Interlocking directorates (IDs hereafter) arise when two or more corporate boards have one or more members in common. The presence of directors on the boards of competing firms raises concerns that IDs may facilitate collusion, particularly through the exchange of sensitive pricing information (Harrington and Skrzypacz 2007). As a consequence, most advanced countries forbid or limit IDs among competing firms. In the US, Section 8 of the Clayton Act forbids IDs between companies “that are […] competitors such that the elimination of competition by agreement between them would constitute a violation of any of the antitrust laws.” Japan and South Korea also forbid IDs if they harm competition. In Europe, where pro-competition policies have a more recent tradition than in the US, IDs are not specifically regulated, but they are managed under general competition law. In practice, however, limits to IDs and outright bans are not strictly enforced. For example, during the 2010s, 9.8% of all directors in the sample of S&P 1500 firms sat on at least two boards within the same SIC classification (Nili 2023). In recent years, antitrust authorities in the US have intensified efforts to enforce the ban on IDs (Department of Justice 2022, Feinstein 2017), sparking a lively debate among practitioners.
Estimating the effects of interlocking directorates: The case of the Italian corporate loan market
Despite the relevance of ID regulation, we know little about the causal effects of board interconnections on competition. Mayer et al. (2020) offer evidence that IDs increase firm value, possibly due to reduced competition. In a recent paper (Barone et al. 2025), we provide direct causal evidence on the effects of board interconnections on competition. This is not an easy task, as board connections are typically endogenous, and individual prices are often difficult to observe. We overcome these difficulties by focusing on the Italian corporate lending sector in the early 2010s. This setting represents an ideal testing ground for three reasons. First, an unanticipated regulatory change in 2011 that forbade IDs among competing banks – breaking board connections – represents an exogenous change in the structure of bank interconnections. The policy was part of the ‘Save Italy’ decree – an emergency package hastily drafted by a newly appointed technocratic government tasked with managing the impact of the sovereign debt crisis on Italy. Second, not only does each bank lend to multiple firms, but firms also typically borrow from multiple banks, which allows us to flexibly control for both fixed and time-varying unobserved heterogeneity between treated loans – loans of interlocked banks before the reform – and controls. Third, a database with detailed information on individual firm-bank lending contracts allows us to observe firm-bank individual loan prices and their evolution before and after the reform, and to test whether prices of treated loans decreased after the reform relative to controls. We also show that the structure of the Italian banking sector is comparable to that of other major economies, which supports the external validity of our analysis.
Following the antitrust authority, we use Italian provinces as the relevant geographical market for business lending (NUTS 3 units, broadly comparable to a US county). Within the province, we define a network of interlocked banks as a set of banks with IDs and with sufficiently high market share to exert market power. We then label loans of the banks belonging to a network as treated. Our empirical strategy compares the change in the interest rate on treated and control loans before and after the reform, taking into account potential confounders at the bank-time, firm-time, and firm-bank levels (thanks to the fact that firms typically borrow from multiple banks; Khwaja and Mian 2008). Importantly, a bank’s loans can be treated in some provinces – where other interlocked banks operate – and serve as controls in others, where no such interlocks exist. This geographical variation allows us to rely solely on within-bank-period variation, thereby controlling for any shocks affecting the bank, something important during the turbulent period we consider.
Severing interlocking directorates improves access to finance
We find that the severance of IDs has a competition-enhancing effect. Figure 1 plots the estimated difference in the interest rate between treated and control loans over time (vertical dashed bars indicate 95% confidence intervals). The base quarter is the last quarter of 2011 and is identified by a vertical dashed bar. We find no evidence of a pre-trend in the interest rate on treated relationships: all the differences for the pre-period are small and statistically insignificant. There is also no effect in the first quarter after the reform was passed, when the law was effective, but banks had four months to comply with it. After that, the coefficients become negative and progressively larger in absolute value. On average, in the post-reform period, the interest rate on treated relationships drops by 14 basis points relative to the control group (1.5% of the average interest rate on treated loans before the treatment). In terms of size, the effect is comparable to the impact of common ownership on airline ticket prices (Vives and Azar 2022), as well as to the effect of bold expansionary monetary policy measures by the ECB on interest rates after the financial crisis (Sette et al. 2021).
Figure 1 The effect of banning interlocking directorates
The drop in the interest rate is consistent with pre-reform collusive behaviour. In line with the theory of collusion, which predicts that prices are less dispersed in a collusive equilibrium than under competition, we show that interest rates on previously interlocked relationships are less dispersed before the reform and that dispersion increases after it.
Figure 2 shows the differential effects according to firm and local market characteristics. Positive/negative values indicate lower/higher reductions in the interest rate; ‘Low’ indicates observations for which a given characteristic is below the median (vertical bars indicate 95% confidence intervals). In Panel A, we find that less creditworthy firms observe a smaller drop in rates. We interpret this result as indicating that, once the market becomes more competitive following the reform, these firms have worse outside options and therefore renegotiate loans less aggressively. In Panel B we find that the drop in the interest rate is larger for loans whose network of connected banks had a higher pre-reform market share, thus potentially higher market power.
Figure 2 Heterogeneity results
Finally, we look at the real effects of the reform: firms with a larger share of credit from treated relationships experience higher investment rates, employment growth, and sales growth in the post-reform period. Overall, we conclude that the reform was instrumental in improving the performance of the corporate sector.
Some policy insights
Our results provide robust causal evidence of the anti-competitive effects of IDs. They support the recent emphasis by antitrust authorities around the world on enforcing bans on IDs. Moreover, they complement the recent literature suggesting that common ownership (which may give rise to IDs) can also reduce competition (e.g. Azar et al. 2018, He et al. 2017, Koch et al. 2021).
There are a couple more general lessons to take home. First, supply-side, microeconomic reforms can promote growth in stagnating, high-debt economies. Their proper design depends on accurate counterfactual evaluation. Second, structural reforms are often nearly costless in fiscal terms, but they can require a significant investment of political capital with uncertain electoral returns. The political and economic context of the Italian economy in the second half of 2011 – when the country was in the midst of the sovereign debt crisis – drastically shifted policymakers’ incentives and enabled the passage of a reform that would have been unlikely under normal circumstances. In more typical times, raising public awareness of the long-term benefits of such reforms is crucial to incentivise politicians to adopt growth-enhancing supply-side policies.
Source: cepr.org
