECE 0.0594 SOUTH AFRICA 0.1956
INDONESIA 0.4323 SOUTH AMERICA 0.0170
IRAN 0.5221 SRI LANKA 0.1424
IRAQ 0.1954 SUDAN 0.0249
ITALY 0.0981 TAIWAN 0.0577
JAPAN 0.2013 THAILAND 0.2314
JORDAN 0.0452 TURKEY 0.0448
KOREA SOUTH 0.2133 VIETNAM 0.1623
KUWAIT 0.3310 YEMEN 0.1024
MALAYSIA 0.6044 ZIMBABWE 0.1854
MAURITIUS 0.2236 TOTAL TRADE 0.1172
MOROCCO 0.0027
MOZAMBIQUE 0.1147
July 2000 47
Table 4. Distances to Australia, US and EU
Country Distance to (km) Distance Advantage

over Australia
US Australia EU US EU
(France) (France)
USA 0 15,958 6,169 15,958 9,789

Germany 6,406 16,557 400 10,152 16,157
Japan 10,910 7,966 9,723 -2,945 -1,757
United.Kingdom 5,904 17,004 341 11,100 16,663
France 6,169 16,943 0 10,774 16,943
Italy 7,222 16,239 1,108 9,016 15,131
Canada 734 16,123 5,652 15,389 10,470
China 11,154 9,025 8,222 -2,129 803
Netherlands 6,198 16,648 428 10,450 16,220
Hong.Kong 13,129 7,396 9,642 -5,733 -2,247
Belgium-Lux 6,221 16,741 261 10,520 16,480
Singapore 15,561 6,222 10,749 -9,339 -4,527
Spain 6,096 17,593 1,055 11,497 16,538
Korea.RP.(S) 11,174 8,430 8,975 -2,744 -544
Mexico 3,039 13,183 9,207 10,145 3,976
Taiwan 12,659 7,325 9,835 -5,334 -2,510
Switzerland 6,607 16,615 441 10,008 16,174
Austria 7,130 15,936 1,036 8,806 14,900
Sweden 6,641 15,632 1,544 8,991 14,088
Thailand 14,169 7,489 9,455 -6,679 -1,966
Australia 15,958 0 16,943 -15,958 -16,943
Malaysia 15,350 6,540 10,434 -8,810 -3,894
USSR 7,827 14,504 2,487 6,676 12,017
Saudi.Arabia 10,857 12,647 4,693 1,791 7,954
Turkey 8,733 14,516 2,602 5,782 11,913
Indonesia 16,371 5,408 11,591 -10,962 -6,183
Brazil 6,799 14,066 8,733 7,268 5,333
Denmark 6,518 16,057 1,028 9,539 15,030
Norway 6,238 15,998 1,343 9,759 14,655
Portugal 5,742 18,074 1,454 12,332 16,620
Ireland 5,448 17,261 778 11,812 16,483
Israel 9,452 14,062 3,284 4,610 10,779
Argentina 8,402 11,738 11,064 3,337 674
Finland 6,938 15,233 1,912 8,295 13,322
Phillipines 13,793 6,300 10,757 -7,493 -4,457
United.Arab.Emirates 11,355 11,971 5,246 616 6,725
Greece 8,261 15,224 2,100 6,963 13,123
South.Africa 12,723 10,779 9,352 -1,944 1,427
India 12,051 10,374 6,589 -1,678 3,784
Poland 7,183 15,581 1,368 8,398 14,212
July 2000 48

over Australia
(France) (France)
Yugoslavia 7,597 15,614 1,450 8,016 14,163

Venezuela 3,317 15,460 7,625 12,143 7,835
Hungary 7,344 15,736 1,250 8,392 14,486
Iran 10,190 12,826 4,219 2,636 8,607
Czechoslovakia 6,907 16,057 887 9,150 15,170
Chile 8,079 11,327 11,659 3,247 -333
Panama 3,341 14,298 8,671 10,958 5,627
Colombia 3,829 14,435 8,639 10,606 5,796
Pakistan 11,389 11,020 5,933 -368 5,087
Egypt 9,358 14,284 3,216 4,926 11,068
New.Zealand 14,098 2,324 19,007 -11,774 -16,684
Morocco 6,109 17,897 1,885 11,788 16,012
Algeria 6,800 16,997 1,347 10,197 15,650
Romania 7,985 15,168 1,874 7,183 13,294
Nigeria 8,737 15,286 4,722 6,549 10,563
Tunisia 7,347 16,369 1,486 9,022 14,883
Kuwait 10,533 12,794 4,410 2,260 8,384
Liberia 7,536 16,195 4,911 8,659 11,284
Peru 5,671 12,861 10,264 7,190 2,597
Dominican.Republic 2,376 15,763 7,189 13,387 8,574
Oman 11,681 11,559 5,616 -122 5,942
Jordan 9,540 13,956 3,372 4,416 10,585
Cyprus 9,122 14,309 2,955 5,187 11,354
Syrian.Arab.Republic 9,445 13,991 3,280 4,546 10,711
Ecuador 4,357 13,709 9,366 9,352 4,343
Bangladesh 12,943 9,026 7,919 -3,917 1,106
Uruguay 8,488 11,788 10,970 3,300 818
Bahamas 1,538 15,460 7,211 13,921 8,249
Costa.Rica 3,300 13,967 8,925 10,667 5,042
Sri.Lanka 14,402 8,627 8,524 -5,775 103
Malta 7,729 15,968 1,753 8,239 14,215
Paraguay 7,421 12,721 10,170 5,300 2,551
Guatemala 3,110 13,628 9,084 10,518 4,543
Bulgaria 7,920 15,361 1,762 7,440 13,599
Bahrain 10,968 12,398 4,838 1,430 7,560
Jamaica 2,326 15,124 7,732 12,798 7,393
Reunion.Islands 15,095 9,015 9,375 -6,081 -360
Yemen 11,450 12,340 5,320 890 7,020
El.Salvador 3,049 13,751 8,984 10,702 4,767
Bolivia 6,235 13,050 10,037 6,815 3,012
Honduras 2,936 13,954 8,808 11,018 5,146
Trinidad.And.Tobago 3,501 15,892 7,224 12,391 8,668
July 2000 49

over Australia
(France) (France)
Zimbabwe 12,835 11,194 7,966 -1,641 3,229

Qatar 11,109 12,263 4,980 1,154 7,284
Ghana 8,488 15,471 4,828 6,983 10,644
Cote.D'Ivour 8,175 15,709 4,884 7,533 10,825
Iceland 4,518 16,770 2,234 12,253 14,537
Kenya 12,152 11,942 6,491 -211 5,451
Myanmar 13,796 8,053 8,893 -5,743 -840
Mauritius 15,224 8,882 9,424 -6,343 -542
Senegal 6,379 17,373 4,213 10,994 13,160
Zambia 12,400 11,628 7,601 -772 4,028
Tanzania 12,759 11,348 7,163 -1,412 4,185
Angola 10,653 13,289 6,507 2,636 6,782
Ethiopia 11,530 12,451 5,577 921 6,874
Cameroon 9,622 14,450 5,084 4,828 9,366
Gabon 9,686 14,332 5,440 4,647 8,892
Mozambique 13,428 10,497 8,856 -2,931 1,641
Papua.New.Guinea 14,560 2,884 14,567 -11,676 -11,683
Sudan 10,541 13,414 4,608 2,873 8,806
Benin 8,669 15,345 4,718 6,676 10,627
Barbados 3,345 16,232 6,882 12,887 9,349
Suriname 4,288 15,911 7,196 11,624 8,716
Congo 10,515 13,512 6,046 2,996 7,465
Zaire 10,519 13,508 6,050 2,989 7,457
Guinea 7,050 16,677 4,628 9,627 12,049
Iraq 9,978 13,293 3,866 3,315 9,427
Nicaragua 3,115 13,916 8,901 10,801 5,015
Mauritania 6,282 17,619 3,800 11,337 13,819
Nepal 12,396 9,702 7,243 -2,693 2,460
Madagascar 14,291 9,815 8,770 -4,476 1,045
Togo 8,572 15,418 4,758 6,846 10,660
Guyana 4,043 15,805 7,317 11,762 8,488
Malawi 12,781 11,299 7,651 -1,482 3,648
Fiji 12,511 3,460 16,595 -9,052 -13,135
Haiti 2,312 15,553 7,362 13,241 8,191
Afghanistan 11,153 11,370 5,590 218 5,780
Mali 7,328 16,623 4,142 9,295 12,481
Belize 2,670 13,998 8,664 11,328 5,334
Laos 13,704 7,702 9,245 -6,002 -1,542
Djibouti 11,674 12,208 5,596 533 6,612
Gambia 6,535 17,210 4,308 10,676 12,903
Niger 8,146 15,939 3,935 7,793 12,005
Uganda 11,679 12,423 6,131 744 6,292
July 2000 50

over Australia
(France) (France)
Burkina.Faso 7,908 16,134 4,078 8,226 12,056

Seychelles 15,095 9,015 9,375 -6,081 -360
Guinea.Bissau 6,730 17,002 4,440 10,272 12,562
Mongolia 10,363 10,175 7,053 -188 3,122
Sierra.Leone 7,162 16,553 4,724 9,391 11,829
Rwanda 11,629 12,481 6,256 852 6,225
Somalia 12,600 11,401 6,622 -1,199 4,780
Comoros 13,442 10,667 7,845 -2,775 2,823
Burundi 11,670 12,438 6,376 768 6,062
Chad 9,351 14,757 4,256 5,406 10,501
Central.African.Repu 10,196 13,911 5,187 3,715 8,724
Solomon.Islands 13,514 3,084 15,174 -10,430 -12,091
Bhutan 12,513 9,375 7,568 -3,138 1,808
7.2 The Model
In order to model the determinants of the price premium we

employ a multiple regression model. This method will allow us to identify
what proportion of the variation in price premiums is due to location of
the market, the grade of the wheat sold to market and the ability of the
AWB to extract a profit due to the institution of the single desk. The
advantage of a multiple regression framework is that we can condition the
premium on all the other factors that might have an influence on the
premium. This form of model is a well-established method for
determining the value of a characteristic of a market that cannot be found
directly. This type of analysis is often called an hedonic price model and it
is formed by specifying the market clearing price of a good or service as a
function of the good or service.46
46 A survey of the use of hedonic price models can be found in chapter 4 of The Practice of Econometrics:
Classic and Contemporary, by Ernst R. Berndt, Addison Wesley Publishing, 1990.
July 2000 51
Hedonic price analysis is used in cases were the good or service

sold has a market that only exists for that sale. A good example is in the
sale of houses. No two houses are really the same. Thus the value of a
house will depend on the characteristics of the house and the nature of the
demand and supply when the house is sold. What is often done to value a
home is to look at how the market values the features of the house. For
example, what value is an extra bedroom or an extra bath? How is the
proximity of the house to neighbourhood amenities valued? Thus even
though a particular house may not have been on the market for some time
it is possible to use current sales of other houses and their characteristics
to estimate the components of the sales price given current supply and
demand. In many countries real estate evaluations for tax purposes are
based on such models. Another example of the use of hedonic price
models is in the determination of price indexes for goods that change in
character over time. Computers a good example of such an item. If one
tried to match the price of the same computer over time it would be hard
to find the same model for sale more than a year after they were
introduced. Thus the usual method of determining the price of a “typical
basket of goods” over time to construct the price index would not work.
This problem can be solved by determining the value of the characteristics
of the computers at one point in time and then estimating what a
particular computer would be priced if it was available at another point in
time.
In the case of the impact of the single desk operations of AWB we

are interested in modelling the characteristics of each sale that determined
the size of the % price premium. As shown in Table 1 we observe on
average a 4.35% price premium per contract however to what extent is this
due to location advantages given the origin of shipments in Australia
versus the US or the EU that are not accounted for by the FOB prices. In
addition, one would expect that seasonal variation in Australian supply
and customer demand will be important as well. Based on Table 2 we also
know that average % premiums vary by grade of wheat in some dramatic
ways although the markets for these high value grades are small unless
we account for different grades of wheat we may not account for sufficient
variation in the price premiums.
In the analysis conducted here we limit our attention to linear

models that predict of the % premium. We use the % premium so that the
influence of the nominal prices is accounted for. We define this as the %
July 2000 52
over the US Pacific Northwest FOB price of wheat. The linear model we
specify has the following form:
42 2
yi = α + ∑ λ j D ji + β14243
1 x i +β 2xi + φ11 z14
14 i 244 z22i + φ
+ φ12 3 21 z24
14 i 244
2
+ φ22 z3
2i + ρ{Ji
j =1
1424 3 US FOB DA for US DA for EU Japan
Date
8
+ ∑ θ j G ji + π1q i + π2qi2 + γ1 si + γ 2 si2 + ψ1 mi + ψ 2 mi2 + εi
j =1 1444424444 3 14 4244 3
1 424 3 Size Market Share
Grade
where the yi are the % premiums
The Dji are dummy variables for each month (Dji = 1 when the
month is month j and zero otherwise), xi is the US FOB price, z1i is the
distance advantage for the US from Table 4, z2i is the distance advantage
for the EU from Table 4 and z12i = sign ( z1i ) z12i , z22i = sign( z2 i ) z22i , these are
the square of the distance advantage that retains the same sign, and Ji is a
dummy variable for the when the contract is with Japan (Ji = 1 when the
contract is for a Japanese trader and zero otherwise). The preceding
variables are all factors that affect the international wheat from the
perspective of the Australian producer with out regard for any factor
controlled by the AWB.
The Gji are the dummy variables for the Grade (Gji=1 when the
contract is for grade j and zero other wise), q is the total number of tonnes
shipped for the contract, si is the number of separate shipments used, and
mi is the market share for Australian Wheat as given in Table 3. The
α, β, φ, λ, π, γ , ρ , ψ, and θ are the parameters of the model to be
estimated.
This specification is similar to the form of models used in empirical

international trade literature. Specifically the use of quadratic forms in the
distance differences matches the usual justifications for a non-linear effect
of distance on trade. Given the lack of specific freight charges we assume
that freight costs will be quadratic in nature and the estimation bears this
out in that the squared term is always statistically significant. The
quadratic terms used for the relationship for size, and US FOB price are
also based on the logic that these variables will have some non-linear
effects. The dummy variable for Japan was introduced after it was found
that trade with Japan was quite different from that to most other nations.
July 2000 53
Thus the distance variables don’t capture enough of this difference. Other
specifications were tried where the agents were used as dummy variables
although this method resulted in a higher proportion of the variation
explained by the model it also eliminated the possibility of decomposing
the influence of trading procedures of the AWB.
7.3 Results
The resulting estimates from the model described above can be

used to determine the degree to which the total variation in the % price
premiums are accounted for by those factors over which the AWB does
not have an influence namely the date of the contract, the US FOB price,
and the distances to the customer. Those factors that the AWB can
influence are the grades that it classifies and the size of the shipment that
it can coordinate. When we fit this model we found that the % of the total
variation in the % price premium that was accounted for by our model
was 60.84%. Thus, a little less than 40% of the premium was explained by
the error.
In Table 5 one can see that a large proportion of the models ability
to explain the variation in % price premium comes from the distances and
the grade of wheat. Note that this is a sequential analysis of variance
(ANOVA). This means that the amount explained by each factor is after
the others have been determined. Thus the 18.33% explained by the
distance advantage is after the Date and the US FOB price level has been
accounted for. Note that aside for the Size effect all the other factors are
significant at least the .01 level. This means that we would have a low
probability of making an error if we concluded that these factors influence
the variance of the % price premium.
July 2000 54
Table 5. The Contribution of Each Factor in a Sequential ANOVA
Part of Model 47 % of Total

Variation
Explained
Date 6.51
US FOB 0.81
Distance Advantages 18.33
Japan 26.11
Grade of Wheat 8.57
Size .218
Market share .286
Total Model 60.84
The full ANOVA results for these sequential additions (Type I

Sums of Squares) and the alternative ANOVA where each factor is
introduced as the last (Type III Sums of Squares) are given in Table 9
below. The full set of parameters cannot be estimated for the dummy
variables due to the exhaustive set of dummies. However the program for
these results (SAS) uses a generalized inverse routine that drops the last
dummy automatically and we have provided the tables of the parameter
estimates in Table 10.
In addition to being able to establish what the effect of these

various effects are we can also estimate an average % price premium for
each grade of wheat. This is done by using the estimated parameters from
the model above and the average values for US FOB, the distances, and
the measures of the size of the shipment. For the date dummy variables it
is assumed that each has an average effect thus each coefficient has a
weight of 1/42. The results of this estimation is given in Table 6.
47 All of these effects are shown to be statistically significant at at lease the .05 level.
July 2000 55
Table 6. The Average Mean Of The % Premium
Grade Computed Estimated Standard

mean Error
A Gen Pur 1.83 0.59
A Hard 4.57 0.38
A P Wht 3.90 0.38
A Prime Hard 5.33 0.45
A S Wht 4.35 0.31
A S Wht 8.09 0.79

Noodle
Durum 28.52 1.26
Feed 0.13 1.59
Soft 5.61 0.43
In addition to average effects of grade differences we can also

determine the degree of variation in the % price premiums by month. The
?’s estimated in the model and as reported in Tables 9 and 10 provide an
indication of those factors that vary in the international wheat market by
month. These are determinants of the price premiums that are not a
function of the Pacific NW FOB prices that are already in the model. They
may be interpreted as a combination of seasonal factors of demand and
supply and the actions of the AWB. At this point we have no means of
decomposing these to influences. The time plot of these means is given
below excluding those observations prior to 1997 is given in Figure 2 and
the estimates and their standard errors are given in Table 7. Note that the
estimated standard errors for these predicted values are relatively small.
July 2000 56
Figure 2. Predicted % Price Premiums by Month since January, 1997
10
8
Predicted % Price Premium
0
JUL 97
JUL 98
JUL 99
MAY 97
MAY 98
MAY 99
MAR 97
MAR 98
MAR 99
JAN 97
JAN 98
JAN 99
SEP 97
NOV 97
SEP 98
NOV 98
SEP 99
Month
July 2000 57
Table 7. Estimated Average % Premiums by Month
Month Estimated SE Feb-98 6.95 0.63

Mean Mar-98 7.21 0.78
Jan-96 8.63 5.32 Apr-98 6.77 0.87
Feb-96 9.93 5.32 May-98 6.84 0.76
Jun-96 16.89 5.37 Jun-98 5.73 0.69
Aug-96 12 3.1 Jul-98 6.46 0.75
Sep-96 10.69 1.33 Aug-98 4.89 0.78
Oct-96 9.85 0.99 Sep-98 4.51 0.83
Nov-96 7.53 0.76 Oct-98 4.8 0.74
Dec-96 7.22 0.72 Nov-98 4.03 0.64
Jan-97 6.95 0.74 Dec-98 4.25 0.68
Feb-97 7.89 0.73 Jan-99 4.73 0.66
Mar-97 8.96 0.81 Feb-99 4.25 0.73
Apr-97 7.45 0.93 Mar-99 3.51 0.82
May-97 10.14 0.81 Apr-99 3.97 0.66
Jun-97 6.8 0.73 May-99 4.37 0.7
Jul-97 8.57 0.77 Jun-99 4.88 0.76
Aug-97 8.44 0.74 Jul-99 6.23 0.78
Sep-97 8.11 0.67 Aug-99 4.24 0.84
Oct-97 9.09 0.71 Sep-99 5.92 0.89
Nov-97 8.77 0.74 Oct-99 3.07 1.18
Dec-97 9.13 0.81
Jan-98 7.7 0.65
An alternative model was fit without the market share and size of
shipment measures to determine what the price premiums for the various
grades would be without these influences on the results. Table 8 given
below provides the price premiums from that case. Note that this model
results in almost identical values for these grades.
July 2000 58
Table 8. The Average Mean Of The % Premium Using A Model

Without The Size Or Share Of Market Factors
Grade Computed Estimated Standard

mean Error
A Gen Pur 1.24 0.58
A Hard 4.49 0.37
A P Wht 3.87 0.39
A Prime Hard 5.29 0.45
A S Wht 4.26 0.31
A S Wht 8.23 0.79

Noodle
Durum 28.55 1.25
Feed 0.48 1.58
Soft 5.82 0.43
July 2000 59
Table 9. The Full Set of ANOVA and Parameter Estimates for the Model
Estimated
Source DF Squares Mean Square F Value Pr > F

Model 63 96755.3590 1535.7993 55.11 <.0001
Error 2235 62280.0143 27.8658
Corrected Total 2298 159035.3733
R-Square Coeff Var Root MSE p_prem Mean

0.608389 120.4149 5.278805 4.383847
Source DF Type I SS Mean Square F Value Pr > F

yr_c*mon_c 42 10358.24371 246.62485 8.85 <.0001
us_fob 1 1144.04336 1144.04336 41.06 <.0001
us_fob*us_fob 1 156.08191 156.08191 5.60 0.0180
D_AU_US 1 5.70776 5.70776 0.20 0.6509
d_au_us2 1 10405.16839 10405.16839 373.40 <.0001
D_AU_EU 1 8185.35959 8185.35959 293.74 <.0001
d_au_eu2 1 10570.03751 10570.03751 379.32 <.0001
japan 1 41509.23675 41509.23675 1489.61 <.0001
NGRADE 8 13615.22110 1701.90264 61.08 <.0001
tot_s 1 42.19381 42.19381 1.51 0.2186
tot_s*tot_s 1 209.88148 209.88148 7.53 0.0061
n_ship 1 94.51949 94.51949 3.39 0.0656
n_ship*n_ship 1 2.30782 2.30782 0.08 0.7735
a_share 1 25.45181 25.45181 0.91 0.3393
a_share*a_share 1 431.90453 431.90453 15.50 <.0001
yr_c*mon_c 42 4002.53779 95.29852 3.42 <.0001
us_fob 1 339.90603 339.90603 12.20 0.0005
us_fob*us_fob 1 23.14439 23.14439 0.83 0.3622
D_AU_US 1 1192.23868 1192.23868 42.79 <.0001
d_au_us2 1 451.31450 451.31450 16.20 <.0001
D_AU_EU 1 259.37781 259.37781 9.31 0.0023
d_au_eu2 1 165.14071 165.14071 5.93 0.0150
japan 1 33238.79195 33238.79195 1192.82 <.0001
NGRADE 8 12683.82326 1585.47791 56.90 <.0001
tot_s 1 229.70469 229.70469 8.24 0.0041
tot_s*tot_s 1 179.62967 179.62967 6.45 0.0112
n_ship 1 7.12271 7.12271 0.26 0.6132
n_ship*n_ship 1 0.30161 0.30161 0.01 0.9171
a_share 1 456.02613 456.02613 16.37 <.0001
a_share*a_share 1 431.90453 431.90453 15.50 <.0001
July 2000 60
Table 10: Estimates
Parameter Estimate Stnd Error t Value
Intercept 12.51673576 4.7656474 2.63

yr_c*mon_c 1996 1 9.16669922 5.94761015 1.54
yr_c*mon_c 1996 2 10.46567541 5.94882235 1.76
yr_c*mon_c 1996 6 17.42221217 6.01675765 2.9
yr_c*mon_c 1996 8 12.53440456 4.10296359 3.05
yr_c*mon_c 1996 9 11.223484 2.99961196 3.74
yr_c*mon_c 1996 10 10.37892479 2.86094035 3.63
yr_c*mon_c 1996 11 8.06517728 2.7856163 2.9
yr_c*mon_c 1996 12 7.75538413 2.77247794 2.8
yr_c*mon_c 1997 1 7.48113187 2.77552482 2.7
yr_c*mon_c 1997 2 8.42496415 2.77693481 3.03
yr_c*mon_c 1997 3 9.49053212 2.79922252 3.39
yr_c*mon_c 1997 4 7.98701478 2.84085663 2.81
yr_c*mon_c 1997 5 10.67621849 2.79991433 3.81
yr_c*mon_c 1997 6 7.32928949 2.76120252 2.65
yr_c*mon_c 1997 7 9.10452725 2.7621868 3.3
yr_c*mon_c 1997 8 8.97650031 2.76395264 3.25
yr_c*mon_c 1997 9 8.6410192 2.74620581 3.15
yr_c*mon_c 1997 10 9.62321558 2.7575556 3.49
yr_c*mon_c 1997 11 9.29851539 2.76304347 3.37
yr_c*mon_c 1997 12 9.66031208 2.77789607 3.48
yr_c*mon_c 1998 1 8.23616059 2.7317299 3.01
yr_c*mon_c 1998 2 7.47947507 2.72476481 2.74
yr_c*mon_c 1998 3 7.74113826 2.76293809 2.8
yr_c*mon_c 1998 4 7.30124663 2.78337947 2.62
yr_c*mon_c 1998 5 7.36999384 2.74470254 2.69
yr_c*mon_c 1998 6 6.26205738 2.72379031 2.3
yr_c*mon_c 1998 7 6.99481575 2.73138187 2.56
yr_c*mon_c 1998 8 5.4242801 2.73138637 1.99
yr_c*mon_c 1998 9 5.03956609 2.74613896 1.84
yr_c*mon_c 1998 10 5.33259676 2.74295939 1.94
yr_c*mon_c 1998 11 4.55920127 2.70782008 1.68
yr_c*mon_c 1998 12 4.78113328 2.72270142 1.76
yr_c*mon_c 1999 1 5.26011869 2.71381768 1.94
yr_c*mon_c 1999 2 4.77836616 2.73014281 1.75
yr_c*mon_c 1999 3 4.04502468 2.75443652 1.47
yr_c*mon_c 1999 4 4.50001585 2.70529216 1.66
yr_c*mon_c 1999 5 4.90551507 2.71516352 1.81
yr_c*mon_c 1999 6 5.41296134 2.73501307 1.98
yr_c*mon_c 1999 7 6.75768563 2.72844486 2.48
yr_c*mon_c 1999 8 4.76834832 2.75852114 1.73
yr_c*mon_c 1999 9 6.45119357 2.77837207 2.32
yr_c*mon_c 1999 10 3.60429031 2.86931071 1.26
July 2000 61
Parameter Estimate Stnd Error t Value
yr_c*mon_c 1999 11 0. .
us_fob -0.16944941 0.04851725 -3.49
us_fob*us_fob 0.00013235 0.00014523 0.91
D_AU_US -0.00139007 0.00021252 -6.54
d_au_us2 0.00000005 0.00000001 4.02
D_AU_EU 0.00044905 0.00014719 3.05
d_au_eu2 0.00000002 0.00000001 2.43
japan 16.16989869 0.46818799 34.54
NGRADE A Gen Pur -3.78390754 0.65687868 -5.76
NGRADE A Hard -1.04151945 0.47181353 -2.21
NGRADE A P Wht -1.70875796 0.49734812 -3.44
NGRADE A Prime Hard -0.28024978 0.5791735 -0.48
NGRADE A S Wht -1.26736164 0.40431127 -3.13
NGRADE A S Wht Noodle 2.48257395 0.86139373 2.88
NGRADE Durum 22.9066306 1.30433535 17.56
NGRADE Feed -5.48556312 1.58430372 -3.46
NGRADE Soft 0. .
tot_s -0.00001352 0.00000471 -2.87
tot_s*tot_s 0 0 2.54
n_ship 0.14537308 0.28753912 0.51
n_ship*n_ship -0.00103798 0.00997697 -0.1
a_share 10.63799146 2.62966461 4.05
a_share*a_share -10.68253821 2.71341464 -3.94
July 2000 62
Attachment A: Recent Research
Selected Recent Research Papers and Publications

by Joshua Gans48
Competition Policy and Regulation:
Joshua Gans and Stephen King, “Options for Electricity Transmission Regulation in
Australia,” Australian Economic Review, June 2000 (forthcoming).
Joshua Gans, “The Competitive Balance Argument for Mergers,” Australian Economic
Review, Vol.33, No.1, March 2000, pp.83-93.
Teresa Fels, Joshua Gans and Stephen King, “The Role of Undertakings in Regulatory
Decision-Making,” Australian Economic Review, Vol.33, No.1, March 2000, pp.3-16.
Joshua Gans and Richard Scheelings, “Economic Issues Associated with Access to
Electronic Payments Systems,” Australian Business Law Review, Vol.27, No.5,
December 1999, pp.373-390.
Joshua Gans and Philip Williams, “Efficient Investment Pricing Rules and Access
Regulation,” Australian Business Law Review, Vol.27, No.4, August, 1999, pp.267-
279.
Joshua Gans and Philip Williams, Access Regulation and the Timing of Infrastructure
Investment,” Economic Record , Vol. 79, No.229, June 1999, pp.127-138.
Joshua Gans, “Regulating Private Infrastructure Investment: Optimal Pricing for Access
to Essential Facilities,” Working Paper, No.98-13, 1998, Melbourne Business
School.
Telecommunications:
Joshua Gans, Stephen King and Graeme Woodbridge, “Numbers to the People:
Ownership, Regulation and Local Number Portability,” Working Paper, No. 2000-
08, Melbourne Business School.
48 Many of these papers can be accessed on the web through www.mbs.unimelb.edu.au/jgans/research.htm.

Attachment A Recent Research
Joshua Gans and Stephen King, “Using ‘Bill and Keep’ Interconnect Arrangements to
Soften Network Competition,” Economic Letters, 2000 (forthcoming).
Joshua Gans and Stephen King, “Mobile Network Competition, Customer Ignorance and
Fixed-to-Mobile Call Prices,” Information Economics and Policy , 2000,
(forthcoming).
Joshua Gans, “Network Competition and Consumer Churn,” Information Economics and
Policy , 2000, (forthcoming).
Joshua Gans and Stephen King, “Termination Charges for Mobile Phone Networks:
Competitive Analysis and Regulatory Options,” Working Paper, No.99-19,
Melbourne Business School, December, 1999.
Joshua Gans and Stephen King, “Regulation of Termination Charges for Non-Dominant
Networks,” Working Paper, No.99-20, Melbourne Business School, December,
1999.
July 2000 64
Attachment B Consultant Profiles
Attachment B: Consultant Profiles
"Incumbency and R&D Incentives:

Licensing the Gale of Creative
JOSHUA GANS Destruction," (with Scott Stern)
Journal of Economics and
Joshua Gans graduated in 1994 with a Management Strategy, 2000
Ph.D. in Economics from Stanford (forthcoming).
University. His thesis concentrated on the
"Options for Electricity Transmission
determinants of economic growth and the
Regulation in Australia," (with Stephen
role of technological progress. He returned
King) Australian Economic Review,
to Australia as a lecturer in the School of
33, June 2000 (forthcoming).
Economics, University of New South
Wales. In 1996, he moved to Melbourne to “The Role of Undertakings in Regulatory
take up a position as an Associate Decision-Making,” (with Teresa Fels
Professor at the Melbourne Business and Stephen King), Australian
School, University of Melbourne. At Economic Review, 33, March 2000,
present, Joshua is Economics Editor of the pp.3-16.
Australian Journal of Management and on
"The Competitive Balance Argument for
the editorial board of Information
Mergers," Australian Economic
Economics and Policy.
Review, Vol.33, No.1, March 2000,
Joshua teaches MBA students introductory pp.83-93.
microeconomics and also several
“First Author Conditions,” (with Maxim
specialised subjects on innovation, game
Engers, Simon Grant and Stephen
theory and incentives and contracts. He
King), Journal of Political Economy,
has adapted a leading US textbook for the 107, 1999, pp.859-883.
Australasian context; Principles of
Economics (co-authored with Stephen "Efficient Investment Pricing Rules and
King, Greg Mankiw and Robin Stonecash; Access Regulation," (with Philip
published by Harcourt-Australia). Williams) Australian Business Law
Review, 27, August 1999, pp.267-279.
Joshua has also had extensive consulting
experience working on competition issues "Why Referees Don't Get Paid (Enough)"
for several law firms, private utilities and (with Maxim Engers), American
government agencies including the Economic Review, 88, December
Australian Competition and Consumer 1998, pp.1341-1349.
Commission. His work spans a wide
"Growth in Australian Cities" (with
variety of industries including Rebecca Bradley), Economic Record,
telecommunications, electricity, gas,
74, September 1998, pp.266-278.
pharmaceuticals, banking and financial
services. "Industrialisation with a Menu of
Technologies: Appropriate
On the research-side, Joshua specialises
Technologies and the Big Push",
in industrial organisation and applied game
Structural Change and Economic
theory. His particular interest is on the link
Dynamics, 9, 1998, pp.63-78.
between competition and innovation. He
has also researched on the publication "Contracts and Electricity Pool Prices"
process in academia and his papers are (with Danny Price and Kim Woods),
part of a forthcoming book Publishing Australian Journal of Management,
Economics (to be published by Edward 23, June 1998, pp.83-96.
Elgar in 2000). Some examples of his "Driving the Hard Bargain for Australian
publications include:
R&D,” Prometheus, 16, March 1998,
"Network Competition and Consumer pp.47-55.
Churn," Information Economics and
Policy, 2000, (forthcoming).
July 2000 65
“Bootstrapping Confidence Intervals for

Linear Programming Efficiency
JOSEPH HIRSCHBERG Scores: With an Illustration Using
Italian Banking Data”, 1997, (with G.
Joe Hirschberg received his Ph.D. in D. Ferrier), Journal of Productivity
Economics in 1984 from the University of Analysis, 8, pp 19-33.
Southern California. His thesis
concentrated on the impact of time-of-use "Detailed Patterns of Intra-industry Trade
electricity pricing. Since then he has held in Processed Food", (with J. R.
appointments at Southern Methodist Dayton), in Industrial Organization and
University and more recently as an Trade in the Food Industries, edited
Associate Professor in the Department of by Ian M. Sheldon and Phillip C.
Economics, University of Melbourne. Abbott, Westview Press, 1996, pp
141-159.
Joe teaches post-graduate and under-
graduate econometrics and statistics as "Industrial Determinants of International
well as specialised subjects in forecasting Trade and Foreign Investment by
and urban and regional economics. Food and Beverage Manufacturing
Firms", (with D. R. Henderson and P.
Joe has also had extensive consulting
R. Voros), in Industrial Organization
experience working on estimate issues for
and Trade in the Food Industries ,
several law firms, private firms and
edited by Ian M. Sheldon and Phillip
government agencies including the
C. Abbott, Westview Press, 1996, pp
Australian Competition and Consumer
197-215.
Commission and Productivity Commission.
His work spans a wide variety of industries "Fiscal Stress and the Production of Public
including railways, pharmaceuticals and Safety: a Distance Function
supermarkets. Approach", 1995, (with K. J. Hayes
On the research-side, Joe specialises in and S. Grosskopf), Journal of Public
developing technical econometric methods Economics, 57, pp. 277-296.
for the analysis of productivity in firms and "An Empirical Bayes Approach to
elsewhere as well as broader economic Analyzing Earnings Differentials for
relationships. Some of this work is Various Occupations and Industries",
reflected in his 1991 book (with D. J.
1994, (with D. J. Slottje), Journal of
Slottje, G. W. Scully, and K. J. Hayes)
Econometrics, 61, pp. 65-79.
Measuring the Quality of Life Across
Countries. His main published work "Occupational Segregation in the
includes: Multidimensional Case", 1994, (with
D. Boisso, K. J. Hayes, and J. Silber),
“Modelling time of day substitution using
Journal of Econometrics, 61, pp. 161-
the second moments of demand”,
171.
2000, Applied Economics, 22, 979-
986. "A New Method for Detecting Individual
and Group Labor Market
“Does The Technology Of Foreign-
Discrimination", 1994, (with D. J.
Invested Enterprises Spill Over To
Slottje, K. J. Hayes, and G. W. Scully),
Other Enterprises In China? An
Journal of Econometrics, 61, pp. 43-
Application Of Post-DEA Bootstrap
64.
Regression Analysis”, (with P. J.
Lloyd), in Modelling the Chinese "An Analysis of Bilateral Intra-industry
Economy, edited by P. J. Lloyd and X. Trade in the Food Processing Sector",
G. Zang, Edward Elgar Press, 2000, 1994, (with J. R. Dayton and I. M.
(forthcoming). Sheldon), Applied Economics, 26, pp.
159-167.
"Can We Bootstrap DEA Scores?, (with G.
D. Ferrier), 1999, Journal of "Simulating Changes in Conjectural
Productivity Analysis, 11, 81-92. Variations: The Case of Tariff and
Quota Nonequivalence in Food
“Earnings of Public Sector and Private
Processing", 1993, (with I. M. Sheldon
Sector Employees in Australia: Is
and S. McCorriston), American
There a Difference?”, 1998, (with Jeff
Journal of Agricultural Economics, 75,
Borland and Jenny Lye), The
pp. 1221-1225.
Economic Record, 74, 36-54.
.
July 2000 69
July 2000 1

Competition and Single Desk Selling of Wheat From Australia: An Economic Analysis

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Competition and Single Desk Selling of Wheat From Australia: An Economic Analysis

Загружено:

Авторское право:

Доступные форматы

Competition and Single Desk Selling of Wheat

from Australia: An Economic Analysis

6th July, 2000

This report was commissioned and funded by AWB (International) Limited.

2 Evaluating the Single Desk..................................................7

2.2 Marketing Advantages of a Single Desk ....................... 9

2.3 Logistics Advantages of a Single Desk.........................20

2.4 Overall Effect of Non-exclusivity on AWB...................24

3 Effect on Growers ................................................................25

4.2 International Purchasers...............................................30

4.3 Other Export Traders....................................................32

4.4 Overall Impact ..............................................................33

6 Technical Appendix A: Theory.........................................39

6.2 Contestable Trading......................................................41

6.3 Few Traders..................................................................42

7 Technical Appendix B: Empirical Analysis (with Joseph

7.2 The Model .....................................................................51

Attachment A: Recent Research....................................................63

Attachment B: Consultant Profiles...............................................65

In 2000, the Federal government announced a National

The WMA contains several sections that impose restrictions on the

• Export Controls: There are several ways in which the WMA

• Receival Standards: the WMA gives AWB Ltd the right to

The first set of provisions effectively imposes single desk selling of

receival standards ‘too high.’ In each case, the potential exclusionary

It should also be noted that the WMA contains some pro-

• AWB Purchase Obligation: Under s84 of the WMA, AWB

• WEA Regulatory Powers: AWB’s performance is subject to

As will be discussed later in the paper, the AWB’s purchase obligation is

Also of relevance for the economic analysis is the corporate

Second, and related to the first, is that the corporate objectives of

A final important feature of AWB for the purpose of economic

3 Constitution of AWB Limited, Article 2.

The paper proceeds as follows. First, the single desk itself is

2 Evaluating the Single Desk

The objective of the WMA is to “maximise the net returns to

Nonetheless, it is worth emphasising that broader issues of

• What determines the net returns to growers in the current

• What will change if the single desk provisions were removed?

Answering these questions involves describing the AWB’s current role, in

intrinsic advantages in marketing Australian wheat that cannot easily be

2.1 Classifying AWB’s Functions

An individual grower considering selling wheat internationally

2. The price paid by export traders (p) is high.

In actuality, a grower’s main concern is the price they receive, p; however,

AWB operates as an intermediary primarily focussed on export

• Marketing: this is AWB’s role as a negotiator and seller of Australian

• Logistics: this is AWB’s role as a coordinator of the international

2.2 Marketing Advantages of a Single Desk

A selling price, such as P (the international price paid for

2.2.1 Substitutes Available to Buyers

An exclusive right to sell, by definition, reduces the substitutes

• Location: other countries are located further from those buyers.

• Differentiation: Australian wheat supply is more reliable or better suited

Of course, in this circumstance that AWB and other traders would

an effective control system. These traders, including AWB, would contract

It is also important to emphasise the role of the single desk in

To counter this, AWB requires a credible threat that it either has

In summary, a single desk has greater bargaining power compared

2.2.2 Options Available to the Single Desk

AWB is also able to negotiate a higher price, P, if it can potentially

In the absence of a single desk, AWB would have more difficulty in