CODING DOCUMENTATION FOR IFS CD-ROM DATA SET
A. Introduction
A.1. INTERNATIONAL FINANCIAL STATISTICS (IFS) is the International
Monetary Fund's principal statistical publication and has been
published monthly in printed form since January 1948. In 1972 a 9-
track magnetic tape subscription was introduced, and in June 1991 the
monthly publication of IFS on CD-ROM was inaugurated. Each CD-ROM
edition is closely keyed to the same month's printed edition and
contains data sets for most Fund members, as well as for Aruba and the
Netherlands Antilles.
A.2. The CD-ROM subscription provides users with time series data
covering approximately 26,000 economic concepts which show major
economic aggregates used in the analysis of economic developments
and includes data on the following topics: balance of payments,
banking and financial systems, employment, exchange rates, Fund
position, government finance, interest rates, international
liquidity and banking, national accounts, population, prices,
production, and trade.
A.3. In the printed publication, data expressed in domestic currency for
countries that have undergone periods of high inflation (e.g., Argentina,
Bolivia, Brazil, Peru, Nicaragua, and Israel) are presented in different
magnitudes on the same line. Users of the printed publication may refer to
midheaders on the country pages that indicate the magnitude changes. The
practice of expressing different magnitudes on the same line was adopted in
order to present historical values that are nonzero. On the CD-ROM, the
data are stored in a scientific notation with eight significant digits for
all time periods. Therefore, users may view historical as well as current
values by using the display option for scientific notation available on the
CD-ROM. To view data in scientific notation, select menu item F3 (Format),
choose "Display Format," and then choose "TS Display III" (or IV or V).
B. Contents of the CD-ROM
B.1. The time series that are to be found in the CD-ROM publication
include the following:
1) all series appearing on the IFS country pages;
2) all series published in the IFS world tables, except for the
daily exchange rates appearing in the exchange rate tables;
3) the following exchange rate series as available: aa, ac, ae,
af, ag, ah, b, c, de, dg, ea, eb, ec, ed, g, rb, rd, rf, rh,
sa, sb, sc, sd, wa, wc, we, wf, wg, wh, xe, xf, ye, yf, nec,
rec, aat, aet, rbt, rft, neu, reu, ahx, and amx;
4) Fund accounts time series, all in SDR terms: 2eb, 2eu,
2ey, 2tl, 1b.s, 1c.s, 2c, 2dus, 2ees, 2egs, 2ehs, 2eqs,
2ers, 2ets, 2f.s, 2h.s, 2krs, 2kxs, and the following series which
have the source code T--1bf, 2fz, 2kk, 2kl, 2lk, 2ll, and
2tl;
5) money index series (34..i and 34..x);
6) export and import series (70 and 71);
7) export and import unit value index series (74 and 75);
8) balance of payments data on other goods, services, and income
(series 77add and 77aed), other long- and short-term capital
(series 77bcd and 77d.d), and details, as available, of
exceptional financing (series 79a*d).
The series referred to in item 3 above are for all IMF members, plus
Aruba, Hong Kong, and Netherlands Antilles. The series referred to in items
4 through 8 are for all countries for which data are available, although
some are not published in the printed IFS publication. All series on the
IFS CD-ROM contain control source code F except for the previously mentioned
Fund accounts series which contain the code T (see item 4 of paragraph
B.1.).
B.2. The CD-ROM contains data in three frequencies (annual,
quarterly, and monthly). When the data are available, the annual
entries generally begin in 1948, with quarterly and monthly entries
beginning in 1957. All data items are stored internally in a
compressed form of scientific notation which maintains 8 significant
digits of accuracy for every observation. The data may be presented
either in scaled decimal numbers (e.g., units, millions, billions,
etc.) or in scientific notation.
B.3. Each time series in the CD-ROM publication contains a number
of meta data items that describe some aspect of series. Searching
capabilities are provided for most of these meta data fields,
although some of them may not be viewed in any of the display
screens. However, one of the print/export screens (the screen
titled "Upload") does allow all of these meta data items to be
exported to DOS files. A brief description of all of these meta
data items is given in this document, and a list of these items
follows:
base year partner country name
beginning period qualification code
break/footnote indicator region code
classification code 1 scale
classification code 2 section code
control/source code source
country/area code source code
country/area name splice code
coverage (number of years) subsection code
data type code time series code
decimal code topic
English descriptor topic (subject) code
equation level update date
frequency update day
magnitude update month
magnitude code update year
national language descriptor version code
partner country code
C. Description of Time Series Code
C.1. Time series in the CD-ROM publication are uniquely identified
by a thirteen-character time series code which is associated with
all time series in the IFS database. A detailed description of the
time series code is given below. Briefly, the time series code,
CCCTTTTTVSPPP, is constructed as follows:
CCC - Three-character country code (always numeric)
TTTTT - Five-character topic code (usually 2 numerics
followed by 3 alphabetics)
V - Single-character version code (always alphabetic)
S - Single-character control source code (always
alphabetic)
PPP - Three-character partner country code (usually
numeric; sometimes alphanumeric)
As a matter of convention, a period (i.e., the punctuation
mark ".") is always used in the CD-ROM publication as a position
holder to indicate a blank in the time series code (e.g.,
"001.7xrdzf900", "111..AEZF...", or "738.4a.d...").
C.2. Country/Area Code in Time Series Code
C.2.1. The country code identifies the country or area to which the
time series data relate. The term 'country,' as used in this
publication, does not in all cases refer to a territorial entity
that is a state as understood by international law and practice; the
term also covers some territorial (nonsovereign) entities, for which
statistical data are maintained and provided internationally on a separate
and independent basis.
C.2.2. The country code always consists of three numeric digits.
The first digit typically provides regional grouping as follows:
0 world or all country aggregates
1 industrial countries
2,3 Western Hemisphere
4 Middle East
5,8 Asia
6,7 Africa
9 centrally managed or formerly centrally managed and other
economies
C.2.3. A complete list of the country codes is provided on the
final pages of this document, first in numerical order, and then in
alphabetical order. In addition, page viii of each issue of IFS contains
country code lists for that publication. Data pertaining to an area provide
aggregated data for countries that comprise that area.
C.3. Topic (Subject) Code in Time Series Code
C.3.1. The topic code is a five-position alphanumeric code
indicating the subject or topic for each time series. In the IFS
printed publication, the topic code may be found at the outer side
of the page. The first two characters are always numeric or blank;
the next three characters are always alphabetic or blank. In the
printed publication a "." or period is inserted as a place holder
when the position is blank, except at the beginning or end of the
code. In the CD-ROM publication, the "." is always shown when a position
would otherwise be blank. In addition, the printed publication provides a
key to commodity codes (positions 4 and 5 of the subject code); the key
appears on page viii of each issue.
C.3.2. The first position of the topic code (sometimes referred to
as the section code) identifies the institutional, topical, or
subject area of this series. The section codes used in IFS are as
follows:
. exchange rate, Fund position, or international liquidity
1 monetary authorities
2 deposit money banks
3 monetary survey
4 other banking/nonbank financial institutions
5 banking survey/financial survey
6 interest, prices, production, or employment
7 international transactions
8 government finance
9 national accounts or population
C.3.3. The second through fourth positions of the topic code are
also known as the subsection code, classification code 1, and
classification code 2 respectively. Their usage is not sufficiently
standardized to warrant further discussion.
C.3.4. The fifth position of the subject code (sometimes referred
to as the qualification code), when specified, provides information
about special characteristics of the data. It may indicate a
statistical procedure, an institutional identity, a level of
government, or the currency of the data observation. The list below
summarizes these qualification codes:
. - national currency (n.c.) data or unadjusted data
A - seasonal factors
B - seasonally adjusted by IMF
C - seasonally adjusted by reporting country
D - U.S. dollars
E - transactions in convertible currencies
F - U.S. dollars, seasonally adjusted
P - constant prices (unadjusted)
Q - constant prices (seasonal factors)
R - constant prices (seasonally adjusted by reporting country)
S - SDRs
T - indicates a comparable time series that requires more than
one magnitude
U - SDRs, seasonally adjusted
X - exchange rate index
Y - price quotation or unit value in n.c. units
Z - price quotation or unit value in U.S. dollars
C.3.5. These qualification codes are also used depending on subject
area as follows:
a. For deposit money banks (section code 2), topic code
position 5 provides institutional breakdowns. These
codes may be found on individual country pages in the
corresponding monthly publication where the nonstandard
nature of the institution is explained in the country
notes.
b. For nonbank financial institutions (section code 4), the
code is used to distinguish types of institutions which
are generally as follows:
F - finance and securities
companies
G, H, I, J - savings institutions
K, M, N, O - development institutions
L, S - insurance companies
X - trustee accounts
C.4. Version Code in Time Series Code: A one-position alphabetic
code used to identify the various revisions of concept and content
of a time series. The first version is identified by Z and each
successive version (created by revision) is identified by earlier
letters in successive sequence Y, X, W.... For the IFS CD-ROM
publication there are only Z versions. All time series have been
butt spliced except index number time series and some currency
conversion factor (exchange rate) time series. Butt splicing
creates a continuous time series unadjusted from the various
versions, giving priority to the latest version. Series that have
been spliced are marked with B (indicating the splicing formula) in
the break/footnote indicator of the data observation at the point of
splice.
C.5. Control Source Code in Time Series Code: An alphabetic code
that identifies either the source of the data or the publication
containing the series. All series on the IFS tape contain control
source code F except for certain Fund account series, which contain
the source code T (for IMF Treasurer's Department), i.e., 1bf, 2fz, 2kk,
2kl, 21k, 2ll, and 2tl. See also paragraph D.15.
C.6. Partner Country Code in Time Series Code: A three-position
numeric or alphanumeric code that is most commonly used to designate the
partner country in those time series that specify transactions between
countries. For IFS data, a partner code is included in only a
limited number of cases. When it exists, it usually is shown in the
printed publication in either the midheading (see world tables on
international banking and real effective exchange rate indices) or
in the notes (see notes to world table on commodity prices, in which case
the three-position partner code that is designating the commodity market is
alphanumeric).
D. ADDITIONAL META DATA ITEMS
D.1. Base Year: An alphanumeric field that identifies the base
year for index number time series, for which all data have been
adjusted accordingly. This field applies to derivative series only.
D.2. Beginning Period: The period (i.e., year, year and quarter,
or year and month) of the first observation in a series.
D.3. Break/Footnote Indicator: An alphabetic break/footnote
indicator is associated with each data point in the CD-ROM
publication. It may contain the following codes:
B - Identifies the point at which multiple time series
versions are linked by butt splicing*
D - Identifies data observations estimated from lower
frequency data by distribution routines*
E - Identifies observations that are estimated by economic
staff
H - Identifies unreported data (observation contains zero)*
I - Identifies data observation estimated by interpolation
routines*
J - Identifies the point at which preceding or subsequent
data are not applicable for estimation routines. This
indicator is used mainly to suppress estimation of data
during calculation of area summations
K - Identifies a break in the analytical comparability of
data
L - Identifies a break in the technical comparability of
data
M - Draws attention to a specific fact that should assist in
the interpretation of the data
N - Draws attention to specific data that are of technical
interest only
P - Indicates that data are in whole or in part provisional,
preliminary, or projected
R - Identifies the point at which multiple time series
versions are linked by ratio splicing using first annual
overlap*
S - Identifies the point at which multiple time series
versions are linked by ratio splicing using all annual
overlaps*
T - Identifies the point at which multiple time series
versions are linked by ratio splicing using all annual,
quarterly, and monthly overlaps*
U - Indicates that original data value is not to be overlaid
with subsequent calculated data
V - Indicates data that are new or have changed since
the previous issue of the publication; does not indicate
changes in dimension, unit, decimal, or the addition of
countries to IFS*
W - Indicates that data value is not available and should
not be calculated from other data
X - Identifies data observations estimated from
lower-frequency data by extrapolation routines*
_________
* These indicators are automatically generated.
D.4. Country/Area Name: A field which names in English the
reporting country for which the data pertain.
D.5. Coverage (number of years): The number of years for which
data observations are available for a given series.
D.6. Decimal Code: A one-position numeric code that indicates the
number of decimal places assumed from the right-most position of the
data for that time series. The valid codes are 0 - 6.
D.7. English Descriptor: A field which describes in English the
subject of the time series.
D.8. Equation Level: An alphabetic letter that specifies the order
of calculation of nested equations. Allowable values are A through
Z, where A level equations are calculated first, followed by other
level equations in alphabetical order.
D.9. Frequency: The time structure of a series. Valid values are
A, Q, and M for annual, quarterly, and monthly data respectively.
D.10. Magnitude: A field which describes in English the magnitude
of the time series (e.g., thousands of n.c. units, billions of SDRs,
etc.).
D.11. Magnitude Code: A one-position alphanumeric code which
indicates the unit of measurement (weights, volumes, currencies,
etc.) and the scale (units, thousands, millions, etc.) of a series.
The valid codes are as follows:
A - (specified national currency [n.c.] units) per U.S. dollar
B - (specified n.c. units) per U.S. cents
C - U.S. dollar per (unit of specified n.c.)
D - U.S. cents per (unit of specified n.c.)
E - (specified n.c. units) per SDR
F - SDRs per (unit of specified n.c.)
G - Metric tons
H - Thousand metric tons
I - Thousand barrels
J - Kilograms
K - Thousand long tons
L - Long tons
M - Millions fine troy ounces
N - Thousand cwt (100 pounds)
O - Thousand cubic meters
P - Thousand cwt (112 pounds)
Q - 100 pounds
R - 1000 pounds
S - Millions of SDRs
T - Billions of SDRs
U - U.S. dollars per metric ton
V - U.S. dollars per (unit of unspecified n.c.)
W - Fine kilograms
X - Thousand fine kilograms
Y - Thousand fine ounces
Z - Fine ounces
0 - Unspecified units (generally named in descriptor)
1 - Millions of U.S. dollars
2 - Millions of (specified n.c. units)
3 - Billions of U.S. dollars
4 - Billions of (specified n.c. units)
5 - Percent per annum
6 - Index number
7 - Thousands of U.S. dollars
8 - Thousands of (specified n.c. units)
9 - Millions
> - Trillions of (specified n.c. units)
$ - U.S. cents per pound
* - U.S. dollars per short ton
# - U.S. dollars per barrel
& - U.S. dollars per bushel
; - U.S. dollars per fine ounce
- - Millions of metric tons
/ - U.S. dollars per pound
) - Million barrels
( - SDRs per fine ounce
_ - U.S. dollars per 100 yards
= - Thousand gallons
{ - ECUs (European currency units) per (unit of specified n.c.)
} - (specified n.c. units) per ECU
D.12. National Language Descriptor: A description of a time series
usually given in a national language (e.g., French, Spanish, etc.).
D.13. Scale: The scale factor used for the default presentation of
the data in decimal format (i.e., if the default presentation is in
millions of units, the scale factor will be 6.)
D.14. Source: A field that describes in English the source of the
time series (see examples of the source in D.15. below).
D.15. Data Source Code: A one-position alphabetic code used to identify
either the IMF publication in which the data are published or place
from which the data originated. This code is separate and is
additional to the control source code included in the time series
code. This code generally indicates the source of the data, whereas
the control source code generally indicates the IMF publication. The
following list identifies the valid data source codes and control source
codes:
Central bank annual report A
Central bank bulletin B
IMF area departments or Research Department C
Direction of Trade Statistics (DOTS) D
Organization for Economic Cooperation and Development E
International Financial Statistics (IFS) F
Government Finance Statistics Yearbook (GFSY) G
Exchange and Trade Relations Department H
IBRD commodities (original sources) J
IBRD commodities (calculated from J) K
International Labor Organization L
Data supplied directly to IMF by national authorities N
Balance of payments series (national source) P
Balance of Payments Statistics (BOPS) Q
Balance of payments (historical data) R
Statistical office's publications S
IMF Treasurer's Department T
United Nations U
National Bureau of Economic Research Z
D.16. Splice Code: A code that indicates the type of splicing that
was employed to create a series. Valid codes are B for butt
splicing and R for ratio splicing.
D.17. Topic: A field that describes in English the topic
(subject) of the series (e.g., prices, national accounts, etc.).
D.18. Type Code: A one-position alphanumeric code that identifies
the data in the time series as stock, flow, or average. This code
also defines the way the months relate to the quarters and the way
the months and quarters relate to the years. That is, this code
indicates whether lower-frequency data can be or are derived from
higher-frequency data. It also identifies the end of the fiscal
year of the time series data. The following table provides
information on this code ("conditionally" means derivation is
performed only if the desired observation does not already exist).
Type End of Quarters Derived Annual Derived
Code of Data Fiscal Year from Months from Quarters
----------------------------------------------------------------
A Average December Always Always
B Average March Always Always
C Average June Always Always
D Average September Always Always
F Flow December Always Always
G Flow March Always Always
H Flow June Always Always
I Flow September Always Always
J Average December Conditionally Never
K Average December Always Never
N Flow December Conditionally Never
P Flow December Always Never
R Any March Never Never
S Stock December Always Always
T Stock March Always Always
U Stock June Always Always
V Stock September Always Always
W Stock December Conditionally Never
X Stock December Always Never
1 Average December Conditionally Conditionally
2 Average March Conditionally Conditionally
3 Average June Conditionally Conditionally
4 Average September Conditionally Conditionally
5 Flow December Conditionally Conditionally
6 Flow March Conditionally Conditionally
7 Flow June Conditionally Conditionally
8 Flow September Conditionally Conditionally
9 Stock December Conditionally Conditionally
blank Any Any Never Never
D.19. Update Date: The date on which the time series was last updated in
the form YYMMDD, where YY is the year, MM is the month, and DD is the day
of this date.
D.20. Update Day: The day on which the time series was last updated in
the form DD.
D.21. Update Month: The month in which the time series was last updated
in the form MM.
D.22. Update Year: The year in which the time series was last updated in
the form YY.
Back to IFS User Guide