CPSS — Q2 2026 Financial Report Analysis

Record loan buying lifts profit 30%, and this time without an accounting boost

Consumer Portfolio Services (CPS) lends to car buyers with damaged or thin credit histories ("subprime" borrowers) by buying their loan contracts from car dealers. In the second quarter of 2026 (April to June) it bought a record $757.7 million of new contracts, 75% more than a year earlier, and net income rose 30% to $6.2 million, or $0.27 per diluted share versus $0.20. The more telling detail is underneath: last year's quarter included a $3.0 million one-time accounting gain on the value of its loans, and this quarter included none. Strip that out and pre-tax profit more than doubled, from about $4.0 million to $9.0 million, driven by a bigger loan book and cheaper borrowing.

At a glance

• $757.7 million of new loans bought (+75%): the most in the company's history, pushing the total portfolio past $4 billion to $4.31 billion (+16%). Growth is now the main story.
• Pre-tax profit $9.0 million vs $7.0 million (+30%): but against roughly $4.0 million once last year's $3.0 million valuation gain is removed, so underlying earnings improved far more than the headline suggests.
• Annualized net charge-offs 7.28% vs 7.45%: the share of the portfolio written off as unrecoverable, on a yearly pace, edged down, and loans 31+ days late (including repossessed cars not yet sold) fell to 12.16% from 13.14%.

Key figures

CPS doesn't report a conventional "operating margin"; pre-tax income divided by total revenue is used here. Figures are from the company's 10-Q for the quarter ended June 30, 2026 and its August 4, 2026 earnings release.

Takeaway: The 30% profit gain understates the quarter. Last year's comparison was padded by a $3.0 million valuation gain; this year had none, so underlying pre-tax profit more than doubled. The engine was a 13.7% bigger average loan book plus cheaper funding: CPS's average borrowing cost fell to 6.6% from 7.0% while what it earns on its loans held near 11.3%. The open question is whether a 75% jump in lending to subprime borrowers keeps the credit quality that today's falling charge-offs suggest.

How CPS makes money, briefly

CPS buys a car loan from a dealer, earns the interest the borrower pays, and funds those loans mostly by bundling them into bonds sold to investors ("securitization"), using short-term bank credit lines ("warehouse lines") to hold loans until the next bundle is sold. Its profit is essentially the gap between what borrowers pay and what CPS pays its own lenders, minus losses on loans that go bad and the cost of running the business.

One accounting detail matters for reading the numbers: CPS records its loans at "fair value". Instead of setting aside a separate provision for bad loans each quarter as most lenders do, it books interest income already reduced for the losses it expects. That's why there's no "provision for credit losses" line, and why the 11.3% yield it reports is lower than the rates borrowers actually pay. When CPS changes its view of future losses or earnings on the existing book, it records a "mark", a gain or loss on the loans' value, which is what produced the $3.0 million gain a year ago.

What drove the quarter

Volume, mostly. Interest income rose $12.8 million (+12.1%). The company's own breakdown attributes $13.8 million of that to the larger loan book (the average portfolio was 13.7% bigger) and a $1.0 million drag to a slightly lower yield (11.3% vs 11.4%).

Cheaper money. Interest expense rose only $5.5 million (+9.5%) despite much more debt, because rates fell: the average cost of its securitization bonds dropped to 6.1% from 6.5%, and its warehouse lines to 7.9% from 9.3%. The two most recent securitizations priced at a blended 5.18% (January 2026) and 5.51% (April 2026), versus 5.88% and 5.96% in early 2025. By CPS's own volume-versus-rate breakdown, the $7.2 million increase in net interest spread split almost evenly: $3.6 million from growth and $3.6 million from lower rates.

Costs held back. Employee costs fell 3.9% to $23.4 million (the 10-Q credits lower average pay among servicing staff, the people who collect payments) even as headcount rose to 994 from 932, mostly sales staff (160 from 118). Sales commissions rose $2.6 million to $8.3 million with the volume surge, and general and administrative costs rose $2.3 million to $14.7 million.

A new source of other income. Other income more than doubled to $3.3 million from $1.4 million, of which $1.8 million came from "dealer recoveries": a third party CPS hired spots cases where a repossessed car's value was misrepresented by the selling dealer and collects the difference from that dealer. There were none a year earlier.

Credit: improving, with caveats

For a subprime lender, credit losses are the business. On the reported measures, the quarter was better than a year ago:

• Annualized net charge-offs were 7.28% of the average portfolio versus 7.45%, and below the full-year 2025 rate of 7.8%.
• Recovery rates (what CPS gets back when it sells a repossessed car at auction, as a share of the loan balance) rose to 33.3% from 30.4%.
• Delinquencies plus repossession inventory fell to 12.16% of the portfolio from 13.14% a year ago and 14.77% at December 31, 2025.

The first half as a whole was less clean: the six-month charge-off rate was 7.90% versus 7.49% a year earlier, which means the first quarter of 2026 ran noticeably hotter before the second quarter improved.

What the headline numbers hide

• The comparison flatters last year, not this year. Q2 2025 included a $3.0 million upward mark on the loan portfolio (the first half of 2025 included $6.5 million). Q2 2026 had no mark. Excluding it, revenue grew 13.7% rather than 10.6%, and pre-tax income rose from about $4.0 million to $9.0 million. This is the rare case where the GAAP headline understates the improvement.
• Some of the gain is new and may not repeat at the same size. The $1.8 million in dealer recoveries didn't exist a year ago. Even without it, pre-tax profit would have been about $7.2 million, still well above last year's roughly $4.0 million excluding the mark, but it's worth watching whether this income stream is steady.
• Fast growth makes credit ratios look better for a while. When a portfolio grows 16% in a year, a large share of its loans are only a few months old, and losses on subprime auto loans tend to show up well after the first months. Charge-off and delinquency percentages are divided by a swelling base of young loans. The real test of the record 2026 purchases comes later.
• CPS itself assumes slightly higher lifetime losses. The key assumption behind the fair value of its loans, cumulative net losses over their life, rose to 16.35% at June 30, 2026 from 16.02% at December 31, 2025. That's a small move, but in the opposite direction from the improving quarterly ratios.
• Delinquency figures are measured after payment extensions. CPS lets struggling borrowers defer payments ("extensions"). $1.79 billion of the $4.31 billion portfolio, about 42%, has had at least one extension, and the delinquency figures reflect those extensions. The company argues extensions keep loans paying; they also mean the reported delinquency rate understates how many borrowers have struggled at some point.
• EPS got a small lift from buybacks; the tax rate didn't help. Diluted shares fell 2.9% to 23.5 million (CPS bought back 231,134 shares in the quarter), so a few points of the 35% EPS gain came from fewer shares and the rest from the 30% net income increase. The effective tax rate was 31% in both years.
• Cash flow isn't a useful check here. Operating cash flow of $189.9 million for the first half dwarfs net income of $11.8 million, but that's because fair-value accounting adds back $139.7 million of non-cash interest accretion. For a lender, the meaningful cash story is funding, below.
• Growth is funded with borrowed money, and liquidity is thin. Warehouse borrowing more than doubled to $679.9 million from $324.9 million at year-end. Total debt was about $4.0 billion against $319.2 million of shareholders' equity, roughly 12.6 times. At June 30 CPS had $7.5 million of unrestricted cash and $43.4 million of undrawn warehouse capacity. That's normal for a securitization-funded lender between deals, but it leaves little cushion if bond markets close; the 10-Q says the company manages liquidity by keeping its pace of loan purchases matched to available capital.

Outlook

CPS does not give earnings guidance. The CEO's read in the earnings release is that the record volume came "without compromising our credit underwriting standards."

Since quarter-end CPS has lined up the funding it needs to keep growing: it renewed its Citibank warehouse line on July 9, 2026, sold a new securitization backed by about $734.5 million of receivables (CPS Auto Receivables Trust 2026-C) on July 22, which should pay down much of the warehouse balance, and amended its Capital One warehouse line on October 9.

Our read: the near-term earnings trajectory looks favorable. The company bought more loans than ever, it is funding them at lower rates than a year ago, and operating costs are rising more slowly than the book. The net interest spread has held at about 5.1% of the portfolio, so if credit holds, profit should keep climbing as the record 2026 volume earns interest. The risk sits on the other side: CPS grew purchases 75% in a year in a market of borrowers with weak credit, its own lifetime-loss assumption ticked up, and the 2026 loans haven't yet reached the age where subprime auto losses typically build.

What to watch in the third-quarter results (expected in early November 2026): whether quarterly charge-offs stay near 7.3% or drift back toward the first quarter's hotter pace, whether the cumulative-loss assumption moves again, whether purchases stay near the $750 million level, and how much the July securitization reduced the warehouse balance.