Bankstats article
By Alistair Strathern
Additional detail on holdings of securities by monetary financial institutions collected since 2014 has facilitated an improvement in the way transfers of securities are treated in the calculation of transactions data. This will allow for exclusion of the effects of asset transfers and therefore bring their treatment more in line with the Bank’s approach to transfers of loans. The changes will be implemented for June 2015 data, available in the next Bankstats publication. This will result in revisions to a number of securities and lending net flow and growth rate series. Whilst there are significant revisions to data in some months, the overall effect of revisions across the period affected (February 2014 to May 2015) is small. The Bank will also implement the results of a review of the seasonal adjustment of M4L measures.
Introduction
The Bank of England publishes a range of data on monetary financial institutions’ (MFIs) holdings of securities. As with all data published by the Bank, these are subject to ongoing review. Additional detail on holdings of securities by monetary financial institutions collected since 2014 has facilitated an improvement in the way transfers of securities are treated in the calculation of transactions data. This will allow for exclusion of the effects of asset transfers and therefore bring their treatment more in line with the Bank’s approach to transfers of loans. Alongside this, the Bank is also making changes to the seasonal adjustment of related series. The changes to be implemented are as follows:
- From February 2014 data onwards, transfers of quoted shares will be omitted from net securities flow series, and therefore net M4 lending (M4L) flow measures. The resultant contribution of quoted shares to these measures will typically be zero. A full list of series affected by this change is available in the Annex to this article.
- Following a recent review of seasonal adjustment (SA) of M4L measures, the Bank will move to an indirect seasonal adjustment of some measures, adjusting the loans and securities components separately.
The changes will be implemented with the publication of June 2015 data, available in the next Bankstats publication. This will result in revisions to a number of securities and lending net flow and growth rate series. Whilst there are significant revisions to data in some months the overall effect of revisions across the period affected (February 2014 to May 2015) is small.
Statistical treatment of net flows
In Bank of England credit statistics, a distinction is made between changes in amounts outstanding of financial instruments that are due to lending transactions and those that are due to other changes in the value of assets (OCVAs). This latter category includes items such as revaluation effects, reporting changes and asset transfers. The Bank’s statistical net credit flows data are constructed to reflect lending transactions as far as possible using the difference in amounts outstanding between reporting periods, less the value of identified OCVAs.
In the case of securities data a number of OCVAs are currently captured, with revaluation effects and reporting changes all routinely identified or estimated. However, asset transfers have proved challenging to identify as the example below sets out.
Consider a situation where an MFI is looking to purchase a security (such as a quoted share) of a given company. It has two options: 1) to purchase it directly from the company at the point of issuance, or 2) to buy a previously issued security via a secondary market. The first channel would be classified as a lending transaction, while the second should be classified as an OCVA. MFIs are currently unable to distinguish between the two channels in their statistical reporting of securities holdings to the Bank. Therefore, currently both lending transactions and secondary market asset transfers feature in the Bank’s published statistical data on net securities flows.
Recent analysis, as part of the Bank’s commitment to review its published data regularly, has identified improvements to this approach drawing upon new breakdowns introduced to the Bank’s statistical data collection. The Bank began collecting statistical data on MFIs’ holdings of securities by instrument type (quoted shares, unquoted shares, bonds and other debt securities, and other securities), by sector, from January 2014 data. While reporting institutions remain unable to separate out transferred assets themselves, Bank analysis has identified that the most of movements reported by MFIs for the quoted share component reflect secondary market asset transfers. Internal analysis based on capital issuance data and confidential counterparty information provided by reporters.
Given this, the Bank will now move to a position of including net quoted share flows only where transactions relating to new issuance and buyback activity can be identified. This improvement will come into effect from February 2014 data (the point from which net flows data for the new breakdown are available). The component change within net securities flows is summarised in Table 1.
This will affect published net flows and growth rates for quoted shares, as well as the securities and lending measures these feed into. This change will introduce a break from February 2014 data onwards in both the relevant net flow measures and their respective growth rates. The respective ‘amounts outstanding’ series will be unaffected by this change.
Table 2 provides a high level summary of the M4L series that will be revised as a result of this improvement.
Changes to seasonal adjustment
Concurrent with this change in securities treatment, the Bank will also be implementing the results of a review into the seasonal adjustment of M4L measures. Recent analysis has indicated that the private non-financial corporations (PNFCs) M4L series would be better suited to indirect seasonal adjustment, whereby the total series is the sum of a seasonally adjusted loan component and a non seasonal securities component.4 Intermediate other financial corporations and non-intermediate other financial corporations M4L tested under their current treatment of direct adjustment remain non seasonal. The suitability of these series to indirect seasonal adjustment will be assessed once a sufficient back run of the separate securities data (available from 2014) is available.
Materiality of the changes
Both the change in treatment of securities and in seasonal adjustment will lead to revisions over the back run of the affected series. Table 3 shows revisions to a number of series that will result from both changes being implemented.
The magnitudes of the mean revisions presented in Table 3 reflect the fact that over the period affected, the revisions are largely offsetting. Consequently, the mean absolute revisions are materially larger for the series shown.
Taking PNFC M4L as an example, the following charts illustrate the materiality of revisions for both the monthly net flows and twelve-month growth rate series.
Chart A shows the net flows of PNFC M4L, non seasonally adjusted, with the new series showing significant revisions changes in some months, but less volatility over the affected period and a small cumulative effect overall. The mean absolute revision is £2.5 billion.
Chart B shows the twelve-month growth rates. These are significant in some months, with an average revision to the twelve-month growth rate of 0.6 percentage points.
For seasonally adjusted data, the new measure will also reflect the changes to the method of seasonal adjustment outlined earlier. Charts C and D show the combined effect these changes have on the net flows and twelve-month growth rate of PNFC M4L, seasonally adjusted.
These charts both show that the majority of the changes to seasonally adjusted data from February 2014 onwards are due to the improvement in securities treatment. As shown in Table 3, the mean absolute revision to the net flows of PNFC M4L, seasonally adjusted, is £2.0 billion, due to these changes.
Next steps
The changes outlined in this article will take effect from the next edition of Bankstats, published on 29 July 2015.
To view all related charts, tables and Annexes, please download the full article:
Download PDF
For questions relating to this article please contact dsd_ms@bankofengland.co.uk or call +44 (0) 20 3461 5356.