A loan-level data collection for buy-to-let lending - Bankstats article

This article summarises the responses received to the consultation on the loan-level data collection for buy-to-let lending, published in the April edition of this publication. It also outlines planned changes to the initial proposals in light of the feedback received.

Published on 29 July 2016

Bankstats article

By Zeeshan Akhtar and Ross Meader

Overview

The Bank of England published a consultation on proposals to develop a new buy-to-let data collection on 29 April 2016.

The loan-level collection will support the needs of the Monetary Policy Committee, the Financial Policy Committee and the Prudential Regulation Authority (PRA) Board and aims to combine Bank of England statistical and regulatory data requirements into one collection for the first time.

The Bank invited comments on the proposed list of attributes, the potential difficulties in obtaining data on lending to corporates, and the proposed timing of the collection. Ten consultation responses were received, with some of these from industry bodies representing a wider group of respondents.
The consultation closed on 10 June 2016. This article summarises the key themes from the consultation responses and explains changes to the proposals.

Timing

A number of responses highlighted issues with meeting the proposed first period of data collection of 2017 Q2 and favoured delaying the collection. Respondents provided a number of reasons:

There were concerns about collecting new data fields on retail buy-to-let, where the applications for loans can be made up to six months before completion.

Therefore front-end systems changes that are required for loans completed in April 2017 would need to be in place by October 2016;

There were concerns about being able to report data for buy-to-let lending held in their corporate lending books by 2017 Q2, given its greater complexity;
The Council of Mortgage Lenders (CML) currently collects loan-level data on buy-to-let lending from 17 lenders. Firms that do not currently report to CML may have difficulties delivering systems changes.

In response, the Bank will delay the first period of data collection by one quarter, and implement the collection in three phases:

Phase 1 - delivered in October 2017 for Q3 2017 data: core retail book attributes required for the most pressing policy purposes together with some existing attributes within the Product Sales Data (PSD), reported to the Financial Conduct Authority (FCA). Mostly these are already collected by the CML from larger reporters. There are four attributes that are optional for phase 1. These are loan regulation, type of dwelling, and date of birth (and an accompanying borrower reference). They are included because they are currently included in existing reporting to the CML and will be optional until phase 2 reporting starts.
Phase 2 - delivered in April 2018 for Q1 2018 data: the remaining retail book attributes.
Phase 3 - delivered in October 2018 for Q3 2018 data: data attributes required for buy-to-let lending in the corporate book.

The Bank has today published details of the data collection relating to phase 1. Respondents suggested aligning the publication of the details of the data collection to the publication of the PRA’s final policy on buy-to-let underwriting standards. The PRA is reviewing responses to the consultation paper on underwriting standards, and will publish its final policy in due course. The Bank intends to publish details of phase 2 alongside the PRA’s Policy Statement on underwriting standards.

The Bank welcomes the CML suggestion to set up a working group with firms to refine the collection of data relating to the corporate book, where there are greater complexities. Details of the collection of corporate data will be published once that work is complete.

Reporting thresholds

The Bank had proposed that all firms with gross flows of new lending exceeding £20mn annually would be required to report. Respondents raised concerns around the proposed population coverage of the data collection, particularly for some specialist firms that carry out a small number of high value loans. The Bank is also mindful of the burden of reporting for smaller entities within a wider banking group.

To avoid disproportionate reporting burden, only firms or groups of firms carrying out new buy-to-let mortgage lending exceeding £20mn annually across at least 60 loans will be required to report. Individual entities that form part of a group are expected to report, except where an entity lends less than £20m and 60 loans per year and contributes less than 2% of the number of new loans within a group. This will be monitored using the Mortgage Lenders and Administrators Return (MLAR), for firms that report it, and market intelligence.

Data sharing and data protection

In addition to the Bank and the PRA having access to the data collection, the Bank anticipates sharing data with the Financial Conduct Authority (FCA). Some respondents questioned the purpose of this data sharing and sought clarity on the circumstances under which this will occur.

The Bank anticipates sharing data with the FCA, to inform the FCA’s considerations of how to advance its statutory objectives under the Financial Services and Markets Act 2000 (FSMA). This includes the FCA considering whether and how it can advance its market integrity objective, and whether activity in the buy-to-let market might have any consequential effects on consumers of regulated products and its other statutory responsibilities.

The Bank envisages routine data sharing with the FCA in accordance with Schedule 7 of the Bank of England Act 1998. This permits the Bank to share data collected under Section 17 of the 1998 Act with the FCA in circumstances when the Bank considers that disclosure would enable or assist the FCA to discharge its functions under legislation, including FSMA.

The data collected includes a limited amount of data relating to individual borrowers that could potentially, in combination with other data, allow individuals to be identified – in particular, the Bank requires borrower date of birth and postcode attributes (as part of phase 2). These are required for two purposes:.

To match to other datasets to understand broader trends in the housing market to inform monetary policy; and
To match across borrowers and identify the total borrowing across a borrower’s entire portfolio to support micro and macroprudential policy.

As with all collections containing personal data, the Bank will process this in accordance with the requirements of the Data Protection Act 1998. As such, we anticipate that where possible, use of the data will be restricted to a less detailed version of the data.

Respondents highlighted that firms may need to make their customers aware that certain personal data are being used for purposes other than the mortgage approval process. The Bank considers that it is likely that the provision of this data to the Bank and PRA by firms and any onward disclosure by the Bank to the FCA would be on the basis that it is necessary for the performance of their respective statutory functions, and so would not require consent from individuals. Nevertheless, firms may wish to consider whether it is appropriate for them to amend the terms of their agreements or privacy notices with customers to make clear that such information will be made available to the regulators for the identified purposes.

Detail of the collection

Following discussions with industry groups during the consultation the Bank recognises that the proposed collection requires flexibility to collect attributes relating to multiple borrowers and multiple properties. Some attributes will be collected for each borrower and property relating to the loan. The collection will be limited to loans that are wholly secured by buy-to-let property or the borrower’s main residence.

Given similarities with the PSD attributes, firms would prefer definitions to be aligned where a common attribute is collected even though they are for different sets of loans. The Bank has endeavoured to do this where possible. Differences will exist where data requirements are driven by the PRA supervisory statement, or in order to be comparable to other existing data sources collected by the Bank because this collection needs to support multiple purposes.

In response to concerns from respondents where there are multiple tranches of loans, the requirement will be to report based on the largest part of the loan rather than each separately. This is consistent with PSD definitions and avoids potential difficulties with having to report each tranche separately.

Firms have different practices in how they treat remortgages resulting in additional borrowing where the borrower refinances with the same lender. Some firms treat this as a further advance, while others redeem the original loan and raise a new loan. Discussions with industry groups during the consultation concluded that, for the purpose of this data collection, in both cases this should be reported as a further advance for the additional borrowing only.

Firms reported issues in being able to deliver data relating to interest rate switches. The Bank has weighed these difficulties against the benefits of collecting this information and has decided to remove interest rate switches data from the collection.

A final updated list of phase 1 attributes can be found in Appendix 1.

Plausibility checking

Validation rules applied to submissions will ensure errors are kept to a minimum. The Bank will also carry out 'plausibility' checking on the data and raise queries with reporters about the data as appropriate. Broadly speaking, validations relate to ensuring accounting identities are met and plausibility relates to comparing, for plausibility, the latest reported data point with its time series and other related information, both current and historical. The Bank’s interest in collecting granular data does not lie in wanting to examine individual loans. Instead, it is to be able to better understand the distribution of characteristics for lending within the aggregate data. For this reason, the Bank does not envisage the plausibility checking process being more intense for reporters than traditional statistical forms once the new data has settled down.

To view all related Annexes, please download the full article:

Download PDF.