Dye and Sridhar (1995)

 This paper is about multi-firm voluntary disclosure. But it allows firms to choose which period (first or second) they would like to disclose. It also assumes that not all firms receive information. It's not exactly a game theoretic model. It provides cutoff values that a firm would disclose or not as equilibrium values. I probably would never do such a model. It's not my type of paper.

Dastidar (2003)

This paper is another extension of Brander and Lewis (2003). It's published because it turns over the conclusion of Brander and Lewis (1986). While Brander and Lewis (2003) shows firms choose positive debt levels, this papers shows firms always choose zero debt. I think the model set-up is mechanical. Instead of sequential moves in choosing debt levels and output levels, this paper's model makes firms choose them simultaneously by fixing capacity at the first stage. The strategic effect of debt disappears also because Dastidar (2003) features the market value of debt in the first stage and thus maximize equity value instead of total value in the first stage.

Showalter (1995)

This comment comes 9 years later than the original paper. It helps with understanding of the original paper and offers something new. It shows that not only the type of competition matters but also the type of uncertainty counts. The paper derives an opposite result when the uncertainty is about cost: firms don't use debt at all. It's a nice catch and warrants an AER publication.

Brander and Lewis (1986)

This is the one paper that starts studies that link up financial decisions and output market decisions. It's a simple but beautiful idea. The state variable z is randomly distributed as f(z) on support [z_low, z_up] and there is a cut-off point z_hat below which the firm is bankrupt. Thus, the limited liability effect kicks in. It links up the product market and the financial market.
This is a two stage game in which the firm decides on debt levels in the first stage and output levels in the second stage. Although the title is Oligopoly and Financial Structure, the analysis of the paper focuses on symmetric Cournot duopoly. It analyses the second stage first, taking debt levels as given. It includes both equity maximization and debt maximization. For the first stage, it concludes both positive and zero debt levels are possible, depending on the sign of R_iz.
There are precursors and companion papers that I need to read.

Jean-Baptiste and Riordan (2003)

This is not the first paper I read after Riordan (2003) but it doesn't hurt to talk about it first. This paper has never been published and Michael Riordan won't say why. But I guess the paper cannot conclude that capital markets constrain industry scale and its conclusion that competition doesn't change with concentration but depends on entrepreneur's inside equity is unrealistic and not supported by empirical evidence.
The paper assumes n entrepreneurs each with limited inside equity e who need to borrow K at cost r from the credit market. The entrepreneurs can use inside equity and debt on production or perquisites at value of t<r, which means the perquisites are inefficient. The authors further assume the perquisites are not verifiable or recoverable upon bankruptcy. The authors then show that in the monopoly, symmetric oligopoly and asymmetric oligopoly cases, the entrepreneurs produce min(q_hat, q_star) which is the minimum of the amount allowed by the capital market so that the debt doesn't allow the entrepreneurs to steal and the utility maximizing amount of the classical monopoly, symmetric oligopoly and asymmetric oligopoly cases. The authors don't restrict their production function and cannot say unequivocally that q_hat is smaller than q_star and do not derive conditions where it is smaller maybe because they cannot. So the model isn't saying much.
The conclusion that competition doesn't change with concentration but depends on inside equity is a mechanical fact from the model structure that's unrealistic and unsupported by empirical evidence.

Riordan (2003)

Riordan (2003) is the first paper I come across when I decide to study the interaction of capital market and product market competition. It's a speech given at an academic conference, which is later published in Review of Industrial Organization. Its value lies in the fact that it cites a number of theoretical studies on the interaction of capital market and product market competition before 2003. I wrote to Michael Riordan for more recent papers on the topic but he didn't reply.
Riordan (2003) gives an example of how capital market affects product market competition. There is one firm looking to invest in a project. There are two technologies. Technology A is efficient in the sense that its variable cost is lower. Technology B is inefficient but grants the entrepreneur private benefit. The entrepreneur needs to borrow K from the capital market with an expected return of (1+r). Riordan (2003) graphs the entrepreneur's revenue using two technologies respectively against the amount of debt and shows it's easy to construct two curves where the meeting point of the two curves is to the left of the profit maximizing point using technology A. This means if the capital market doesn't ration its supply of funds, the entrepreneur will pick the inefficient technology B, which is not profit maximizing for investors and not socially optimal.
So far as I know, studies on the interaction of capital market and product market competition have mostly focused on capital structure and product market competition for corporate finance and industrial organization scholars. For accounting scholars, it has mainly been disclosure and competition.

Hilary and Hsu (2013)

This one is innovative. It examines the standard deviation of forecast errors (consistency) instead of forecast accuracy (absolute forecast errors).
It says consistent forecasts are more informative. Maybe I should cite this article.