${r.comparison_window} ${year}: ${fmt(v)} charges

2d:null 2e:["$","title",null,{"children":"Q3 charge counts fell. The first nine months rose. | Matt Lenzie | Observable"}] 33:T466,# Q3 charge counts fell. The first nine months rose. **Matt Lenzie | Construction Capital Research | 7 October 2026** A quarterly comparison and a year-to-date comparison can point in different directions without contradicting each other. In Construction Capital's curated specialist development cohort, recorded charges fell from **210 to 207** between Q3 2025 and provisional Q3 2026, a decline of **1.4%**. Across January to September, the count rose from **645 to 681**, an increase of **5.6%**. The reason is visible in the first quarter: **66 additional charge records** outweighed the reductions of **27 in Q2** and **three in Q3**. The result describes when recorded activity appears in this frozen dataset. It does not establish that lenders supplied more money or that developers secured more new facilities. Use the controls to compare like-for-like calendar windows. The wider property-funding cohort is available as a separate view; it overlaps the specialist cohort and the two must not be added together. All calculations use the existing [public charge study and data](https://doi.org/10.7910/DVN/EK56EQ). 34:T26f7,// MIT licence, Matt Lenzie / Construction Capital, 2026. Source: public aggregate release. const chargeTimingData = [{"cohort": "specialist", "quarter": "2023-Q2", "charges": 249}, {"cohort": "specialist", "quarter": "2023-Q3", "charges": 213}, {"cohort": "specialist", "quarter": "2023-Q4", "charges": 257}, {"cohort": "specialist", "quarter": "2024-Q1", "charges": 176}, {"cohort": "specialist", "quarter": "2024-Q2", "charges": 249}, {"cohort": "specialist", "quarter": "2024-Q3", "charges": 237}, {"cohort": "specialist", "quarter": "2024-Q4", "charges": 276}, {"cohort": "specialist", "quarter": "2025-Q1", "charges": 176}, {"cohort": "specialist", "quarter": "2025-Q2", "charges": 259}, {"cohort": "specialist", "quarter": "2025-Q3", "charges": 210}, {"cohort": "specialist", "quarter": "2025-Q4", "charges": 267}, {"cohort": "specialist", "quarter": "2026-Q1", "charges": 242}, {"cohort": "specialist", "quarter": "2026-Q2", "charges": 232}, {"cohort": "specialist", "quarter": "2026-Q3", "charges": 207}, {"cohort": "ecosystem", "quarter": "2023-Q2", "charges": 9344}, {"cohort": "ecosystem", "quarter": "2023-Q3", "charges": 10224}, {"cohort": "ecosystem", "quarter": "2023-Q4", "charges": 10530}, {"cohort": "ecosystem", "quarter": "2024-Q1", "charges": 10943}, {"cohort": "ecosystem", "quarter": "2024-Q2", "charges": 12555}, {"cohort": "ecosystem", "quarter": "2024-Q3", "charges": 13324}, {"cohort": "ecosystem", "quarter": "2024-Q4", "charges": 14942}, {"cohort": "ecosystem", "quarter": "2025-Q1", "charges": 14916}, {"cohort": "ecosystem", "quarter": "2025-Q2", "charges": 14009}, {"cohort": "ecosystem", "quarter": "2025-Q3", "charges": 15784}, {"cohort": "ecosystem", "quarter": "2025-Q4", "charges": 16427}, {"cohort": "ecosystem", "quarter": "2026-Q1", "charges": 16779}, {"cohort": "ecosystem", "quarter": "2026-Q2", "charges": 17115}, {"cohort": "ecosystem", "quarter": "2026-Q3", "charges": 15575}]; function chargeTimingExplorer(data) { const root = document.createElement('section'); root.className = 'cc-charge-timing'; root.innerHTML = `

Cohort Comparison window

20252026

Check the windows side by side

The nine-month row is a subtotal of Q1, Q2 and Q3. Do not add it to those quarter rows. Swipe the table on a narrow screen.

Source: Construction Capital, Harvard Dataverse study CC-DF-2026Q3-02, version 1. Q3 2026 provisional. Cohorts overlap and must not be added together. Charge counts are not lending amounts, new facilities or unique developments.

`; const $ = s=>root.querySelector(s); const cohort = $('select[aria-label="Cohort"]'); const windowSelect = $('select[aria-label="Comparison window"]'); const lookup = new Map(data.map(d=>[d.cohort+':'+d.quarter,d.charges])); const windows = {'Q1':[1],'Q2':[2],'Q3':[3],'January to September':[1,2,3]}; const fmt=n=>n.toLocaleString('en-GB'); const sign=n=>(n>0?'+':'')+n.toFixed(1); let current=[]; function render(){ const c=cohort.value,selected=windowSelect.value; current=Object.entries(windows).map(([label,quarters])=>{ const a=quarters.reduce((s,q)=>s+lookup.get(c+':2025-Q'+q),0); const b=quarters.reduce((s,q)=>s+lookup.get(c+':2026-Q'+q),0); return {cohort:c,comparison_window:label,charges_2025:a,charges_2026:b,change:b-a,change_pct:100*(b/a-1)}; }); const d=current.find(d=>d.comparison_window===selected); $('.summary').textContent=`${selected}: ${fmt(d.charges_2025)} charges in 2025 and ${fmt(d.charges_2026)} in 2026. Change: ${d.change>0?'+':''}${fmt(d.change)} (${sign(d.change_pct)}%).${windows[selected].includes(3)?' Includes provisional Q3.':''}`; $('.charttitle').textContent=c==='specialist'?'Three quarters explain the nine-month result':'Wider cohort: the same calendar comparison'; $('.chartnote').textContent='Registered-charge counts. Every bar starts at zero. The selected comparison window is highlighted.'; const width=Math.max(340,Math.min(root.clientWidth||800,900)),height=315,left=58,right=20,top=32,bottom=265; const max=Math.ceil(Math.max(...current.slice(0,3).flatMap(d=>[d.charges_2025,d.charges_2026]))/(c==='specialist'?50:5000))*(c==='specialist'?50:5000); const y=v=>bottom-v/max*(bottom-top),group=(width-left-right)/3,bar=Math.min(45,group*.27); let svg=''; for(let v=0;v<=max;v+=(c==='specialist'?50:5000)){ svg+=`${c==='specialist'?Math.round(v):Math.round(v/1000)+'k'}`; } current.slice(0,3).forEach((r,i)=>{ const cx=left+group*(i+.5),isSelected=windows[selected].includes(i+1); if(isSelected)svg+=``; [[r.charges_2025,'#8997a5',-bar-2,2025],[r.charges_2026,'#147b83',2,2026]].forEach(([v,color,offset,year])=>{ const x=cx+offset;svg+=`${r.comparison_window} ${year}: ${fmt(v)} charges${width<450&&c==='ecosystem'?(v/1000).toFixed(1)+'k':fmt(v)}`; }); svg+=`${r.comparison_window}${i===2?'*':''}`; }); const alt=current.slice(0,3).map(r=>`${r.comparison_window}: ${r.charges_2025} charges in 2025 and ${r.charges_2026} in 2026`).join('; '); $('.chart').innerHTML=``; const ys=current[3]; $('.arithmetic').textContent=`Quarterly differences: ${current.slice(0,3).map(r=>(r.change>0?'+':'')+fmt(r.change)).join(' ')} = ${ys.change>0?'+':''}${fmt(ys.change)} charges across January to September. *Q3 2026 is provisional.`; $('.tablewrap').innerHTML=`${current.map(r=>``).join('')}

${cohort.selectedOptions[0].textContent}: 2025 versus 2026
Window	2025	2026	Difference	Change %
${r.comparison_window}	${fmt(r.charges_2025)}	${fmt(r.charges_2026)}	${r.change>0?'+':''}${fmt(r.change)}	${sign(r.change_pct)}%

`; } cohort.addEventListener('change',render);windowSelect.addEventListener('change',render); $('button').addEventListener('click',()=>{ const keys=Object.keys(current[0]);const csv=[keys.join(','),...current.map(r=>keys.map(k=>JSON.stringify(r[k])).join(','))].join('\n'); const url=URL.createObjectURL(new Blob([csv],{type:'text/csv;charset=utf-8'}));const a=document.createElement('a'); a.href=url;a.download=`construction-capital-${cohort.value}-charge-windows-2025-2026.csv`;a.click();setTimeout(()=>URL.revokeObjectURL(url),1000); }); render();new ResizeObserver(render).observe(root);return root; } display(chargeTimingExplorer(chargeTimingData)); 35:T10e2,## Why the nine-month total changes the story The specialist cohort's Q1 count rose from 176 to 242. Q2 fell from 259 to 232, and Q3 fell from 210 to 207. Adding the differences gives **+66 − 27 − 3 = +36**. Dividing 36 by the 2025 nine-month total of 645 gives 5.58%, rounded to 5.6%. That nine-month increase does not mean activity was strengthening at the end of the period. Nor does the Q3 decline establish that the whole year was down. There are no Q4 2026 observations here. The calendar window belongs in the headline, together with the measure and the provisional status. The wider property-funding cohort illustrates a different path: Q1 adds 1,863 charges, Q2 adds 3,106, and Q3 subtracts 209. Its January-to-September count rises from 44,709 to 49,469, or 10.6%, while its Q3 count falls 1.3% year on year. This broader cohort is not a total for the Development Finance market and is not a control group for the specialist series. ## What can be added, and what cannot The quarter charge counts can be added for the nine-month arithmetic in this release because they refer to separate calendar-quarter bins. The January-to-September row is their subtotal, not a fourth period to add again. The two overlapping cohorts must remain separate. Quarterly borrower counts require different treatment. A company may appear in more than one quarter, so adding quarterly borrower counts would not produce a count of unique nine-month borrowers. This notebook deliberately does not do that. The [charge-intensity briefing](https://view.genially.com/6ac51adb6d35187f30ef0d24) examines charge counts and borrower counts within quarters. ## Using the result in a Development Finance discussion The [Q3 Development Lending Monitor](https://constructioncapital.co.uk/market-reports/development-lending-monitor-q3-2026-review) provides the broader reporting context. When referring to these findings, specify “recorded charge counts in the curated specialist cohort”, then state the period. A charge registration is not a loan amount or necessarily a new loan. These aggregates do not measure pricing, loan-to-cost, available facilities or funding appetite. Construction Capital's [Development Finance guide](https://constructioncapital.co.uk/services/development-finance/) describes the separate funding process. An actual proposal still needs its own scheme evidence and proposed terms. Planning, charge and sales aggregates cannot trace the same development through consent, finance and disposal. The [planning-denominator article](https://constructioncapitaluk.blogspot.com/2026/10/blog-post.html) explains a related problem with comparing unmatched quarterly flows. ## Source, reproduction and limits [Download the backing package](https://drive.google.com/file/d/10r2_tB8IXu9p7WUL_LvLmc2rlyR40V3L/view) for the 28-row source table, eight window comparisons, source-count JSON, analysis, Python reproduction script, notebook code, dictionary, provenance and licence. Run `python3 reproduce.py` to rebuild the derived tables using the Python standard library. The notebook computes its displayed values from the same aggregate counts and lets you download the selected cohort's comparisons. The source is Construction Capital's Companies House-derived aggregate release, frozen at repository commit `44ec7821f8bac827de198d09b4cca66725eae055`, and published as [Harvard Dataverse study CC-DF-2026Q3-02](https://doi.org/10.7910/DVN/EK56EQ). Reproduction starts from those aggregates. It cannot rebuild the private lender mapping, deduplication or upstream extraction. Source metadata records `snapshotDate` as 1 October 2026 and `generatedAt` as 3 October; the maximum charge date is 2 October and maximum satisfied date is 3 October. These are distinct fields, not one settled cutoff. Q3 is provisional and later filings can change the comparison. No seasonal adjustment, causal inference or lending-volume estimate is attempted. Construction Capital is a property-finance brokerage. This is an original timing adaptation of its existing study, not a new independent dataset, peer review or platform endorsement. No lender names, borrower identities or individual charge records are disclosed. Original text and analysis are CC BY 4.0; code is MIT; underlying source rights remain with their holders. 36:[] 31:{"id":"ce064f095c9c63ca","slug":"q3-charge-counts-fell-the-first-nine-months-rose","title":"Q3 charge counts fell. The first nine months rose.","thumbnail":"e190acd632b812cd0d94f17a02f66fdf000a18ea71a38043b0e3ed832fcacce9","access_level":"public","create_time":"$D2026-10-07T16:18:00.185Z","update_time":"$D2026-10-07T16:19:12.070Z","view_time":null,"first_public_time":"$D2026-10-07T16:19:20.759Z","has_unpublished_edits":null,"fork":false,"stars":0,"starred":false,"owner":{"type":"team","id":"ea2de9ba2c0a2ed6","login":"matt-lenzie","name":"Matt Lenzie","avatar_url":null,"tier":"free_2026","bio":""},"version":10,"publish_version":10,"role":null,"authored":false,"authors":[{"type":"individual","id":"2a28c9706791cd4f","login":"matt-lenzie","name":"Matt Lenzie","avatar_url":null,"tier":null,"bio":""}],"members":[{"type":"team","id":"ea2de9ba2c0a2ed6","login":"matt-lenzie","name":"Matt Lenzie","avatar_url":null,"tier":"free_2026","role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"titling":"auto","title":"Q3 charge counts fell. The first nine months rose.","cells":[{"id":1,"value":"$33","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2,"value":"$34","pinned":false,"hidden":false,"mode":"ts","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":3,"value":"$35","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W36","stdlib":"2"}} 32:"$31:body"