When broadband comes to banks: Credit supply, market structure, and information acquisition

Banks have long relied upon cutting-edge technologies to deliver products and improve efficiency, but there is limited evidence on the impact of fast internet on banking activities, particularly in lending. This column examines the effect of broadband adoption on lending in Italy from 1998 to 2008. Broadband boosts credit supply and branch efficiency by improving the efficiency with which banks collect, process, and transmit borrower information. Banks can more tightly monitor borrowers and therefore lend more readily to firms without credit scores or prior credit history, which primarily benefits small and medium-sized firms.

High-speed internet has emerged as one of the most disruptive innovations in recent history, profoundly impacting economic activity by improving information access and communication. This transformation is particularly significant for information-intensive sectors like financial intermediation (D’Acunto et al. 2019, Frost et al. 2019, Fuster et al. 2019, Berg et al. 2020, Hvide et al. 2024), where innovations in information technologies have the potential to mitigate information asymmetries and enhance operational efficiency (Mishkin and Strahan 1999).

Banks have long relied upon cutting-edge technologies to deliver innovative products, streamline loan-making processes, and improve their back-office efficiency (Frame et al. 2018). Despite this, there is limited empirical evidence on the impact of fast internet on banking activities, particularly in the context of lending. This reflects both the limited availability of high-quality administrative data and the methodological challenges in addressing the endogeneity inherent in internet connectivity, which is often tied to local economic conditions.

In a recent paper (D’Andrea et al. 2025), we study how broadband adoption influences credit supply and uncover the underlying mechanisms. We focus on Italy during 1998–2008, a time marked by the rapid expansion of fast internet. We combine detailed data on broadband infrastructure with loan-level information from the Italian Credit Register, which includes records of banks’ requests to access borrowers’ credit histories for screening and monitoring purposes. Additionally, we incorporate administrative data on bank branches, assets, liabilities, and workforce.

Measuring the impact of broadband on credit supply presents several identification challenges. We adopt an approach close to that used in Campante et al. (2018), Hjort and Poulsen (2019), and Lin et al. (2021) as we exploit some of the characteristics of the infrastructure built right after WWII for the telephone network, which are arguably unrelated to the drivers of broadband deployment, and we further control for unobservable characteristics of the local areas where banks and firms operate. Specifically, we develop a measure of distance between the municipality of the bank branch and the internet infrastructure that we use as our main predictor in the regressions.

Our results reveal a positive and statistically significant effect of broadband internet on both the intensive and extensive margins of lending. A one standard deviation increase in distance (after the introduction of fast internet) – which indicates a lower likelihood of high-speed connectivity – is related to a decrease of 4.4% in the volume of credit extended and a 3.6% decrease in the probability of a new relationship in the broadband period. Moreover, fast internet has a statistically significant, albeit economically low, effect on credit pricing, with an increase of one standard deviation in distance associated with a 6.5 basis points increase in the average interest rate. These results are confirmed in a difference-in-differences event study, where we consider branches located near the ADSL infrastructure as treated branches.

Our findings reflect equilibrium outcomes shaped by the interplay of both demand and supply factors. To isolate credit supply, we employ high-dimensional fixed effects regressions to control for time-varying factors related to the firm industry-location-size (Degryse et al. 2019) and firm characteristics (Khwaja and Mian 2008). Our estimates reveal that approximately 33% of the effect on credit (63% to 85% for interest rates) can be attributed to supply-related factors (Figure 1).

Figure 1 Extended credit and interest rates at the loan level – disentangling supply

A novelty of our study is the direct evidence on the information channel – specifically, how broadband influences banks’ collection and processing of information. Using unique data on information requests made by banks to the Credit Register, we document that access to fast internet significantly increases the number of inquiries made throughout the credit relationship, indicating tighter monitoring of borrowers. In contrast, broadband has minimal impact on information requests at loan origination, suggesting that the screening stage remains largely unaffected. These results are corroborated by examining the dispersion of interest rates: connected branches adjust interest rates throughout the credit relationship, while showing no evidence of more granular pricing at loan origination.

Building on this, we provide additional insights that characterise the information channel. First, we find that broadband internet improves the information flow between banks’ branches and headquarters, thus reducing functional distance (Levine et al. 2020). Second, our findings are more pronounced for information-sensitive loans such as credit lines (Chodorow-Reich et al. 2022). Finally, we find that broadband enables branches to lend more readily to firms without credit scores or prior credit history, as they anticipate that they can efficiently gather borrower information at later stages.

Next, we document the intermediate channels through which broadband affects credit supply. We find that branch efficiency, as measured by labour productivity and credit quality, improves following the introduction of broadband. Banks’ geographical reach also expands significantly with the availability of fast internet: connected branches are more inclined to originate loans outside their province, resulting in an increase in the average distance between the branch municipality and borrowers. Finally, municipalities closer to the internet infrastructure experience heightened banking competition, as measured both by the number of bank brands within the municipality and by conventional concentration indicators (top 3 and top 5 concentration ratios, Herfindahl–Hirschman Index for deposits).

In our final step, we examine the distributional impact of fast internet among borrowers. The effects of fast internet primarily benefit small and medium-sized firms (SMEs) and, in part, large firms. There is no discernible effect for micro firms. As hard information is typically available for SMEs and large firms, but much less for micro firms, this is consistent with fast internet allowing banks to gather and manage hard information about borrowers. Moreover, we observe stronger effects among riskier firms. These results align with our main findings, which show that broadband internet facilitates the acquisition of codifiable information by banks (Landier et al. 2009).

Overall, our results show that the deployment of broadband led to an expansion of banks’ credit supply and altered the structure of local credit markets by improving the efficiency with which banks collect, process, and transmit borrower information.

Source : VOXeu