This estimator exists for one very specific situation. You are looking at a UK SME book and you want a defensible sense of where unsecured loan arrears and losses probably sit in the current environment, and how those numbers move under different macro and sector mixes.

You do not have a paid data feed or a full internal model. You do have a working brain and a healthy suspicion of numbers that appear without a source.

The estimator takes the main public signals we have about UK SME credit conditions and structures them into something you can actually use: an estimate of arrears, defaults and losses that lines up with the regime the UK is in, and that can be stressed in a controlled way.

It is not a live arrears dashboard. It is a structured way to turn the public record into a usable arrears and loss picture.

The Data That Underpins It

Everything in the tool comes from sources you can point at.

We start with the government loan schemes. HM Treasury and the British Business Bank publish detailed repayment data for BBLS, CBILS and CLBILS, and for the later Recovery Loan Scheme. These tables tell you, by scheme, how much was drawn, how much is in arrears, how much has defaulted and how that has evolved over time. The estimator uses these numbers directly for the "scheme" portion of any portfolio. They give you real arrears rates and real default-from-arrears ratios on tens of billions of SME lending.

Next we bring in the UK company insolvency series. The Insolvency Service publishes annual and quarterly numbers for company insolvencies, including breakdowns by industry. Those tables show which sectors consistently fail more often than others. Construction, hospitality and retail feature heavily at the top, while professional services, health and utilities are further down. From this we derive sector multipliers. High-insolvency sectors carry higher multipliers. Sectors with lower insolvency rates carry lower ones. This makes the model sensitive to sector mix rather than treating a restaurant and a firm of auditors as the same risk.

The Bank of England Financial Stability Report and stress testing material provide the macro frame. The FSR gives the baseline view for GDP and unemployment and defines what "mild" and "severe" scenarios look like in terms of macro shock and system-wide credit losses. The estimator calibrates its macro multipliers so that a severe scenario means roughly "BoE-style severe" and a mild one sits partway between baseline and that point.

We then add international SME NPL benchmarks from sources such as the OECD SME Scoreboard and public bank disclosures, and cross-check against UK SME arrears work from credit bureaux such as Experian. These show where SME non-performing loans and arrears sit as a percentage of SME lending in normal conditions and in stress across advanced economies.

Finally, we look at structural SME behaviour by size, sector and product. These multipliers are calibrated rather than measured. They sit on top of the harder anchors described above and are documented as judgement, not hidden as fact.

As a rule of thumb, if a number can be read directly from an official table, it is treated as measured. If it is derived from several tables with simple arithmetic, it is calibrated. If it is there because it has to be assumed, it is labelled as judgement.

How The Estimator Works

The model follows a simple sequence: base arrears → macro stress → defaults → losses.

For each segment, the estimator first sets a base arrears rate. For scheme-type lending it uses the arrears rate published for the scheme. For other unsecured SME lending it starts from a baseline arrears level (set by the Non‑Scheme Arrears slider) and adjusts using size, sector and product multipliers.

This base arrears rate is then passed through a macro scenario. Each scenario has a multiplier. Baseline leaves things as they are. Mild and severe use multipliers chosen so their severity corresponds to the Bank of England's view of a mild downturn and a severe stress.

The next stage converts arrears into actual defaults. For the schemes this uses the default-from-arrears ratios observed in the repayment data. For the rest of the book it uses a scenario-level base conversion rate, then adjusts for size and sector.

Finally, Loss Given Default is applied. At borrower level the estimator assumes a central unsecured LGD. At bank level it applies the guarantee percentages to show how much of the loss is effectively underwritten by government.

The tool also offers a trade credit view for those working with receivables rather than loans. It combines the portfolio default rate and the sector insolvency rate into a simple probability of distress for customers.

Why This Is Useful

From a user's point of view, the individual pieces of public data are a mess. You can see government scheme performance over time, you can see insolvency rates, you can read the FSR, you can dig out SME NPL ratios and arrears work from Experian and others. It takes time and effort to turn that into a coherent answer.

This estimator does that structuring work. It gives you something you can interrogate and amend instead of a pile of PDFs.

If you are a lender, it gives you a check on whether your internal arrears assumptions are in the same postcode as what public data would imply. If you work in trade credit, it gives you a reasoned way to connect the loan side of the system to the receivables side.

The model will never tell you exactly what your book's arrears were last month. That is your MI's job. What it can do is give you a defended range and a clear link between the external environment and the shape of your arrears and loss profile.