Average tax refund up nearly 11% from a year ago, IRS data shows

Average tax refund through March 20 is $3,571 in the 2026 filing season, up $350 or 10.9% year-over-year, with total refunds exceeding $202B (up 12.9%) and roughly 56.7M refunds issued (up 1.8%). Filing activity is slightly slower than last year (returns received down 0.9% to ~78.9M; processed down 1.1% to ~77.8M), while direct deposit refunds rose 6.5% to ~57.3M with an average direct deposit refund of $3,561 (up 8.4%) and total direct-deposit refunds near $204B (up 15.5%). IRS.gov traffic surged 55.6% to >380M visits, and recent tax-law changes under the One Big Beautiful Bill Act introduced temporary deductions (tips, overtime, auto loan interest, enhanced senior deduction) and new seeded savings 'Trump Accounts,' which may be driving increased taxpayer engagement.

Analysis

A near-term, concentrated liquidity impulse tends to behave like a short-lived stimulus: it lifts goods and services where households have high marginal propensity to consume (MPC) — think discretionary e-commerce, quick-turn home improvement and the used-vehicle market — within 0–3 months, then fades as saving/repayment choices dominate. Expect disproportionate uplift to businesses that convert one-time flows into recurring revenue (subscription services, buy-now-pay-later vendors who capture payment volume, and brokers that convert deposits into AUM).

Faster migration away from legacy disbursement rails amplifies two structural winners: firms that own the on- and off-ramps (API access, card issuance, fraud detection) because incremental government/tax-season flows are sticky from network effects; and software platforms that reduce advisor friction by enabling self-serve filing, since market share gains in tax season are sticky into tax-adjacent financial products. That combination raises both transaction volumes and KYC/fraud exposure — a revenue upside offset by rising compliance and chargeback costs for exposed processors.

Policy-driven creation of new small-dollar custodial savings instruments creates a latent, very long-duration asset source for custodians and brokerages; the per-account balances will be small but highly sticky, producing attractive lifetime value if acquisition costs are managed. Key risks: fraud and consumer-protection regulation could spike after concentrated digital disbursement events, and any quick reversal of supportive fiscal measures would remove the near-term demand bump — these are 0–6 month to 1–3 year horizon risks depending on political developments and regulatory filings.

Key Decisions for Investors

• Pair trade (3–9 months): Long INTU (Intuit) 2% NAV / Short HRB (H&R Block) 1% NAV. Rationale: capture secular share-shift to self-serve tax software and cross-sell monetization. Target 20–30% net spread; stop-loss at 10% adverse move and revisit if new regulatory actions target tax software fees.
• Long FIS (FIS) or FI (Fiserv) equity (6–12 months) 1.5–2% NAV. Rationale: selective increase in government and payroll-related transaction volume benefits processors with issuance and gateway capabilities; payoff asymmetric if volume converts to higher fee-bearing products. Risk: margin compression from competitive bidding; hedge with 6–9 month protective puts sized to limit downside to 8–10% NAV.
• Long SCHW (SCHW) 12–24 months 1.5% NAV. Rationale: ongoing policy-created custodial flows are sticky AUM acquisition opportunities for low-cost brokers; upside via increased assets under custody and allocation into proprietary products. Expect 15–25% upside if adoption meets modest assumptions; catalyst: successive quarterly AUM inflow beats. Monitor regulatory guidance on custodial-fee caps.
• Event hedge (0–6 months): Buy protection on consumer-facing fintechs and card issuers (e.g., 3–6 month ATM/chargeback-related put spreads on GDOT or MA sized 0.5–1% NAV). Rationale: elevated fraud/scam incidents during heavy disbursement windows can trigger outsized chargebacks or litigation; small insurance preserves core long exposure while keeping upside intact.