£12bn is real. The 57% comparison covers two years
REPRODUCE THIS INVESTIGATION

Follow the original wording, identify the years being compared and repeat the arithmetic. Then check the discrepancy that limits how confidently the response versions can be combined.
You will need: A browser and a calculator, or Excel or another spreadsheet app. The optional Python file requires Python 3 and no extra libraries.

ACCESS AND DATA LIMITS
This is a reconstruction from published cells, not raw claimant data. The 2023 original response has not been recovered. The older 2024 response pages have cropped headers. Both DWP reproductions are requester-hosted. Their overlapping 2025 values differ, so the comparison is not a fully restated same-extraction series. If Instagram or the original article is inaccessible, mark that step as not independently repeated.

WORKING FILES
Source notes: https://tallycat01.com/downloads/yusuf-universal-credit-57-percent-source-notes.txt
Timeline: https://tallycat01.com/downloads/yusuf-universal-credit-57-percent-timeline.csv

1. Record the exact claim
Open the original Instagram post. Separate the caption from the image. Write down the sentence assigning 57% to “the year before”, and note that the £7.7bn sentence in the caption describes households.
Keep in mind: An inaccessible post cannot be treated as independently inspected. Our preserved check recorded publication on 20 September 2026.
Open Yusuf’s post: https://www.instagram.com/zia.yusuf/p/Ddhb3x_hau6/

2. Find the comparison periods
Open the September PDF at pages 2 and 3. Find the comparisons with 2023 and 2024. Record the year beside each percentage before looking at the size of the number.
Keep in mind: The report’s 57% period is two years. Its approximately 25% comparison uses the previous year.
Inspect the comparison on page 2: https://centreformigrationcontrol.com/airo-assets/uploads/research-pdfs/1789944912332-e2d35adf5328.pdf#page=2
Inspect page 3: https://centreformigrationcontrol.com/airo-assets/uploads/research-pdfs/1789944912332-e2d35adf5328.pdf#page=3

3. Verify the household spending cells
In the same PDF, inspect page 8 for the 2025 total and pages 7, 10 and 11 for overlap and definitions. Then open the older response’s second image and find £9,533m for 2024.
Keep in mind: These are DWP response reproductions hosted by the requester. Preserve the older page’s missing header and date limitations.
Inspect the 2025 spending table: https://centreformigrationcontrol.com/airo-assets/uploads/research-pdfs/1789944912332-e2d35adf5328.pdf#page=8
Inspect the older overall table: https://substack-post-media.s3.amazonaws.com/public/images/3b8a3f55-b146-4478-90b9-bb9ed8b1a663_496x522.png

4. Calculate the one-year comparison
Download the published-inputs CSV. Keep the rows whose input_id values are annual_2024_older_response and annual_2025_september_response. Both are in £m. Divide 11,920 by 9,533, subtract 1 and multiply by 100. Expect 25.0393%, usually displayed as 25.0%.
=(11920/9533-1)*100
Keep in mind: This repeats the published-version comparison. It does not remove the extraction discrepancy checked below.

5. Calculate the comparison that produces 57%
Keep the annual_2023_cmc_reported and annual_2025_september_response rows. Divide 11,920 by 7,600, subtract 1 and multiply by 100. Expect 56.8421%, rounded to 57%. The years are 2023 and 2025.
=(11920/7600-1)*100
Keep in mind: The £7.6bn input is reported by CMC. Its original 2023 response was not recovered, so this reproduces the report’s arithmetic rather than authenticating that original response.

6. Check the same-period discrepancy
Compare £5,587m for January to June 2025 in the older second page with £5,861m obtained by adding the six months in the later table. Subtract the earlier total. Expect £274m. Divide by £5,587m for about 4.9%.
5861-5587 = 274 million pounds
=(5861/5587-1)*100 = approximately 4.9%
Keep in mind: Rounding alone is too small to explain the difference. The cause is not established. Do not silently merge these versions into a matched trend.
Open the older January to June total: https://substack-post-media.s3.amazonaws.com/public/images/3b8a3f55-b146-4478-90b9-bb9ed8b1a663_496x522.png
Open the later monthly table: https://centreformigrationcontrol.com/airo-assets/uploads/research-pdfs/1789944912332-e2d35adf5328.pdf#page=8

7. Keep households separate from individuals
Read the official parliamentary answer and the later PDF’s notes. Check what the source says about joint awards, individual allocation and overlapping work categories. Compare that with the image and with the more qualified caption.
Keep in mind: The precise later nationality/work-status join is not supplied. Do not invent which household member must be out of work.
Read answer 74001: https://questions-statements.parliament.uk/written-questions/detail/2025-09-01/74001/
Read the later definitions: https://centreformigrationcontrol.com/airo-assets/uploads/research-pdfs/1789944912332-e2d35adf5328.pdf#page=10

8. Write the narrow conclusion
Compare three statements: £12bn is supported as rounding; the report’s 57% comparison spans two years; the exact restated one-year administrative growth remains unresolved. Keep intent and causes in your unknown column.
Keep in mind: No matched tax data, all-benefits total or causal decomposition is provided by this spending table.

ADDITIONAL WORKING FILES
Published numerical inputs and provenance · CSV: https://tallycat01.com/downloads/yusuf-universal-credit-57-percent-published-inputs.csv
Reproduced calculations and conditions · JSON: https://tallycat01.com/downloads/yusuf-universal-credit-57-percent-calculations.json
Recalculate the published inputs · Python: https://tallycat01.com/downloads/yusuf-universal-credit-57-percent-recalculate.py

Evidence checked: 10 October 2026
Research and production assisted by AI