Transcript PPTX - United Nations Statistics Division

Real Sector Division
IMF Statistics Department
IMF Statistical Forum and Big Data
in National Accounts
10th Meeting of the Advisory Expert Group on National Accounts
April 13-15, 2016
Paris, France
Thomas F. Alexander
IMF
Reproductions of this material, or any parts of it, should refer to the IMF Statistics Department as the source.
Real Sector Division
IMF Statistics Department
Big Data and Micro Data
 Third IMF Statistical Forum
• Big data and micro-data links a sub-theme of the Forum
•
•
•
The terms big data and micro-data are often used interchangeably
although they refer to slightly different phenomena
Big data: very large datasets which may be analyzed
computationally to reveal patterns, trends, and
associations. Sources of big data include social media,
transactional data, sensor data, and public data.
Also:


IMF Big Data and Analytics Symposium - November 2015
UN Global Conference on Big Data for Official Statistics - October 2015
2
Real Sector Division
IMF Statistics Department
Big Data and Micro Data
 Micro-data: unit-level data obtained from sample
surveys, censuses, and administrative systems. They
provide information about characteristics of individuals
or units such as households, business enterprises,
facilities, farms or even geographical areas such as
villages or towns (an example would be the wages of
individuals in a given company for a set period).
3
Real Sector Division
IMF Statistics Department
Big Data and National Accounts
Big data in national accounts featured throughout the Forum—
GDP estimates; GDP forecasts; sector accounts
 Session III: Statistics on a From-whom-whom Basis
• Focused on the need for granularity (by counter-party)
• Collection of data on a loan-by-loan, security-by-security,
trade-by-trade, etc. basis
Europe - Centralized Securities Database and the Securities Holdings Statistics
Database; Analytical Credit Dataset (anacredit)

•
Session V: Measuring the Material Conditions of Households
Examined
the need to go beyond estimates of GDP to examine household
well-being.
 Household accounts compiled from micro-data in Thailand (Samphantharak)
 Use of administrative data as a replacement for survey data (Guiso)
4
Real Sector Division
IMF Statistics Department
Big Data and National Accounts
 Session I: The Relevance of Micro-data for Evidenced Based
Policymaking
 Collection of micro data to measure the heterogeneous
expectations of individuals and the usefulness of point
estimates as opposed to probabilistic expectations (Manski)
•
Comparisons of point predictions of GDP growth with the
subjective probability distributions held by forecasters


Forecasters report point predictions that give a more favorable view of
the economy than do their subjective means/medians/modes
The deviations between point predictions and the central tendencies of
forecasters’ subjective distributions tend to be asymmetric
5
Real Sector Division
IMF Statistics Department
Micro Data Hubs
 Session II: Big data and micro-data hubs
 Examined the importance of micro-data hubs to make
available to users, granular data that may be confidential
or otherwise not shared at the national level
•
•
•
The BIS Data Hub: Usefulness in capturing balance sheet
linkages
Proper aggregation requires underlying granularity
“exposure” gross vs net exposures
Corporate hierarchy (who owns whom?)


But how do data hubs address the sharing of data across countries
Even if data are shared across countries; how to address different metrics at
the granular level
6
Real Sector Division
IMF Statistics Department
Big Data and Analytical Use
 The availability of micro data to a broad spectrum of
users has enhanced analyses in a wide range of areas
including unconventional monetary policy, financial
stability, natural resource management, fiscal policy, and
monitoring household wellbeing.
 Understanding interconnectedness and spillovers across
countries and institutions—crucial for monitoring
financial stability—requires granular and real time data.
7
Real Sector Division
IMF Statistics Department
Role of Statistics Agencies
 The relationship between official statistics and big data is
evolving.
•
•
Statistical agencies are well positioned to exploit big data
while maintaining their role as providers of the official
macroeconomic statistics, with assurances of strong
internationally accepted methodological foundations and
integrity.
Statistical agencies need to establish mechanisms to facilitate
the sharing of underlying micro data with users, including for
the monitoring of the digital economy.
8
Real Sector Division
IMF Statistics Department
Role of Statistics Agencies
 The legal and privacy constraints of using micro data

must be addressed. Statistical agencies and policymakers
urged to establish new legal frameworks to support
micro-data access while preserving confidentiality. Micro
data provided by private institutions, including social
media, also need legal underpinning. Once available,
micro data can be powerful tools for policy formulation.
Micro-data viewed as key to timely analysis and policy
response for everything from unconventional monetary
policy, financial stability, natural resource management,
fiscal rules, and household well-being, income and
wealth distribution.
9
Real Sector Division
IMF Statistics Department
Key Takeaways
 New technologies are key components of micro-data
provision, dissemination, and analysis. They include IT,
efficient data transmission via SDMX, and the associated
further harmonization of the accounting concepts and
legal identifiers.
 Collaboration and coordination of the national and
international statistics communities, as well as academic
and policy communities, are key to building "micro-data
highways".
10
Real Sector Division
IMF Statistics Department
Key Takeaways
 We may be at a defining period for micro-data, yet the
road ahead is challenging. Increased access to microdata brings a wealth of opportunity, but also raises
questions concerning privacy, reliability, and coordination
- IMF DMD Min Zhu.
 Basic principles of statistics still hold whereby
trustworthy, informative and accessible data are
fundamental to good policy decisions. Statistics support
accounting and independence through transparency -STA
Director LM Ducharme.
11
Real Sector Division
IMF Statistics Department
Comments!
12