Session 12: Banking and the Deposit Franchise

We investigate the dynamics of household deposits using account-level data from 12 million accounts across 154 U.S.credit unions. Significant skewness in the retail deposit distribution— with 10% of depositors controlling 70% of total deposits—means large-balance accounts drive aggregate retail deposit flows. On average, high-balance accounts become large after significant one-time inflows and are more likely to experience large, idiosyncratic drawdowns. Unlike low-balance households, high-balance retail depositors are not sensitive to interest rate shocks. Additional evidence suggests that overall retail deposit stickiness is driven by high-balance accounts that are used as medium-run liquidity stores rather than for interest income.

We show that depositor inattention gives banks deposit market power, explaining incomplete monetary pass-through and generating interest rate exposure that changes sign over the monetary cycle. We present a dynamic deposit-pricing model in which banks trade off current deposit spreads against future deposit base, with inattention dampening spread-sensitive outflows. Empirically, we measure inattention using differential responses to scheduled versus unscheduled income and show that inattentive depositors withdraw less following rate hikes. The data confirm that banks with more inattentive depositors have lower deposit rates, weaker pass-through, and less spread-sensitive outflows. Our calibration quantifies how inattention shapes pass-through and financial stability.

We use transaction-level data from a major Icelandic bank—servicing about one third of the population—to study households’ portfolio allocation within liquid, short-term, safe assets that are identical in risk, maturity, and liquidity but offer different yields. These assets are the main source of liquid financial wealth for households and the largest source of funding for banks. Despite all assets being equally safe and liquid, with instantaneous and free transfers, we find substantial heterogeneity in holdings: households tilt portfolios toward high-return assets as wealth rises, yet portfolios are largely unresponsive to interest-rate fluctuations, except among the wealthy. Even in this frictionless environment, we document large costs of inaction, with forgone interest income for wealthy households reaching about 2.5 percent of annual consumption. We find that aggregate deposit fluctuations are driven mainly by wealthy households’ portfolio rebalancing. Finally, we show that a calibrated portfolio-choice model with standard adjustment frictions replicates cross-sectional holdings but substantially overstates sensitivity to interest rates.

This paper studies how banks compete through a novel lens: the content of advertising videos. Using video embeddings, we identify three competitive dimensions: pricing, service quality, and trust-building emotional appeals. We build a framework connecting advertising content to franchise value and monetary policy transmission. Empirical evidence is consistent with model predictions. Banks with high local market shares compete on service and trust while downplaying pricing. New entrants compete primarily on pricing. Banks lacking pricing or service advantages lean on emotional appeals. Non-price advertising strengthens the deposit franchise, enabling banks to maintain wider spreads during monetary tightening and shaping monetary transmission.

We examine how banks’ depositor composition shapes lending behavior, using granular supervisory data on deposits, loans, and securities for the largest U.S. banks. Classifying banks by depositor specialization, we find persistent differences in funding that translate to differences in asset allocations. Retail-depositor oriented banks hold longer-maturity loans and conduct more real estate lending, while corporate- and NBFI-oriented banks, whose funding is more volatile, hold shorter loans and liquid securities. Loan-level analyses show that stable funding is associated with lower rates, longer maturities, and larger loans. Growth in deposits is allocated differently depending on the depositor specialization of the bank, something we can explore using exogenous deposit growth during COVID.

Banks’ deposit market power is customer-specific: depositors who hold stocks, bonds, or investment fund shares (market-priced savings, MPS) have stronger outside options and therefore more elastic deposit demand. Using Danish administrative data linking the universe of Danish deposit accounts to each depositor’s complete investment portfolio, we show that banks price this elasticity: MPS holders receive a 6.6 bps larger deposit-rate increase per 100 bps policy-rate increase than comparable non-holders at the same bank in the same year. The premium opens when depositors acquire MPS and fades when they sell these assets. We identify the pricing response using exogenous variation in MPS ownership from inheritances following unexpected parental deaths. Higher pass-through only partially offsets MPS holders’ greater elasticity: under the same tightening, they reduce deposits 2.4 pp more, and the resulting funding pressure leads high-MPS banks to cut lending.