Page 73 - Special Topic Session (STS) - Volume 3

STS515 Steve MacFeely
regression must be included, but so too should some basic, but often
ignored, statistical techniques. For example, data cleaning (including
treatment of outliers, imputation and interpolation) is an essential skill.
Unfortunately, NSOs rarely get to work with the clean datasets favoured by
university courses. Real life data are typically very messy. Another technique,
too often ignored, is seasonal adjustment. Essential for analyzing and
presenting sub-annual time series, yet frequently young career statisticians
don’t know how to seasonally adjust time series or how to test time series
models to ensure they are appropriate. More could be done to examine real
life issues, such as, how to seasonally adjust series in the aftermath of shocks,
such as the 2008 financial crisis. Another area deserving of more attention
is index theory and practice – the various index formula and where it is
appropriate to use them. Fixed weight versus chain linking. These are
essential for statisticians working in price, business and macroeconomic
statistics. Increasingly NSOs and IOs are compiling leading, composite and
sentiment (LCS) indices - the merits and technical challenges of such
9
indices could usefully be discussed. In a ‘big data’ world, with massive
computing power facilitating the linking of records, safeguarding
confidentiality and public key cryptography are becoming greater
challenges. The importance of these subjects deserves a prominent place in
any statistics curriculum. Students should also be introduced to
programming logic, not just how to use the most fashionable software
packages.
4.3 Policy Issues: Beyond technical statistical techniques and tools, universities
could play an important role in highlighting some of policy issues that pose
significant challenges for public policy (and by extension NSOs). For the
purposes of this paper 4 issues are highlighted, but this is by no means
exhaustive. The first is globalization. With the emergence of the internet, the
fall of the Berlin wall and China joining the WTO, properly measuring the
impacts of globalization is now a major issue for statisticians. Economic and
social globalisation is impacting on employment, crisis contagion, trade
policy, protectionism and migration [7]. It has also challenged the relevance
of traditional trade, price and macro-economic statistics. The next issue is
measuring wellbeing (aka progress).
Since the 1970’s there have been several attempts to supplement or
supplant GDP with other indicators of progress . Since the 2008 financial
10

9 Some examples: UNDP Human Development Index; OECD Better Life Index; WEF Global
Competitiveness Index.
10 The Measure of Economic Welfare (MEW); the Genuine Progress Index (GPI); the Human
Development Index (HDI); and index of Gross National Happiness (GNH) to name a few.
62 | I S I W S C 2 0 1 9