Review of Economics 2021; 72(3): 183–197

Fritz Helmedag*
Basic Bidding Formats: Characteristics and
Differences
https://doi.org/10.1515/roe-2021-0014
Received May 1, 2021; accepted October 4, 2021

Abstract: In standard auction theory, the ‘revenue equivalence theorem’ asserts

that the outcomes of the elementary allocation methods coincide. However,
bidding processes differ fundamentally with regard to the decision situation of
the participants: Is it at all imperative to take into consideration the number
of competitors (‘stochastic’ strategy) or not (‘deterministic’ course of action)?
Furthermore, established auction theory neglects the operating modes of pro-
curement alternatives under uncertainty. Apart from the lacking knowledge how
many rivals have to be beaten, tenderers regularly are ignorant of the buyer’s
reserve price. Then it is even more tentative to calculate an offer based on prob-
ability theory. Consequently, the suppliers’ propensity to collude increases.

Keywords: auctions, strategic bidding, tendering

JEL-Classification: D 44

1 Price Competition for a Given Quantity

From an individual’s point of view, making a bargain means: sell expensively, buy
cheaply! In case the article of trade is indivisible, the transaction requires specific
bidding procedures. Yet, despite all dissimilarities, canonical auction theory
assures that the essential methods lead to the same price and are efficient as well,
i.e. the most powerful candidate wins. In textbooks, this ‘revenue equivalence
theorem’ is even celebrated as “the biggest result in auction theory” (Rasmusen
2007, p. 403).
The theorem is based on an analysis where alternative auctions are modelled
as two-stage games (cf. Harsanyi 1967, 1968a, b). At first, the type of a player is
determined by nature (or privately drawn). Accordingly, all contestants form a

*Corresponding author: Fritz Helmedag, Economics Department, Chemnitz University of Technology,

Thüringer Weg 7, D-09107 Chemnitz, Germany, E-mail: [email protected].
https://orcid.org/0000-0002-0748-3303
184 F. Helmedag

(Bayesian) decision function. Then, a (Nash) equilibrium is achieved where every

player choses the best response given the best responses of the other participants.
Alas, that description doesn’t tell the true story how things happen in real life.
The very coexistence of various practices long since provides sufficient reason for
doubt about the supposed congruence of outcomes (cf. Lucking-Reiley 1999).
Consequently, the usual Bayesian-Nash equilibrium logic has to be overcome by
introducing alternative behavioural rules.
The present study probes into the particular characteristics of the basic bid-
ding procedures. For this purpose, it is important to consider separately the pos-
sibilities to cede or obtain a specified item for money. The literature, however,
usually assumes that sales and purchases represent two sides of the same coin:
“The process of procurement via competitive bidding is nothing but an auction,
except that in this case the bidders compete for the right to sell their products
or services” (Krishna 2010, p. 1).1 Correspondingly, established auction theory
centres on methods to vend an object.2
The following enquiry also starts with such transactions. But the subsequent
scrutiny will show that strategic behaviour in the supply of goods suffers from a
systematic deficient knowledge vis-a-vis the corresponding demand processes.
This fact increases the propensity for agreements among bidders. Finally, pro-
posals for regulating invitations to tender are presented.

2 Selling Procedures
2.1 The Standard Types

To begin with, allocation arrangements are examined from the auctioneer’s

perspective who aims to receive the highest price for the article on sale. Each
bidder is supposed to possess an individual limit, which indicates the maximum
willingness to pay for the object. It is of no importance here whether the bargain
hunter ascribes a ‘private value’ with a particular pecuniary equivalent to, say, a
painting, or whether (s)he imputes any ‘common value’ in money terms, e.g. to the
potential exploitation of a particular mine.3 Table 1 portrays the two major types of
auctions and their specific subcategories (cf. Molho 1997, p. 211).

1 This assertion can also be found in the textbook by Monezes and Monteiro (2005, p. 11, fn. 2).
2 Leitzinger (1988) constitutes an exception to established practice by thoroughly addressing
alternative options to find a purveyor.
3 The ‘Winner’s Curse’ (cf. Kagel and Levin 1986), which is not addressed in the present paper,
rests upon an allegedly too optimistic calculation of the revenue that an acquired asset yields.
 Basic Bidding Formats 185

Table : Bidding procedures for sales.

Property Deterministic: no conjecture about the Stochastic: conjecture about the

number of competitors required number of competitors required

Auction name English Vickrey Dutch First price

Method Open (multiple) Sealed single bid Open Sealed single

increase (continuous) bid
decrease

Determination Iteratively from Definitively to Definitively from Definitively to

of award minimum price to highest bidder for highest bid to highest bidder
last bidder the offer of the first bidder for highest offer
second-highest
bidder

Behaviour Overbidding to Bidding the limit Strategy Strategy

the limit

‘Deterministic’ models specify the buyer either publicly via a successive

(sufficiently small) increase of bids in an ‘English-auction’ or in a procedure
suggested by Vickrey (1961). In this variant contenders submit a covert offer,
typically presented on a sheet of paper in an envelope. All participants know that
the highest bidder will receive the good for the nearest lower quotation (‘second-
price-sealed-bid’). According to the principle of indifference (or insufficient
reason) we assume that the limits of the relevant 1, 2, …, n − 1 contenders are
equally distributed between a minimum price B ≥ 0 and the utmost willingness to
pay Ln > Β. Then the object is awarded to buyer n who has to disburse an amount
GD, determined by the second highest limit Ln−1:
n−1
GD = Ln−1 = B + (Ln − B) (1)
n
In both deterministic formats, a participant’s dominant strategy consists simply in
bidding either up to or directly the individual price-cap. The winner’s benefit is the
difference between his or her maximal willingness to pay and the second highest
limit at which the penultimate candidate drops out.4 All other attendees do not
matter for the result of the bidding process, though they may participate in an open
auction at the outset or reveal their willingness to pay in a sealed single bid.
The two ‘stochastic’ allocation methods are far more complicated because they
actually constitute a ‘one-shot’ game under uncertainty.5 In a ‘Dutch auction’, an

4 Depending on the stipulated minimum increment of the bids, the hammer price in an open cry
format is (somewhat) higher than in a Vickrey design.
5 Secondary markets, resale options and (re-)negotiation possibilities are disregarded.
186 F. Helmedag

initial (and obviously) overcharged amount is continuously lowered until the first
customer signals acceptance. Then, e.g. a batch of flowers is knocked down for that
price. In a ‘first-price-sealed-bid’, concealed proposals are submitted and the good
will be sold for the highest offer.
Contrary to an English- or a Vickrey-design, stochastic formats do not permit a
simple ‘mechanical’ behaviour, i.e. bidding gradually up to or directly the indi-
vidual limit. Instead, every decision maker faces the dilemma that a high bid
increases the probability of success but decreases the advantage from the deal.
Thus, an analytic approach is required to make the best of the conflict.

2.2 Strategic Behaviour of Demanders

In the just described situations an applicant neither knows how many rivals there
are nor their definite offers. In order to cope with this uncertainty, standard
auction theory presupposes that the contestants’ bids are random numbers
drawn from probability and density functions which are common knowledge.
Yet, this assumption may be a reasonable substitute for lacking information in
theory, alas not in practice. Actually, one thing is for sure: In order to make a
bargain, every aspirant will spend no more than the individual appreciation of
the item in money terms. Therefore, the selling price definitely won’t exceed the
highest willingness to pay.
Though every participant naturally hopes to make the victorious bid, only one
of them really holds the object on sale in utmost esteem. Let us slip in the shoes of
this candidate n with the maximum limit (Ln). To be sure, (s)he will make an offer
(Gn) which raises the reserve price B by an initially unspecified fraction g (0 ≤ g < 1)
of the potential price spread Δ = Ln − B > 0:
Gn = B + g(Ln − B) = B + gΔ (2)
In case of success, the strategist’s rent (Rn) comes to:
Rn = Ln − Gn = Ln − (B + g(Ln − B)) = (1 − g)(Ln − B) = (1 − g)Δ (3)
Evidently, g coincides with the probability to beat any random bid of a single rival in
the applicable range. Yet, for more than one competitor, bidder n has to develop a
precise idea of how many applicants – who all hope to win (otherwise they would
not participate) – are in the running. It is likely that a relatively high reserve price
reduces the estimated number of relevant contenders and vice versa. If the assess-
ment proves to be correct, the strategist receives the article with probability pn (g):
pn (g) = g n−1 (4)
 Basic Bidding Formats 187

Then, the expected value of the optimizer’s rent E(Rn) is calculated to:

E(Rn ) = pn (g)Rn = g n−1 (1 − g)Δ (5)

A risk-neutral bidder aims to maximize this term.6 The necessary condition reads:
∂E(Rn )
= (n − 1)g n−2 (1 − g) − g n−1 Δ = (n − ng − 1)g n−2 Δ = 0 (6)
∂g
Solving gives:
n−1
g∗ = (7)
n
At g∗ the second derivative of expression (5) is negative for n > 1, so that the
sufficient condition is fulfilled. As a result, the bidder with the supreme limit has
found the best surcharge g∗Δ on the minimum price. Obviously, the offer varies
with the predicted count of attendees n.7
But note the difference: In deterministic procedures the limits of (n − 1) partakers
were equally distributed between the reserve price B and the maximum willingness
to pay Ln, i.e. the price margin Δ.8 Thus, the price-setting second highest limit rises
with the number of contenders. In a stochastic environment, however, the rivals’
bids are taken as random realizations on this line segment. By no means is it thereby
assumed that all vying candidates behave haphazardly. Instead, without further
knowledge, it appears from the decision maker’s point of view as if the assumed
other (n − 1) bids in the relevant interval are drawn arbitrarily.9
After re-substituting the optimal fraction of the price spread (7) into Eq. (4), the
probability to beat all other applicants comes to:
n − 1 n−1
pn (g ∗ ) = ( ) (8)
n
Interestingly, the prospect for success converges towards a positive minimum:
n − 1 n−1
̂ n (g ∗ ) = lim (
p ) = e−1 ≈ 0.368 (9)
n→∞ n

6 Matthews (1987) enquires into the behavior of risk adverse demanders.

7 Levin and Ozdenoren (2004) investigate situations in which participants assign different
probabilities to the respective number of competitors.
8 Bose et al. (2006) analyze how the design of auctions is affected by ambiguity about the
distribution of valuations.
9 Put in the terminology of level-k-reasoning (see e.g. Crawford and Iriberri 2007) the bidder holds
a first-order belief, thinking “… that others select a number at random, and he chooses his best
response to this belief” (Nagel 1995, p. 1313).
188 F. Helmedag

Even with ‘infinitely many’ contenders, the player prevails in more than one third
of all cases. Remarkably, as shown in Figure 1, already with six competitors the
probability to triumph comes close to the lower bound.

Figure 1: Probability of success.

Inserting g∗ (7) in the bidding function Gn (2) yields the strategist’s offer:
n−1
Gn = B + (Ln − B) (10)
n
The expected surplus for candidate n is:
n − 1 n−1 1
E(Rn ) = pn (g ∗ )(1 − g ∗ )Δ = ( ) ( )Δ (11)
n n
If, in a first approximation, ‘many’ potential buyers encompass just n = 6 at-
tendees, the optimizer, employing Eq. (10), increases the bottom price B by 83.3%
of the difference to his or her limit and anticipates, by virtue of formula (11), 6.7% of
the margin as rent. But hope can be deceptive not only because the multitude of
candidates has been erroneously specified.

2.3 The Bids in Comparison

According to the prevailing doctrine, all players behave strategically with the same
n in formula (8) since the number of participants is treated as common knowledge
(cf. Krishna 2010, p. 12). Consequently, the auction-goer with the superior will-
ingness to pay systematically obtains the good. Thus, efficiency is ensured as in
 Basic Bidding Formats 189

deterministic designs. Moreover, for an identical n, a comparison of the strategic

bid Gn (10) with the victorious offer in deterministic formats GD (1) seems to verify
the ‘general equivalence theorem’: Ostensibly, the canonical bidding formats yield
the same outcome.
Yet, the usual account elevates an exception to the rule. Actually, the individual
who values the article on sale most may lose against another contestant, even if this
person employs the same bidding rule (10) but supposes more applicants. For
illustration purposes, consider the following situation. Let the minimum price be
zero, i.e. B = 0. The partaker with the maximum limit Ln = 1 assumes five competitors.
Consequently, according to formula (10), his or her offer is (4/5) · 1 = 0.8. The
potential purchaser with the second highest price-cap Ln−1 = 0.9 uses the same
precept but guesses that 10 rivals are in the running. On condition that all other
offers are lower, (s)he triumphs with the bid (9/10) · 0.9 = 0.81. Clearly, the allocation
contradicts efficiency.
To be sure, in a stochastic environment no attendee can predict the exact
behavioural conduct of every single rival. Thus, it is obvious to treat the relevant
contestants as random bidders. Providing this assumption reflects reality well and
the number of candidates has been precisely anticipated by the competitor with
the supreme limit, it is nonetheless quite possible that (s)he walks away empty-
handed. Then, an inferior outcome also ensues as in the numerical example above.
In order to calculate the successful contender’s mean bid GC, let x denote a part
of the price spread Δ in the interval n−1n ≤ x ≤ 1. The probability that x exceeds the
evenly distributed other (n − 2) fractions of Δ is xn−2 . Hence the following condition
must hold:
1 1
∫ x ⋅ xn−2 dx = g C ∫ xn−2 dx (12)
n−1 n−1
n n

Equation (12) results in:

1 n
∫ x ⋅ xn−2 dx 1 − (n−1
⎜
n−1 ⎛
⎜
⎜ n ) ⎞ ⎟
⎟
⎟
⎜ ⎟
n−1
∗
gC = n
1 =( )⎜
⎜
⎝ ⎟
n−1 ⎟
⎠>g (13)
n
∫ n−2
x dx 1−( n )
n−1
n−1
n

Of course, an arbitrary guess on the number of candidates can turn out wrong.
Therefore, the strategist possibly resorts to the convergence property of the
probability to win. Consequently, (s)he inserts e.g. six potential purchasers n in
expression (7) and calculates g* (6) = 0.833. But it may well be that, according to
formula (10), the ensuing bid Gn is lower than GC since Eq. (13) yields gC (6) = 0.927:
190 F. Helmedag

GC = B + g C Δ = B + 0.927Δ > Gn = B + g ∗ Δ = B + 0.833Δ (14)

Tough the optimizer’s bid Gn (10) corresponds to the result in case of a deter-
ministic format GD (1), the item is now not necessarily awarded to the bidder who
appreciates it most. Perhaps another individual who holds a lower limit triumphs
with a bid GC.
The expected price E (∅G) amounts to the sum of the weighted offers:
n−1 n
E (∅G) = B + (pn (g ∗ )g ∗ + (1 − pn (g ∗ ))g C )Δ = B + ( ) Δ +
n
n n (15)
(1 − n)((n−1
n )
− 1) n − 1 + (n−1
n )
+ Δ=B+ Δ > Gn
n n
From Eq. (15) follows that stochastic formats are on average inefficient, even if the
strategist by some amazing fluke correctly estimates the number of competitors.
Against this backdrop, it seems sensible to ponder on a statuary prohibition of
formats where contenders act under uncertainty. By then the auctioneer will
choose that deterministic or stochastic design which promises to fetch the highest
price irrespective of allocation issues (cf. Milgrom 1989, p. 10 ff.). Anyway, the
possibility to establish the one or the other bidding procedure opens up the scope
for action to sellers. But market power increases all the more when it comes to
acquire a good.

3 Purchasing Procedures
3.1 The Standard Types

In the following, the basic types for selecting a supplier will be examined from the
perspective of the buyer or customer who specifies the good to be provided.10 By
assumption, every competitor knows the minimum price for which (s)he can
deliver the ordered article or service. This amount is often based on the estimated
costs that an acceptance of a commission entails. Table 2 shows that, similarly to
sales procedures, there are four alternatives. Yet, from an economic point of view,
they form two major strands.

10 In German, such procurements were formerly often referred to as ‘Lizitation’, a term that has
now fallen into oblivion. In English, however, all bidding formats, both for sales and purchases,
are called ‘auction’. To avoid misunderstanding, the words ‘tendering’ or ‘submission’ are used to
denote alternative possibilities to find a purveyor.
 Basic Bidding Formats 191

Table : Bidding procedures for purchases.

Property Deterministic: no conjecture about the Stochastic: conjecture about the

number of competitors required number of competitors and
maximum willingness to pay
required

Name Tendering Vickrey submission Dutch tendering Submission (first

(descending (second price) price)
auction)

Method Open (continuous) Sealed single bid Open increase Sealed single bid
decrease

Determination Iteratively from a Definitively to Definitively from Definitively to

of award maximum bid to lowest bidder for a minimum price lowest bidder for
last bidder the bid of second to first bidder the lowest bid
lowest bidder

Behaviour Underbidding to Bidding the limit Strategy Strategy

the limit

Ultimately, in deterministic processes the cheapest attendee gets the order.

This person is the lowest bidder either in an open descending process starting from
an unrealistically high amount or in a sealed Vickrey-submission, where the losing
runner-up determines the payment in return for the provision. The advantage for
the remaining successful candidate consists in the difference to his or her bottom
price. Both mechanical selection techniques lead to the same result and do neither
require considerations regarding optimization nor suppositions on the count of
contenders. Again, in deterministic environments a dominant strategy for aspi-
rants exists.
But the level of uncertainty in stochastic procurement methods has increased
once again in comparison to auctions, where the bidding formula only necessitates
to assess how many rivals are involved. For the first or lowest applicant in an open
procedure (‘Dutch tendering’) or a sealed single bid (‘submission’) an optimal
asking price entails additionally to develop an idea about the purchaser’s
maximum willingness to pay. In this respect, however, a strategically acting ten-
derer is largely left in the dark, because it is now not his or her own limit that forms
an upper bound of an offer, but the buyer’s covert reservation expenditure. Before
discussing potential consequences for competition policy, it is worthwhile to
inquire into the purposeful behaviour of a risk-neutral agent seeking to undercut
his or her rivals.
192 F. Helmedag

3.2 Strategic Behaviour of Suppliers

Initially, in order to detect the formal relationships, a special kind of lottery comes
into focus. A participant has the opportunity to demand a premium between zero
and one hundred cents. The person earns the declared sum if a previously
announced number of (n − 1) random draws without replacement – representing
claims of other contenders – turn out higher. All results are supposed to be located
on a line segment with the length of 1. Contrary to strategic price competition, in
the present circumstances the player knows how many times his or her request has
to remain unbeaten. The probability pn ( f) that the desired reward f prevails is:

pn (f ) = (1 − f )n−1 (16)
Since in case of success the benefit Q equals the payment f, the expected gain E (Qn)
amounts to:

E(Qn ) = pn (f )Q = (1 − f )n−1 f (17)

The necessary condition for an optimum requires:
∂E(Qn ) (f n − 1)(1 − f )n
= (1 − f )n−1 − f (n − 1)(1 − f )n−2 = − =0 (18)
∂f (1 − f )2
Solving for the best prize money gives:
1
f∗ = = 1 − g∗ (19)
n
At f ∗ the sufficient condition is fulfilled for n > 1. Therefore, this wish maximizes the
average payoff. The probability to succeed is calculated to:
1 n−1 n − 1 n−1
pn ( f ∗ ) = (1 − ) = ( ) = pn ( g ∗ ) (20)
n n
As in Eq. (9), expression (20) converges towards a positive minimum:
n − 1 n−1
̂ n ( f ∗ ) = lim (
p ) =p ̂ n (g ∗ ) = e−1 ≈ 0.368 (21)
n→∞ n
The strategist also triumphs in more than one third of occasions, and the proba-
bility to be victorious quickly approaches the lower bound, as Figure 1 shows. The
expected advantage amounts to:
n − 1 n−1 1
E(Qn ) = pn ( f ∗ )f ∗ = ( ) ( ) (22)
n n
Again, if n = 6 is accepted as indicating ‘fierce competition’, the optimizer demands
(rounded down) 16 cents and can hope for 42% of this sum, i.e. 6.7 cents.
 Basic Bidding Formats 193

3.3 The Results in Comparison

It is possible that the strategist loses against a random call fC. Let y denote a request
in the interval 0 ≤ y ≤ n1 . The probability for y to undercut the evenly distributed
other (n − 2) asking prizes comes to ( 1 − y)n−2 . Then, the ensuing condition must
hold:
1 1
n n
n−2 n−2
∫ y ⋅ (1 − y) dy = fC ∫ (1 − y) dy (23)
0 0

Equation (23) leads to:

1
n n
n−2
∫ y ⋅ (1 − y) dy (n−1
n ) (2n
− 1) + 1 − n
fC = 0
1 = n =
n
∫ (1 − y)
n−2
dy n )
n(n(n−1 + 1 − n)
0
(24)
n

⎜
n−1 ⎛
⎜ 1 − (n−1
n ) ⎞ ⎟
⎟
=1−( ⎜
⎜
)⎜ ⎟
⎟
⎟
n ⎝
⎜ ⎠ = 1 − gC
n−1 ⎟
1−( n )
n−1

The average reward E(∅f )amounts to:

n − 1 n−1 1
E ( ∅f ) = pn (f ∗ )f ∗ + (1 − pn (f ∗ ))fC = ( ) +
n n
n
n − 1 n−1 ⎛ 1 − (n−1
n ) ⎞
+ (1 − ( ) )⎜ ⎝ 1 − (n − 1) ⎛
⎜
⎝ ⎟⎟
⎠⎞
⎠= (25)
n n n−1 n−1
1−( ) n
n
1 − (n−1
n )
=
n
Moreover, this time it is possible to calculate the expected rent of the winning
competitor since it is equivalent to the weighted remuneration fC:
n
(n−1
n ) (1
− 2n) + n − 1
E(QC ) = (1 − pn ( f ∗ ))fC = (26)
n(n − 1)
With Eqs. (22) and (26), the strategist’s anticipated share in the overall surplus can
be ascertained:
194 F. Helmedag

n )
E(Qn ) n(n−1
= (27)
E(Qn ) + E(QC ) n
(1 − n )
n)((n−1 − 1)

Just as the probability to come first, the proportion converges for an increasing n to
a positive limit:
n
E (Qn ) n(n−1
n ) 1
lim = lim = ≈ 0.582 (28)
n→∞ E (Qn ) + E (QC ) e−1
n
n→∞ (1 − n)((n−1) − 1)
n

Thus, even with a large number of rivals, the strategist can attract on average
nearly 60% of the (decreasing) payment. But this statement relates to a particular
decision-making situation, where, as in the raffle discussed above, the prize
offered as reward is known. De facto, however, invitations to tender hardly ever
announce the available budget for the procurement. This information deficit im-
pedes the applicant to calculate (and maximize) the expected advantage from the
commission. In auctions, on the other hand, the potential purchaser is able to
confine possible outcomes to the range between the minimum price and the per-
son’s own willingness to pay. Thus, probability theory is applicable. The transfer of
analogous considerations to the behaviour of suppliers in stochastic formats
requires that the costumer is obliged to publicly declare his or her reservation
price.11 Without this notification, the prerequisite to optimize a tender is missing.

4 Rules for Bids

The highest possible profit a purveyor i can earn equals the difference between the
ensuing costs and the client’s revealed maximal budget (M). In order to adapt the
previously reviewed lottery to the situation now prevailing, costs (Ci) are raised by
the fraction f∗ = 1/n times the maximal rent (M − Ci), which was set to unity in the
sweepstake above. Then one gets the quotation (Fi) of a strategically operating
tenderer:
1
Fi = Ci + (M − C i ) (29)
n
Even in this milieu, however, the bidding formula does not at all guarantee the
triumph of the most economical attendee. Of course, a victorious lower asking

11 In order to guarantee the announcement’s credibility, an obligation to buy or a liability for

damages in case of suspending the purchase should be implemented.
 Basic Bidding Formats 195

price is not always just a fluky shot. Instead, it can also be based on precept (29),
when both costs and the number of contenders were estimated higher.12
Whatever the reason may be: If the most potent candidate fails, the allocation
proves to be inefficient. Yet, this is of no significance to those demanders who want
to purchase as cheap as possible. Hence, similar to auctions, the chosen bidding
method to obtain a good also reflects the economic power relation between the
organizing principal and the procuring agents.
Obviously, the deterministic techniques applied to receive or provide an
indivisible object entail advantages for contestants. In contrast to a stochastic
course of action, aspirants do not have to worry about the multitude of rivals.
Entrants simply continue to participate in the bidding process as long as their
willingness to pay or their minimal claim has not been reached yet. At the same
time, these routines correspond to the Pareto criterion: The superior applicant
wins, while the runner-up, though dropping out, fixes the price.
If bidding is executed via the Internet instead of relying on older means
of communication, the arranger can hope for a better result, since, with more
potential buyers (suppliers), the deciding second-best limit probably increases
(decreases). Alas, the mechanical line of action also has its possible pitfalls,
because it is not at all safeguarded against manipulation attempts. According to
rumours front men intervene every now and then in bidding events to move prices
up or down so that the good is sold dearer or bought cheaper. To be sure, these
endeavours should not be carried too far because a missing deal with a third party
could be the consequence.
Contrary to the established opinion, stochastic formats sometimes turn out to
be inefficient. Thus, from the vantage point of competition policy, it seems worth
considering whether such methods should be banned.13 At any rate, there is an
indisputable need to regulate at least acquisitions via first-price-sealed-bids. This
inquiry has shown that strategic bidding practices exhibit a fundamental asym-
metry between sales and purchases. In the case of auctions, every attendee knows
his or her limit. This enables the person, once the number of contenders has been
assessed, to optimize the expected gain from the transaction.
In contrast, to date the customer, when inviting for tenders, keeps a low profile
regarding the maximum expenditure. Prospective purveyors, therefore, literally

12 If n is the actual count of competitors with their costs evenly distributed in the interval [Ci, M[,
then the mechanical price determination coincides with the strategic offer.
13 “Indeed, some hold the view that one essential role of government is to declare that the rules of
certain social ‘games’ must be changed whenever it is inherent in the game situation that the
players, in pursuing their own ends, will be forced into a socially undesirable position” (Luce and
Raiffa 1957, p. 97).
196 F. Helmedag

have no basis to apply probability theory, since they possess little or no infor-
mation about the highest price the opposite party is ready to pay. This uncertainty
furthers the tendency of potential suppliers to engage in prohibited agreements;
behaviour that e.g. in Germany constitutes a criminal offense by now. On the other
hand, the country’s ‘Gesetz gegen Wettbewerbsbeschränkungen’ (Act Against
Restraints of Competition) specifies in Section 97 that at least public procurement
should be operated by ‘transparent procedures’. In order to comply with this
provision, authorities should, in future, reveal their willingness to pay for a specific
good, possibly derived from a preliminary calculation (cf. Helmedag 2004). If the
contract is conferred on a contender who receives a (perhaps just slightly) lower
remuneration, the allegation of overcharging becomes invalid because a reason-
able case for damages can no longer be made.
Nevertheless, it appears to be even better that, instead of a first-price-sealed-
bid, a Vickrey-submission became the norm. This creates the incentive to disclose
the true minimal claims, since tenderers, including the winning and most cost-
effective competitor, are less concerned about accepting a commission which
eventually turns out to be a loss-making deal.

References
Bose, S., Ozdenoren, E., and Pape, A. (2006). Optimal auctions with ambiguity. Theor. Econ. 1:
411–438.
Crawford, V.P. and Iriberri, N. (2007). Level-k auctions: can a nonequilibrium model of strategic
thinking explain the winner’s curse and overbidding in private-value auctions? Econometrica
75: 1721–1770.
Harsanyi, J. (1967). Games with incomplete information played by ‘Bayesian’ players, part I. The
basic model. Manag. Sci. 14: 159–182.
Harsanyi, J. (1968a). Games with incomplete information played by ‘Bayesian’ players, part II.
Bayesian equilibrium points. Manag. Sci. 14: 320–334.
Harsanyi, J. (1968b). Games with incomplete information played by ‘Bayesian’ players, part III. The
basic probability distribution of the game. Manag. Sci. 14: 486–502.
Helmedag, F. (2004). ‘Ausschreibungsbetrug’ im Licht der Gemeinsamkeiten und Unterschiede
von Bietverfahren. Wirtsch. Wettbew. 54: 1000–1012.
Kagel, J.H. and Levin, D. (1986). The winner’s curse and public information in common value
auctions. Am. Econ. Rev. 76: 894–920.
Krishna, V. (2010). Auction theory, 2nd ed. Academic Press: Amsterdam.
Leitzinger, H. (1988). Submission und Preisbildung. Mechanik und ökonomische Effekte der
Preisbildung bei Bietverfahren. Carl Heymanns: Köln.
Levin, D. and Ozdenoren, E. (2004). Auctions with uncertain number of bidders. J. Econ. Theor. 118:
229–251.
Luce, R.D. and Raiffa, H. (1957). Games and decisions. Introduction and critical survey. Wiley: New
York.
 Basic Bidding Formats 197

Lucking-Reiley, D. (1999). Using field experiments to test equivalence between auction formats:
magic on the internet. Am. Econ. Rev. 89: 1063–1080.
Matthews, S. (1987). Comparing auctions for risk averse buyers: a buyer’s point of view.
Econometrica 55: 633–646.
Milgrom, P. (1989). Auctions and bidding: a primer. J. Econ. Perspect. 3: 3–22.
Molho, I. (1997). The economics of information. Lying and cheating in markets and organizations.
Blackwell Publishers: Oxford and Malden.
Monezes, F.M. and Monteiro, P.K. (2005). An introduction to auction theory. Oxford University
Press: Oxford and New York.
Nagel, R. (1995). Unraveling in guessing games: an experimental study. Am. Econ. Rev. 85:
1313–1326.
Rasmusen, E. (2007). Games and information. An introduction to game theory, 4th ed. Blackwell:
Maldon, Oxford and Carlton.
Vickrey, W. (1961). Counterspeculation, auctions and competitive sealed tenders. J. Finance 16:
8–37.