How Fast Does an Unpaid Debt Lose Its Value? The Collectability Curve, Every Source (2026)

The curve that puts an unpaid business debt at about 94% collectable after 30 days, 74% after 90 days and 27% after one year is the Commercial Law League of America's own chart, the version the CLLA publishes today and one of at least five versions of one collection-industry chart; the earliest copy we found was archived in June 2002, and the two editions that describe their method (2004 and 2016) say it comes from a survey of member collection agencies. It is not a US Department of Commerce study: we searched for one and found none, and although the versions disagree on the exact figures, every one shows the chance of collecting falling at every step from the due date to two years.

This page traces every version of the chart we could find, who published it, what each one says about its own method, and how the numbers changed as they were copied from site to site. Every figure links to the document it comes from; all were fetched on 29 October 2026, and the charts from the old CCAA website link to their Internet Archive copies because the site itself is gone. We ran the same kind of source audit on lead response time, where the best-known multipliers turned out to come from a different study than the one they are usually credited to.

Where does the 94% / 74% / 27% collectability curve come from?

The numbers in circulation today match one document: the Commercial Law League of America (CLLA) chart "Probability of Recovering Commercial Debt with Age", linked from the League's resources page as "Collectability Chart PDF". Its eight bars read 98% at the due date, 94% at 30 days, 85% at 60 days, 74% at 90 days, 58% at six months, 43% at nine months, 27% at one year and 14% at two years. The axis is labelled "Probability of Collecting Money", and the only attribution on the page is "Source: Commercial Law League of America". The PDF's own metadata dates the file to 17 July 2025. It prints no survey year, no sample and no method, and the resources page adds no description.

The July 2025 date belongs to the file, not to the figures. Pacific Credit Services was already quoting the chart's 94% and 74%, credited to "data compiled by the Commercial Law League of America", in a post published in November 2020 and modified in 2021, so the figures were public by 2021 at the latest. We could not date them more precisely, so we cannot say whether they are older or newer than the Commercial Collection Agencies of America file dated 2016.

The chart belongs to a family that goes back at least to 2002. The other versions were published by the League's commercial collection agency section, the Commercial Collection Agency Association (CCAA), and later by Commercial Collection Agencies of America. The 2004 edition is explicit about where the numbers come from. Its text reads: "According to the survey results, the probability of full collection on a delinquent account drops dramatically with the length of delinquency. For example, even after only three months, the probability of collecting a delinquent account drops to 69.6%. After six months, collectability drops to 52.1%. And after one year, the probability of ever collecting a delinquent account drops to 22.8%." (CCAA, "Collection Trends", October 2004, archived copy). The survey "polled the members of the Commercial Collection Agency Association, who in aggregate handle about eighty percent of claims placed with professional commercial collection agencies in the United States". It does not say how many agencies answered, or when.

Two points of scope matter before you use any of these numbers. First, the editions that describe their sample (2004 and 2016) measure US commercial claims placed with collection agencies, not consumer debt; the CLLA chart's title says "Commercial Debt" and gives no sample. Second, every version starts its first bar at the due date, not the invoice date; the 2004 chart carries the footnote "For some industries the due date may be several months after the delivery date."

Did the US Department of Commerce publish a debt collectability study?

We found no Department of Commerce study that measures how collectable a debt is by age. The curve traces to a collection-agency survey. On 29 October 2026 we ran five targeted searches pairing "Department of Commerce" with the curve's numbers (69.6%, 93% or 94% at 30 days, 73% or 74% at 90 days) and with phrases such as "chance of collecting a past-due account". None led to a Commerce or Census Bureau study. Every percentage traced back to the CLLA family of charts.

The Commerce documents that did appear were its own debt-management rules. Those rules, 15 CFR Part 19, are about debts owed to the Department: §19.1 defines the debts the part covers, and §19.8 is titled "When will Commerce entities suspend or terminate debt collection on a Commerce debt?" Neither contains any statistic on how collectability changes with age.

If you have seen the curve credited to the US Department of Commerce, we found no such study. The closest government credit we could find is a different curve on Sage's accounts receivable statistics page, credited to the "US Census Bureau" with no link, table or year, and the Census Bureau is part of the US Department of Commerce. That version is covered below.

How fast does an unpaid debt lose its value? Every version of the chart

Here is every version of the curve we could trace to a primary document, plus four related data sets that are quoted alongside it. The first five rows are the same chart family; the last four measure something different or cannot be traced to a primary document, and are listed so you can see why they should not be merged into one curve.

Read across the first five rows and the shape is consistent even where the figures are not. At 90 days the versions range from 68.9% to 74%, so roughly a quarter to a third of the chance of collecting is gone. At one year they range from 21.4% to 28.4%, so roughly three quarters is gone. At two years they read 8.9% to 14%. Every version also puts the due date itself below 100%, between 94.9% and 98.7%. The current CLLA chart matches no other version exactly; it sits closest to the 2002 chart.gif version (72.3% at three months, 28.4% at twelve).

Who ran the survey behind the curve?

The 2004 chart states that "The Commercial Collection Agency Association was established as a separate section of the Commercial Law League of America in 1972." The association's own background page, as archived in 2008, adds: "The CCAA was established in 1972 to improve the quality and reputation of the commercial collection industry. It is part of the Commercial Law League of America (CLLA), the oldest creditors' rights organization in the country, founded in 1895." (CCAA, "Background of the CCAA", archived copy).

In May 2012 the two announced they would separate. insideARM reported on 16 May 2012 that "The Commercial Law League of America (CLLA) and the Commercial Collection Agency Association (CCAA) will become separate entities with a distinct alignment for further business opportunities on behalf of their respective members", after a board vote on 2 May 2012, and quoted Annette Waggoner as the CCAA's executive director. The later charts carry the name Commercial Collection Agencies of America and name "Annette M. Waggoner, Executive Director of Commercial Collection Agencies of America", with survey text that closely follows the 2004 wording. We found no document that says outright that one is the renamed other, so the fair description is: the agency section that later published as Commercial Collection Agencies of America. The chart in circulation now is published by the CLLA itself.

One practical consequence: the old CCAA domain, ccaacollect.com, no longer hosts the association's site (on 29 October 2026 it served an unrelated online-casino page). The 2002 and 2004 charts survive only as Internet Archive copies, and older pages that link to ccaacollect.com, such as Metcredit's 2015 post, no longer lead to the source.

Why do the numbers differ from one source to the next?

Three reasons, and the pages in the attribution trail below show all three.

1. Different editions. There are at least five versions, and the pages we traced quote one without its date. A "69.6% at 90 days" is the 2004 CCAA survey; "68.9%" is the Commercial Collection Agencies of America file dated 2016; "73.1%" is a 2002 chart; "74%" is the current CLLA chart.
2. Different measures. The 2004 and 2016 texts speak of the "probability of full collection"; the CLLA axis says "Probability of Collecting Money"; IC System says accounts are "worth" a share "of their original value"; the FTC measures the price paid per dollar of debt; Sage gives an "uncollectible" share. These are not one curve, and averaging them produces a number nobody measured.
3. Copying errors. The clearest is in the Commercial Collection Agencies of America sheet "The Shrinking Dollar", whose text reads: "After six months, collectability drops to 51.3% and after one year, the probability of ever collecting delinquent accounts drops to 8.9%." But 8.9% is the two-year figure on the association's own chart, and the same sheet's sidebar says "1 Year Overdue... $.21" and "2 Years Overdue... $.09". The one-year figure is 21.4%. At least one agency page, CST Worldwide's collection statistics page, repeats the 8.9%.

Smaller errors pile up the same way. Pacific Credit Services describes the CLLA data as running "from 94% for debts that are still current, to 74% for debts that are 90 days past due", but on the CLLA chart 94% is the 30-day bar and the due date is 98%. Remote Scouts writes that after 90 days "this probability increases to 74%", for a fall from 94%. And Sage's "26% uncollectible after 90 days" is the CLLA's 74% turned around, which Billed then turns the wrong way again: "Roughly 74% become uncollectible per Sage/Census".

How did the curve spread, and who gets the credit?

Below is the attribution trail we traced, all pages fetched on 29 October 2026. These are agency, vendor and press pages: they are listed to show how the numbers travel and who gets the credit, not as sources for the numbers themselves.

Three patterns stand out. First, the parent is usually named correctly (the CLLA, the CCAA or Commercial Collection Agencies of America), but the edition almost never is. The most recent example, NACM North Central's May 2026 article, puts it plainly: "According to CLLA, the probability of recovering accounts falls from 94% at 30 days outstanding to 74% at 90 days, 58% at six months, 43% at nine months and a paltry 27% after one year." Second, credit drifts to whoever repeated the figure last. Dun & Bradstreet's page gives the 2004 numbers and says the one-year figure is "according to one collections agency", without naming it; an April 2026 press release then opens with "According to Dun & Bradstreet research, receivables past 90 days have only a 69.6% probability of being collected". Third, the numbers mutate as they travel: no published version shows "low teens" at one year, or $0.71 at three months.

Where does the "US Census Bureau" version come from?

The nearest thing to a government credit we found is on Sage's "Top Accounts Receivable Stats" page (3 June 2025), which lists: "Aging receivables quickly become uncollectible: 26% uncollectible after 90 days 70% after 180 days 90% after one year (US Census Bureau)". Sage gives no link, table or year for the Census credit, and we could not find a Census Bureau publication with this curve.

The numbers do not hold together as one series either. 26% uncollectible at 90 days is the CLLA's 74% collectable turned around, but 70% at 180 days and 90% at one year match no CLLA, CCAA or Commercial Collection Agencies of America chart, all of which put one-year collectability between 21.4% and 28.4%, not 10%. Billed's May 2026 statistics page repeats the Sage line ("per the U.S. Census-derived curve cited by Sage, 26% of receivables become uncollectible after 90 days past due") and then, in its aging table, inverts it to "Roughly 74% become uncollectible per Sage/Census". Confidence in the Sage curve: Low, because its source cannot be traced.

What do collection agencies' own data show?

At least one collection agency publishes a curve from its own data that looks nothing like the CLLA family. IC System, a US collection agency, wrote in January 2022: "Our research on millions of debts shows that past-due accounts are worth just 87% of their original value after just 90 days past due." and "After 120 days, they’re worth only 33% of their initial value". A 2021 post gives the same 90-day figure: "On average, past-due accounts are worth 87% of their original value after 90 days of delinquency."

Two cautions. The measure is different ("worth X% of original value", not a probability of collection), and IC System does not describe its method or say whether the debts are consumer or commercial. And the drop it reports between 90 and 120 days, from 87% to 33%, is far steeper than anything on a CLLA chart. The published material does not let anyone reconcile the two, so we do not try. Confidence: Medium (vendor data, method undisclosed).

What does the FTC debt-buyer study add?

A regulator's data set points the same way, in a different currency. The US Federal Trade Commission's January 2013 report "The Structure and Practices of the Debt Buying Industry" studied 9 of the largest debt buyers: "more than 5,000 portfolios, containing nearly 90 million consumer accounts" with a face value of $143 billion, bought between 1 July 2006 and 30 June 2009. Its regression found that buyers paid on average 7.9 cents per dollar for baseline debt (a credit card debt less than 3 years old, bought from the original creditor, under a thousand dollars, never sent to a contingency collector), 3.1 cents for debts 3 to 6 years old, 2.2 cents for debts 6 to 15 years old, and that "debt buyers paid effectively nothing for accounts that were older than fifteen years."

The FTC's explanation is the useful part: "Debt buyers presumably pay less for older debts because their expected return from collecting on those debts is lower, likely reflecting the fact that the consumers may be less willing or able to pay the debt or the consumers may be more difficult for debt buyers to locate." These are charged-off consumer debts priced in years, not commercial invoices in months, so they are not a point on the same curve. They are independent confirmation of the direction: older debt is worth less, partly because the people who owe it get harder to find. Confidence: High (primary, regulator).

Which collectability figure should you cite?

If you need one line, cite the document and say what it does not disclose: "The Commercial Law League of America's collectability chart puts the probability of collecting a commercial debt at 74% at 90 days and 27% at one year; the chart gives no survey year or method." If you want a version that states its method, use the 2004 CCAA survey (69.6% at 90 days, 22.8% at one year) or the Commercial Collection Agencies of America file (68.9% and 21.4%), and say that neither gives a response count. Do not credit any of them to the Department of Commerce, to Dun & Bradstreet, or to "an annual survey", and do not blend them with IC System's value figures or the FTC's prices.

When should you contact a debtor?

Every published version, from the 2002 charts to the current CLLA chart, shows the chance of collecting falling at every step from the due date to two years. IC System's own data shows a steep drop between 90 and 120 days. The FTC suggests older debt sells for less partly because people "may be more difficult for debt buyers to locate". None of these sources tests a particular contact schedule, so the advice below is our practical reading of the data, not a sourced percentage:

1. Make first contact at or before the due date. The charts start their clock at the due date, and every version already shows the due date below 100%.
2. Follow a fixed schedule through the first 90 days, for every account. Not only the large balances. Debtor segmentation decides what each account gets, not whether it gets contacted.
3. Decide on escalation before the account crosses 90 days. Whether to keep working an account in-house or place it with an agency is a decision to make early; our guide to first-party versus third-party collections covers the trade-offs.

Knowing this is the easy part. Doing it is a capacity problem: early, systematic contact means every account, on schedule, from day one, which is volume work. An AI voice agent for debt collection can call every account on that schedule without adding collectors, while your people take the disputes, hardship cases and payment plans that need judgement. Reaching the right person matters as much as reaching them early, which is what right-party contact verification handles, and voice is one channel among several: how calls, SMS and email work together is in our omnichannel collections guide. For why collections suits voice agents at all, see why debt collection is a strong use case for AI voice agents.

One limit to keep in mind: the collectability charts measure the effect of time, not of any channel, so they say nothing about how much any tool adds to recovery, and we make no such claim. To build your own business case, use our AI debt collection ROI methodology with your own portfolio data, and see the cost factors for AI debt collection; AInora is priced individually, after a call. To compare platforms, see our AI collections software comparison and AI platforms for collection agencies compared. All the charts above come from US bodies; for Europe, see the EU Late Payment Directive and AI collections and our report on what is actually published about EU debt collection.

If you want to see what contacting every account on schedule would look like for your book, book a call on our contact page.