Print

Current account: change to new balance of payment standard and backcasting of current account time series

Current account: change to BPM6 and backcasting of current account time series

The switch to the new BPM6* standard for the balance of payments has led to various changes in the current account. New terms have been introduced, several components reclassified, the number of components has increased and some of the content has also changed. Particularly important is also the fact that the changes in the balance of payments go hand in hand with the revision of the current account surveys. This revision serves two purposes: first, additional international requirements are hereby met and second, more data will be available to users of the statistics from hitherto neglected areas of foreign trade.

Besides the revision of the surveys, the number of companies required to report data for the current account has also been expanded. In the case of components already reported in the past, the larger number of surveyed companies has led to a considerable increase in the level of receipts and expenses, and thus to a structural break in the data series. In order to make long time series available to users nevertheless, the data are calculated backwards (backcast) wherever possible.

In the following sections, the most important steps in this backcasting procedure are described and the impact on the time series quantified.

*Balance of Payments and International Investment Position Manual Sixth Edition.

1. Classification of components under BPM6

In a first step, various components of the balance of payments that used to be classified differently under the former BPM5 standard have been reclassified retroactively. The following summary shows the most important changes in classification:

  • Trade in precious metals as raw materials (‘bullion’), which was previously recorded in the financial account, has been reclassified as goods trade. Trade in precious metals is the largest single component of the current account and therefore significantly raises the reported turnover in goods. Because of its volatility, it also strongly influences the path of the current account balance.
  • Merchanting is now reported under goods instead of services. Besides investment income, it is merchanting that makes the largest contribution to the current account surplus.
  • Manufacturing services on physical inputs, maintenance and repairs are now all classified under services and no longer under goods.  
  • Postal and courier services have been allocated to transport in the services category.  
  • Telecommunications services are now classified together with computer and information services in a single component.
  • Intellectual property rights arising out of research and development (R&D), such as patents, have been included in the research and development services component. They were previously classified under capital transfers, where they were reported together with purchases and sales of franchises and trademarks. However, historical data on property rights arising out of R&D were not surveyed separately in the past and cannot thus be reclassified.
  • Changes in claims by households against pension funds are now reported under secondary income (formerly referred to as current transfers) and no longer under primary income (formerly labour and investment income).
  • Stamp duty is now reported under secondary income, no longer under services.

2. Backcasting of current account components

As an initial step, the backward calculation was made for the period up to the first quarter of 2000. Backcasting reaching further into the past is planned for the coming year. The degree of detail in the backward calculation – components, sub-components, etc. – was mainly dependent on the availability of the information required. Basically, a distinction was made in the backward calculation procedure between components that were available for the comparable historical time series and components that were not.

Adjustment of components to the historical path of the time series

When components for which comparable data existed in the past were backcast, the new time series was adjusted to the previous path of the time series. Deviation from this procedure occurred only in cases where better information on the past became available than was contained in the old time series. However, this only occurred in exceptional cases. The backcasting was therefore based on the assumption that the historical path of the data series is also representative for the enlarged reporting population.

Backcasting of newly surveyed components without historical data: use of indicators

In the case of both new components and former components that could no longer be compared with the newly surveyed data, the basic question was whether it made sense to backcast them. The following solution to this question was adopted: Backcasting was performed for those time series for which suitable indicators were available as proxies. This means that no backcasting was carried out in the case of components for which no suitable indicators were found. Nor was there any backward calculation performed for new components whose underlying business activity did not exist in the past, or only existed to an insignificant extent.

Selection of indicators

The current account time series that are to be backcast consist of nominal turnover figures from foreign trade in goods and services. It is therefore meaningful to use indicators that also feature the characteristics ‘nominal’, ‘turnover’ and ‘cross-border’.  This applies, in general, to nominal aggregates in the National Accounts (NA) and to the turnover figures in the value added tax statistics. Gross output and intermediate consumption were selected from the NA aggregates, as were total turnover and export turnover from the value added statistics. The advantage of these statistics is that they show nominal turnover figures which also contain the components that are to be backcast. In addition, the degree of detail by industry in these statistics is fairly high, so that the selected industry figures are consistent with the component to be backcast. As a rule, there are also detailed industry figures in the employment statistics. However, these statistics are, in general, less suitable, since employees represent ‘real’ entities, which is why the path of these variables can deviate substantially from the path of the current account turnover figures as a result of developments in productivity. Nevertheless, they were also taken into consideration due to their high degree of industry-by-industry detail.

The indicators based on value-added-tax turnover (total turnover and export turnover), gross output and intermediate consumption by industry, and employment statistics (full-time equivalents by industry) have been compared and selected with the use of statistical tests.  For these tests, components had to be chosen for which both historical time series and indicators were available. This was based on the assumption that an indicator which best fulfilled the requirements in terms of historically available time series would also fulfil the requirements for the backcasting of historically non-available time series. The test results suggest that nominal gross output is the most suitable indicator. It is, however, not available for all the required components, which means that other indicators also had to be used.

The four steps in backcasting

1. Reclassification of time series from BPM5 to BPM6

  • First of all, the historical current account time series are reallocated in line with BPM6 (cf. section 1). This means that certain components formerly within goods are now classified under services and some former services are classified under goods. In rare cases, the reclassification also affects components from primary and secondary income as well as items in the financial account.

2. Identification of structural break

  • The old current account surveys were carried out for the last time in 2011, the new ones with a larger reporting population were introduced in 2012. Since the survey periods do not overlap, the shift in levels between 2011 and 2012 had to be estimated. For components already surveyed in the past, it was possible to continue with the old reporting population by using only the data of companies that were already part of the former reporting population. The difference between the totals for the old and the new reporting populations for the year 2012 represents the shift from one level to the other. The total for the new reporting population was used as reference value for the backcasting. In this way, the time series have been adjusted to the historical path for 25 components.

3. Selection of indicator for new components

  • The indicators were selected based on availability and on the test results. Ranked according to the test results, gross output by industry came first, in principle. However, it was not available for all the required components. It proved applicable in the case of nine components. As far as the research and development time series are concerned, the Swiss Federal Statistical Office’s R&D survey was used both for the receipts and the expenses side. The employment statistics were applied in the case of four time series. For receipts  in construction services, the Eurostat construction index was used. Expenses were backcast using the gross construction output figures. In eight cases, use was made of the goods and services aggregates. Overall, indicators for the backcasting were applied for 25 time series. Four components (pipeline transport, space transport, electricity transmission, health services) were not backcast due to lack of business activity in the past or because turnover was very low. In addition,13 sub-positions of components were not backcast, either. Backcasting for a further 16 components was only possible in combination with one or several other components. This made a total of 29 time series for which indicators were not available to the required degree of detail.

4. Backcasting of annual and quarterly time series

  • Both annual and quarterly figures are calculated backwards using the rate of change against the previous period. First, the annual figures are calculated using the rate of change against the previous year, and subsequently the quarterly figures are calculated based on the rate of change against the previous quarter.

When quarterly figures are backcast based on the changes against the previous period, the sum of the quarterly figures backcast corresponds to its backcast annual counterpart only in exceptional cases. To avoid such inconsistencies, the backcast quarterly time series have been arithmetically adjusted to the corresponding annual figures. Thus, the annual figures remain consistent with the sum of the quarterly figures, and the path of the quarterly time series has been maintained. This means that it was possible to transmit all the information contained in the historical time series to the newly backcast time series. The Chow Lin approach was used to carry out the adjustment taking into account the above-mentioned conditions.*

*For the implementation of this procedure, the R package "tempdisagg" by Christoph Sax and Peter Steiner is applied (http://cran.r-project.org/web/packages/tempdisagg/index.html).

3. Impact on data

Current account

As a result of the changes, receipts and expenses in the current account are substantially higher than before. In 2013, for example, receipts rose by one third after the changeover, amounting to CHF 618 billion as against the previous CHF 459 billion. Expenses increased to a similar extent, from CHF 381 billion to CHF 521 billion. In most cases, the current account balance is slightly higher than it would be under BPM5. However, figures vary greatly from one year to another. For 2013, the balance comes to around CHF 18 billion, whereas the 2009 surplus is CHF 16 billion lower. For both 2009 and 2013, the current account balance changes evident in the revisions were, to a large extent, attributable to the reclassification of precious metals to goods trade.

Goods

In the goods category, turnover is considerably higher as a result of the reclassification of precious metals and merchanting. Growth in the goods surplus was particularly substantial owing to the reclassification of merchanting and the expansion of the current account surveys;   the figure rose from CHF 17 billion to CHF 52 billion in 2013.

Services

Changes in the services category were due to the reclassification of merchanting from services to goods, on the one hand, and to the expansion of the current account surveys on the other. Expenses in particular increased by around 60% as a result of the latter factor. In 2013, expenses came to CHF 79 billion after the changeover, having previously amounted to CHF 49 billion. The increase in receipts due to the changeover was distinctly lower. They advanced by 18% to CHF 104 billion in 2013. The main reason for the comparatively low increase was the reclassification of ‘net receipts from merchanting’ to the goods category. This partly offset the survey-related increase. Since the impact of the changeover was twice as high on the expenses as it was on the receipts, the services surplus declined significantly, decreasing from CHF 40 billion to CHF 25 billion in 2013.

Table: Impact of changes on current account, with breakdown for goods and services

Current account net
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
82
84
-2
-8
5
14
15
2011
39
52
-13
-19
6
7
9
2012
66
57
9
-10
19
11
10
2013
96
78
18
9
9
16
13

Goods receipts
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
287
204
83
83
0
50
36
2011
306
209
97
96
1
52
36
2012
311
212
98
97
1
53
36
2013
346
213
133
136
-4
57
35

Goods expenses
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
252
191
61
69
-8
44
33
2011
281
194
86
95
-8
48
33
2012
274
197
77
86
-8
46
33
2013
294
196
98
107
-9
49
33

Goods net
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
35
13
22
13
8
8
2
2011
25
14
11
2
9
8
2
2012
37
16
21
11
10
8
3
2013
52
17
35
30
6
8
3

Services receipts
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
98
87
11
-21
32
17
15
2011
95
84
11
-21
32
16
14
2012
99
85
14
-21
35
17
14
2013
104
88
15
-20
36
17
15

Services expenses
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
68
38
30
1
30
12
7
2011
70
40
30
1
29
12
7
2012
74
44
30
1
29
12
7
2013
79
49
30
1
29
13
8

Services net
BPM6
BPM5
Difference
Of which reclassification
Of which extension of survey
BPM6 in percent of GDP
BPM5 in percent of GDP
2010
30
49
-19
-22
2
5
9
2011
25
44
-19
-22
3
4
8
2012
26
41
-15
-22
6
4
7
2013
25
40
-15
-21
6
4
7