Every week, thousands of UK companies quietly register new mortgage charges at Companies House — creating a paper trail that most due diligence workflows barely scratch. BORSCH.AI has processed 2,493,063 signals from the Companies House Mortgages dataset (Prod199 Bulk) across 803,752 company records. That works out to an average of 3.1 mortgage charge signals per indebted company — suggesting that for many businesses, secured borrowing is not a one-off event but a recurring financing strategy.
Here’s what that data reveals about the UK business landscape, and why mortgage charge patterns deserve far more attention than they typically get.
The Scale of Corporate Secured Debt in the UK
The Companies House mortgage register — formally the register of charges — is legally required to record every charge a company creates over its assets. That includes fixed and floating charges, debentures, and of course property mortgages. Failure to register within 21 days renders the charge void against a liquidator, which makes this one of the more reliably complete public datasets available.
| Data Source | Signals | Records Fetched | Avg. Signals/Record |
|---|---|---|---|
| Companies House Mortgages (Prod199 Bulk) | 2,493,063 | 803,752 | 3.1 |
| Companies House Charges API (ch_charges) | 32,472 | 129 | — |
| HM Land Registry | 687,373 | — | — |
The gap between the bulk dataset (2,493,063 signals) and the API-based charges feed (32,472 signals) is stark. Analysts relying solely on the Companies House API for charge data are working with a fraction of what’s actually registered. The bulk Prod199 file — the same one BORSCH.AI ingests — is where the complete picture lives.
Out of 16,229,746 companies tracked on the BORSCH.AI platform, 803,752 have at least one mortgage signal. That’s approximately 1 in 20 UK registered companies carrying some form of registered charge — a proportion that rises sharply when you filter to active trading companies with real assets.
Mortgage Signals as a Share of All Financial Intelligence
Mortgage charges don’t exist in isolation. They sit within a broader financial intelligence picture that includes filed accounts, payment practices, and land registry data.
| Signal Category | Total Signals | % of All Signals |
|---|---|---|
| Governance | 32,866,570 | 67.4% |
| Financial | 7,712,132 | 15.8% |
| Risk | 3,728,920 | 7.6% |
| Compliance | 3,428,162 | 7.0% |
| Trade | 405,220 | 0.8% |
| Other categories | 634,615 | 1.3% |
Within the financial category (7,712,132 signals total), the Companies House mortgage dataset alone accounts for 32.3% — making it the single largest source of financial signals on the platform. Cross-referencing these 2.5M mortgage signals with the 687,373 HM Land Registry signals and 5,507,247 Companies House XBRL accounts signals gives analysts a three-dimensional view of a company’s secured asset position that no single source can provide.
Why the 3.1 Average Matters for Risk Assessment
The figure of 3.1 mortgage charge signals per indebted company is worth unpacking. A company with a single charge — typically a floating charge held by its main bank — looks very different from one with five or six registered charges across multiple lenders and asset classes.
Multiple charges signal several things worth investigating:
Layered financing structures. Property developers, asset-heavy manufacturers, and hospitality groups routinely stack charges across different creditors, with priority hierarchies that matter enormously in an insolvency scenario.
Refinancing activity. Satisfied and replaced charges leave traces in the data. A company showing a high signal count relative to current active charges may have refinanced several times — worth exploring whether that reflects opportunistic treasury management or repeated covenant stress.
Cross-collateralisation. Where the same underlying asset secures multiple obligations, the effective recovery position for each creditor is materially different from what any individual charge document suggests.
The BORSCH.AI platform captures all of this because it ingests the full historical bulk file, not just currently active charges returned by a point-in-time API call. That distinction matters for understanding a company’s entire borrowing history, not just its current snapshot.
What Free Sources Don’t Show You
Anyone can search a single company’s charges on the Companies House website. What that gives you is a list of charges for that one company, one at a time, with no ability to query across the population.
The analytical value emerges from the aggregate. With 2.5M signals indexed:
- Pattern detection across sectors: which SIC codes show the highest charge density per company?
- Lender exposure mapping: which charge holders appear most frequently across the 803,752 indebted companies?
- Temporal clustering: are new charges being registered at an accelerating or decelerating rate in specific industries?
- Insolvency correlation: cross-referencing with the 14,405 liquidation signals and 327,872 Gazette notices reveals which charge configurations precede formal insolvency proceedings.
The Gazette dataset (327,872 signals) and the liquidation dataset (14,405 signals) sit on the same platform as the mortgage data, which means these correlations can be computed rather than guessed.
A compliance officer screening a potential counterparty manually would check Companies House for active charges, perhaps run a Land Registry search, and call it done. That process misses the historical charge pattern, the cross-source correlations, and any indication of whether the company’s charge profile resembles others that subsequently failed.
Practical Applications for Due Diligence and Credit Risk
For lenders and credit analysts: The 3.1 signals-per-company average provides a baseline. A company presenting for new facilities with six or seven existing charges warrants deeper investigation into charge priority and asset coverage ratios, particularly when cross-referenced against filed accounts (5,507,247 XBRL signals on the platform).
For M&A and private equity teams: Acquisition targets with dense charge histories may have encumbered assets that don’t appear cleanly on a balance sheet. Understanding whether charges are satisfied or outstanding — and whether the asset base has been pledged multiple times — is foundational to any asset-based valuation.
For insolvency practitioners and restructuring advisors: The combination of mortgage signals, Gazette notices (327,872), and liquidation data (14,405) creates an early-warning layer. Companies entering distress typically show a distinctive pattern of new charges being registered in the 12–18 months before formal proceedings.
For compliance officers: HMRC AML supervision data (27,535 signals) combined with mortgage charge patterns can flag structures where property assets are being used in ways inconsistent with declared business activity — relevant for both AML and beneficial ownership analysis.
The Data Infrastructure Behind the Analysis
The mortgage dataset is updated daily via the Prod201 bulk SFTP feed (ch_mortgages_daily), meaning new charge registrations flow into the platform within 24 hours of Companies House processing them. The full historical bulk load (Prod199) delivered 803,752 matched records with a 100% match rate — every record fetched was matched to a company profile, reflecting the quality of the underlying source data.
This sits alongside 48,775,619 total signals across 53 government and public data sources — a combination that makes it possible to see mortgage activity not in isolation, but in the context of a company’s officers, PSC structure, accounts, payment practices, and regulatory standing simultaneously.
The Takeaway
Mortgage charge data is one of the most information-dense public datasets in the UK corporate registry — but its value is almost entirely unrealised when accessed through standard company searches. The 2,493,063 signals in the BORSCH.AI platform represent a complete, historically continuous, cross-referenced view of secured corporate borrowing that no manual process can replicate at scale.
If your current due diligence workflow treats charge searches as a checkbox rather than an analytical input, you’re leaving a significant portion of the risk picture unread.
Explore the full mortgage charge dataset and cross-source correlations at borsch.ai.
Disclaimers
Disclaimer: This article was generated with AI assistance using data from Borsch.AI’s aggregation of 53 UK government sources. While all statistics are derived from real data, analysis and interpretation are AI-generated and should be independently verified.
Disclaimer: Data presented reflects information available at the time of publication and may not reflect the most current state. Source data is aggregated from public government registers which may contain delays, errors, or omissions.
Comments