cdfi-loan-pricing

CDFI loan pricing model — compute minimum viable loan rates from first principles: cost of capital, expected loss, admin cost, and target ROAA. Built for Community Development Financial Institutions that need transparent, component-driven pricing rather than market-rate following.

Why

CDFIs lend into markets where credit is scarce or expensive. Pricing a loan "to market" often means either leaving money on the table or charging more than borrowers can bear. This library models pricing from the CDFI's own cost structure so you can:

• Find the breakeven rate for any loan from your own cost structure — note that this figure includes a required equity return (target_roae * capital_charge_rate), so it sits above the rate that recovers cash costs alone; see Known issues in CHANGELOG.md
• Set a rate that achieves a target return while staying mission-aligned
• Quantify the cross-subsidy your profitable loans provide to mission loans
• Run sensitivity analysis on any cost assumption

Installation

pip install cdfi-loan-pricing

Quickstart

Every figure below is generated: the code block is the verbatim source of examples/quickstart.py and the output block is that script's captured stdout. tests/test_readme_quickstart.py fails if either drifts.

from cdfipricing import (
    LoanRequest,
    CDFICostStructure,
    recommend_rate,
    compute_breakeven_rate,
    sensitivity_analysis,
    cdfi_standard_scenarios,
    risk_tier,
    loan_profitability,
    portfolio_profitability,
    cross_subsidy_analysis,
    compare_pricing_scenarios,
    market_rate_comparison,
)

# Define a loan to price
loan = LoanRequest(
    loan_amount=750_000,
    term_years=10,
    amortization_years=20,
    sector="small_business",
    ltv=0.75,
    dscr_at_origination=1.25,
    borrower_credit_score=660,
    geographic_distress_level="medium",
)

# Define the CDFI's cost structure
cost_structure = CDFICostStructure(
    cost_of_funds=0.0300,       # 3.00% weighted average cost of debt
    target_roaa=0.0075,         # 0.75% target return on average assets
    target_roae=0.0450,         # 4.50% target return on average equity
    admin_cost_pct=0.0300,      # 3.00% annual operating cost
    loan_loss_reserve_rate=0.0150,  # 1.50% annual loan loss provision
    capital_charge_rate=0.15,   # 15% equity capital requirement
    fee_income_pct=0.0050,      # 0.50% net fee income
)

# Get the full pricing recommendation.
# Rates print as percentages, dollar amounts as currency — each field's unit
# is declared, never inferred from its Python type.
result = recommend_rate(loan, cost_structure)
print(result.summary())

# Check breakeven alone
be = compute_breakeven_rate(loan, cost_structure)
print(f"\nBreakeven rate: {be:.2%}")

# Which risk tier did the loan land in?
print(f"Risk tier: {risk_tier(loan)}")

# Sensitivity to cost of funds
print("\nSensitivity to cost of funds:")
rows = sensitivity_analysis(loan, cost_structure, "cost_of_funds", [0.02, 0.03, 0.04, 0.05])
for r in rows:
    print(f"  CoF {r['value']:.1%} -> recommended {r['recommended_rate']:.2%}")

# Use pre-built scenarios for different CDFI types
print("\nPre-built CDFI scenarios:")
scenarios = cdfi_standard_scenarios()
comparisons = compare_pricing_scenarios(loan, scenarios)
for c in comparisons:
    print(f"  {c['scenario']:22s}  {c['recommended_rate']:.2%}")

# Single-loan profitability at an actual rate
prof = loan_profitability(loan, cost_structure, actual_rate=0.085)
print(f"\nNet income: ${prof['net_income']:,.0f}   ROAA: {prof['roaa']:.2%}")

# Portfolio analysis
loan2 = LoanRequest(500_000, 7, 15, "affordable_housing", 0.60, 1.40, 710, "low")
port = portfolio_profitability([loan, loan2], [0.085, 0.065], cost_structure)
print(f"Portfolio ROAA: {port['portfolio_roaa']:.2%}")

# Cross-subsidy: which loans subsidize which?
subsidy = cross_subsidy_analysis([loan, loan2], [0.085, 0.055], cost_structure)
print(f"Self-sustaining: {subsidy['is_self_sustaining']}")
print(f"Annual subsidy: ${subsidy['total_subsidy_amount']:,.0f}")

# Compare against market rate
mkt = market_rate_comparison(loan, cost_structure, market_rate=0.115)
print(
    f"Mission discount: {mkt['mission_subsidy_bps']:.0f} bps "
    f"(${mkt['annual_subsidy_dollars']:,.0f}/yr)"
)

Output:

CDFI Loan Pricing Summary
========================================
  Recommended rate : 9.8250%
  Breakeven rate   : 8.5750%
  Target rate      : 9.8250%

Rate Components:
  cost_of_funds                2.5500%
  expected_loss                2.8500%
  admin_cost                   2.5000%
  capital_charge               0.6750%
  target_return                0.7500%
  risk_tier_premium            0.5000%

Profitability Metrics:
  net_interest_margin          7.2750%
  spread_over_breakeven        1.2500%
  annual_gross_income          $73,687.50
  annual_loss_provision        $21,375.00
  net_income_estimate          $9,375.00
  estimated_roaa               1.2500%
  meets_target_roaa            True

Breakeven rate: 8.58%
Risk tier: tier_2

Sensitivity to cost of funds:
  CoF 2.0% -> recommended 8.97%
  CoF 3.0% -> recommended 9.83%
  CoF 4.0% -> recommended 10.68%
  CoF 5.0% -> recommended 11.53%

Pre-built CDFI scenarios:
  affordable_housing      8.22%
  urban                   9.29%
  healthcare              10.02%
  rural                   10.95%
  small_business          12.14%

Net income: $8,250   ROAA: 1.10%
Portfolio ROAA: 0.84%
Self-sustaining: False
Annual subsidy: $9,187
Mission discount: 168 bps ($12,562/yr)

Key Features

• Component-driven pricing — cost of funds, expected loss, admin cost, capital charge, and target return each contribute a transparent, auditable slice of the rate
• Sector-specific default rates — 10 sectors (small business, affordable housing, healthcare, agriculture, etc.) with calibrated expected-loss assumptions
• Geographic distress premiums — four distress levels (low / medium / high / severe) with risk add-ons
• Risk-tier classification — DSCR, LTV, and credit score combine into four pricing tiers, exposed as risk_tier(loan) and risk_tier_premium(loan)
• Pre-built CDFI scenarios — rural, urban, healthcare-focused, affordable housing-focused, small business-focused
• Sensitivity analysis — vary any cost or loan parameter and see rate impact across a range
• Cross-subsidy analysis — identify which loans fund mission-rate lending in your portfolio
• Market rate benchmarking — quantify mission discount in bps and dollars
• Declared units — every field of a PricingResult declares whether it is a rate, a dollar amount or a flag (cdfipricing.UNIT_REGISTRY), so PricingResult.summary() cannot label its dollars as percentages. The registry is keyed by field name alone and covers recommend_rate's output only: loan_profitability returns admin_cost and cost_of_funds as dollars under names the registry declares as rates, so do not call cdfipricing.render() on anything but a PricingResult. See the CHANGELOG's Known issues.

Use Cases

• Loan officers pricing individual credits against the CDFI's actual cost structure
• CFOs modeling the blended portfolio rate needed to sustain net asset growth
• Board/CDFI Fund reporting — documenting the mission subsidy embedded in below-market loans
• Grant applications — quantifying the subsidy value of concessionary capital
• Scenario planning — how does a 100 bps rise in cost of funds affect minimum viable rates?

License

MIT © Jay Patel