Skip to search results

Section 05

Who Flips Your Loan

Lender funding model helps explain why similar SBA loans price differently. The SBA guarantees up to 85% of the loan, and that guaranteed strip trades on a secondary market at roughly a 9-point premium. Some lenders sell nearly everything they originate; others retain nearly all of it.

How this works

The 7(a) guaranty is tradeable. A Treasury-backed income stream sells at a premium on the secondary market — roughly $109 for every $100 of guaranteed principal at FY25 volume-weighted averages. On a $1M loan with an 85% guaranty, that premium is about $76,500, booked by the originating lender at sale. Loan-level sale dates and per-loan premiums are not published in any SBA release, so the premium dollars are an estimate over aggregated distributions; the binary sold-flag itself is OBSERVED at loan level.

The 7(a) program is built around a liquid secondary market. The guarantee turns ten-year illiquid paper into a security and broadens the capital base that funds the program. Whether a given lender chooses to sell is a function of how that lender is funded.

Deposit-funded banks can hold a 7(a) loan for a decade against checking-account liabilities. Selling is optional. Lenders without deposit funding — non-depository SBLCs in particular — operate off warehouse lines and the proceeds of each prior sale. For those lenders the sale is how the funding cycle completes. HYPOTHESIS: gain-on-sale economics appear central to the business model of high-sale lenders, and the pricing patterns in the rest of this section are consistent with that framing, though the public data does not distinguish funding sources directly.

The FY25 league table ranks lenders on volume alone. It does not reveal which funding model produced the loan.

Top-12 lenders, two funding profiles

0% vs 99%retained vs sold, FY23–25 pooledvariable-rate 7(a), top-12 FY25 lenders
6 of the top-12 7(a) lenders by FY2025 loan count sold 0.0 to 0.9% of their variable-rate book. 4 others sold 98.6 to 99.9%. Live Oak and Newtek Bank NA sit between those poles at 64.6% and 74.8%. The volume-ranked league table does not distinguish between these funding profiles.
(7(a), FY23–25, n=63,586, funded variable; Top-12 FY25 lenders by count)
“Lender identity explains 69% of the variance in whether a 7(a) loan ends up on the secondary market.”

The univariate R² of the sold-flag across eight candidate factors comes out at 0.689 for lender identity. Subprogram explains 0.282. Rate structure, 0.077. State, 0.057. Term, 0.097. Size band, 0.018. Lender identity dominates every borrower-side input tested by a factor of seven or more (OBSERVED). (7(a), FY25, n=65,529, univariate R² on sold-flag, 8 candidate factors)

INTERPRETATION: the lender’s funding model appears to shape pricing before the borrower sees a final quote. The same underlying deal can end up on a held balance sheet or in a pool certificate depending on which lender writes it, and the channel is a strong correlate of the rate the borrower ultimately pays.

Chart 6A

Chart 6A

Volume rank does not distinguish retained-book lenders from originate-to-distribute lenders.

Huntington funded 6,524 7(a) loans in FY2025 and sold 148 out of 16,406 three-year variable originations. Lendistry funded 1,774 and sold 2,840 of 2,843. Both appear on the same league table. One runs a retained-book model; the other runs a high-sale, originate-to-distribute model. Row order alone does not surface that distinction.

Retained book (≤5%)Mixed (5–90%)High sale (≥90%)
Top-12 FY2025 7(a) lenders by loan count, rows in rank order. Bar = FY23–25 pooled sale rate on funded variable-rate 7(a). Funded-variable pool only; fixed-rate 7(a) sells at 3.0% across 22,196 FY23–25 loans and sits in a separate market.
(7(a), FY23–25, n=63,586, funded variable; status IN (EXEMPT, PIF, CHGOFF))
The sorting mechanism

The sorting mechanism

Retained-book lenders hold the loan and collect coupon income for up to ten years — the 7(a) cap on most working-capital and acquisition paper. High-sale lenders book a one-time gain-on-sale at or shortly after funding, retain a thin servicing strip, and redeploy capital into the next origination. Two revenue-model orientations sit inside the same program and the same league table.

Balance-sheet structure appears to drive the split. Non-depository SBLCs have no deposits and operate off warehouse lines, so sales complete the funding cycle. Deposit-funded banks can choose, and the money-center franchises in the data mostly choose to retain. The bifurcation in Chart 6A is consistent with a charter-and-funding constraint rather than a borrower characteristic.

Inside every size band, the sold cohort pays 51 to 77 basis points more than the held cohort on comparable deals (OBSERVED). (7(a), FY25, n=56,741, variable funded; 6 size bands) INTERPRETATION: the gap does not appear to be a credit premium — collateral presence shifts the rate by two points and the spread by 201 basis points, a much larger effect. A candidate explanation is that sold-channel pricing tracks pool investors’ coupon demand while held-channel pricing tracks the lender’s own cost of funds. Both channels quote a single term sheet to the borrower.

FY25 estimated up-front premium captured by originating lenders on 7(a) guaranteed-portion sales is ~$1.15 billion, roughly 46 percent of FY25 Year-1 industry lender revenue on the full 7(a) book. For Newtek, Readycap, Lendistry, Northeast, and BayFirst, the estimated premium share runs past 70%. The revenue composition is the clearest quantitative signature of the originate-to-distribute model — 10-Q disclosures remain the primary external confirmation for the specific firms.

Chart 6B

Chart 6B

Within every size band, sold loans price higher than held loans.

The gap persists across size bands and across credit controls — collateral presence shifts the rate by two points but shifts the spread by 201 basis points, a much larger effect. INTERPRETATION: pool investors treat the SBA guaranty as covering collateral-shortfall risk and bid on coupon, so a Prime+3.00% variable 7(a) is a more valuable cash-flow stream than an identical Prime+1.50% and commands a higher premium. Sold-channel lenders book that premium at closing; held-channel lenders earn net interest on the coupon over the loan’s life. The borrower signs one term sheet either way.

Size bandn soldn heldSold vs held spreadSoldHeldGap (bps)
$0-$150K9,59018,329P+4.03%P+3.37%+65
$150K-$350K4,5716,201P+3.16%P+2.40%+77
$350K-$500K3,0352,973P+2.62%P+1.96%+66
$500K-$1M2,7852,569P+2.24%P+1.73%+51
$1M-$2M1,8841,528P+2.09%P+1.49%+60
$2M-$5M1,8381,438P+1.89%P+1.14%+75
Sold to secondaryHeld on balance sheet
Simple per-loan average spread within cohort (not volume-weighted), 7(a) FY2025 variable funded, n=56,741. Gap = sold-cohort mean spread minus held-cohort mean spread, same size band.
(7(a), FY25, n=56,741, variable funded; sold vs held within size band)
The revealed preference

Five observations from the pooled data

First.The secondary-market sale rate appears to saturate near 60% of originations inside a homogeneous bucket. In the $500K–$1M variable 7(a) Guaranty cohort (FY25, n=5,210), sold-rate climbs from 3% at P+<0.5% to 60% at P+1.5–2.0%, then flattens across the P+1.5% to P+3.5% band. Above P+3.5% it does not rise further. INTERPRETATION: pool investors price prepayment — 52% of P+2.75%+ loans pay off before half their term — and the premium bid caps out. Originations above that ceiling end up retained by the originating lender.

Second. Live Oak Bank originates two distinguishable cohorts inside one charter. Sold and held cohorts price 65 to 106 basis points apart in every size band (OBSERVED). (7(a), FY23–25, n=4,506, Live Oak funded variable; 6 size bands)

Size bandSold spreadHeld spreadWithin-lender gap
$0-$150KP+3.17%P+2.10%+107 bps
$150K-$350KP+2.86%P+1.90%+96 bps
$350K-$500KP+2.36%P+1.67%+69 bps
$500K-$1MP+1.45%P+0.61%+84 bps
$1M-$2MP+1.27%P+0.51%+76 bps
$2M-$5MP+1.31%P+0.23%+108 bps

INTERPRETATION: Live Oak’s sold cohort appears priced to clear the secondary market, while its held cohort appears priced closer to deposit funding costs. The 65- to 106-basis-point gap is consistent with a deposit-funded bank that routes a growing share of its book through the secondary-market channel.

Third.Newtek Bank NA’s sold-cohort pricing clusters at exactly P+3.00% median across all six size bands (OBSERVED). That sits right below the P+3.5% ceiling at which prepayment risk appears to cap the premium bid. The held cohort is the residual: sub-$150K loans at P+5.14% that did not clear the pool market at prevailing bids. Live Oak and Newtek show two distinct pricing fingerprints consistent with the same mechanism — the secondary-market bid sets a coupon ceiling, and lenders price their sold cohorts against that ceiling in different ways.

Fourth.FA$TRK (SBA Express) sits outside this secondary-market economy. FY25 variable 7(a) sale rate: Guaranty 61.9% (n=37,904), FA$TRK 0.7% (n=18,382). FA$TRK loans run smaller (~$150K vs ~$500K+) and concentrate in two deposit-funded banks — Huntington (19,765 FY23–25 FA$TRK variable, 0.8% sold) and M&T (5,223, 0.2% sold). Pool investors do not materially bid on small-balance single-obligor tranches. The premium-pricing framework in this section does not apply to FA$TRK originations.

Fifth. The aggregate sale rate rose from 37.4% in FY23 to 47.7% in FY25 — ten points in two years (OBSERVED), concentrated in the sub-$150K band where non-depository SBLCs scaled (Readycap, Lendistry, BayFirst, Celtic, and Northeast all run 95%+ sub-$150K sale rates). INTERPRETATION: the rise tracks growing non-depository share. If non-depository SBLCs acquire bank charters and deposit funding, the bifurcation may compress; without that shift, the sale rate is likely to stay elevated. (7(a), FY23 and FY25, n=91,847, funded variable; FY23 n=42,275 / FY25 n=49,572)

Taken together: the secondary market pays for coupon, the sale rate saturates near 60% inside homogeneous buckets, and non-depository SBLC growth explains most of the recent aggregate rise. Live Oak, Newtek, and Huntington occupy three different positions in the same program because their balance-sheet structures differ.

Chart 6C

Chart 6C

The aggregate sale rate rose ten points between FY23 and FY25.

The FY22 → FY23 trough coincides with the Fed’s fastest tightening. INTERPRETATION: investor premium bids compressed as spread-to-alternatives narrowed, and deposit-funded banks retained a larger share of what they wrote. FY24–25 reverses that pattern as non-depository SBLC originations scaled. The most likely explanation for the ten-point rise is supply-side composition — growing non-depository share — with a demand-side component as investor appetite recovered.

42.9%
FY2020
n=27,537
44.4%
FY2021
n=36,187
38.5%
FY2022
n=34,693
37.4%
FY2023
n=42,275
45.0%
FY2024
n=51,126
47.7%
FY2025
n=49,572
Share of funded variable 7(a) originations with secondary-market sale flag, by approval fiscal year. FY24–25 rise from 37.4% to 47.7% is the current cycle’s most visible shift; the FY22–23 trough coincides with the Fed’s fastest tightening.
(7(a), FY20–FY25, n=241,390, funded variable; status IN (EXEMPT, PIF, CHGOFF))
The revenue model

Industry revenue decomposition

Estimated $1.15 billion in FY25 up-front premium, about 46% of Year-1 industry revenue.

Estimated FY25 up-front premium captured by originating lenders on 7(a) guaranteed-portion sales comes to about $1.15 billion, roughly 46 percent of FY25 Year-1 industry lender revenue on the full 7(a) book (honest range 40–52%). Per-loan premiums are not published, so this is a model output over the aggregated SBA 7(a) Premium Tables distribution; the finding holds across the honest range. (7(a), FY25, n=26,166, $12.46B sold guaranteed at vol-wtd ~109.2 per SBA 7(a) Premium Tables PDF)

For Newtek, Readycap, Lendistry, Northeast, BayFirst, and Celtic, estimated gain-on-sale premium runs past 70% of Year-1 revenue. HYPOTHESIS: these lenders operate an originate-to-distribute model in which premium is the dominant Year-1 revenue line; the pricing patterns in Chart 6A and 6B are consistent with that framing, and Newtek’s public 10-Q disclosures provide external confirmation for at least one entity in the set.

FY25 sector — Year-1 7(a) lender revenue

Gain-on-sale premium$1.15B
Net interest margin$0.750B
Packaging + closing fees$0.450B
Servicing$0.125B
Year-1 total$2.48B
Premium share46%

Newtek — FY2024 worked example

Gain-on-sale premium$0.079B
Net interest margin$0.017B
Servicing$0.008B
Year-1 total$0.10B
Premium share76%
Left: FY25 industry-level Year-1 lender revenue on the full 7(a) book. Right: Newtek FY24 origination cohort (all figures in $B). Premium estimate uses SBA 7(a) Premium Tables PDF vol-weighted averages applied to sold guaranteed-portion volume. Honest range on the sector premium share is 40–52%; finding holds across the range.
FYSold guaranteedLoans soldVol-wtd premiumEst. up-front premium
FY2021$10.82B13,101~113.0~$1.41B
FY2022$12.47B16,713~111.5~$1.43B
FY2023$9.25B15,815~108.0~$0.74B
FY2024$10.73B23,188~108.3~$0.89B
FY2025$12.46B26,166~109.2~$1.15B
Chart 6E

Chart 6E

In FY25 H2, the sale rate on $1M+ variable 7(a) loans fell twelve points versus H1.

The prior three years held H1/H2 parity within four points. FY25 breaks that pattern: H1 62.0%, H2 49.9%, a 12.1 pp gap on $1M+ variable 7(a) (OBSERVED). Two candidate mechanisms appear in the data. INTERPRETATION: rate-cycle inflection is the leading candidate — the Fed cut 100 bps from Sep24 to Dec24, and high-coupon paper retains more balance-sheet value when the investor bid is falling. Data-recency artifact (public release as of 2025-12-31, late-FY25 originations have had less time to settle) explains roughly 2–3 points based on prior-year H1/H2 differentials. The residual roughly nine points is consistent with a behavioral shift toward retention through the cut cycle.

FY2022

61.9%
H1
65.0%
H2

Gap -3.1 pp

FY2023

64.5%
H1
68.1%
H2

Gap -3.6 pp

FY2024

68.9%
H1
65.7%
H2

Gap +3.2 pp

FY2025

62.0%
H1
49.9%
H2

Gap +12.1 pp

$1M+ gross amount, funded variable 7(a). FY25 H1 n=3,855; FY25 H2 n=4,599. Prior years held H1/H2 parity on the sale rate within three points.
(7(a), FY22–FY25, n=24,044, $1M+ gross amount; funded variable; FY25 H1 n=3,204, H2 n=3,834)
Named lenders

Top-12 high-sale lenders — FY23–25 funded variable 7(a)

Ranked by sale rate with n ≥ 50 floor. Spread and size are raw cohort averages across the FY23–25 funded-variable pool. (7(a), FY23–25, n=41,512, funded variable; n ≥ 50)

#LenderTypenSale %Avg spreadAvg size
1Lendistry SBLC, LLCnonbank2,84399.9%P+4.44%$205K
2Northeast Bankbank5,18099.1%P+2.97%$137K
3Readycap Lending, LLCnonbank6,70498.8%P+4.48%$366K
4BayFirst National Bankbank6,94798.6%P+4.29%$158K
5Newtek Small Business Finance, Inc.nonbank54998.5%P+3.28%$582K
6Harvest Small Business Finance, LLCnonbank96195.7%P+2.51%$500K
7Celtic Bank Corporationbank2,91792.7%P+2.84%$290K
8SouthState Bank, N.A.bank1,06384.7%P+2.59%$1.07M
9Newtek Bank, National Associationbank5,68574.8%P+3.01%$420K
10Byline Bankbank1,16171.7%P+2.67%$793K
11Live Oak Banking Companybank4,50664.6%P+1.92%$1.37M
12KeyBank, N.A.bank2,10630.7%P+2.29%$648K

Top-12 retained-book lenders — FY23–25 funded variable 7(a)

Every row below sold zero loans or sub-1% of its FY23–25 funded-variable pool. Money-center franchises tend to run small-ticket Express programs as deposit-customer acquisition tools; regional banks tend to run larger-ticket 7(a) as a relationship product. Both operate in the sub-1%-sale bucket but with different commercial orientations. (7(a), FY23–25, n=31,306, funded variable; flip ≤ 0.2%; n ≥ 50)

LenderHQnSale %Avg rateAvg size
The Huntington National BankOH16,4060.9%10.22%$205K
TD Bank, National AssociationNY7,9750.0%11.61%$99K
Manufacturers and Traders Trust CompanyNY4,4990.2%11.89%$128K
JPMorgan Chase Bank, National AssociationCA3,8850.0%12.00%$205K
Wells Fargo Bank National AssociationCA3,4850.0%13.42%$28K
U.S. Bank, National AssociationCA2,8960.0%10.68%$208K
Zions Bank, A Division ofUT1,7800.0%10.71%$221K
First Bank of the LakeCA1,4194.4%11.13%$681K
Bank of America, National AssociationCA1,0180.0%9.76%$338K
PNC Bank, National AssociationTX9620.0%12.22%$64K
Banco Popular de Puerto RicoPR8460.0%12.03%$130K
Comerica BankTX4880.0%9.17%$586K

Cohort summary — bimodal distribution

Lenders with n ≥ 50 funded variable 7(a) in FY23–25, grouped by sale-rate cohort. Lenders in the 90%+ sale cohort price 64 bps higher on average than hold-to-maturity peers and originate at 2.2× the mean loan size. The mixed middle (5–50%) is the thinnest cohort, consistent with the bimodal distribution visible in Chart 6A.

Cohort (sale %)LendersTotal loansMean rateMean spreadMean avg size
0-5% (hold-to-maturity)8954,5719.88%P+1.908%$456K
5-50% (mixed)3912,6519.88%P+1.930%$626K
50-90% (mostly flip)9929,99510.19%P+2.244%$941K
90%+ (pure flipper)4434,96710.46%P+2.547%$1.02M
Two-cohort structure

Two-cohort structure

Newtek, Live Oak, Columbia — three distinct sold/held pricing patterns.

Newtek Bank NA’s 74.8% aggregate sale rate is the volume-weighted average of two distinct book shapes. The sold cohort prices at P+3.00% essentially flat across size bands; the retained cohort holds the sub-$150K tail — 1,419 loans at an average P+5.14% that did not clear the pool market at prevailing bids.

Newtek size band (FY23–25, n=5,685)Sale %Sold spreadUnsold spreadRate gap (pp)
$0-$150K60.6%P+3.01%P+5.49%+2.48
$150K-$350K82.5%P+3.01%P+3.88%+0.87
$350K-$500K91.0%P+3.01%P+3.02%+0.01
$500K-$1M87.5%P+3.00%P+3.05%+0.05
$1M-$2M86.7%P+3.01%P+3.04%+0.03
$2M-$5M86.0%P+3.01%P+2.96%-0.05

Newtek’s aggregate 74.8% blends an Express book (52.5% sale rate) with an above-$350K standard book (75–85% sale rate). One charter, two cohorts with distinct pricing characteristics.

Columbia Bank shows a similar shape at smaller scale: 12.8% overall sale rate (n=1,347), sold loans average $1.07M at 10.03%, retained loans average $69K at 12.51%. The sold cohort is a million-dollar-plus acquisition product; the retained cohort is a small-ticket working-capital product that did not clear the pool market. Same program line, two products.

Live Oak’s pattern differs again. 72.9% of its 4,502 funded variable-rate loans sit below cluster median (per Section 3); 38% sit below cluster p25. Volume-weighted, Live Oak’s book is below the 7(a) median even at a 64.6% aggregate sale rate. Newtek’s flat sold-cohort ceiling and Live Oak’s left-skewed pricing distribution are the two clearest within-lender shapes in the book; Columbia’s two-product split is a smaller-scale echo of the Newtek pattern.

Cross-section handoffs
Methodology

Methodology & data

Source data
SBA’s public 7(a) lending disclosure, loan-level, through 2025-12-31. Secondary-market flag is binary Y / blank. No loan-level premium is published in any SBA release. Premium distributions come from the SBA 1086 Premium Heat Maps PDF (gross-loan bands × price bands × FY). Any premium-per-loan stat is a model output over the aggregate distribution, shipped with its estimation sidebar.
Par sold on a 1086
Equals gross_approval × guaranty_pct.gross_amount alone overstates by 25–50% (the retained unguaranteed portion). Small Loans (85% guaranty) and the Feb–Sep FY21 90% window (Economic Aid Act, expired Oct 1 2021) change the ratio; we use the actual guaranty_pct.
PPP excluded
Separate 7(a) authority, CARES Act 100% guaranty, never sold on 1086s. The heat-map dataset filters PPP at source.
Settlement vs approval windows
PDF counts FY25 1086 settlements at 26,166 loans / $12.46B par. Our loan-level approval-cohort counts variable-rate funded 7(a) with sold-flag set at 23,929 loans FY25 / ~$12.2B guaranteed. Both are correct under their windows — PDF = settlement date, loan-level = approval date. Expected gap 5–10%; ~9% observed.
What the data does not show
Sale date is not published at loan level — we can flag which loans sold but not when. Per-loan premium is not published. Broker-channel flag is not in the public release at loan level (Form 159 is filed but not published). Lender concentration of FY25 sold volume is not separately disclosed.
Full methodology
Methodology index carries the full rule set, regulatory citations (13 CFR §§ 120.600–660; § 120.612; § 120.223; § 120.425), and the SHARED-METHODOLOGY cross-reconciliation table.

What to watch in next year's data

Four data points worth watching as FY26 originations settle.

  1. Whether the non-depository 97% sale rate compresses if SBLCs acquire bank charters and deposit funding. Newtek Bank NA’s 74.8% already sits below Newtek Small Business Finance’s 98.5%. A Lendistry, Readycap, or BayFirst charter event would be the first external signal of convergence between bank and non-depository sale rates.
  2. Whether the sold-cohort spread gap (Chart 6B) widens or compresses as rates fall. The gap widened from roughly 6 bps in FY21–22 to roughly 75 bps in FY24–25. Easing rates could compress the premium share of industry Year-1 revenue mechanically.
  3. Whether the FY25 H2 collapse on $1M+ variable 7(a) was rate-cycle-driven or a data-recency artifact. FY26 H1 settlement data resolves this.
  4. Whether SBA publishes Form 1086 settlement data at the line-item level, including premium price by originating lender. Form 1086 is already filed and settled through Guidehouse; publication is a disclosure-policy choice. Until then, the ~$1.15B sector figure carries its honest range and lender-level premium stays at distribution-aggregate level.

Up next

06

Lender League Table

Combined, 7(a)-only, and 504 CDC league tables. Ranked by funding rate, with median spread and size mix for 151 lenders that cleared the reporting floor.

Read section