Deposit betas measure the sensitivity of your deposit costs to market rate changes. A beta of 0.60 means: for every 100 basis points the market moves, your deposit costs move 60 basis points. This single number shapes your net interest margin, your ALM hedge ratios, your funding strategy, and ultimately your earnings stability.
Yet deposit betas are maddeningly hard to estimate and unstable. Historical betas from 2019 are worthless in 2023. Banks that estimated betas on pre-2022 data took massive margin surprises when rates rose and deposits repriced faster than model assumptions. The 2022–2024 rate cycle revealed this starkly: regional banks that had betas of 30–40% in calm markets discovered they had effective betas of 80–90% when spreads widened and competition intensified.
This module teaches you how to estimate deposit betas rigorously, how to track them in real time, and how to build the estimation logic that works across rate cycles. By the end, you'll have a working framework that captures structural changes in your deposit franchise without requiring you to rebuild your model every 18 months.
One simple example: Assume you have $100B in deposits, a 5-year loan portfolio averaging 4.5% yield, and a 50% asset-to-deposit ratio in shorter-term debt (CDs, wholesale funding). Your net interest margin is:
NIM = (Loan Yield × Loan Assets + Short-Debt Yield × Short Debt) / Total Assets - (Deposit Beta × Market Rate Change × Deposit Base / Total Assets)
If the Fed raises rates 100 bps and your deposit beta is 0.40, your deposit costs rise 40 bps on average—a $400M earnings impact per year. If your beta is actually 0.70 (as it was for many banks in 2023), the hit is $700M—a $300M surprise miss. On a typical regional bank's pretax income of $1.5–2B, that's a 15–20% earnings shock.
This is why accurate beta estimation is not an academic exercise. It directly drives your interest rate risk hedging strategy, your funding cost forecasts, and your earnings guidance to investors.
Estimating deposit betas—the sensitivity of your deposit rates to market rate changes—is both crucial and challenging. No single method provides perfect estimates; rather, practitioners combine multiple approaches to triangulate reasonable estimates, update them continuously as new data emerges, and stress test around the uncertainty.
The Historical Regression Approach: Simple but Regime-Dependent
The classical approach involves regressing your bank's average deposit rate against a risk-free benchmark rate (federal funds, SOFR, or Treasury rates) over a historical period. The resulting regression coefficient becomes your estimate of deposit beta. This method yielded useful estimates during the calm 2015–2021 period. JPMorgan's consumer deposits, for example, showed estimated betas of 0.32 when regression analysis was applied to 80 quarters of data from 2015–2019. That 0.32 beta held reasonably well through 2021. However, when the rate environment shifted dramatically in 2022–2023, a simple regression on the same historical period became essentially useless. Historical betas estimated from the zero-rate era provided almost no predictive power in a 400+ basis point rate cycle. When practitioners ran regressions on just 2022–2023 data, JPMorgan's deposit beta revealed itself to be 0.65+, nearly double the historical estimate. For every 100 basis points the Fed raised, JPMorgan's deposit rates rose 65 basis points.
The key insight is that deposit betas are regime-dependent. Historical regression excels for extrapolating normal periods but fails spectacularly during structural breaks. Its strengths are simplicity and reliance on observable data. Its critical weakness is the assumption that past relationships hold in the future—an assumption that collapses during major regime shifts.
Product-Level Beta Analysis: Capturing Heterogeneity
Rather than estimating a single bank-level beta, sophisticated practitioners segment deposits by product category and estimate separate betas for each. Non-interest-bearing checking accounts typically show betas of just 5–15% (very sticky, minimal competition). Savings accounts and money market accounts show higher betas, typically 50–80%, reflecting their price sensitivity and the ease of switching. Retail certificates of deposit show betas of 70–90% (mature product, highly competitive pricing). Commercial demand deposits and money market accounts show the highest betas, typically 60–85%, because commercial treasury managers actively shop rates and quickly move funds to higher-yielding alternatives. Institutional deposits essentially trade at market rates and show betas near 100%.
To estimate your bank-level beta, you calculate the weighted average of product-specific betas, weighted by the proportion of total deposits held in each product. For a concrete example, assume BankX deposits comprise 30% non-interest-bearing checking (5% beta), 40% savings (70% beta), 20% certificates of deposit (85% beta), and 10% commercial accounts (80% beta). The blended beta becomes: 0.555 (calculated as (0.05 × 0.30) + (0.70 × 0.40) + (0.85 × 0.20) + (0.80 × 0.10)).
This product-level framework is more robust than single-equation regression because it acknowledges that different deposits compete in different markets and face different competitive pressures. It also enables powerful scenario analysis: "If our mix shifts to 50% savings deposits (higher beta), what becomes our all-in deposit beta?" The framework allows practitioners to stress assumptions about mix evolution and repricing dynamics separately. The strength of this method is that it forces discipline: you must understand your actual deposit mix and separately justify each product's repricing sensitivity. The weakness is data availability—many smaller banks don't maintain the granular segment-level repricing data necessary to estimate product betas reliably.
Opportunity-Cost-Adjusted Betas: Real-Time Tracking
The most sophisticated approach models deposit betas as a function of the opportunity cost gap rather than absolute market rates. The opportunity cost gap measures the difference between the risk-free rate available to depositors (T-bills, money market funds) and the rate your bank is paying. When the gap is small (50 basis points), deposits are sticky and require less rate movement to retain. When the gap is large (300+ basis points), you must raise rates aggressively or lose deposits.
The formulation expresses deposit rate changes as a combination of base repricing and rate-gap elasticity. Base repricing captures structural repricing—you always raise 30 basis points when the Fed raises 100 basis points, all else equal. The elasticity component captures how much additional repricing is needed per 100 basis points of opportunity cost gap. Real data from 2022–2023 showed that when the opportunity cost gap was small (50 basis points), banks raised deposit rates by just 20 basis points of a 100 basis point Fed move (20% base repricing). When the gap widened to 200 basis points, banks raised 55 basis points (20% base plus 35% elasticity response). When the gap reached 400 basis points, banks raised 75 basis points (20% base plus 55% elasticity).
This non-linear model captures something fundamental: you have pricing power when alternatives are weak, and you're forced into competitive rate wars when alternatives are strong. During Q4 2022, when regional banks faced opportunity cost gaps of 300–400 basis points, the model predicted deposit betas of 70–85%. Actual repricing and deposit outflows confirmed those predictions. JPMorgan and other large banks, with more stable deposit bases and lower sensitivity to basis risk, showed betas of 50–65%, also matching opportunity-cost-adjusted expectations.
Peer Benchmarking and Comparables: External Validation
When your own historical data is sparse or unstable, peer benchmarking provides external validation. Most banks disclose in their 10-K filings the interest paid on deposits divided by average deposit balances, which yields the average deposit cost. By collecting this data for peer institutions and comparing it to your own, you can infer relative betas. If JPMorgan (large franchise, strong) shows a beta of 0.50 and you show 0.75 on the same market moves, the difference signals either that your deposits are less sticky, your competitive position is weaker, or you're losing market share to rate-setting competitors. Peer comparisons are useful for sanity-checking internal estimates and identifying outliers.
The 2022–2024 cycle taught an important lesson: deposit betas are highly regime-dependent, and historical estimates become obsolete quickly when the environment changes. A beta estimated from 2015–2021 (zero-rate period) tells you almost nothing about 2023–2024 behavior (high-rate period). Practitioners now track betas in real time and update assumptions frequently.
The mechanical tracking process involves calculating a rolling 12–18 month regression of your deposit rate changes against Fed funds changes, comparing the fitted beta to your forward-looking assumption, and investigating large deviations. If your model assumes a 0.50 beta for savings deposits and a monthly rolling regression reveals an actual 0.62 beta, that 12 basis point difference suggests your model is underestimating repricing. If the deviation persists for 2–3 months, you should raise your beta assumption for the next quarter's funding forecast. Conversely, if your model assumes 0.65 but actual data shows 0.50, your deposit base may be stickier than assumed, or your franchise may be stronger. Either way, the discrepancy deserves investigation.
SVB's Beta Mistake: The Concentration Risk
SVB's internal asset-liability management models reportedly assumed deposit betas around 0.25, estimated from 2015–2021 historical data. When the Fed raised rates 400+ basis points in 2022–2023, SVB's deposits repriced at a realized beta of 0.90+, nearly dollar-for-dollar with the rate increase. SVB's actual deposit cost (interest paid divided by average deposits) surged from 0.35% to 2.80%—a 245 basis point increase on a 400+ basis point rate move, implying a beta well above 0.60.
Why did SVB's beta shock so dramatically above even the elevated levels other banks experienced? SVB's deposit base was heavily concentrated in venture capital and tech-related companies. These depositors are typically young, highly financially sophisticated, hyper-responsive to rate incentives, and not emotionally sticky to any particular bank. When Treasury bills offered 5% with zero credit risk, they left. SVB had never competed in a high-rate environment and lacked any deposit-pricing discipline from prior cycles. The lesson is clear: concentration of depositor types matters enormously. A deposit base weighted toward rate-sensitive corporate treasurers and venture capital firms will show much higher betas than a diverse retail base. SVB's management learned too late that beta assumptions must account for the composition of the deposit base, not just historical averages.
Huntington's Beta Shock: The Competitive Acceleration
Huntington Bancshares entered the 2022 hiking cycle expecting cumulative deposit betas of 25–27%, based on careful analysis of the 1994–1995 and 2004–2006 rate cycles where betas had been relatively modest. By year-end 2023, actual betas had reached 40%+ dramatically overshooting forecast. In 10-K disclosures, Huntington attributed the overshoot to competitive pressure from other banks and money market funds that intensified throughout 2022–2023. The bank's own pricing responses accelerated as deposit losses mounted, creating a feedback loop where competitive pressure forced faster repricing than historical models predicted.
Huntington's experience revealed an important dynamic: deposit betas are non-linear and accelerate as rate cycles progress. Early in a hiking cycle, banks hope to retain deposits with minimal rate moves (low betas). By mid-cycle, after losses mount, banks realize they must match market rates more aggressively (higher betas). By late cycle, competitive desperation pushes betas to their maximum. The acceleration in competitive pressure was particularly strong in the second half of 2023 as opportunity cost gaps widened beyond historical precedent.
JPMorgan's Franchise Effect: The Power of Relationships
JPMorgan, the largest U.S. bank, maintained betas of 0.45–0.55 throughout the 2022–2023 cycle despite a 400+ basis point rate move—notably lower than many regional competitors. JPMorgan's structural advantage is an extraordinary deposit franchise: over 50 million retail customers with checking and savings linkage, 10 million credit cardholders (switching to a new primary bank is friction-intensive), and 4 million investment management customers whose deposits are bundled with wealth management services. JPMorgan's commercial banking presence dominates corporate payroll processing—moving payroll is a genuine operational and relationship cost for CFOs.
These relationship characteristics manifest in lower deposit betas. Even when JPMorgan paid 1.5% and money market funds offered 4.5%, many depositors stayed because moving their entire payroll setup, credit products, and investment accounts elsewhere required substantial effort. JPMorgan's 0.50 beta during the cycle reflected this lower sensitivity. Simultaneously, JPMorgan's deposits actually grew 1% in 2023 while smaller competitors lost 5–8%, demonstrating the franchise power in an adverse environment.
Estimating betas for forward-looking projections requires layering multiple analyses into a coherent forecast framework. Start with current-period realized betas (last 12 months of actual data) as your baseline. Adjust upward or downward based on your depositor composition relative to peer averages—a retail-heavy franchise should assume lower betas than a commercial-heavy book. Adjust based on your franchise strength (measured by loan-to-deposit ratios, funding costs relative to peers, and stock price performance relative to peer averages). Adjust for the expected rate regime—in high-rate environments (above the long-term neutral rate), assume higher betas; in low-rate environments, assume lower betas. Build out multiple scenarios: a base case where rates stabilize and betas gradually mean-revert toward long-run levels; a stress case where rates spike and betas stay elevated; and an upside case where rates fall and depositors flock back to banks, allowing betas to compress.
Sensitivity testing is critical. For every 50 basis points of beta misestimation on $100 billion of deposits, simulate the earnings impact across a full 100 basis point rate move. The result tells you how much of your earnings forecast depends on accurate beta estimation—often revealing surprises about forecast risk concentration.
Successfully managing deposit betas requires intellectual humility: no single estimation method is perfect, and historical patterns frequently break down. Use multiple methods simultaneously. Combine historical regression, product-level analysis, opportunity-cost adjustment, and peer benchmarking into a triangulation that reduces single-method bias. Update monthly, don't wait for quarterly results. Track actual repricing in real time and compare it to forecasts. When you observe persistent deviations between forecast and actual, investigate the cause and adjust. Acknowledge that some quarters will show wild beta spikes (200+ basis point opportunity cost gaps, competitive crises) and stress your earnings and liquidity forecasting for these scenarios. Finally, remember that franchise quality matters enormously. Banks with strong deposit franchises (JPMorgan, BofA) can sustain betas of 0.45–0.50. Banks with weak franchises or concentrated depositor bases (SVB, many regional banks) face betas of 0.75+. Understand where you sit on the franchise strength spectrum and why, and build beta assumptions that reflect this reality.