key: cord-0997569-lcy2hh0d
authors: Goerke, Laszlo
title: A political economy perspective on horizontal FDI in a dynamic Cournot-oligopoly with endogenous entry
date: 2020-05-01
journal: Eur J Polit Econ
DOI: 10.1016/j.ejpoleco.2020.101897
sha: 3f196e9de6de86f4fe8efbefcea64483407197d3
doc_id: 997569
cord_uid: lcy2hh0d

Abstract Entry in a homogeneous Cournot-oligopoly is excessive if and only if there is business-stealing (Amir et al., 2014). The excessive entry prediction has been derived primarily for closed economies and using a welfarist benchmark. We extend this framework and allow for (1) horizontal FDI in a multi-period setting and (2) interest group-based government behaviour. Opening the market to greenfield investments from abroad tends to aggravate the entry distortion. Moreover, market opening may reduce welfare if a more pronounced entry distortion dominates the gain in consumer surplus. Finally, a government, which places sufficiently little weight on the interests of consumers, will object to market opening, even if welfare rises.

Between 2002 and 2016, the stock of world-wide Foreign Direct Investments (FDI) as a share of GDP has risen by about 60% (Carril-Caccia and Pavlova 2018) . While this development has not been reflected in according variations in flows, greenfield investments have picked up again in 2018 (UNCTAD 2019). Moreover, many (consumer) goods have become more homogeneous, in that increasingly similar items are produced and sold in different countries.

This implies that knowledge about production techniques or distribution networks gained in one country can be utilised in another location as well. In consequence, the fixed costs of setting up a further production site abroad are likely to be lower than the costs of entering the first, domestic market. Hence, the incentives to undertake horizontal FDI have increased.

Whether this development will continue after the coronavirus pandemic has subsided is hard to predict at the moment.

The increased relevance of FDI occurred at the same time at which goods markets have become less competitive and more oligopolistic. Head and Spencer (2017) , for example, commence their call for greater attention to oligopolies in the analysis of trade by asserting that "Oligopoly is pervasive in our daily live." (p. 1415). This statement can be accentuated by findings that concentration of industries has increased considerably in the United States in recent decades (Autor et al. 2017 , Grullon et al. 2019 , and hardly declined in European Union (Gutiérrez and Philippon 2019) . Accordingly, markets are far from being competitive and many industries feature oligopolistic characteristics. If big firms can cope with the consequences of the crisis commencing in 2020 more successfully than smaller firms, which may also receive less government aid, oligopolistic markets will become even more prevalent within the next years.

In our subsequent analysis we combine these aspects and enquire how horizontal FDI affects outcomes in oligopoly. We focus on a long-run, free entry setting which allows for adjustments also at the extensive margin to accommodate the fact that oligopolistic markets are not static but change their composition. In order to ease comparability with earlier investigations, we assume a homogeneous Cournot-oligopoly with identical firms, linear demand and cost schedules and fixed costs of market entry. In such a setting, the number of entrants will be excessive in a closed economy if entry reduces output per firm, that is, if there is businessstealing (von Weizsäcker 1980 , Mankiw and Whinston 1986 , Suzumura and Kiyono 1987 , Amir et al. 2014 . We extend the analysis to a multi-country, two-period framework. In period one, there are only domestic competitors. At the beginning of period two, the market may be opened up to horizontal FDI. Entry costs arise in each period, whereas additional set-up costs 3 only occur in the first period of activity and for one market. Therefore, incumbents that undertake FDI do not have to incur set-up costs again. This captures the idea that horizontal (greenfield) FDI activities can utilise knowledge gained in earlier domestic production, such that firms spread fixed costs over more markets and, thereby, greater quantities.

In order to focus on the effects of FDI, we consider two settings: In the first, entry is not regulated. Firms take up production, as long it is profitable. As FDI affects contemporaneous payoffs, market opening alters entry choices in the second period and, if anticipated, also first period decisions. In the second setting, the government regulates entry, for example, by granting costless entry licences. Because such regulation is no longer effective if market opening occurs in period two, it also changes the government's first period choice. In contrast to earlier analyses, we assume that the government maximises a political support function, and not necessarily welfare. This modification enables us to investigate how the nature of the government's objective affects choices. For both settings, the unregulated market equilibrium and the one in which entry is restricted, we initially consider a closed economy. This provides a benchmark, to which we compare outcomes, which will result if there is horizontal FDI.

Our analysis shows that the number of firms in a closed economy will be constant over time, though profits vary. This is true in market equilibrium and likewise if the government determines the maximum number of entrants. In this way, unit production costs are minimised.

Moreover, and irrespective of the government's objective, entry usually is excessive, not only in the closed economy, but also if horizontal FDI takes place. The intuition is as follows:

Opening the market in period two raises the number of firms in market equilibrium in that period and reduces it in period one. These adjustments result in an increase in total output and, hence, welfare. The expansion comes about because production costs per firm decline and the cost effect of the drop in the number of firms in period one dominates the impact of additional entry in period two. If the government regulates entry, the change in the number of firms will exhibit qualitatively the same features as in market equilibrium. Therefore, entry tends to be excessive also in the presence of FDI. We further show that market opening may actually decrease welfare if entry is regulated. Such outcome can occur if the number of firms is relatively high prior to market opening. Therefore, the increase in output and consumer surplus due to market opening is limited and does not compensate the reduction in profits. Finally, we demonstrate that even if market opening raises welfare, the government's payoff may decline.

This can be the case if the fall in profits because of intensified competition is substantial, while the rise in consumer surplus is not valued sufficiently by the government. 4 This survey of results indicates that we can add to the literature in four ways: First, we clarify that the excessive entry prediction, derived almost exclusively for closed-economy settings, can also result if there is FDI. Second, we demonstrate how an intertemporal optimisation process of firms can affect outcomes. Third, we illustrate new channels by which international integration can alter welfare. Fourth, we evaluate in how far the government's incentives to regulate entry depend on whose interests it pursues.

The above summary clarifies that our investigation is related primarily to two strands of literature, namely analyses of Cournot-oligopolies with endogenous market structure and of oligopolies in international trade. The excessive entry prediction has initially been derived by von Weizszäcker (1980), Perry (1984) , Brander and Spencer (1985) , Mankiw and Whinston (1986) , and Suzumura and Kiyono (1987) . Amir et al. (2014) generalise assumptions and show that "free entry leads to an excessive (resp., insufficient) number of firms relative to secondbest planning if and only if a "business stealing" (resp., "business enhancing") effect is present … ." (p. 113) The robustness of the excessive entry prediction has been investigated for various extensions, such as imperfectly competitive input markets (Okuno-Fujiwara and Suzumura, 1993 , Ghosh and Morita, 2007a ,b, Mukherjee 2009 , 2013 , de Pinto and Goerke 2020 , R&D investments (Okuno-Fujiwara and Suzumura, 1993 , Haruna and Goel, 2011 , Mukherjee, 2012a , Chao et al. 2015 , Mukherjee and Ray 2014 ,asymmetric costs Saha 2007, Mukherjee 2012a) , and alternative firm objectives (Varian 1995 , Suzumura 1995 , chap. 8, Hamada et al. 2018 ).

Analyses of the excess entry result have occasionally been expanded to open economy settings. Mukherjee (2013, 2015) and Mukherjee (2013) assume that there is one foreign firm which enters the domestic oligopoly. They show that if trade costs decline, so do the foreign firm's costs, such that it produces a greater amount. This, in turn, reduces profits of domestic firms and deters their entry. Moreover, the number of firms in market equilibrium may be insufficient. The rationale for this outcome is that a government, which maximises domestic welfare, takes into account that additional entry raises domestic consumer surplus at the expense of foreign profits. However, this cross-country effect plays no role for the firms' entry decision in market equilibrium. 1 If input prices are given by wages determined by a trade union, lower transport cost may reduce welfare and entry is once again insufficient Mukherjee 2013, 2015) . The latter effect arises because wages deter entry, but do not have a direct welfare effect (see de Pinto and Goerke (2020) for a similar approach). Miyagiwa and Sato (2014) consider a two-country setting in which firms produce locally and can also export, facing linear trade costs. Domestic and foreign governments regulate entry by taxing operating profits of local firms. Miyagiwa and Sato (2014) show that entry is excessive, since domestic taxation fosters entry abroad. Moreover, if entry costs are high, there are few competitors. Allowing for trade intensifies competition substantially. This trade effect exceeds the negative consequences of excessive entry. Hence, trade raises welfare. If market entry costs are low, the gains from trade are also relatively moderate, such that the excessive entry distortion dominates and welfare declines. 2

The second relevant strand of literature considers trade between countries with oligopolistic markets in which the number of firms is determined by a zero-profit constraint. Trade raises welfare in comparison to autarky despite the occurrence of trade costs because fixed costs of entry can be distributed across a greater quantity (see, for example, Krugman 1983, Venables 1985) . Under additional conditions, which hold for a linear demand schedule, the total number of suppliers rises, but not necessarily of domestic firms (Venables 1985 , see also Anderson et al. 1989) . Tanaka (1993) considers a two-country free-trade setting. The introduction of a tariff raises welfare if demand is strictly concave. This effect occurs because the number of firms decreases, as Ikeda (2007) clarifies. In partial contrast, Amir et al. (2019) show that free trade can raise welfare and reduces the number of firms in a world without trade costs, such that international competition results in fully integrated markets. Finally, Stähler (2006) assumes a two-period setting in which domestic and foreign firms can enter their home market in period one and export part of their production. In period two, only incumbents can undertake horizontal FDI. He shows that exports and FDI may co-exist. In this case, the number of firms shrinks with FDI, while welfare rises, for example, if demand is linear.

In sum, the studies on excessive entry outcomes in Cournot-oligopolies have ignored foreign competition via greenfield FDI, in particular, if the government does not maximise welfare. In order to analyse this issue, the further paper proceeds as follows: In Section 2, we develop the model and describe output decisions and entry choices. Section 3 delineates the market equilibrium and the government's preferred outcome in an isolated, i.e., closed economy.

Section 4 considers market opening, both for the market equilibrium and if the government determines maximum entry. Section 5 concludes. Most proofs are relegated to an appendix.

There are m, m > 1, identical countries. In each of them, there is an oligopolistic market, in which an endogenously determined number of firms produce a homogeneous good. We consider a two-period setting, t = 1, 2, without discounting. In the first period, markets are isolated. Therefore, firms have to decide whether to set up one production site domestically. In market equilibrium this decision is determined by the level of profits. If the government regulates entry, a firm additionally requires a licence to take up production. At the beginning of period two, firms choose whether to continue production domestically or whether to take it up, if they had not done so in the previous period. Moreover, markets may be opened up. In this case, firms can undertake horizontal FDI, i.e. enter foreign markets by greenfield investments.

While a two-period framework is a stark simplification, it suffices to analyse the impact of horizontal FDI not only at a specific moment, but also over time.

The good under consideration is not tradeable and cannot be stored, such as it is the case with certain services, which, for example, require the physical presence of customers. Accordingly, the good is consumed in the country and period of production and markets are segmented.

Firms compete in quantities and take the choices of competitors as given (Cournot-Nash behaviour). In each country, the (inverse) demand schedule equals p(X ) = 1 − X , X = x + X _ , where p is the price, X the aggregate quantity, x the quantity produced and sold by firm i, and X _ the quantity sold by all other firms. 3 Variable costs equal cx , 0 ≤ c < 1. In order to serve the market in period t, a firm has to make an investment k , k > 0. In addition to these periodic market entry costs, k , the firm incurs fixed set-up costs, k, which are also sunk and for which k > |k − k | holds. Therefore, set-up costs, k, are sufficiently high to dominate any difference in periodic market-entry costs, k and k . Such set-up costs can, for example, be viewed as the expenditure for acquiring the knowledge about production or for purchasing the necessary patents. They arise only once, either domestically in period one or, for a late entrant, in period two in at most one market (see Horstmann and Markusen (1992) for a similar approach). If a firm does not produce, its payoff is zero.

The timing is as follows: First, all agents learn whether market opening will take place.

Second, at the beginning of period one firms decide upon entering the domestic market. Total 7 fixed costs of an entrant equal k + k . In market equilibrium, firms are not restricted in their decisions. If the government decides about entry, it grants costless entry licences and can, thereby, determine the maximum number of (domestic) entrants. 4 Third, entrants decide on output. Fourth, each firm which entered the market decides whether to exit at zero costs at the beginning of period two, or to remain active and incur fixed costs, k . Fifth, previously inactive firms can enter the market at the beginning of period two at costs k + k . Moreover, in case of market opening, firms can set up at most one production site in each foreign market, incurring fixed costs k in every case. Sixth, active firms choose second period output.

Given the assumptions outlined above, we can focus the analysis on one country. We treat n , the number of firms, n > 1, as a continuous variable and ignore the integer constraint (see, for example, Ghosh and Morita, 2007a ,b, Marjit and Mukherjee, 2013 , and Seade, 1980 . 5 Moreover, we consider an equilibrium in pure strategies.

Firm i chooses the quantity, x , to maximise operating profits, π , in period t.

Since all firms face the same cost and demand conditions, they behave identically and we subsequently omit the firm index, i. Output per firm and aggregate output can be written as:

Denoting the period of entry by τ, τ = 1, 2, resulting profits are given by

and Richardson (1999) interprets a reduction in the maximum number of entrants as looser competition policy. 5 Mankiw and Whinston (1986) have shown that profits may be higher if there are n competitors, and n is an integer, than if there are n + 1 firms, while welfare is higher for n +1 producers, assuming business stealing. Thus, entry may be insufficient in the presence of business-stealing by at most one firm in the presence of the integer constraint. An analysis of the interaction of the integer constraint and market opening on the extent of excessive or insufficient entry is beyond the scope of this contribution because such investigations of the integer constraint usually rely on alternative specifications of cost and demand conditions (cf. Galera and Garcia-del-Barrio 2011), which are as simple as possible in the present set-up. Therefore, this issue warrants a separate investigation. 8

In market equilibrium, each firm decides whether to take up production. It will become active if total operating profits weakly exceed the sum-of set-up and entry costs, that is, if π + π ≥ 0. If the firm enters only in one period, π (τ) ≥ 0, t = 1, 2, has to hold. Since n is a continuous variable, profit-constraints hold as equalities. In addition, we consider a setting in which the government determines the maximum number of entrants. 6

Investigations of the excess entry theorem usually assume that the government maximises the sum of consumer surplus and profits. However, government actions are not restricted to the creation of wealth, but also affect its allocation across groups. Accordingly, the weights of consumer surplus and profits in the government objective may not be the same. If, for example, entry regulation is determined by a government subject to, for example, the influence of lobbying groups, their relative size may determine the value of β. Restricting the number of entrants constitutes a public good (or bad) for consumers and firms (if n > 1). Since small groups can overcome a free-rider problem more easily (Olson 1965) , the impact of firms on the government's payoff may exceed that of consumers (Hillman 1989 ). However, Amir et al.

(2019) forcefully argue that competition authorities may pursue a "populist" objective which consists of the sum of welfare and consumer surplus. 7 Moreover, in our setting consumption takes place domestically, whereas there is no restriction on the ownership of firms. As firms will make profits if entry is restricted, the full profit effect of such constraint will not be realised domestically, if some firms are foreign-owned. 8

To incorporate such considerations, we assume that the government maximises a weighted sum, W & , of consumer surplus over two periods, CS + CS , and domestic profits, n π + n π , and interpret W & as political support function in the spirit of Peltzman (1976) and Hillman (1982) . The relative weight of profits is denoted by β, 0 < β, and independent of outcomes.

W & (n , n ) = CS (n ) + CS (n ) + βn π (n ) + βn π (n ) 6 von Weizsäcker (1980) and Suzumura and Kiyono (1987) , for example, consider first-best settings in which the government can also affect output per firm directly. 7 See also their reference to Schmalensee (2004) who presumes that the firms' payoff is irrelevant and explicitly discards the objective on which many of the analyses of excessive entry are based: "In what follows I accept the objective of consumer welfare, because it is the goal that U.S. antitrust policymakers have chosen. … (Thus, I do not think the interesting definitions of … C. C. von Weizsäcker, which are based on total welfare considerations, are useful for U.S. antitrust.)" (p. 472) 8 Marjit and Mukherjee (2013, 2015) implicitly assume such an objective in a setting with one foreign and n -1 domestic firms. The former is profitable, but does not figure in the definition of welfare. Furthermore, Doganoglu and Wright (2006) consider compatibility issues in a framework with unit demand, such that price increases benefit firms but do not affect consumption quantities. To capture the possibility that higher prices lower welfare in such setting, the weight of firms in the government objective is weakly smaller than that of consumers.

= + (X (n )) 2 , + β + n -(1 − X (n ) − c)x (n ). , − βκ(n , n ) (4)

In equation (4), κ(n , n ) represents total fixed costs, specified in more detail below. If β = 0 holds, the government is interested only in consumer surplus. If β = 1, the government is a welfare maximiser. The higher the parameter β is, the more important the firms' payoffs become. If β is sufficiently large, the government predominantly cares about profits.

We, finally, presume that the government can restrict entry but that it has no instruments at its disposal with which it can entice firms to take up production. However, as long as entry is profitable, the maximum allowed by the government will also represent the actual number of entrants. If, in contrast, the preferred number of firms makes production unprofitable, the zeroprofit condition binds and the government faces a constrained optimisation problem. 9

3. The Benchmark: Isolated Markets

The market equilibrium features a constant number of entrants. We subsequently characterise the equilibrium and derive the resulting payoffs, which are denoted by a '*'. Moreover, in Appendix A.1 we show that this is the only equilibrium in pure strategies.

If all firms, which enter the market in period one, remain active in period two, while no additional entry takes place (τ = 1), the sum of profits over two periods can be obtained by adding (3a) and (3b). Setting this sum equal to zero, the equilibrium number of firms is (see Mas-Collell et al. 1995, p. 405 ff and Etro 2014, inter alia, for a one-period setting)

where k 3 ≔ (k + k + k )/2 defines the average of set-up and periodic entry costs over two periods. Moreover, n * > 1 requires (1 − c) > 4k 3 . Using equations (2) and (5), output per firm can be computed as x * = 2 k 3 , while aggregate output equals X * = 1 − c − 2 k 3 . Profits, respectively losses, are given by k > π (n * ) = 0.5(k + k − k ) = −π (n * ) > 0.

Accordingly, welfare, W, and the government's payoff, W & * , coincide and amount to: W = W & * = βn * -π (n * ) + π (n * ). 788888988888:

,; + 2 (X * (n * )) 2 = (1 − c) − 2(1 − c) 2 k 3 + k 3 (6) Because fixed costs vary over time, profits cannot be zero in each period. The number of firms which ensures zero profits in period one would be too low to guarantee that outcome in period two as well, because set-up cost, k, only arise in period one. However, it is not profitable to enter the market only in period two, because the fixed costs of taking up production for the first time, k + k , which a new entrant would incur, exceed operating profits. This is due to the assumption that k < k + k (see Appendix A.1). Accordingly, profits are maximised by spreading set-up costs over two periods, that is, by entering the market as early as possible.

In this sub-section, we characterise the government's preferred outcome, denoted by the superscript 'opt', given that it can costlessly restrict entry. Using (2), (3a) and (3b), the government's objective is given by:

W & > (n , n |β) = (X (n )) + (X (n )) 2 + β-n π (n )+n π (n ) − κ(n , n ).

= ? n 2 + β@ n 1 − c 1 + n + ? n 2 + β@ n 1 − c 1 + n − βκ(n , n ),

where total fixed costs equal κ(n , n ) = B n (k + k + k ) + (n − n )(k + k ) if n ≥ n n (k + k + k ) + (n − n )k if n < n (8)

In Appendix A.2 we show that the optimal number of firms will be the same in both periods (n > = n = n ). This is the case for two reasons: First, a constant number of firms minimises aggregate set-up costs, for a given average number of firms. Moreover, set-up costs reduce the government's payoff directly (see equation (7)) or indirectly via consumer surplus. Second,

given an interior solution, consumer surplus is increasing and strictly concave in the number of firms. Hence, the government cannot increase its payoff by shifting output from one period to the next. Finally, the government has no incentive to alter its choice in period two. Thus, a constant number of firms is also time-consistent and ex-post optimal (see Appendix A.2).

When choosing the optimal number of firms, the government has to observe the non-negative profit constraint. We do not explicitly incorporate the restriction, but take it into account when interpreting the first-order condition. Maximisation of W & > for n = n = n yields:

The derivative in (9) decreases in the number of firms, as n ≥ 1 holds, such that the secondorder condition is fulfilled in case of an interior solution. Moreover, it declines in the relative weight of profits, β. Hence, the government's preferred number of firms shrinks in the weight of their payoff in its objective.

We can rewrite (9) by using a common denominator for both summands and then express it as:

where (10) is positive if n = 0. It will surely become negative for a sufficiently high number of firms, n, and β > 0, as only the linear term, B, may additionally be positive, whereas the summands including quadratic and cubic terms in n have negative signs.

Hence, there is a unique solution to (10) for n > 0, which can basically be obtained by applying the cubic formula to (10). This optimal number of firms from the government's perspective is straightforward to compute for special cases.

If, for example, the government maximises welfare (β = 1), we have A = −β, B = −3A, and C = (1 − c) − k > 0, such that the solution to (9) and (10) becomes (see Mas-Collell et al. 1995, p. 405 ff and Etro 2014, inter alia, for a one-period setting):

This establishes the excess entry prediction for a welfare-maximising government. Substituting (11) into (2), we can calculate output, and using these findings in (7), we obtain:

The second special case assumes that the government cares only about consumers (β = 0), implying that A = C = 0 and B = (1 − c) > 0. Since the derivatives in (9) and (10) are then unambiguously positive, the government will not restrict entry. In the absence of instruments that encourage firms to take up production, the resulting number of competitors is implicitly defined by the zero-profit constraint, i.e., the unrestricted market outcome (n > (β = 0) = n * ).

The greater the relevance of the firms' payoff is, the more likely it becomes that the detrimental impact of more firms on profits affects the government's choice. Hence, by solving (9) for n = n* (as defined in equation (5)), we observe that n > (β) = n * holds for any β ≤ β QR , : = 1 2 ,

and n > (β) < n * if β > 0.5.

The third special case assumes that the government predominantly cares about firms. For any β ≥ β QR , , equations (9) and (10) are unambiguously negative for n = 1 and the government establishes a monopoly, where this critical value is given by:

since (1 − c) > 4k 3 for n* > 1 according to equation (5). Output amounts to X > β ≥ β QR , = 0.5(1 − c), while profits equal π > β ≥ β QR , = 0.25(1 − c) − k 3 .

Summarising the above, we obtain:

Proposition A Suppose, the government determines the maximal number of firms.

a) This number of firms does not vary over time.

b) If β ≤ β QR , , the government will not restrict entry, such that the market equilibrium results. c) If β = 1, n > (β = 1) = (1 − c) /J /k 3 /J − 1 firms enter the market. d) If β ≥ β QR , , the monopoly outcome results and n > β ≥ β QR , = 1. e) If β QR , < β < β QR , , we have 1 < n > (β) < n * , such that output and welfare levels lie in between the two cases implicitly described in parts b) and d).

Proof: See Appendix A.2 for part a) and Appendix A.3 for the remaining parts.

If the weight of the consumers' payoff is sufficiently high, the government wants to maximise aggregate output. Since production increases in the number of firms, the government will not restrict entry. Given a firm's opportunity not to enter, the number of firms is determined by the unregulated market equilibrium. Interestingly, in our setting with linear costs the critical value of β, β QR , = 0.5, results from a "populist" objective (Amir et al. 2019) , according to which consumer surplus has twice the weight of profits. If competition authorities have such objective, or one in which consumer surplus is assigned an even higher importance, they will not restrict entry into homogeneous Cournot oligopolies. Conversely, if the relevance of consumers is sufficiently low, the government's choice is determined by the fact that aggregate profits decline in the number of firms. Hence, their number is minimal. Finally, for 13 intermediate values, the government will allow more than one firm to enter, but fewer than in market equilibrium.

Almost universally, the standard used to establish excessive entry has been the number of firms a welfare-maximising government would choose (Suzumura 2012) . Given our more general government objective, we could redefine excessive entry as a situation in which the number of competitors in the absence of intervention is greater than the number preferred by the government. In this case, the benchmark for evaluating the market equilibrium would depend on the nature of the political process and the mechanism by which preferences are aggregated or affect the government's actions. In the present setting, no decision with respect to the appropriate standard has to be taken. This is because the benchmark for establishing excessive entry plays no role for appraising the market outcome. Proposition A clarifies that there will never be fewer competitors in market equilibrium than desired by the government, irrespective of the combination of profits and consumer surplus in its objective. Therefore, the number of entrants in market equilibrium is never less and usually (for β > 0.5) greater than the government's preferred extent of competition, also in a setting in which the standard for evaluating the market outcome is not the sum of consumer surplus and profits. 10 This assessment, however, will no longer hold if the weight of consumers in the government's objective were sufficiently high and entry were not restricted by the zero-profit constraint. This would be the case because the government would prefer more than n* competitors if it could subsidise entry and maximise (7) without having to observe the profit restriction. 11

In this section we investigate the impact of market opening, first, for the market equilibrium and, second, if the government unilaterally restricts entry. These analyses constitute the open economy counterpart to the investigation in Section 3. Market opening is interpreted as the possibility for firms from all m -1 foreign countries to enter the domestic market by undertaking horizontal FDI. For simplicity, we assume that this applies to domestic firms also. This is without impact on results if the number of countries, m, is high enough. Hence, market opening effectively abolishes the government's ability to regulate entry. 10 Etro and Colciago (2010) analyse a DSGE-model and compare the market outcome with a situation in which steady-state consumption is maximised. Though their model differs substantially from the much simpler one outlined above, they also find entry to be excessive. 11 I am grateful to an anonymous referee for bringing this issue to my attention.

14 One could also imagine less radical forms of market opening. FDI could, for example, be restricted to firms which had already been active in period one, or be selective and exclude firms from some countries, or necessitate an international agreement on domestic entry in period one or, finally, be subject to a quit pro quo rule, according to which entry from foreign firms requires domestic firms to invest in the respective countries as well. All these types of market opening can be viewed as less comprehensive than the variant we analyse below. While some of them will yield the same outcomes, for example, if entry is restricted to incumbents and all firms enter in period one anyhow, others may result in less pronounced adjustments in entry, output and payoffs. 12 Therefore, our variant of market opening indicates its maximal consequences.

In market equilibrium, entry in period two in case of market-opening takes place, as long as operating profits, π , weakly exceed entry costs k . As indicated above, we assume that the number of domestic and foreign entrants is sufficient to ensure zero profits in that period. 13

Hence, when deciding whether to enter the market in period one, firms compare operating profits in that period with the sum of set-up and market entry costs, k + k .

Setting equations (3a) and (3b) equal to zero (for τ = 1), the number of domestic entrants, n * T , in period t, can be calculated, where the superscript 'm' indicates that FDI is feasible.

where n * T > 1 implies that (1 − c) > 2((2 − t)k + k ). Output per firm equals x * T = 2(2 − t)k + k , while aggregate output is given by:

m ≥ m QR = W k + k k 7898: X 1 − c − 2k 1 − c − 2k + k 7888898888: X If the number of countries were less than m QR , profits in period two would be positive and the number of domestic firms entering in period one would be higher than n * T because the marginal firm in period one could earn positive profits in period two. A similar conclusion can be drawn if FDI involves higher costs than k (or k + k ), as in Helpman et al. (2004). 15 Therefore, the government's payoff, W & * T , and welfare, W * T , equal: W * T = W & * T = βn * T -π (n * T ) + π (n * T ). 78888889888888:

,; + 0.5((X * T ) + (X * T ) )

This yields:

Suppose there are m, m ≥ n * T /n * T , countries and there is market opening at the beginning of period two.

The market equilibrium is characterised by n * T = (1 − c)/2(2 − t)k + k − 1 firms, n * T < n * < n * T , which together produce X * T = 1 − c − 2(2 − t)k + k units and obtain zero profits in each period. Aggregate output over both periods and welfare rise with market opening.

Proof: See Appendix A.4 and below.

Market opening implies that the number of firms in period two rises and output increases to above the level that results in the absence of market opening. The difference in the number of firms in period two due to market opening, n * T − n * , can be interpreted as the extent of FDI. It rises with set-up costs, k, and market entry costs, k , in period one, and declines with entry costs, k , in period two. This is the case because these costs determine the extent to which operating profits differ across periods in the absence of market opening and, hence, the incentives to enter in period two, once FDI becomes feasible.

Turning to the comparison of outcomes in period one, we can observe that the number of firms and aggregate output decline, relative to a world without FDI. This change comes about because firms cannot distribute set-up costs over two periods, as it is feasible in a closed setting. Accordingly, operating profits in period one have to equal the sum of set-up costs, k, and market entry costs, k . Since this sum is higher than operating profits in the absence of market opening, incentives to enter in period one decline with FDI activities in period two.

Contrasting outcomes over both periods, the feature that firms can enter the foreign market in period two at costs k implies that total fixed costs decline. Hence, the average number of firms rises, output per firm falls and aggregate domestic output goes up. 14 Because profits are 14 The reduction in total fixed costs is given by n * T (k + k ) + n * T k − 2n * k 3 = (1 − c) Y2k + k + 2k − zero, irrespective of whether there is market opening or not, the welfare increase and the rise in consumer surplus coincide. Additionally, they entirely accrue to the domestic economy. Thus, the government's payoff rises as long as consumer surplus figures in its objective.

Our results can be compared to those which consider a reduction in trade costs in Cournotmodels. Brander and Krugman (1983) show that welfare rises due to trade on account of the decline in average costs. Venables (1985) also finds a non-negative welfare effect and the number of firms to rise with trade if demand is linear. Hence, with respect to these outcomes, trade and FDI appear to have similar consequences.

Note, finally, that results are unaffected by an alternative timing of market opening. If it took place in period one, profits would be zero in both periods. Because set-up costs could be distributed over more than one market, they would drop to below k. Hence, aggregate output would rise beyond X * T and welfare with market opening would exceed W * T .

A government, which regulates entry in a closed economy, can do so in both periods. If market opening takes place at the beginning of period two, entry in period one continues to be limited by the government. In period two, firms from all other countries can become active on the domestic market. Once again, we assume that there are enough of them for the profit constraint to bind (see equation (15)). 15

Given n * T active firms in period two, aggregate output, X * T = x * T n * T , and consumer surplus, 0.5(X * T ) = 0.5(1 − c − 2k ) , are independent of period one outcomes. Accordingly, the government effectively maximises its period one payoff and its objective is given by:

W & > ,T (n |β) = n 2β + n 2 1 − c 1 + n − βn (k + k ) + 0.5(X * T )

Maximisation with respect to the number of firms in period one yields:

Moreover, aggregate output rises, as X * T + X * T − 2(X * ) = 2(1 − c) − 2k + k − 2k − 2 ?1 − c − 2 k 3 @ = 220.5(k + k + k ) − 2k + k − 2k > 0. Therefore, the average number of firms rises. 15 If this assumption did not hold, the level of profits in period two would depend, inter alia, on the number of countries and market entry costs k . Moreover, the government would take into account that a rise in the number of firms in period one reduced profits in period two. A similar conclusion would arise if market opening were not absolute in that all foreign firms could undertake FDI without restrictions. Suppose, in line with this argument and as suggested by an anonymous referee, that domestic entry regulations in period one determine the extent of FDI in period two. This could be the case if for each domestic entrant a fixed number of foreign firms per country could enter. Once again, the incentives to let firms enter the market in period one would be affected by the repercussions of the ensuing change in the number of competitors in period two. Such restrictions would be without impact, relative to the findings derived below, if they do not bind. Otherwise, the effects of market opening may be more moderate than for the encompassing version of market opening considered.

∂W & > ,T (n |β) ∂n = (1 − c) -(1 − β)n + β.

(1 + n ) J − β(k + k ) = 0 (19)

Since the derivatives of (19) with respect to n and β are negative (given n ≥ 1), the government's preferred number of firms declines with the relevance of the firms' payoff in its objective also in the presence of horizontal FDI. Following the same approach as in Subsection 4.1, we compute n > ,T for special cases

If β ≥ β QR ,J holds, the monopoly outcome arises. If, in contrast, the consumers' payoff is sufficiently prominent, β ≤ β QR , , the government mimics the market equilibrium. For any value of the parameter β such that β QR , < β < β QR ,J holds, 1 < n > ,T < n * ,T obtains. Given the unique relationship between output and the number of firms, we have X > ,T β ≥ β QR ,J > X > ,T (β = 1) > X > ,T β ≤ β QR , . We can, hence, summarise our findings as follows: e) Finally, if β QR , < β < β QR ,J , we have 1 < n > ,T < n * T , such that outcomes lie in between the two cases described in parts b) and d).

For the proof, see Appendix A.5.

If the difference in the number of firms in period two due to market opening, n * T − n > (β), describes the extent of FDI, its magnitude rises with the weight of consumers in the government's objective. This is the case since n * T is determined by the zero-profit outcome, while the optimal number declines with β if the government determines entry. Hence, FDI can be argued to be less prevalent the greater the political importance of consumers is and the less relevant firms are. The reason for this prediction is that consumers prefer as large an aggregate output as possible. Thus, FDI is less beneficial than if aggregate output and the number of firms were lower prior to market opening. 16 Moreover, as in the case of the market equilibrium, FDI rises both with set-up costs, k, and market entry costs, k , in period one because they reduce the optimal number of firms in a world without FDI.

Comparing the government's preferred outcome in a closed economy to that resulting if market opening occurs at the beginning of period two, we obtain: (2002) is based on total FDI inflows as dependent variable and does not provide a direct proxy of β, it cannot empirically substantiate our theoretical prediction.

To provide intuition for part b) of Corollary D, note that the alteration in profits in period two requires the governments to make firms more profitable in period one, by restricting their number. This argument is valid, as long as the number of firms in period one is not already minimal, i.e. a monopoly exists. If all countries reduce the number of entrants in period one, the total number of active firms (in the world) will drop from mn * T to mn > ,T . 17

Part c) indicates that the government may object to market opening although welfare rises. This will be the case if the weight of firms in the government's objective is sufficiently large. If β > β QR , (> β QR ,J ) holds, the government prefers a monopoly in both periods in a closed setting and in period one if market opening takes place (cf. Propositions A and C). Market opening reduces period two profits to zero, while consumer surplus increases. This raises welfare.

However, if consumer surplus affects the government's payoff, W & , to a sufficiently small extent relative to profits, W & will decline. Consequently, we can establish a further argument why detrimental protectionist policies may not be abolished.

Finally, part d) of Corollary D indicates that welfare may change in either direction with market opening if entry is regulated. On the one hand, the mechanism is the same as in the equilibrium. FDI allows firms to spread set-up costs, such that average costs decline.

Therefore, aggregate output rises in period two. The increase in welfare due to the resulting ascent in consumer surplus in period two is partially balanced by a decline in period one. It occurs because set-up costs cannot be distributed over both periods, such that entry costs in period one rise. In consequence, the government lowers the number of entrants in that period.

This results in a decline in aggregate output. Appendix A.6 provides one example for a setting in which welfare rises and one in which it declines.

Comparing the findings for the market equilibrium with those for a setting in which the government determines entry, we can state:

Assume that there is market opening at the beginning of period two. Entry is excessive in period one if the firms' payoff is sufficiently important, i.e., β > β QR , .

Proof: The government's preferred number of firms is n * T if β ≤ β QR , and less otherwise. ■

If the importance of consumers in the government's payoff is sufficiently high, the government desires aggregate output to be maximal. Since output rises in the number of firms, the market 17 See Venables (1985) and Richardson (1999) for similar findings in settings with trade. 20 outcome and the preferred number of firms will coincide. If the weight of firms in the government's objective is higher than β QR , = 0.5, it will prefer fewer firms in period one, such that excessive entry results. Moreover, the government in a closed economy prefers at most n * firms in period two (cf. Proposition A). Because n * T > n * , there will surely be excessive entry. As in a closed economy, the number of entrants in equilibrium is never less and usually (for β > 0.5) greater than the government's preferred extent of competition, irrespective of the standard used to evaluate the market outcome. This statement holds, like in a setting without FDI, as long as the government cannot overcome the zero-profit constraint.

We analyse a two-period Cournot-model with costly entry, in which fixed costs of production vary over time. As a benchmark, we investigate a closed-economy. Our main contribution is to extend the analysis to an open economy setting in which firms can undertake greenfield investments abroad. Such market opening allows firms to spread fixed cost over more markets and, thus, to lower unit production costs. This beneficial impact of horizontal FDI is mitigated or dominated by the increase in the number of firms. Such increase entails a welfare loss due to the business stealing externality. In particular, we compare the unregulated market equilibrium with a situation in which governments can restrict entry at least in the first period. Going beyond previous contributions in a further dimension, we compare the market equilibrium not only to the (second-best) welfare-maximum, but also to the outcome which arises if the government maximises a political support function.

Within this set-up, we have established a number of findings. Three of them are particularly noteworthy. First, market opening surely aggravates the excess entry problem in period two and has the same effect in period one, unless the government primarily maximises consumer surplus. In consequence, providing regulatory bodies with a sufficient set of instruments to restrict entry is a more pressing problem in an open than in a closed economy.

Second, market opening has beneficial welfare effects in market equilibrium, but may have detrimental ones if the government regulates entry. In a closed economy, the number of firms in market equilibrium is too high. Therefore, welfare is less than maximal because of the excessive entry distortion and the firms' market power. If the market is opened up, the number of firms in market equilibrium in period one declines, while the increase in period two is relatively moderate because there had already been excessive entry. Therefore, total fixed costs born by domestic firms decline. Moreover, consumer surplus rises. The latter effect ensures 21 that market opening raises welfare in market equilibrium. If, however, the government determines the number of firms in a closed economy, the entry distortion is much less pronounced, unless the weight of consumers in the government's objective is sufficiently large.

Market opening brings about too many competitors in period two. Hence, welfare declines on account of the business-stealing externality and higher market entry costs in period two. This effect is slightly mitigated by the fall in the number of firms producing in period one. The rise in consumer surplus will only dominate if the excessive entry distortion introduced by market opening is not too large. Moreover, the countervailing effects in a situation in which the government determines entry also explain why welfare may decrease. Put differently, the feature that the welfare loss due a free-entry Cournot-oligopoly in a closed economy is relatively small if entry is regulated, reduces the gross gain from market opening and raises the negative impact. Hence, market opening may have a negative welfare impact.

Third, welfare and the payoff of a government maximising a political support function may vary in different directions. This is the case because an equivalent variation of profits and consumer surplus in the opposite direction will leave welfare constant, whereas this will generally not be the case for the government's payoff. Because market opening tends to reduce profits if entry is regulated, a government is more likely to oppose it the greater the firms' weight in the government's objective is. Hence, our analysis indicates a further reason why protectionist tendencies may result if political activities of interest groups determine government behaviour (see Ethier and Hillman 2019 for a recent survey).

The conclusions above and many of our other findings may depend on a number of modelling features. In future work, it may be worthwhile to explore in how far they affect the theoretical predictions. As one example, the weight of the consumers' and firms' payoffs in the government's objective is exogenous and constant. However, it could be argued that the respective parameters are endogenous and depend on the payoffs obtained. This would imply that market opening affects the relative importance given to consumer surplus and profits when determining competition policy (see Hillman and Ursprung (1993) for an according approach idea in the context of trade policy). Since market opening tends to reduce aggregate profits, the consumers' interests may then become more important for entry regulations. As a second example, we have not allowed for exporting activities. However, if the good under consideration could be transported across borders easily, this restriction can no longer be justified. The findings of Miyagiwa and Sato (2014) suggest that there will also be excessive entry if firms can export their products, but not undertake FDI. The combination of both activities, exporting and investing abroad may have different consequences if interaction 22 effects occur (see Helpman et al. 2004 and Stähler 2011 To prove that the outcome described in Sub-section 3.1 is a unique equilibrium, we show, first, that n = n = n * is a stable outcome and, second, that it is the only one in pure strategies.

Note for that purpose that if k ≥ k , the constraint k > |k − k | entails k + k > k and k + k > 2k − k ≥ 2k − k = k . The same inequalities can be derived for k < k .

To establish the first part, we show that neither entry nor exit is profitable. Exit at the beginning of period two yields a payoff of zero, whereas producing the profit-maximising quantity generates profits of π (n * ) = 0.5(k + k − k ) > 0. Hence, no firm will exit in period two. Entry in period two of a firm which had not produced in period one will at most generate losses of π (n * ) − k = 0.5(k − k − k) < 0, as k > |k − k |, and because profits decline in the number of firms. Turning to period one, an additional entrant would incur higher contemporaneous losses and lower profits in period two than an incumbent, since profits decline in the number of firms. Thus, there is no additional entry. Not entering at all would yield zero profits, that is, no improvement relative to the equilibrium described. Not entering in period one, but then entering in period two would generate a payoff of π (n * ) − k < 0. Hence, waiting to enter does not pay. In consequence, the outcome constitutes a stable equilibrium. In addition, profits decrease in the number of firms if firms choose output optimally (cf. equations (3a) and (3b)). Therefore, if n < n*, profits will be positive and firms have an incentive to enter, whereas firms will no longer enter if n > n*. Because, furthermore, all firms are identical and the resulting equilibrium is symmetric, it is the only one in pure strategies.

Maximisation of W & > (as defined in (7)) with respect to n , t = 1, 2, yields:

If the terms in square brackets are both negative, the derivatives in (A.1) and (A.2) will be negative as well, and n and n attain the smallest feasible values of unity in both periods. Suppose an interior solution next. Since κ, as defined in (8), is linearly increasing in n and n , the expression in square brackets in (A.1) and in (A.2) is positive and the first terms are decreasing in the number of firms. If n > n , the first term in (A.1) is thus smaller than the first term in (A.2), while the second term, which is deducted, is larger in absolute value in (A.1) than in (A.2). Hence, if n is chosen such as to warrant (A.2), (A.1) is negative for any n > n . Therefore, any choice of the number of firms which (1) implies interior solutions to the maximisation problem and (2) that more firms are allowed to enter in period one than in 24 period two, or vice versa, cannot be optimal. Finally, it may be the case that the term in square brackets is positive in (A.1) and negative in (A.2), or vice versa. In the former case, the optimal number of firms in period two would be minimal, i.e. one. If the optimal number of firms in period one is unity, the two-period monopoly case discussed above would result. If, however, the optimal number of firms in period one were greater than one, (1 − β)n + β < (1 − β)n + β would hold because n > n by assumption and since (1 − β)n + β < 0 requires 1 − β < 0. This, however, is in contradiction to the assumption that the term in square brackets in (A.1) is positive and negative in (A.2). Hence, a choice of the number of firms such that n > n = 1 cannot be optimal for a government, irrespective of its exact objective. A comparable argument establishes that 1 = n < n cannot maximise the government's payoff either. Note, finally, that the sign of (A.2) is independent of n . Thus, the government has no incentive to alter its choice in period two ex-post and optimal choices n > = n = n are also time-consistent.

The arguments prior to Proposition A establish part c). Combining them with the results of Sub-section 3.1 proves part b). Part d) follows straightforwardly, since the government establishes a monopoly. Finally, the optimal number of firms exceeds unity but falls short of n * . In addition, output per firm and profits decline in the number of firms, while aggregate output rises in n (cf. equation (2)). Moreover, welfare is strictly concave in n and maximal at n = n * . Since 1 < n > β QR , < β < β QR , < n * , we have proven part e).

We have to establish that (1) there are no incentives for firms to deviate from the outcome characterised in the Proposition and (2) this is the only equilibrium in pure strategies.

In period two, a domestic firm which leaves the market obtains a payoff of zero. Hence, it is not beneficial to exit. The same is true for foreign firms which have undertaken FDI and entered the domestic market. Additionally, profits decline in the number of firms. Hence, an additional entrant would obtain a negative payoff. Accordingly, no firm has an incentive to deviate from its choice in period two. In period one, profits decline with entry as well. Moreover, entrants make zero profits. Thus, no additional firm has an incentive to enter the market in period one. Furthermore, postponing entry until period two will result in a loss, since set-up costs would be incurred in period two. Moreover, not entering at all would yield a payoff of zero. Therefore, the outcome described in Proposition B is locally stable. Applying the same argument as in Appendix A.1, we can establish that there is no other equilibrium.

Aggregate output over both periods rises, as 25 X * T + X * T − 2(X * ) = 2 W k + k + k 2 − 2k + k − 2k > 0 (A. 3)

Since welfare equals the sum of consumer surpluses, while the government's payoff increases linearly in welfare (unless consumer surplus does not figure in its objective), both rise with market opening.

A.5: Proof of Proposition C If m ≥ n * T /n > ,T , profits will be zero in period two. The number of firms in period two is given by (5), replacing k 3 by k . This proves part a). The proofs of parts b) and d) follow those of Proposition A. In particular, if β ≥ β QR ,J output will be given by X > ,T = 0.5(1 − c), and profits equal π > ,T = 0.25(1 − c) − (k + k ). The findings in part c) can been derived, taking into account X > ,T (β = 1) = 1 − c − x > ,T (β = 1) = 1 − c − ((1 − c)(k + k )) /J and π > ,T (β = 1) = (k + k ) n > ,T + 1 /J − (k + k ). Finally, the outcomes and payoffs summarised in part e) for β QR , < β < β QR ,J , are given by: 1 < n > ,T < n * T , 2k + k = x * T < x > ,T (β) < x > ,T β ≥ β QR ,J = 0.5(1 − c), and 0.5(1 − c) = X > ,T β ≥ β QR ,J < X > ,T (β) < X * T = 1 − c − 2k + k , 0 < π > ,T (β) < π > ,T β ≥ β QR ,J = 0.25(1 − c) − (k + k ). They follow from the unique relationship between the number of firms, n , and output, respectively the parameter β, and the negative effect of n on profits.

A.6: Proof of Corollary D ad a) The difference in the number of firms in period two can be computed using (5) and (15) The first inequality in (A.4) results because k + k > k , such that 2k 3 = k + k + k > 2k . The results for output follow immediately from (2). ad b) From Propositions A and C we know that the government will establish a monopoly in the closed market and also after market opening in period one, if β > β QR , . Moreover, for all β < β QR , , the number of firms in a closed setting is given by n * , whereas the respective number, if market opening takes place, is n * T < n * (cf. (5) and (15)). In addition, the government's preferred number of firms is one if there is market opening also for β QR ,J < β < β QR , , whereas it exceeds one in the case of a closed market (cf. Propositions A and C).

Finally, the government's preferred number of firms declines linearly in β, as the derivatives of 26 (9) and (19) clarify. Hence, we have n > ,T < n > for all β < β QR , and n > ,T = n > = 1 otherwise. The relationship between the number of firms and output follows from (2).

ad c) and d) We prove parts c) and d) by providing according examples (see also Goerke 2020) . Suppose c = 0.1, k = k = 0.01, and k = 0.04. In this case, the government's preferred number of firms in a closed-economy is n > = 2, resulting in an aggregate output of X > = 0.6, a consumer surplus of 0.36 over both periods and profits of 0.12 per firm over two periods. Adding aggregate profits for two firms and consumer surplus yields W > = 0.6. In case of market opening the number of firms is given by n > ,T = 1.53 and n * ,T = 8. Aggregate output equals X > ,T = 0.544 and X * T = 0.8, resulting in a consumer surplus of 0.468 over both periods. Profits are zero in period two and amount to 0.076 per firm in period one, yielding a welfare level of W > ,T = 0.585. Hence, welfare falls with market opening. If we assume k = 0.09 instead, but retain the other parameter values, the respective welfare levels are given by W > = 0.513 and W > ,T = 0.522, demonstrating that market opening enhances welfare. This establishes part d).

Continuing the example for k = 0.09, but assuming that β > β QR , > β QR ,J , for β QR , = (1 − c) /(8k 3 ) = 0.81/0.44 = 1.841, will ensure that the government prefers a monopoly both in a closed economy and also if market opening takes place.

In a closed economy, n > β ≥ β QR , = 1 yields an aggregate output of X > = 0.45 per period, a consumer surplus of 0.2025 over two periods and profits of 0.295. Hence, welfare equals W > = 0.4975, whereas the government's payoff amounts to W & > = 0.2025 + 0.295β.

In case of market opening, there will be a monopoly in period one, while n * ,T = 8. This yields consumer surpluses of 0.10125 and 0.32 and a profit level of 0.10125 in period one, while profits are zero in period two. We obtain W > ,T = 0.523, whereas the government's payoff amounts to W & > ,T = 0.42125 + 0.10125β. Accordingly, market opening raises welfare, whereas the government's payoff will fall for any β > 1.14. This restriction will hold, given the assumption β > β QR , . This example establishes part c).

Free Entry versus Socially Optimal Entry

Improving on Free Trade in Cournot Oligopoly

Is International Trade Profitable to Oligopolistic Industries?

Concentrating on the Fall of the Labor Share

A 'Reciprocal Dumping' Model of International Trade

Tacit Collusion, Free Entry and Welfare

Foreign Direct Investment and its Drivers: A Global and EU Perspective

Welfare-enhancing Trade Unions in an Oligopoly with Excessive Entry

Multihoming and Compatibility

The Politics of International Trade

The Theory of Endogenous Market Structure

Endogenous Market Structures and the Business Cycle

Ecessive Entry and the Integer Constraint with Many Firms: A Note

Free Entry and Social Efficiency under Vertical Oligopoly

Social Desirability of Free Entry: A Bilateral Oligopoly Analysis

Excess Entry in the Absence of Scale Economies

Horizontal FDI in a Dynamic Cournot-Oligopoly with Endogenous Entry

Are US Industries Becoming More Concentrated?

How EU Markets Became more Competitive than US Markets: A Study of Institutional Drift

Insufficient Entry of Employee-controlled Firms in a Free-entry Oligopoly

Do Civil and Political Repression really Boost Foreign Direct Investments?

Oligopoly in International Trade: Rise, Fall and Resurgence

Export versus FDI with Heterogeneous Firms

Declining Industries and Political-Support Protectionist Motives

The Political Economy of Protection

Multinational Firms, Political Competition, and International Trade Policy

Endogenous Market Structure in International Trade (Natura Facit Saltum)

Does Tariff Really Enhance Welfare?

Foreign Competition and Social Efficiency of Entry

Endogenous Market Structure, Trade Cost Reduction, and Welfare

Endogenous Market Structure and Foreign Market Entry

Microeconomic Theory

Free Entry and Regulatory Competition in a Global Economy

Excessive Entry in a Bilateral Oligopoly

Endogenous Cost Asymmetry and Insufficient Entry in the Absence of Scale Economies

Social Efficiency of Entry with Market Leaders

Endogenous Domestic Market Structure and the Effects of a Trade Cost Reduction in a Unionised Industry

Entry, Profit and Welfare under Asymmetric R&D Costs

Symmetric Cournot Oligopoly and Economic Welfare

The Logic of Collective Action

Toward a More General Theory of Regulation

Scale Economies, Imperfect Competition, and Public Policy

Trade and Competition: Concordia Discors?

Sunk Costs and Antitrust Barriers to Entry

On the Effects of Entry

Market Entry and Foreign Direct Investment

Excess Entry Theorems after 25 Years

Entry Barriers and Economic Welfare

Reciprocal Tariff Imposition and Welfare in a Free Entry Oligopoly

World Investment Report

Entry and Cost Reduction

Trade and Trade Policy with Imperfect Competition: The Case of Identical Products and Free Entry

A Welfare Analysis of Barriers to Entry

Do Industrial and Trade Policy Lead to Excess Entry and Social Inefficiency?

Highlights A Political Economy Perspective on Horizontal FDI in a Dynamic Cournot-Oligopoly with Endogenous Entry

Entry in a homogeneous Cournot-oligopoly is excessive if there is business-stealing • Allowing for horizontal FDI aggravates the excess entry problem • Nonetheless welfare rises in market equilibrium, but may fall if entry is restricted • A government maximising a political support function may restrict FDI