FISIM
February 2005
Methodological improvements since June 2004
This note summarises progress made in calculation and allocation of FISIM between June 2004 and February 2005. It is based on Member States responses to mini-questionnaires in January 2005. Areas (FISIM components) of significant progress are marked “YES” in capital letters on the attached table 1 (see file Table 1 Feb05.xls)
General points
Belgium
Since June 2004, the major improvements introduced in the calculation of FISIM are linked to the recalculation of accrued interest by sector for s122, having an impact on FISIM allocation among the domestic user sectors (point c) and on calculation of FISIM intermediate consumption and final consumption by households (point f). Export of FISIM has been revised especially for the beginning of the period (years 1995-1998).
Apart from improvements made to the calculation of FISIM, my work since June 2004 has been mainly concentrated on three issues: calculation of quarterly FISIM, calculation of backdata and calculation of breakdown of imports and exports between EURO area and ECB, non-EURO area EU member States and EU institutions and extra EU countries.
Quarterly FISIM
My first concern was to develop a simplified methodology based on the estimation of the production then to allocate it based on the previous year results. As the consumption pattern of FISIM is volatile from one year to the other the estimated impact on the GDP would have been overestimated for 2003 by some 290 millions Euro, about 17 p.c. of the "annual" impact on GDP.
Another way to calculate the quarterly FISIM is to make the full calculation for each quarter. These calculations imply the compilation of quarterly interest matrix (or at least one line and one column of that matrix, for s122). The production of FISIM by s123 and the allocation of that production are still estimated based on last year results. This method is more accurate but is also much heavier.
CzechRepublic
Now we calculate External reference rate and FISIM export and import. In this moment we have data for the year 2002 and we are preparing calculations for other years. We have also data of FISIM at constant prices and during the next 3 month we will calculate allocation among industries. In May and in the beginning of June we have our deadline for yearly and quarterly NA.
Cyprus
The collection of data needed for the calculation of internal and external reference rates is well under way. The eventual calculations, however, are not yet completed due to delays experienced in the data collection phase.
Poland
Annual and quarterly data needed for calculation and allocation of FISIM were collected for the period 4th quarter 1996 – 4th quarter 2004. Data for 1995 (annual as well as quarterly) and 1996 (1st, 2nd, 3rd quarters) are incomplete, therefore, some estimations needs to be made in cooperation with the National Bank of Poland. Several meetings with the National Bank of Poland took place in previous year in order to discuss the availability of data and methodology of estimation lacking information. The quarterly calculations of FISIM were made for 2003.
a.Calculation of internal reference rate
Cyprus
The only remaining item to be collected is the data on inter-bank stocks and interest of loans and deposits concerning the Co-operative Societies and Credit Institutions (Co-ops). Since the data in the requested analysis is not available at the aggregate level of all Co-ops, the primary data is collected (from the Co-ops Audit Service) and processed by Cystat for each individual case. The data collection phase is now expected to be completed by mid-March 2005.
Latvia
We have made some corrections in internal and external reference rate calculations, using specifications for inter-bank data of loans and deposits and also of interest of loans and deposits. Information was more reliable on S.122 than on S.123, therefore estimates were needed in some cases.
Portugal
Exclusion Money Market Funds from S.122 accordingly to the conclusions of the FISIM workshop of June 2004
Slovenia
Internal and external reference rate could be calculated, when data from new reporting system will be available, which is expected in the first half of 2005.
b.Calculation of external reference rate
Cyprus
All data needed for the calculation of external reference rate is now available and it appears that there is no problem whatsoever to complete the calculation in the next few days.
Latvia
(see above point a)
Poland
The feasibility of calculation of FISIM with the breakdown of S.2 sector into Euro, Non-Euro and Extra-EU is currently being analysed. Only data on stocks of loans and deposits are available in mentioned above breakdown for monetary and non-monetary institutions from 1st quarter of 2004.
Portugal
Calculation of two distint external reference. External reference rate for exports and external reference rate for imports. New detail of data base, distinguish the loans from the deposits, on the assets side and on the liabilities side.
Slovenia
(see above point a)
c.FISIM allocation among the domestic user sectors
Belgium
Interests received on loans and paid on deposits are given by the profit and loss account of banks but without information about the breakdown by counterpart sectors. The breakdown by domestic user sectors are estimated based on a model applying rates to stocks of loans and deposits by sector of the financial accounts. The rates previously used were implicit rate based on profit and loss data divided by the corresponding item of the balance sheet. Those implicit rates was increased or decreasedto take into account the different rates applied by the banks according to their clients (business, retail) in order to have higher rates for households loans compared to corporations loans, and conversely lower rates for deposits.
The results of MIR statistics for 2003 show that the rate used so far was not enough differentiated between corporations and households. Therefore, interest rates applied to stocks have been revised for the whole period based on that result.
Cyprus (relevant also for the following item d, r, f, h)
The data needed for the completion of the work for the above mentioned subjects is being collected, while for the cases where no data appears to be available, estimates will be made. For the cases that concern missing data that should have been provided by Central Bank of Cyprus, a joint working team was set-up a few weeks ago to investigate alternative data sources and make estimates.
Latvia
We used supplementary information to derive more detailed FISIM allocation among the domestic user sectors. As a result, we recorded no negative net FISIM margins on loans and deposits for all sectors.
Eurostat: Please provide more information on the supplementary information used.
Lithuania
No direct data on interest of banks distributed by institutional sectors are available. We estimate them by multiplying stocks on loans and deposits of each institutional sector by average interest rates. On the FISIM workshop in June 2004 the expert from Germany Mr. Wolfgang Eichmann presenting the paper on reconciliation of interest mentioned that interest rate on loans to Government could be considered lower (about 1 percentage point) than that to enterprises. With respect of that idea after consultation with Eurostat’s expert Mr. Leonidas Akritidis and with the Germany’s expert Mr. W. Eichmann we improved our calculation. Results of FISIM allocation among the domestic user sectors were changed but not very considerably.
Portugal
Loans from the S.122 and S.123 to S.13 : The amount of interests pay by General Government is directly obtain from direct source of General Government.
Slovenia
FISIM allocation among the domestic user sectors have been done on the annual and quarter basis for the period 2000-2003. Allocation for period 1995-1999 is in preparation.
d.Allocation of FISIM exports
Belgium
Calculation of accrued interest for loans and deposits of non-resident non-FI has been revised. It has appeared that the interest rate used for estimating those accrued interest were not plausible, both rate on deposits and loans being nearly identical and quite different from those for resident client.
Netherlands
We made no further progress. For the years 2001 and 2002 we only calculated an amount for the total of S.2. We didn’t make a geographical allocation or allocation among non-resident FI and non-resident N-FI. We are planning on doing this in the near future.
Poland
The individual negative FISIM exports will be reclassified into the corresponding FISIM imports with a positive sign.
e.Allocation of FISIM imports
Estonia
Imports of FISIM calculations have been improved as BoP can offer somewhat better breakdown. Some FISIM is allocated to households as consumers both on loans and on deposits side. Although the figures are underestimated as stocks of loans and deposits include only transactions made via international money order.
Eurostat: Please see discussion on methodological problems and solutions, section e
Netherlands
Imports are allocated using the ERR and the balance sheets of S.2. An assumption is made to calculate for the part of loans and deposits that is held by NR-FI and NR-NFI.
Slovenia
FISIM import data source used are as follows:
- we use BoP data from Bank of Slovenia
- data on loans are available by sectors
- total amount is available for deposit
We will contact colleagues from Bank of Slovenia and ask them again for breakdown of deposit by sector. As soon as we receive their answer we will inform you.
Data from “new reporting system” will be available for the year 2005 and onward.
f.Calculation of FISIM intermediate and final consumption by households
Belgium
Once again, the exploitation of MIR statistics allows having a better breakdown of accrued interest between interest paid/received by households as consumers, households as owners of dwellings and households as owner of unincorporated enterprises. This has a noticeable impact on GDP since the accrued interest on consumer loans was previously underestimated.
Latvia
According to requirement of the Eurostat, we have improved the calculation of intermediate consumption of FISIM on deposits by S.14 sector. Thus, this sector shows no negative FISIM on both loans and deposits and also no negative net FISIM margins.
Netherlands
No further progress. We still use our original method to split the FISIM consumption by households in final and intermediate consumption.
Eurostat: Information on the above is available on circa in the document FISIM 02.
Poland
FISIM on loans for households as owners of dwellings is calculated by using the banking statistics on loans by purpose.
g.Calculation of FISIM at constant prices
Cyprus
The data needs and methods for the calculation of FISIM at constant prices are still being examined
Latvia
FISIM at constant prices have revaluated on both the loans and depositsside, using the implicit price deflator for domestic final demand.
Netherlands
We use the method as described in the ‘Handbook on volume and prices indicators’, which is based on a variety of output indicators.
Poland
Calculation of FISIM at constant prices is being currently analysed. The method of the base period margin described in the FISIM Regulation will be applied, using as a general price index the overall CPI.
h.Allocation of FISIM among industries
Latvia
We started the allocation of FISIM among industries. For this purpose Method II - based on the output for each industry will be used.
Lithuania
FISIM has been allocated among industries using Method II based on the output for each industry.
Netherlands
We allocate FISIM among industries based on the output of industries.
1