Transcript The collection of monetary, financial and market statistics by the

BANK OF GREECE
STATISTICS DIVISION
JOINT NATIONAL BANK OF THE REPUBLIC OF
MACEDONIA/ECB SEMINAR
From ITRS to Direct Reporting
3 OCTOBER 2013
Alexandros Milionis
1
External Sector Statistics
• Balance of payments and International
investment position
• FDI and related statistics
• International reserves and foreign
currency liquidity templates
2
DATA COLLECTION SYSTEMS (1)
PREHISTORY:
• Up to the1990’s data collection was based on an
exchange control framework and banking statistics
(aggregated data)
HISTORY
• Since 1999 a new collection system in effect following
the methodology of BPM5 of IMF resident/non-resident
transaction by transaction.
• For a detailed presentation of the methodology,
sources and output for BoP data for all EU countries ,
see ECB’s publication ‘European Union balance of
payments/international investment position statistical
methods’ May 2007
3
DATA COLLECTION SYSTEMS (2)
TODAY
Present data collection system at the Bank
of Greece - Mixed:
• Main data source: International transactions
reporting system (ITRS) –
• Direct Reporting: For the following BOP items
only:
4
DATA COLLECTION SYSTEMS (3)
Direct Reporting sources of information
1. Frontier travel survey (travel item)
2. Oil refineries (oil account)
3. Mutual funds
4. Investment companies
5. Stock exchange firms
6. Custodians and end-investors stock data (s-b-s)
7. Central Securities Depository
8. Annual FDI survey
9. Annual survey on external assets and liabilities
10. Residents non-residents transactions without the
intermediation of resident banks
5
DATA COLLECTION SYSTEMS (4)
Other sources of Direct Reporting data:
1. Data from BoG’s departments for:





current transfers
general government
reserve assets
electronic secondary securities market (HDAT) – bonds/MMI
Consolidated Balance Sheet of MFI’s
2. General Accounting Office (Ministry of Finance)
3. ELSTAT, National Accounts and Foreign Trade
4. BIS (household’s data)
6
REASONS FOR ITRS
ABANDONMENT
 PRAGMATIC


Insufficiency (transactions without a resident-MFI
intermediation)
Inaccuracy (including miscoding, funnel effect, etc.)
 LEGAL
Exemption threshold of declaration at European
level (Reg EC 2560/2001)
– 2002: 12 500 €
– 2010: 50 000 € (Reg EC 924/2009)
– 2016: full exemption (Reg EC 260/2012)
7
TWO STRIKING EXAMPLES OF
INEFFICIENCY/INACCURACY OF ITRS
 LOANS GRANTED BY NON RESIDENT MFIs TO
RESIDENT ENTERPRISES (further details later)
 TRAVEL RECEITS
TRAVEL RECEITS-GREECE
2500
2001
MIL. EURO
2000
2002
1500
1000
500
0
JAN
FEB
MAR
APR MAY
JUN
JUL
AUG
SEP
OCT
NOV
DEC
8
DIRECT REPORTING
POPULATION OF POTENTIAL DIRECT
RESPONDENTS (1)
 FINANCIAL SECTOR
• MFIs
• Unit trusts
• Insurance Companies
• Pension funds
i.e. small and well defined population sweeping
9
POPULATION OF POTENTIAL DIRECT
RESPONDENTS (2)
NON FINANCIAL SECTOR
much larger and widely dispersed
populationsampling
OTHERS
•
•
•
•
Public Sector
Individuals (households)
Embasies
Notaries
10
METHODOLOGICAL ISSUES
• Two characteristics of practical importance:
1) Amount of BOP transactions “concentrated” in
a relatively small number of transactors. This is
more evident in financial companies.
2) “Concentration” quite stable over time.
Therefore complete sweeping may not be
imperative even for financial companies.
11
Non-financial companies -sampling
• Take advantage of some stylized facts and form
selection criteria.
• For non financial companies the sector of activity
(e.g. NACE classification) and size should be
taken into account for the selection of the
sample.
• For companies of congenial activities the
probability of performing BOP transactions as
well as the amount of transactions are related to
companies’ financial data.
12
METHODOLOGICAL ISSUES
• Define a dichotomous dummy variable,
valued 1= BOP transactions, 0=no BOP
transactions, as the dependent variable
and financial data of companies as causal
factors (independent variables).
• Use two-group discriminant analysis for
(possible):
Discrimination
Forecasting
13
Discriminant analysis
Geometric representation
14
AN EXAMPLE ANALYSIS
LOANS GRANTED TO RESIDENT COMPANIES
BY NON-RESIDENT MFIs
• Sample size: 1050
• Response rate: 88%
15
AN EXAMPLE ANALYSIS (2)
• Some Results
Hotelling’s T-squate statistic =0.993
(statistically significant at 1%)
That means discrimination is possible
16
AN EXAMPLE ANALYSIS (3)
• So total assets may be used as the
primary selection criterion
17
AN EXAMPLE ANALYSIS (4)
• A simple model
18
AN EXAMPLE ANALYSIS (5)
• Basic Reference:
Karapappas A. P. and Milionis A. E.
(1999): Estimation and Analysis of external
debt in the private sector, Economic
Bulletin, 14, pp43-53, Bank of Greece,
19
• Existing ITRS data with transactors very
useful!
The initial “core” of the direct reporters
registry.
Assessment of the quality of statistical
models.
Use the statistical models to update the
registry.
20
POTENTIAL DIFFICULTIES WITH DIRECT REPORTING
•
•
•
•
•
•
•
•
TRANSACTIONS BY PHYSICAL ENTITIES IN GENERAL
GOODS NOT CROSSING THE FRONTIERS
TRANSPORTATION SERVICES – SEA TRANSPORT
INCOME-COMPENSATION OF EMPLOYEES
CURRENT TRANSFERS- WORKERS’ REMITTANCES
INVESTMENTS IN REAL ESTATE
DEPOSITS OF GREEK RESIDENTS ABROAD
DEPOSITS OF NON-RESIDENTS IN GREECE
21