Sampling is the audit-style approach the Claim funding for mentor training (Claim) team will take to ensure the grant funding is being claimed and used correctly.
We must sample a minimum of 10% of the total claim value. This sample needs to be fair and agnostic regarding school demographics. Due to the funding amount and the level of post-claim assurance we have implemented, the sampling will be completed after payment, which, for the Private Beta, we expect to commence in October 2024.
Sampling will involve working with the accredited ITT providers to validate the claims for which they are named. In some instances, further clarification may be required.
For individual claims, where fraud is potentially detected, the claim may be subject to internal clawback processes. This is where the Department recuperates the claimed funds.
For the sampling workflow, depending on the nature and level of the fraud, we would need to increase the sample size and re-run the process, excluding those claims that have already been assured.
How it works
Identifying the sample
We will identify the sample size from the paid claims. The sample will calculate a set percentage of the total value of claims paid.
The minimum sample size is 10%. However, we are considering increasing this to account for the new service and this being the first claim window.
We have highlighted the need to adapt the sample size that will be identified easily. This will give us the flexibility to increase the size if we want to, but crucially the ability to increase should we detect fraud within the initial sample.
Identification will be managed via an algorithm we specify and working with colleagues in the Data and Insights (D&I) team to develop and test. Some of the key considerations include:
- the ability to increase the sample size
- ensuring sample identification is agnostic of school demographics
- ensuring sample identification is fair, transparent, and representative
- a larger proportion of claims submitted by a mentor who is part of the claim are sampled
- rules to prevent undue re-identification from a range of scenarios
Having identified the sample, the support team will change the status of those claims to indicate that they require evidence. This will result in an email sent to the schools to notify them that their provider is being asked for evidence, and they may need to support this process.
Creating the artefacts
After identifying the sample, the support team will be notified via email and Slack. The team will then use the export feature to refine the claim result and create the necessary exports, which will be downloaded as Excel CSV files.
Adobe Sign will be used to create an electronic declaration. This declaration will require accredited providers to confirm the claims listed against them are valid. The extracted Excel files will be attached to the declaration and sent to the accredited providers for their review and electronic signature.
Working with accredited providers
After receiving notification of the intended sampling, the accredited provider can access the declaration and annexe data extract. They will review the extract and confirm its correctness by electronically signing the declaration.
The accredited provider may need to contact the school to help verify some details. This process will be managed solely by the accredited provider.
We have elected to work directly with accredited providers for this process because:
- the declaration needs to be independent of the school claiming the grant
- any supporting evidence will also need to be independent of the school claiming the grant
- there are no set evidential type schools universally produce that will reach the evidential burden
- credited providers already have existing relationships with the schools surrounding mentor training
- accredited providers will hold their own independent records
As the service is in Private Beta, the Department will conduct further User Research regarding the sampling process with schools. This research will seek to understand, on a lower level, the artefacts schools produce regarding training and internal processes.
In addition, the Claim team will explore ways to use existing data and services to help identify and substantiate the relationship between the school, the accredited provider, and the mentor.
Further considerations
Clawbacks
A clawback is the process by which overpaid funds or funds without sufficient supporting evidence can be recuperated. Clawbacks are typically run against other scheduled payments to the school, where the total outstanding amount is offset against an upcoming payment. If there are no scheduled payments, the school will be issued an invoice for the outstanding amount. There are a range of ways a clawback can work, this depends on the amount that needs to be returned to the Department.
This service will initiate a clawback in the following scenarios:
- the accredited provider is unable to verify the claim made by the school
- The department is unable to verify the claim by working with the accredited provider and/ or the school
In both above instances, verification of the claim will have been attempted by either the accredited provider or a combination of the accredited provider and the Department for Education.
The Claims team must provide the ESFA with a signed-off list if a clawback is required. The ESFA will seek to recover these through negative offsetting scheduled payments. Alternatively, if there are no scheduled payments, the ESFA will raise an invoice to the school. This is not something the service will be directly involved with following the initial referral. However, we intend that the service will show a status indicating that the claim is ‘Awaiting clawback.’
Depending on the level and value of clawbacks initiated, the service may need to re-run the sampling process. In this instance, the identification algorithm will be re-run to exclude the initial sample and increase the sample percentage. The flow regarding declarations and annexed Excel files sent to the accredited provider for verification will remain the same.
Flow diagram
