Over the past decade, two major digital trends have affected large parts of the population: the digitalisation of banking services, including the ability to manage deposit accounts remotely; and the spread of social media as digital communication platforms where users can easily share information and other content (Basel Committee for Banking Supervision 2024). The proliferation of remote banking services and increased digitalisation have enabled depositors to react more rapidly to price differences in deposit markets and to changes in market conditions. Depositors now find it easier to compare interest rates across different banks or money market funds and to transfer funds to higher yielding accounts (Jiang et al. 2023, Koont 2023 Rose, 2023). At the same time, the growing adoption of social media has been a contemporaneous development that has accelerated the diffusion of information (Cornelli et al. 2024). 

In a new paper (Brei et al. 2026), we examine how the interaction between these two major digital trends has affected the pricing of retail deposits in the US. We focus on two key dimensions of retail deposit pricing: (i) the level of deposit rates; and (ii) the sensitivity of deposit rates to changes in policy rates. 

Identifying digital banks and measuring social media activity

We identify digital banks through a cluster analysis. Our identification strategy is based on the premise that digital banks differ from traditional banks in terms of branch networks, investment profiles, and expenditure structures. More specifically, digital banks should operate through a more limited physical branch network, spend more on IT and advertising and less on fixed assets, rely more on virtual services with limited account options and investment services, and differ in the way they price their deposits across the country.

We proxy social media activity at the county level by the ratio of the number of resident Twitter users to the total population in each county over the period 2016–2019. The data are based on more than one billion geotagged tweets from the Twitter Streaming API. The most repeated location among the tweets sent by each individual during a year is used to identify the residence of users, with non-resident users being those who tweeted outside of their residence county. Figure 1 shows that social media usage, as measured by proportion of resident Twitter users across counties, is not uniform across the United States.

Figure 1 Twitter users per county

Note: The figures show median values of Twitter users per capita by county over the period 2016-2019. The colouring of the shaded areas indicates the intensity of Twitter activity based on the ratio of Twitter users in percentage of the total population of a given county (darker areas indicate more Twitter activity). “Resident Twitter users” are defined based on the most repeated location among the tweets sent by each individual in a given year.
Sources: Martin et al. (2021); U.S. Census Bureau; authors’ calculations.

Deposit rates at digital versus traditional banks

A comparison of deposit rate levels across digital and traditional banks supports the notion that the former offer higher deposit rates as a result of fiercer online competition.  Figure 2 suggests that, on average, digital banks paid higher rates compared to their traditional peers, particularly in the higher-yielding savings and small time deposit market segments. 

In some periods, the differences in the level of deposit rates are substantial. For instance, while the median rate of digital banks’ savings deposits amounted to 1.44% per annum during 2002-07, it amounted to only 1.25% for small banks and 1.10% for large banks. In the wake of downward trending interest rates over the sample period, the differences became smaller over time in absolute terms, but not necessarily in relative terms. For instance, the median savings deposits rate of digital banks during the most recent period is 0.20% per annum compared to 0.15% and 0.10% for small and large banks, respectively.

Figure 2 Deposit rates and spreads across banks 

Note: The figures present quarterly median values of annual deposit rates for (i) digital banks, (ii) banks with median assets above $1 and below $250 bn, and (iii) banks with median assets below $1 bn. Saving deposits are money market deposit accounts with an account size of $25,000; small time deposits are 12-month certificates of deposit with an account size of $10,000; and checking deposits are interest-bearing checking accounts with a minimum balance of $2,500. The shaded areas correspond to the period 2007Q2-2009Q1 (Global Financial Crisis) and 2019Q4-2021Q3 (Covid Pandemic).
Sources: RateWatch, Call reports, SoD; authors’ calculations.

A comparison of median deposit rates in counties with high versus low Twitter activity suggests that digital banks offer significantly higher deposit rates when Twitter activity is high (see Table 8 in Brei et al. 2026), in line with the notion of heightened competition in such an environment. Specifically, we find that digital banks set significantly higher rates in the savings and small-time deposit market segment when their branches are located in counties with high Twitter activity, compared to branches located in counties with less Twitter activity. 

Policy rate pass-through at digital versus traditional banks

Digital banks adjust deposit rates faster and more strongly than traditional peers. Figure 3, which reports the results from dynamic pass-through regressions for up to eight quarters ahead, suggests that the difference is economically and statistically significant.

Figure 3 Branch-level responses of deposit rates to changes in the Fed funds rate

Note: The estimated responses at the branch-level are based on equation (5) in Brei et al. (2026) and measured in basis points. The figures show the local projection (LP) responses (i.e., cumulative differences) of deposit rates to a 100 bp increase in the federal funds rate (for checking and savings deposits) and a 100 basis point increase in the 12-month T-bill rate (for small time deposits). Saving deposits are money market deposit accounts with an account size of $25,000; small time deposits are 12-month certificates of deposit with an account size of $10,000; and checking deposits are interest-bearing checking accounts with a minimum balance of $2500. The shaded area indicates 99% confidence intervals based on standard errors clustered at the branch level. 
Sources: RateWatch, Call reports, SoD; authors’ calculations.

Deposit rate sensitivity differs across products: checking accounts are the least responsive, while small time deposits are the most responsive. While a 100 basis point increase in the policy rate translates into an 89 basis point pass-through to small time deposit rates at traditional banks, the corresponding effect at digital banks reaches 96 basis points. This implies that digital banks nearly fully compensate their customers in this deposit market segment after a year and a half.

High social media activity further accelerates and strengthens policy rate pass-through at digital banks. Concerning social media activity, our within-bank estimations following Drechsler et al. (2017) for the period 2016–2019 suggest that digital banks align small time deposit rates more strongly with the policy rate in counties with higher social media activity (Table 12 in Brei et al. 2026). This suggests that faster information flows and more attentive depositors increase price competition.

Policy implications

Overall, our findings suggest that digitalisation and social media make retail deposits more sensitive to interest rate changes and can accelerate the transmission of monetary policy to banks’ funding costs, especially for higher-yielding deposit products. This has implications for banks’ funding strategies, deposit franchise values and the speed of monetary policy transmission. For policymakers, this implies that the pass-through of policy rates could become faster but also more volatile as digitalisation and social media adoption continue to expand.

References

Basel Committee for Banking Supervision (2024), Digitalisation of finance.

Brei, M, G Cornelli, L Gambacorta and B Hofmann (2026), “The digitalisation of banking and social media: Implications for deposit pricing”, BIS Working Paper 1357 (also published as CEPR Discussion Paper 21609). 

Cornelli, G, J Frost, C Velásquez, J Warren and C Yang (2024), “Retail fast payment systems as a catalyst for digital finance”, BIS Working Paper 1228.

Drechsler, I, A Savov and P Schnabl (2017), “The deposits channel of monetary policy”, Quarterly Journal of Economics 132(4): 1819–1876.

Jiang, E X, G Matvos, T Piskorski and A Seru (2023), “Monetary tightening and U.S. bank fragility in 2023: Mark-to-market losses and uninsured depositor runs?”, NBER Working Paper 31048.

Koont, N (2023), “The digital banking revolution: Effects on competition and stability”, SSRN Working Paper.

Martin, Y, Z Li, Y Ge and X Huang (2021), “Introducing Twitter daily estimates of residents and non-residents at the county level”, Social Sciences 10(6).

Rose, J (2023), “Understanding the speed and size of bank runs in historical comparison”, Federal Reserve Bank of St. Louis Economic Synopses 12.