RETAIL PRICE REGULATION AND THE OPTION TO DELAY

Abstract

This paper examines a two-period model of an investment decision in a network industry characterized by demand uncertainty, economies of scale and sunk costs. In the absence of regulation we identify the market conditions under which a monopolist decides to invest early as well as the underlying overall welfare output. In a regulated environment, we consider a monopolist who faces no downstream (final good) competition but is subject to retail price regulation. We identify the welfare-maximizing regulated prices when the unregulated market outcome is set as the benchmark. We show that if the regulator can commit to ex post regulation - that is, regulated prices that are contingent to future demand realization - then regulated prices that allow the firm to recover its total costs of production are welfare-maximizing. Thus, under ex post price regulation there is no need to compensate the regulated firm for the option to delay that it foregoes when investing today. We argue, however, that regulators cannot make this type of commitment and, therefore, price regulation is often ex ante - that is, regulated prices are not contingent to future demand. We show that the optimal ex ante regulation, and the extent to which regulated prices need to incorporate an option to delay, depend on the nature of demand uncertainty. Copyright © 2009 The Authors Journal compilation © CIRIEC 2009.

Full text

Page 1

Annals of Public and Cooperative Economics 80:3 2009
pp. 451–468
RETAIL PRICE REGULATION
AND THE OPTION TO DELAY
by
Fernando T. CAMACHO and Flavio M. MENEZES∗
University of Queensland, School of Economics, Australia
ABSTRACT∗∗:
an investment decision in a network industry characterized by
demand uncertainty, economies of scale and sunk costs. In the
absence of regulation we identify the market conditions under
which a monopolist decides to invest early as well as the
underlying overall welfare output. In a regulated environment,
we consider a monopolist who faces no downstream (final good)
competition but is subject to retail price regulation. We identify
the welfare-maximizing regulated prices when the unregulated
market outcome is set as the benchmark. We show that if the
regulator can commit to ex post regulation – that is, regulated
prices that are contingent to future demand realization – then
regulated prices that allow the firm to recover its total costs of
production are welfare-maximizing. Thus, under ex post price
regulation there is no need to compensate the regulated firm
for the option to delay that it foregoes when investing today.
We argue, however, that regulators cannot make this type of
commitment and, therefore, price regulation is often ex ante –
that is, regulated prices are not contingent to future demand. We
show that the optimal ex ante regulation, and the extent to which
regulated prices need to incorporate an option to delay, depend
on the nature of demand uncertainty.
This paper examines a two-period model of
∗
∗∗
resumen al final del art´ ıculo.
Emails: f.camacho@uq.edu.au; f.menezes@economics.uq.edu.au
R´ esum´ e en fin d’article; Zusammenfassung am Ende des Artikels;
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009. Published by Blackwell Publishing Ltd. 9600 Garsington Road,
Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

Page 2

452FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
1Introduction
A central issue in the current debate surrounding economic
regulation across the world is the need to provide correct investment
incentives for regulated businesses. This need has shifted the focus of
price regulation from promoting static efficiency towards supporting
dynamic efficiency.
The tension between price regulation and investment incen-
tives is highlighted, for example, by the current debate on the deploy-
ment of fibre-optic infrastructure and so-called Next Generation
Networks (NGNs). This debate is characterized by firms’ requirement
of regulatory certainty prior to investment. This certainty is sought
in order to avoid circumstances in which the incumbent would
be required to provide access to the new infrastructure at prices
that would yield a zero net present value if the new service were
successful, but access seekers would not share the losses if the new
service fails.
It is also anticipated that the relationship between price regu-
lation and investment incentives will become increasingly important
in a low carbon emissions world where substantial amounts of
renewable and gas-fired micro generation will be introduced into the
electricity system. The achievement of such change will necessitate
significant new investment to adapt and expand existing electricity
networks, mainly because renewable energy involves site-specific
power plants.1
It is widely accepted that a regulatory framework should
not provide firms with incentives to make particular investments.
However, regulated prices should remunerate the risks faced by firms
when investing in a new network facility. These risks are related
to the combination of two characteristics: demand uncertainty and
irreversibility.
According to real options theory,2the combination of uncer-
tainty, irreversibility and investment timing flexibility provides the
building blocks of the option to delay. This theory simply states that
a firm will invest in a project today if the expected payoff (NPV) is
higher than or equal to the expected payoff of investing at anytime
in the future. It follows that a firm may reject a project with positive
1
be only built in one specific location due to natural conditions, for example,
a hydro power plant or wind farm.
2See, for example, Trigeorgis (1996).
A site-specific power plant is an electric generating facility that can
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 3

RETAIL PRICE REGULATION AND THE OPTION TO DELAY 453
NPV – as the relevant comparison is the expected payoff of investing
in the future, which might be higher than the positive NPV of
investing today.
The implication for price regulation of the existence of an
option to delay is clear. The standard (ex ante) approach of setting
price caps to ensure a zero NPV might fail to provide appropriate
investment incentives. When investing today a regulated firm fore-
goes an option to delay its investment. Regulated prices might need
to compensate the firm for this foregone option if investment is not
to be distorted.
Regulators have by and large accepted the existence of an
option to delay. For instance, in the context of telecommunications,
the New Zealand Commerce Commission stated that ‘the obligation
to provide interconnection services removes the option for access
providers to delay investment in their fixed Public Switched Tele-
phone Networks.3If this option has a value, the costs of foregoing
the option are a cost that should be reflected in interconnection
prices’ (Commerce Commission 2002). In its latest cost of capital
consultation, the telecommunications regulator in the UK proposed
that ‘Ofcom should begin to develop a framework by which regulatory
policy might reflect the value of these options (real options)’ and ‘a
key area identified by Ofcom as being one in which the value of
wait and see options might be significant was that of next generation
access networks’ (Ofcom 2005).
Thus, the aim of this paper is to provide a framework to
determine when a regulated firm should be compensated for fore-
going an option to delay and the extent of this compensation. Our
work extends existing research that suggests that regulated prices
should always fully remunerate such option. This research includes
Hausman (1999) and Hausman and Myers (2002), who focus on
access pricing methodologies in the telecommunications and rail
industries and they point out that current regulatory frameworks
provide asymmetric rights between incumbents and entrants. More-
over, they argue that incumbents are forced to provide to entrants a
free option, where such option is the right but not the obligation to
purchase the use of the incumbent’s network. The authors conclude
that incumbents should be compensated for this option value.
3
phone system. Also referred to as the ‘landline’ network, it uses a copper
wire network to carry voice and data.
A Public Switched Telephone Network (PSTN) is the traditional
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 4

454 FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
In the same vein, Pindyck (2004, 2005) addresses the impact of
the network sharing arrangements in the telecommunications indus-
try. Pindyck also suggests that regulated prices should incorporate an
option to delay value to compensate incumbents for the asymmetric
risk. For instance, Pindyck (2005) develops a method to adjust the
cost of capital in the US Telco access pricing formula to account
for the option to delay value. Pindyck shows that this adjustment
is always positive and lies on average between 1.2 and 4.5 per cent.
This paper instead derives the welfare-maximizing regulated
prices from a simple two-period model of an investment decision in
a network industry characterized by demand uncertainty, economies
of scale and sunk costs. It considers a monopolist who faces no
downstream (final good) competition but is subject to retail price
regulation. Unlike the previous literature we explicitly consider the
process by which a regulator sets prices and identify the welfare-
maximizing regulated prices using the unregulated market outcome
as the benchmark.
We show that if the regulator can commit to ex post
regulation – that is, regulated prices that are contingent to the
demand that is realized in the future – then regulated prices that
allow the firm to recover its total costs of production are welfare-
maximizing. Thus, under ex post price regulation there is no need
to compensate the regulated firm for the option to delay that it
foregoes when investing today. We argue, however, that regulators
cannot make this type of commitment and, therefore, price regulation
is often ex ante – that is, regulated prices are not contingent to
future demand. We show that the optimal ex ante regulation, and
the extent to which regulated prices need to incorporate an option to
delay, depend on the nature of demand uncertainty.
2 The investment decision by an unregulated firm
This section develops a two-period model framework to investi-
gate the role of the option to delay in investment decisions in network
industries. Our framework encompasses four common characteristics
of network industries: timing flexibility when making the investment
decision, demand uncertainty, investment irreversibility and natural
monopoly.
We consider an unregulated firm’s decision regarding whether
to build a network in order to provide a new service. The network is
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 5

RETAIL PRICE REGULATION AND THE OPTION TO DELAY455
built instantaneously, so the firm can build the network at t = 0 or
at t = 1. If the firm does not invest at t = 0, it has the right but not
the obligation to invest at t = 1. We assume that when indifferent as
to investing, the firm invests and when indifferent between investing
at t = 0 or at t = 1, the firm invests at t = 0. Also, the investment
outlay to build the network is constant over time and equal to I.
That is, the real cost of investment decreases over time. Moreover,
the investment is sunk and there are no maintenance or operational
costs to run the network.
At t = 0 the inverse demand function is characterized by a
choke price equal to¯P. At any price below or equal to¯P the demand,
denoted by q0, is equal to Q. The demand at a price above¯P is always
equal to zero. At t = 1 the inverse demand function is characterized
by a choke price equal to γ¯P (where γ > 0). At any price below
or equal to γ¯P the demand, denoted by q1, will be either equal
to uQ (where u > (1 + r) and r is the risk-free interest rate) or
equal to dQ (where 0 < d < 1) with probabilities θ and (1 − θ),
respectively. The demand at a price above γ¯P is always equal to
zero.
Note that in this setting, the gross value of future cash inflows
will fluctuate in line with the random fluctuations in demand.
Moreover, the benefits from delaying the investment decision until
t = 1 arise from both the resolution of the demand uncertainty and
the decline in the real cost of investing. The cost of delaying the
investment is the first period cash flow.
The network is used to provide a service to final consumers. In
order to produce the final good the firm needs one unit of the network
and one unit of an input with unit price c. Note that the provision of
the network service constitutes a natural monopoly.
We proceed to calculate the NPV from investing at t = 0. It is
assumed that financial markets are efficient and as a consequence
there are no arbitrage opportunities. In this sense, there is a port-
folio of traded assets that replicates the investment’s network cash
flow. The cost of this portfolio equals the NPV (risk-neutral pricing
formula).
Without any loss of generality, it is also assumed that there
exist two assets: a one period risk-free bond with a price equal to 1
and a payoff of (1 + r) after one period; and a risky asset with price
equal to 1 and a payoff after one period equal to u (with probability θ)
and d (with probability (1 − θ)). These two assets are used to build a
portfolio to replicate the investment’s network cash flow at each node.
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 6

456FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
The NPV of this investment decision, denoted by NPV, is equal to
NPV =??¯P − c?+?γ¯P − c??Q − I
The risk-neutral methodology is also used to calculate this
investment decision as a call option, that is, if the firm does not
invest at t = 0 it has the right but not the obligation to invest at
t = 1.4Thus, the expected return on the option, denoted by OD, must
also equal the risk-free rate in a risk-neutral world, that is,
OD =pOD++ (1 − p)OD−
or
(1)
1 + r
OD =
pMax??γ¯P − c?uQ − I;0?+ (1 − p) Max??γ¯P − c?dQ − I;0?
where p =(1+r)−d
equality between the cost of the replicating portfolio and the cost of
the project’s cash flow holds (no arbitrage opportunities).
From (2), the option to delay only has value when γ¯P > c. Since
our goal is to investigate the relation between the option to delay
and regulation we assume throughout the paper that this inequality
holds. Note from (2) that when considering OD as a function of Q,
there are three ranges that play an important role in our analysis. In
the first range, both states of demand, high and low, yield negative
payoffs. In this case OD is equal to zero. In the second range only
the high demand scenario yields a positive payoff and the slope of
the function is
positive payoffs and the slope of the function is (γ¯P − c).
In order to decide whether, and when, to invest in the network
facility the unregulated firm must compare the values of NPV and
OD which are given by (1) and (2), respectively. The comparison
between the market value of the NPV and the OD at t = 0 depends
on (¯P − c), the term that drives the first period net revenue. By
4 The rationale for using the risk-neutral methodology is provided by
Teisberg (1994) who points out that in an option pricing model the value of
the investment opportunity is derived from the market value of the project.
This implies that the riskless rate, rather than the cost of capital, should
be used in the valuation of the investment as the risk of the project is
incorporated in the market valuation of the project. It follows then that the
cost of capital is exogenous and any changes in its value are captured by
the market value of the project.
1 + r
(2)
u−d
is the risk-neutral probability. p is such that the
pu
1+r(γ¯P − c). In the third range, both scenarios yield
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 7

RETAIL PRICE REGULATION AND THE OPTION TO DELAY457
Q
______
NPV
Payoff
____
OD
Fig. 1 – Case 2
taking the NPV and OD as functions of Q we have two different
cases when (¯P − c) > 0.5
In Case 1, there is a Q such that NPV = OD = 0, that is, the
NPV function crosses the OD function in its first range. In Case 2,
there is a Q such that NPV = OD > 0, OD+> 0 and OD−= 0, that
is, the NPV function crosses the OD function in its second range (see
Figure 1). The investment decision outcomes for each case are listed
in Lemma 1.
Lemma 1: Table 1 summarizes the unregulated monopolist in-
vestment decision outcomes as a function of market
conditions.
We can also compute the total welfare when the unregulated
firm invests at t = 0 (NPV ≥ OD). This is given by:
WM= α NPV
where α < 1 denotes the weight assigned by the social planner to
firm’s profits. Equation (3) is the benchmark that the regulator will
(3)
5
never invest at t = 0.
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009
If (¯P − c) ≤ 0 we will have OD > NPV for all Q, that is, the firm will

Page 8

458FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
Table 1 – Unregulated market outcomes
The firm never
invests if
The firm invests at t = 1
when q1= uQ if
n/a
OD+> 0, OD−= 0
and NPV < OD
The firm invests at
t = 0 if
NPV ≥ OD ≥ 0
OD+> 0, OD−≥ 0
and NPV ≥ OD
Condition
Case 1
OD = 0 and NPV < OD
OD = 0
Case 2
try to improve upon. Finally, note that given the values of the
parameters, the NPV and OD functions are fixed at NPV and OD.
In the next section we will calculate the changes in the NPV and OD
when retail prices are set by the regulator.
3Retail regulation
This section investigates the effect of retail price regulation on
the incentives to invest. More specifically, we consider a monopolist
firm that faces no downstream competition and is subject to a
price regulation on the downstream retail (final good) market. The
regulator and the firm both observe the choke prices (¯P and γ¯P) and
are fully informed about the nature of demand uncertainty and the
cost function. The role of the regulator is to set regulated prices PR
and PR
to maximize total welfare:
0
1that will prevail at t = 0 and at t = 1, respectively, in order
Max WR= CS+ απ
(4)
where CS denotes consumer surplus, and π is the firm’s profit.6Next
we investigate the investment incentives and overall welfare under
ex post (i.e., demand-contingent) and ex ante (i.e., non-demand-
contingent) price regulation.
3.1Ex post regulation
For the moment, we will assume that the firm invests at t =
0 and the regulator allows the firm to recover all its costs by setting
6
reduction in the firm’s profit (through lower prices) would be equal to an
equivalent increase in the consumer surplus. Thus, when α = 1 in our
model, a regulator cannot improve upon the outcome of the unregulated
market.
In our set up the horizontal demand implies that when α = 1 any
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 9

RETAIL PRICE REGULATION AND THE OPTION TO DELAY459
ex post contingent demand prices that will prevail at t = 0 and at
t = 1. As prices are not known at the time of the investment decision,
ex post regulation requires a credible commitment by the regulator
that it will set prices so as the firm recovers its costs taking into
account the realized demand. This means that prices might be higher
under low demand and lower under high demand. We will argue
below that such commitment by regulators is not credible. Indeed,
most of if not all price regulation is ex ante in nature.7
It is assumed, without any loss of generality, that regulated
prices are set at each period to recover
neutral world, a regulator committed to full cost recovery would set
the following ex post contingent demand prices:
⎧
⎪⎪⎪⎩
t = 0:
WR=??¯P − c?+?γ¯P − c??Q − I = NPV
The difference between (6) and (3) is equal to (1 − α)NPV > 0. Thus,
we have established the following result:
Proposition 1: When the regulator commits to set ex post demand-
contingent prices that allows the firm to recover its
full costs, regulated prices given by (5) yield higher
overall welfare than an unregulated industry.
The rationale for Proposition 1 is as follows. Under the as-
sumption of credible commitment, the firm always invests at t = 0
I
2+ c. In particular, in a risk-
⎪⎪⎪⎨
PR
0=
I
2q0
+ c at t = 0
PR
1=(1 + r) I
2q1
+ c at t = 1
(5)
These prices yield the following expected welfare function at
(6)
7
regulation is a form of ex post regulation where the regulator commits to
allowing the firm to recover all its costs. Such analogy is complicated by
the fact that rate of return regulation has evolved over time with courts
setting up precedents that constrain the ability of the regulator to renege
on its commitment to full cost recovery. Of course, cost of service regulation
was largely avoided by the new regulatory regimes set up around the world,
following privatization and deregulation, due to the perceived moral hazard
problem; firms have very little incentives to reduce costs under cost of
service regulation. Price cap regulation, which is ex ante in nature (i.e.
prices are set before demand or costs are realized) became the standard
around the world.
Perhaps one can argue that cost of service (i.e., rate of return)
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 10

460FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
and the regulator is able to extract the firm’s entire profit, which
is transferred to consumers. As α < 1, this regulation provides the
correct investment incentives and generates higher welfare than an
unregulated industry. By committing to full cost recovery, the regula-
tor transfers the volume risk associated with demand fluctuation to
consumers. Therefore, there is no reason for the regulated firm to be
compensated for a risk that it does not bear.
The difficulty, however, is that in practice regulators cannot
commit to full cost recovery. For example, if low demand eventuates,
full cost recovery requires the regulator to set high prices. However,
doing so might bring a considerable amount of public dissatisfaction.
A regulator that is concerned with his or her career might find that
dissatisfaction unacceptable.
Importantly, a credible commitment by the regulator entails
the firm assigning zero probability to the event that it will be
expropriated under low demand. The converse, which is a somewhat
more subtle point, is that it suffices for the regulated firm to attach
a small probability that the regulator will not stick to its demand
contingent prices for the regulated prices specified in (5) to lead to
suboptimal investment decisions. The example below illustrates this
point.
In this example, we assume that there is a probability λ that
the regulator will set a price equal to PR
demand is low, whereas the price remains equal to PR
when demand is high (the price at t = 0 is still PR
establishes that in this scenario the firm never invests at period
t = 0.
In this setting, the firm only invests at t = 1 if the regulator
sets regulated prices that allow full cost recovery. These prices are
listed in Table 3.
The regulated prices listed in Table 3, which are the low-
est prices that result in the firm investing at t = 1, yield the
1=(1+r)I
2Q
+ c at t = 1 when
1=(1+r)I
0=
2uQ+ c
I
2Q+ c). Table 2
Table 2 – Investment at t = 0
Price at t = 1
1=(1 + r)I
2Q
1=(1 + r)I
2uQ
DemandFirm’s profit at t = 1
(d − 1)I
dQ
PR
+ c or(1 + r)I
2dQ
+ c
2< 0 or 0
uQ
PR
+ c
0
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 11

RETAIL PRICE REGULATION AND THE OPTION TO DELAY461
Table 3 – Investment at t = 1
Price at t = 1
1=(1 + r)I
dQ
1=(1 + r)I
uQ
Demand Firm’s profit at t = 1
dQ
PR
+ c
0
uQ
PR
+ c
0
following expected welfare function at t = 0:
WR=?γ¯P − c?Q − I
(7)
The difference between (7) and (3) is equal to [(γ¯P − c)Q − I] −
αNPV. The first term of this expression is the increase in the
consumer surplus while the second term is the decrease in the firm’s
profit under the proposed regulated prices. On one hand the firm has
a profit equal to NPV in an unregulated market and 0 under the reg-
ulatory contract. On the other hand, the consumer surplus is 0 in an
unregulated market and equal to [(γ¯P − c)Q − I] under the proposed
regulated prices. If [(γ¯P − c)Q − I] < αNPV, the proposed regulated
prices generate lower welfare than an unregulated industry.
This example illustrates that when the regulator cannot cred-
ibly commit to set appropriate ex post demand-contingent prices
that are consistent with full cost recovery, in general it is not
possible to ensure that ex post regulation can do better than an
unregulated market. Next we show that ex ante regulation, in which
regulators set prices prior to the realization of the demand can
generate outcomes that are welfare superior to the unregulated
market outcome.
3.2 Ex ante regulation
Under this type of regulation, regulated prices are fixed at time
t = 0. That is, the regulator sets ex ante non-demand contingent
maximum prices PR
at t = 0 and at t = 1, respectively. The firm then decides whether to
invest at t = 0 or at t = 1 (if demand turns out to be high). As we
shall demonstrate below, the optimal price cap regulation depends on
the impact of the regulated prices on the comparison between the
NPV and OD, which in turn depends on the unregulated market
conditions. As seen in Section 2, there are two cases to consider
(Cases 1 and 2). The optimal price caps for each case are listed in
0and PR
1that the firm will be allowed to charge
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 12

462FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
Table 4 – Optimal ex ante price caps
Retail prices
Investment at t = 1
with prob. pCasesConditions Investment at t = 0
PR
1= C + 2c − γ¯P and PR
PR
PR
1
NPV = OD
NPV > OD
1=¯P and PR
2= γ¯P
n/a
n/a
PR
2= γ¯P
2
NPV = OD
NPV > OD
1=¯P and PR
1=¯P and PR
2= γ¯PPR
PR
1=¯P and PR
1=¯P and PR
2= CH+ c
2= CH+ c
2= CH+ c + M
Proposition 2. We use the following terminology, C =
average fixed cost of investing at t = 0, CH=
cost of investing at t = 1 under high demand, and CL=
average fixed cost of investing at t = 1 under low demand.
Proposition 2: Table 4 summarizes the optimal price caps as a
function of market conditions.
I
Qdenotes the
I
uQis the average fixed
I
dQis the
Proof: First, suppose that the unregulated market conditions are
such that Case 1 is valid, that is, the NPV function crosses the OD
function in its first range.
If Q is such that NPV = OD = 0 there is no need for regulation
as the best the regulator can do is to replicate the unregulated
market outcome by setting PR
regulator sets regulated prices below market levels the firm will not
invest. If Q is such that NPV > OD ≥ 0 then the regulator can set,
for instance, PR
Q
have NPV = OD = 0.8In this case the firm invests at t = 0 and total
welfare is equal to:
?
The welfare obtained with this regulatory policy must be com-
pared to the unregulated market welfare. The difference between (8)
and (3) is equal to (1 − α)NPV > 0. Note that all the firm’s surplus
is extracted and the firm still invests at t = 0. This is the optimal
regulation since α < 1.
Second, suppose that demand characteristics are such that
Case 2 is valid, that is, the NPV function crosses the OD function
1=¯P and PR
2= γ¯P. Indeed, if the
1=¯P −NPV
= C + 2c − γ¯P and PR
2= γ¯P such that we
WR=
¯P −
?
¯P −NPV
Q
??
Q
(8)
8
PR
In this case, Proposition 2 holds for any regulated prices such that
1+ PR
2= C + 2c.
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 13

RETAIL PRICE REGULATION AND THE OPTION TO DELAY463
in its second range. When Q is such that NPV = OD, OD
and OD−= 0, the minimum regulated prices that induce investment
at t = 0 are PR
any price setting below market levels induces the firm to invest at
t = 1 if demand turns out to be high or even to not invest. In this case
the overall welfare is equivalent to the unregulated market welfare,
that is, αNPV. On the other hand, the minimum regulated prices
that induce investment at t = 1 if demand turns out to be high are
PR
c then the firm would not invest. In this case the overall welfare is
equal to:
p?γ¯P − (CH+ c)?uQ
Thus, this price regulation is optimal only if p >αNPV
+> 0
1=¯P and PR
2= γ¯P. Indeed, it is easy to see that
1=¯P and PR
2= CH + c. If the regulator were to set PR
2< CH +
WR=
(1 + r)
= pπH
t=1
(9)
πH
t=1.
When Q is such that NPV > OD, OD+> 0 and OD−≥ 0, the
minimum regulated prices that induce investment at t = 0 are PR
¯P and PR
As OD+> 0 and OD−= 0, we have CH + c < PR
there is a M > 0, such that PR
OD. The total welfare at t = 0 is given by:
WR=?γ¯P − (CH+ c + M)?Q + α???¯P − c?+ ((CH+ c + M) − c)?Q − I?
On the other hand, the minimum regulated prices that induce
investment at t = 1 if demand turns out to be high are PR
PR
Then, the price setting PR
when
?γ¯P − (CH+ c + M)?Q + α???¯P − c?+ (CH+ M)?Q − I?
t=1
Proposition 2 can be understood intuitively as follows. When
demand characteristics are such that Case 1 prevails, even when
Q is such that NPV > OD > 0 the regulator is able to set prices
PR
invests at t = 0. This price regulation increases the overall welfare
because by setting prices below the choke prices without distorting
the incumbent’s decision of investing at t = 0, the positive impact on
consumers’ surplus is higher than the negative impact on the firm’s
profit (as α < 1). This retail regulation is optimal since all the surplus
is extracted from the firm and transferred to consumers.
1=
2< γ¯P such that NPV = OD < OD, OD
+> 0 and OD
2< CL + c. Then,
−= 0.
2= CH+ c + M < CL+ c and NPV =
(10)
1=¯P and
2= CH+ c. In this case the overall welfare is given by (9).
1=¯P and PR
2= CH + c is optimal
p >
πH
.
?
1and PR
2such that NPV = OD = 0 and the incumbent still
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 14

464FERNANDO T. CAMACHO AND FLAVIO M. MENEZES
When demand characteristics are such that Case 2 holds, the
regulator is unable to set regulated prices below a certain level
without distorting the firm’s investment decision. While for some
parameter values, it is best to ensure that the firm invests at
t = 0, for other parameter values society is better off with lower
prices but investment occurring only at t = 1. Therefore, the optimal
price caps will take one of the two following forms: (i) the minimum
PR
t = 0 – under these prices the firm will have a positive ex-
pected payoff; or (ii) the minimum prices PR
(PR
1 only if the high demand eventuates - under these prices the firm
will have a zero payoff.
In summary, there is a trade off between paying higher prices
and having the service provision with certainty earlier (t = 0) or
paying lower prices and having the service later (t = 1) only if
the high demand eventuates, that is, there is a probability (1 − p)
that the service will not be provided to consumers. Thus, if p, the
probability of the high demand state, is sufficiently high, the optimal
regulation will be (ii). Otherwise, it will be optimal to set the price
caps such that the firm will have a positive payoff to compensate for
the option to delay.
This result differs from the earlier literature that states that
the incumbent firm should always receive a positive compensation to
account for the option to delay value. Instead, Proposition 2 shows
that the extent of compensation, if any, will depend on market
conditions. However, when optimal price caps do include a positive
compensation for the option to delay value (that is, the firm will
earn a positive NPV), traditional regulation that sets prices so that
the NPV is equal to zero, fails to provide the correct investment
incentives.
1and PR
2such that NPV = OD > 0 and the incumbent invests at
1and PR
2such that
2− c)uQ = I, OD+= 0, NPV < 0 and the firm invests at t =
4Conclusion
In this paper we examine a simple two-period model in a
network industry characterized by demand uncertainty, economies of
scale and sunk costs. In this model a firm may invest in the first
period or wait until the second period to decide whether to invest in
the network.
In the absence of regulation we identify the market conditions
under which a monopolist decides to invest early as well as the
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009

Page 15

RETAIL PRICE REGULATION AND THE OPTION TO DELAY465
underlying overall welfare output. In a regulated environment, we
consider a monopolist who faces no downstream (final good) competi-
tion but is subject to retail price regulation. We identify the welfare-
maximizing regulated prices when the unregulated market outcome
is set as the benchmark. We show that if the regulator can commit
to ex post regulation – that is, regulated prices that are contingent
to future demand realization – then regulated prices that allow the
firm to recover its total costs of production are welfare-maximizing.
Thus, under ex post price regulation there is no need to compensate
the regulated firm for the option to delay that it foregoes when
investing today. We argue, however, that regulators cannot make
this type of commitment and, therefore, price regulation is often ex
ante – that is, regulated prices are not contingent to future demand.
We also examine ex ante regulation. Differently from earlier
literature, we explicitly determine regulated prices that account for
demand uncertainty and the irreversibility of investments. In addi-
tion, we show that whether optimal retail prices should incorporate
a positive amount to compensate for the option to delay value will
depend on demand conditions.
REFERENCES
COMMERCE COMMISSION, 2002, Application of a TSLRIC Pric-
ing Methodology – Discussion paper, December. Available at www.
comcom.govt.nz/IndustryRegulation/Telecommunications/Intercon
nectionDeterminations/TotalServiceLongRunningIncrementalCost/
ContentFiles/Documents/TSLRIC%20discussion%20paper%202%
20July%2020020.pdf.
HAUSMAN J. and MYERS S., 2002, ‘Regulating the US railroads:
the effects of sunk costs and asymmetric risk’, Journal of Regula-
tory Economics, 22:3, 287–310.
HAUSMAN J., 1999, ‘The effect of sunk costs in telecommunica-
tions’, in J. Alleman and E. Noam, eds, The New Investment
Theory of Real Options and its Implications for the Costs Models
in Telecommunications, Kluwer Academic Publishers.
OFCOM, 2005, Ofcom’s approach to risk in the assessment of the
cost of capital.
PINDYCK R. S., 2004, ‘Mandatory unbundling and irreversible in-
vestment in telecom networks’, NBER Working Paper No. 10287.
C ?2009 The Authors
Journal compilation C ?CIRIEC 2009