What Are Aging Buckets?

Aging buckets are bands that group overdue loans by how many days late they are — typically 1–30, 31–60, 61–90, and over 90 days. Each delinquent loan sits in exactly one bucket, and the buckets together add up to your total delinquent portfolio.

They exist because "we have 180,000 in arrears" is not a useful sentence. Arrears that are four days old and arrears that are four months old are different problems with different odds of recovery, and they need different responses. Aging buckets are how you tell them apart.

The standard bands

| Bucket  | Days past due | What it usually means                                                                       |
| ------- | ------------- | ------------------------------------------------------------------------------------------- |
| Current | 0             | Performing normally                                                                           |
| 1–30    | 1 to 30       | Late, mostly recoverable. Forgetfulness, cashflow timing, a payment that missed the cut-off.   |
| 31–60   | 31 to 60      | Genuine difficulty. The borrower knows they are behind.                                       |
| 61–90   | 61 to 90      | Deteriorating. Recovery rates drop sharply through this band.                                  |
| 91+     | Over 90       | Usually classified non-performing. Often headed for legal recovery or write-off.               |

Some lenders add a 1–7 or 1–14 band at the front to catch fresh delinquency early, and split the tail into 90–180 and 180+ because a loan six months late is in a very different position from one just past ninety.

The bands are conventions, not law. What matters is that they are consistent across branches, consistent over time, and that everyone reading the report knows which convention is in use.

How a loan's age is measured

Days past due are counted from the due date of the oldest unsatisfied instalment, not from the last payment received.

This trips people up. A borrower who has missed three instalments but paid something last week is still 90 days past due, because the instalment from three months ago has never been fully satisfied. Measuring from last payment date would show them at 7 days and hide the problem entirely.

Two things have to be right for this to work: the schedule of due dates, and the allocation waterfall that decides whether an instalment counts as satisfied. Get the allocation wrong and a partial payment can clear an instalment on paper that is still short in reality — which resets the loan's age and drops it out of the bucket it belongs in.

Aging buckets vs. portfolio at risk

These two get confused constantly, and mixing them in one table is a reliable way to produce a number nobody trusts.

Aging buckets are exclusive. A loan 45 days late appears in the 31–60 bucket only. The buckets sum to total delinquent portfolio.

PAR thresholds are cumulative. PAR 30 includes every loan 30 or more days late — the 31–60 loans, the 61–90 loans, and the 90+ loans together. PAR 30 ⊇ PAR 60 ⊇ PAR 90. They do not sum to anything meaningful.

The relationship between them:

PAR 30  =  (31–60 bucket) + (61–90 bucket) + (90+ bucket)
PAR 60  =  (61–90 bucket) + (90+ bucket)
PAR 90  =  (90+ bucket)

Both views come from the same underlying data. Use buckets when you want to see where the delinquency is concentrated and where it is moving. Use PAR when you want a single headline figure to report to a funder or regulator, or to compare yourself against industry benchmarks.

Label them plainly on any report that shows both. "Loans 31 to 60 days late" and "payments overdue by more than 30 days" are unambiguous. "60 days" on its own is not.

Roll rates: the number that actually predicts trouble

A single month's bucket distribution is a snapshot. The useful signal is what percentage of each bucket moves into the next one the following month. That is the roll rate.

| Month    |   1–30 |  31–60 |  61–90 |    90+ |
| -------- | -----: | -----: | -----: | -----: |
| January  | 90,000 | 40,000 | 25,000 | 60,000 |
| February | 95,000 | 55,000 | 22,000 | 70,000 |

The headline here is that total arrears grew from 215,000 to 242,000. The real story is that the 31–60 bucket jumped 38% while 61–90 fell. Loans are rolling forward out of the front bucket faster than collections can hold them — which means the January 1–30 balance was not recovered, it aged.

Rising roll rates from 1–30 into 31–60 are the earliest reliable warning that collections is losing ground. They show up a full month or two before PAR 30 moves, and two or three months before your provisioning expense does.

Low roll rates with a large 90+ bucket is a different picture entirely: the historic problem is real but contained, and the current book is behaving. That distinction is invisible in a single PAR figure.

What to do with each bucket

The point of bucketing is that each band gets a different treatment.

1–30 days. High recovery, low cost. An SMS or a phone call resolves most of it. The majority of loans in this bucket are not in trouble — they are late. The failure mode is not chasing them at all, because they don't look urgent, and letting them age into a band where recovery takes real work.

31–60 days. Field officer contact, in person where possible. Establish whether this is a timing problem or an ability-to-pay problem, because the answer determines whether you push for collection or start a restructuring conversation.

61–90 days. Formal demand, guarantor contact, group intervention if it is a group loan, collateral review. Recovery economics get thin through this band — decide whether the recovery cost is justified by the outstanding balance.

90+ days. Legal recovery, collateral realisation, or write-off. This bucket needs a decision, not more phone calls. Loans left sitting here indefinitely inflate your balance sheet with assets that are not assets.

Cut the buckets by branch, by loan officer, by product, and by group as well as at portfolio level. A book-level 1–30 bucket of 90,000 that turns out to be one officer's entire caseload is a management problem, not a credit problem.

Aging buckets and provisioning

Most lenders drive their loan loss provisioning directly off the aging buckets, applying a provision percentage to each band that rises as the loans get older. A typical matrix looks like 1% on current, 5–10% on 1–30, 25% on 31–60, 50% on 61–90, and 100% on 90+.

This makes bucket accuracy a financial reporting issue, not just an operational one. If a loan is in the wrong bucket, your provision is wrong, your reported profit is wrong, and your capital position is wrong. Regulators examining a lending book usually start by testing whether the aging is correctly calculated, precisely because everything downstream depends on it.

Where aging goes wrong

Aging measured from last payment date rather than oldest unsatisfied instalment. Understates delinquency systematically, and worst for the borrowers who are furthest behind.

Restructured loans reset to current. Rescheduling a 90-day loan and starting the clock at zero moves it from the worst bucket to the best overnight. Track restructured loans separately so the improvement is visible for what it is.

Buckets recalculated only at month-end. A loan that went 31 days late on the 3rd sits in the wrong bucket for four weeks. Aging is a daily calculation.

Posting date used instead of payment date. A payment made on the 20th but keyed on the 25th ages the loan five extra days. On a book with thousands of payments, this quietly shifts the whole distribution rightward.

Buckets that only exist in a report. If the aging lives in a spreadsheet someone rebuilds monthly, no one is acting on the 1–30 bucket — which is the only bucket where action is cheap.