Yasunori Fujita
Professor of Economics, Keio University, Tokyo, Japan.
DOI: 10.4236/tel.2026.162029   PDF    HTML   XML   119 Downloads   626 Views   Citations

Abstract

This paper examines irreversible market entry under uncertainty in a Cournot oligopoly framework. While standard static Cournot models imply that increased competition monotonically reduces deadweight loss, this paper shows that intensified competition can amplify dynamic welfare losses when entry is irreversible and demand conditions are uncertain. Using a real-options approach, we derive closed-form expressions for privately and socially optimal entry thresholds and demonstrate that the wedge between them increases with the number of firms. Although greater competition improves static allocative efficiency by lowering prices, it reduces private entry incentives more rapidly than social benefits, leading firms to delay entry inefficiently. As a result, dynamic welfare losses arising from delayed market creation grow even as static inefficiency vanishes. These findings highlight a fundamental tension between static and intertemporal welfare effects of competition.

Keywords

Dynamic Entry, Cournot Competition, Real Options, Entry Timing, Welfare Loss

Share and Cite:

1. Introduction

In static Cournot models, an increase in number of firms reduces market power and monotonically decreases deadweight loss. In the limit, the Cournot equilibrium converges to the perfectly competitive outcome, a result known as the Cournot limit theorem. This classical insight, originating in Cournot (1838) and formalized in modern treatments such as Vives (1999) and Tirole (1988), firmly establishes competition as welfare improving in a static sense.

A large literature has extended this static framework to study market structure and welfare. Early contributions include Dixit (1979) and Salop (1979), who emphasized excess entry and product differentiation, while subsequent work such as Mankiw and Whinston (1986) analyzed entry decisions under perfect foresight and certainty. In these models, welfare losses arise from excessive duplication of fixed costs, but the timing of entry itself plays no role.

More recently, a growing literature has incorporated uncertainty and irreversibility into entry and investment decisions using real options theory. Seminal contributions by McDonald and Siegel (1986) and Dixit and Pindyck (1994) demonstrate that uncertainty creates an option value of waiting, leading firms to delay irreversible investment. These insights have been applied to market entry, trade liberalization, and industrial organization (e.g., Dixit, 1989; Leahy, 1993; Grenadier, 2002). In these studies, delayed entry is privately optimal but may be socially inefficient.

A related strand of research examines the interaction between competition and investment timing. Abel and Eberly (1996) and Caballero (1991) show that competitive pressures affect investment thresholds, while Fudenberg and Tirole (1985) and Grenadier (1996) emphasize strategic preemption and waiting in oligopolistic settings. However, much of this literature focuses on strategic interactions among a small number of firms or on preemption games, rather than on the welfare implications of entry timing under intensified competition.

This paper contributes to these literatures by highlighting a novel tension between static and dynamic welfare effects of competition. While increased competition reduces static deadweight loss in Cournot markets, we show that, under uncertainty and irreversible entry, intensified competition can amplify dynamic welfare losses arising from delayed entry. The key mechanism is a growing wedge between the privately and socially optimal entry thresholds, which expands monotonically with the number of firms. Importantly, this result does not contradict the Cournot limit theorem; rather, it complements it by showing that competition can simultaneously mitigate static inefficiency and exacerbate intertemporal inefficiency.

The remainder of the paper is organized as follows. Section 2 introduces the market environment. Section 3 derives the Cournot equilibrium and static welfare. Sections 4 and 5 characterize private and socially optimal entry timing, respectively. Section 6 analyzes the dynamic welfare loss from delayed entry and relates our findings to the Cournot limit theorem. In Section 7, we show that the excess entry theorem (e.g., Mankiw & Whinston, 1986) holds in the setting of this study as well. Section 8 concludes.

2. Market Environment

Consider a market where n symmetric firms exist. $x_i$ and P denotes firm i’s output and market price, respectively. We specify the inverse demand function as

$P=a-b_t{\displaystyle {\sum }_{i=1}^n{x}_i},$ (1)

where $a$ is a positive constant. We also specify each firm’s marginal cost as $c<a$ .

The demand slope parameter $b_t$ is assumed to evolve stochastically according to a geometric Brownian motion

$d{b}_t=-\mu b_{t}dt+\sigma b_{t}d{z}_t,$ (2)

where $\mu >0$ captures the expected rate at which market conditions improve (i.e., demand becomes less price sensitive), $\sigma >0$ measures uncertainty, and $z_t$ is a standard Brownian motion. The discount rate is $\rho >0$ .

Entry into the market is irreversible. When firms enter, each incurs a sunk cost $I>0$ . After entry, no further fixed costs are incurred and Cournot competition prevails forever.

We assume that firms are identical and follow identical threshold strategies. Consequently, entry occurs simultaneously when the state variable reaches the common threshold. This justifies the aggregate sunk cost term $nI$ that appears in section 5.

3. Cournot Equilibrium and Static Welfare

Given $n$ active firms and a realization of $b_t$ , the standard Cournot analysis yields firm i’s output, total output and price in the Cournot equilibrium,

$x_i^{*}=\frac{a-c}{b_t\left(n+1\right)};$ (3)

$Q^{*}=\frac{n\left(a-c\right)}{b_t\left(n+1\right)};$ (4)

$P^{*}=\frac{a+nc}{n+1}.$ (5)

Thus, firm i’s profit, ${\pi }_i^{*}\left(b_t,n\right)$ , consumer surplus, $C{S}^{*}\left(b_t,n\right)$ and total surplus, $T{S}^{*}\left(b_t,n\right)$ are determined, respectively, as

${\pi }_i^{*}\left(b_t,n\right)=\frac{{\left(a-c\right)}^2}{b_t{\left(n+1\right)}^2};$ (6)

$C{S}^{*}\left(b_t,n\right)=\frac{n^2{\left(a-c\right)}^2}{2b_t{\left(n+1\right)}^2};$ (7)

$T{S}^{*}\left(b_t,n\right)=\frac{n\left(n+2\right){\left(a-c\right)}^2}{2b_t{\left(n+1\right)}^2}.$ (8)

From (8), we see that for a given $b_t$ , total surplus increases with $n$ , and static deadweight loss decreases monotonically, converging to zero as $n\to \infty$ .

4. Entry under Uncertainty: Private Incentives

If a firm i enters when the state variable equals $b_t$ , it obtains a perpetual profit flow ${\pi }_i^{*}\left(b_t,n\right)$ . Present discounted value of entry for an individual firm evaluated at the entry time is therefore $V_P\left(b_t\right)=\frac{{\pi }_i^{*}\left(b_t,n\right)}{\rho +\mu }-I$ , which is rewritten as

$V_P\left(b_t\right)=\frac{K_P}{b_t}-I,$ (9)

where

$K_P=\frac{{\left(a-c\right)}^2}{{\left(n+1\right)}^2\left(\rho +\mu \right)}.$ (10)

Prior to entry, on the other hand, firm i holds an option to wait. The characteristic equation is obtained from the homogeneous part of the Bellman equation:

$\frac{1}{2}{\sigma }^2{b}_t^2{V}^{″}+\mu b_t{V}^{\prime }-\rho V=0,$ (11)

which yields the standard quadratic equation for the exponent.

Standard real-options arguments imply that the optimal entry policy is characterized by a threshold $b_P^{*}$ : the firm enters when $b_t\le b_P^{*}$ .

Let $\underset{\_}{\beta }$ denote the negative root of

$\frac{1}{2}{\sigma }^2\beta \left(\beta -1\right)-\mu \beta -\rho =0,$ (12)

which is

$\underset{\_}{\beta }=\frac{\frac{1}{2}{\sigma }^2+\mu -\sqrt{{\left(\frac{1}{2}{\sigma }^2+\mu \right)}^2+2\rho {\sigma }^2}}{{\sigma }^2},$ (13)

Then the privately optimal entry threshold is

$b_P^{*}=\frac{K_P}{I}\left(1+\frac{1}{\underset{\_}{\beta }}\right).$ (14)

We assume that $b_P^{*}>0$ , which requires $\underset{\_}{\beta }<-1$ ; for this purpose, we impose $\rho >\mu +{\sigma }^2$ .

From (14), we see that an increase in the number of firms reduces $K_P$ and therefore lowers $b_P^{*}$ , implying that firms optimally delay entry as competition intensifies.

5. Socially Optimal Entry Timing

From a social perspective, entry generates total surplus, $T{S}^{*}\left(b,n\right)$ , each period but requires the payment of the total sunk cost $nI$ . Present discounted value of social surplus evaluated at the entry time is $V_S\left(b_t\right)=\frac{T{S}^{*}\left(b_t,n\right)}{\rho +\mu }-nI$ , which is rewritten as

$V_S\left(b_t\right)=\frac{K_S}{b_t}-nI,$ (15)

where

$K_S=\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)}.$ (16)

Applying the same optimal stopping logic, the socially optimal entry threshold $b_S^{*}$ is determined as

$b_S^{*}=\frac{K_S}{nI}\left(1+\frac{1}{\underset{\_}{\beta }}\right).$ (17)

A direct comparison of (14) and (17) yields

$b_S^{*}=\frac{n+2}{2}{b}_P^{*}.$ (18)

From (18), we have the following lemma.

Lemma. $b_S^{\ast }>b_P^{\ast }$ for any $n\ge 1$ .

This lemma implies that society prefers entry to occur earlier (at a higher value of $b$ ) than firms do privately.

6. Dynamic Welfare Loss from Delayed Entry

The welfare loss associated with delayed entry is defined as the difference between sum of discounted present value of social welfare after entry occurs at the socially optimal threshold and sum of discounted present value of social welfare after entry occurs at the privately optimal threshold.

When entry occurs at $b_t=b_S^{*}$ , the present discounted value of total surplus thereafter, evaluated at the time $b_t=b_S^{*}$ , is given by

$W_S\left(b_S^{*}\right)=\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)b_S^{*}}-nI,$ (19)

which is equivalent to (15) with (16) and $b_t=b_S^{*}$ substituted into it.

On the other hand, when entry occurs at $b_t=b_P^{*}$ , the present value of total surplus thereafter, evaluated also at the time $b_t=b_S^{*}$ , is given by

$W_P\left(b_P^{*},b_S^{*}\right)=\left(\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)b_P^{*}}-nI\right){\left(\frac{b_S^{*}}{b_P^{*}}\right)}^{\underset{\_}{\beta }},$ (20)

where $\underset{\_}{\beta }$ is the parameter defined in (13).

Therefore, the difference between (19) and (20) is

$\begin{array}{c}{W}_S\left(b_S^{*}\right)-W_P\left(b_P^{*},b_S^{*}\right)=\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)b_S^{*}}-nI\\ -\left(\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)b_P^{*}}-nI\right){\left(\frac{b_S^{*}}{b_P^{*}}\right)}^{\underset{\_}{\beta }},\end{array}$ (21)

which is rewritten as

$W_S\left(b_S^{*}\right)-W_P\left(b_P^{*},b_S^{*}\right)=\frac{n{\left(a-c\right)}^2}{{\left(n+1\right)}^2\rho }{b}_P^{*}\left(1-{\left(\frac{n+2}{2}\right)}^{\underset{\_}{\beta }+1}\right)-nI\left(1-{\left(\frac{n+2}{2}\right)}^{\underset{\_}{\beta }}\right).$ (22)

by using (18).

Substituting (10) and (14), and introducing the notation

$\alpha =-\underset{\_}{\beta },$ (23)

we obtain

$W_S\left(b_S^{*}\right)-W_P\left(b_P^{*},b_S^{*}\right)=nI\frac{\alpha -1}{\alpha }\left[1-\alpha {\left(\frac{2}{n+2}\right)}^{\alpha -1}+\left(\alpha -1\right){\left(\frac{2}{n+2}\right)}^{\alpha }\right].$ (24)

Here, let

$g\left(n\right)=1-\alpha {\left(\frac{2}{n+2}\right)}^{\alpha -1}+\left(\alpha -1\right){\left(\frac{2}{n+2}\right)}^{\alpha }$ , (25)

and let the right-hand side of (24) be denoted as a function of $n$ by $f\left(n\right)$ . Then we have

$f\left(n\right)=\frac{\left(\alpha -1\right)I}{\alpha }ng\left(n\right)$ , (26)

and therefore,

$f\left(0\right)=0$ , (27)

$f^{\prime }\left(n\right)=\frac{\left(\alpha -1\right)I}{\alpha }\left(g\left(n\right)+n{g}^{\prime }\left(n\right)\right)$ . (28)

Here,

$g^{\prime }\left(n\right)=\frac{\alpha \left(\alpha -1\right)n}{{\left(n+2\right)}^2}{\left(\frac{2}{n+2}\right)}^{\alpha -1}$ , (29)

and thus, for $n>0$ and $\alpha >1$ ,

$g^{\prime }\left(n\right)>0$ . (30)

On the other hand, let $t=\frac{2}{n+2}$ . Then for $n>0$ , we have $0<t\le 1$ , and $g\left(n\right)$ can be written as

$g\left(t\right)=1-\alpha t^{\alpha -1}+\left(\alpha -1\right)t^{\alpha }$ . (31)

Now,

$g^{\prime }\left(t\right)=\alpha \left(\alpha -1\right)t^{\alpha -2}\left(t-1\right)\le 0$ , (32)

so $g\left(t\right)$ is monotonically decreasing on $0<t\le 1$ . Moreover,

$g\left(1\right)=1-\alpha +\left(\alpha -1\right)=0$ . (33)

Therefore, for $0<t\le 1$ , that is, for $n>0$ ,

$g\left(n\right)\ge 0$ . (34)

Hence,

$f^{\prime }\left(n\right)=\frac{\left(\alpha -1\right)I}{\alpha }\left(g\left(n\right)+n{g}^{\prime }\left(n\right)\right)>0$ (35)

is obtained.

Thus, we have the following proposition.

Proposition. While static deadweight loss vanishes as $n$ increases, the dynamic welfare loss arising from delayed entry becomes more severe. More precisely, when welfare loss is evaluated as the difference between social value under entry at the socially optimal threshold $b_S^{\ast }$ and social value under entry at the privately optimal threshold $b_P^{\ast }$ , its absolute magnitude is strictly increasing in $n$ .

The Cournot limit theorem concerns static allocative efficiency: for a given market, increasing the number of firms reduces price distortions and deadweight loss. The welfare loss identified in this paper is fundamentally different. It is intertemporal and arises from delayed market creation under uncertainty and irreversibility.

Stronger competition reduces private entry incentives more rapidly than it increases social benefits, because firms do not internalize consumer surplus. Consequently, competition mitigates static inefficiency while amplifying dynamic inefficiency. The two effects coexist and operate in opposite directions.

7. Excess Entry in the Present Setting

In this section, we show that the excess entry theorem holds in the setting of this paper.

Let $n$ denote the number of firms. Suppose that all firms enter when $b_t=b_P^{*}$ . Let $V\left(n\right)$ be the present discounted value of total surplus evaluated at the initial state $b_t=b_0$ . Then,

$V\left(n\right)=\left(\frac{n\left(n+2\right){\left(a-c\right)}^2}{2{\left(n+1\right)}^2\left(\rho +\mu \right)b_P^{*}}-nI\right){\left(\frac{b_0}{b_P^{*}}\right)}^{\underset{\_}{\beta }},$ (36)

where $\underset{\_}{\beta }$ is the parameter defined in (13).

Substituting (10) and (14) into (36) yields

$V\left(n\right)=\left(\frac{n\left(n+2\right)\rho I}{2\left(\rho +\mu \right)}\frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}-nI\right){\left(b_0\cdot \frac{{\left(n-1\right)}^2\rho I}{{\left(a-c\right)}^2}\cdot \frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}\right)}^{\underset{\_}{\beta }}.$ (37)

Define the positive constant (independent of $n$ )

$L=I{\left(b_0\cdot \frac{\rho I}{{\left(a-c\right)}^2}\cdot \frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}\right)}^{\underset{\_}{\beta }}.$ (38)

Then

$V\left(n\right)=Ln\left(\frac{\rho }{2\left(\rho +\mu \right)}\cdot \frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}\left(n+2\right)-I\right){\left(n+1\right)}^{2\underset{\_}{\beta }}.$ (39)

Let

$f\left(n\right)=n\left(\frac{\rho }{2\left(\rho +\mu \right)}\cdot \frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}\left(n+2\right)-I\right){\left(n+1\right)}^{2\underset{\_}{\beta }}.$ (40)

Maximizing $V\left(n\right)$ is equivalent to maximizing $f\left(n\right)$ . Taking logs and differentiating yields

$\frac{f^{\prime }\left(n\right)}{f\left(n\right)}=\frac{1}{n}+\frac{A}{An+2A-1}+\frac{2\underset{\_}{\beta }}{n+1},$ (41)

where

$A=\frac{\rho }{2\left(\rho +\mu \right)}\cdot \frac{\underset{\_}{\beta }}{\underset{\_}{\beta }+1}.$ (42)

The first-order condition $f^{\prime }\left(n\right)=0$ implies

$\frac{1}{n}+\frac{A}{An+2A-1}+\frac{2\underset{\_}{\beta }}{n+1}=0,$ (43)

which can be rewritten as

$\rho \underset{\_}{\beta }\left(\underset{\_}{\beta }+1\right)n^2+\left(\underset{\_}{\beta }+1\right)\left(\left(2\underset{\_}{\beta }+1\right)\mu -\rho \right)n+\left(\underset{\_}{\beta }+1\right)\mu -\rho =0.$ (44)

Hence, the socially optimal number of firms is determined as

$n^{*}=\frac{-\left(\underset{\_}{\beta }+1\right)\left(\left(2\underset{\_}{\beta }+1\right)\mu -\rho \right)+\sqrt{\text{Δ}}}{2\rho \underset{\_}{\beta }\left(\underset{\_}{\beta }+1\right)},$ (45)

where

$\text{Δ}={\left(\underset{\_}{\beta }+1\right)}^2{\left(\left(2\underset{\_}{\beta }+1\right)\mu -\rho \right)}^2-4\rho \underset{\_}{\beta }\left(\underset{\_}{\beta }+1\right)\left(\left(\underset{\_}{\beta }+1\right)\mu -\rho \right).$ (46)

Therefore, free entry (i.e., $n\to \infty$ ) is not optimal. We conclude that the excess entry theorem holds in the present setting.

8. Conclusion

This paper has analyzed irreversible market entry under uncertainty in a Cournot environment. Using a linear inverse demand system and constant marginal costs, we derived closed-form expressions for privately and socially optimal entry thresholds. The analysis reveals a systematic wedge between private and social entry incentives, reflecting the fact that firms internalize only their own profits, whereas society also values consumer surplus generated by entry.

Unlike the preemption literature (e.g., Fudenberg & Tirole, 1985; Grenadier, 1996), we abstract from strategic asymmetries and focus on symmetric threshold behavior. Incorporating preemption would likely lead to earlier entry and reduce the delay identified here.

A key result is that intensified competition has opposing effects on static and dynamic efficiency. Consistent with the Cournot limit theorem, an increase in the number of firms reduces static deadweight loss and brings market outcomes closer to perfect competition. At the same time, stronger competition lowers private entry incentives more rapidly than social benefits, thereby widening the gap between socially and privately optimal entry times. This gap gives rise to a dynamic welfare loss associated with delayed entry, which is conceptually distinct from static allocative inefficiency.

The paper therefore highlights that competition can simultaneously improve static efficiency and worsen dynamic efficiency when entry is irreversible and subject to uncertainty. This insight suggests that welfare evaluations based solely on static considerations may be incomplete. Competition policy, deregulation, and market-opening reforms should take into account their impact on entry timing, as intensified competition may exacerbate welfare losses arising from socially inefficient delays. Future research could extend the analysis to settings with endogenous firm numbers, strategic preemption, or open-economy interactions.

Conflicts of Interest

The author declares no conflicts of interest regarding the publication of this paper.

References

[1]	Abel, A. B., & Eberly, J. C. (1996). Optimal Investment with Costly Reversibility. The Review of Economic Studies, 63, 581-593. [Google Scholar] [CrossRef]
[2]	Caballero, R. J. (1991). On the Sign of the Investment-Uncertainty Relationship. American Economic Review, 81, 279-288.
[3]	Cournot, A. A. (1838). Recherches sur les principes mathématiques de la théorie des richesses. Hachette.
[4]	Dixit, A. K. (1979). A Model of Duopoly Suggesting a Theory of Entry Barriers. The Bell Journal of Economics, 10, 20-32. [Google Scholar] [CrossRef]
[5]	Dixit, A. K. (1989). Entry and Exit Decisions under Uncertainty. Journal of Political Economy, 97, 620-638. [Google Scholar] [CrossRef]
[6]	Dixit, A. K., & Pindyck, R. S. (1994). Investment under Uncertainty. Princeton University Press. [Google Scholar] [CrossRef]
[7]	Fudenberg, D., & Tirole, J. (1985). Preemption and Rent Equalization in the Adoption of New Technology. The Review of Economic Studies, 52, 383-401. [Google Scholar] [CrossRef]
[8]	Grenadier, S. R. (1996). The Strategic Exercise of Options: Development Cascades and Overbuilding in Real Estate Markets. The Journal of Finance, 51, 1653-1679. [Google Scholar] [CrossRef]
[9]	Grenadier, S. R. (2002). Option Exercise Games: An Application to the Equilibrium Investment Strategies of Firms. Review of Financial Studies, 15, 691-721. [Google Scholar] [CrossRef]
[10]	Leahy, J. V. (1993). Investment in Competitive Equilibrium: The Optimality of Myopic Behavior. The Quarterly Journal of Economics, 108, 1105-1133. [Google Scholar] [CrossRef]
[11]	Mankiw, N. G., & Whinston, M. D. (1986). Free Entry and Social Inefficiency. The RAND Journal of Economics, 17, 48-58. [Google Scholar] [CrossRef]
[12]	McDonald, R., & Siegel, D. (1986). The Value of Waiting to Invest. The Quarterly Journal of Economics, 101, 707-727. [Google Scholar] [CrossRef]
[13]	Salop, S. C. (1979). Monopolistic Competition with Outside Goods. The Bell Journal of Economics, 10, 141-156. [Google Scholar] [CrossRef]
[14]	Tirole, J. (1988). The Theory of Industrial Organization. MIT Press.
[15]	Vives, X. (1999). Oligopoly Pricing: Old Ideas and New Tools. MIT Press.