Page 287 - Invited Paper Session (IPS) - Volume 2

IPS246 Tiziana Laureti et al.
In this paper we present an application of uncertainty in measuring economic
well-being at local level, a crucial aspect that impacts both the design and
evaluation of policies. We focus on the issue of measuring the accuracy of at
risk of poverty rate (AROP) which is defined as the proportion of people with
an equivalised total net income below 60% of the national median income and
estimated from the European Union Statistics on Income and Living
Conditions (EU-SILC) surveys. After reviewing the literature devoted to the
computation of standard errors of traditional measures of monetary poverty
and to the assessment of non-sampling errors, we estimate AROP standard
errors for Italy taking into account the complex survey design and using
linearization methods. Moreover, we provide estimates of standard error of
AROP at regional level (NUTS-2) by also using equivalised income adjusted for
price differences in order to better inform policy making. We use the 2017
wave of EU-SILC in which detailed information on the sample design are
provided.

Keywords
Data uncertainty; well-being; poverty measures

1. Introduction
Despite the seminal studies of Kuznets (1948) and Morgenstern (1950)
addressing the issue of uncertainties in economic statistics, the
communication of data uncertainty is still a widely neglected problem.
National Statistical Offices (NSOs) and other statistical agencies often publish
official economic statistics as point estimates making little mention of
uncertainty in the reported estimates. However, pretending an accuracy that
is unrealistic may lead us to a misinterpretation of the statistical results and
erroneous conclusions by the users. As Manski (2019) underlined, point
estimates may be viewed as true but are not necessarily true. Thus, in the
absence of agency guidance, some users of official statistics may simply
assume that errors are small and inconsequential. Contrastingly, users of
official statistics who understand that statistics are subject to error may
conjecture the error magnitudes, thus misinterpreting the information that the
statistics provide.
Therefore, in late 2018 Eurostat launched the COMUNIKOS project
(COMmunicating UNcertainty in Key Official Statistics) which aims to conduct
an in-depth methodological and empirical evaluation of a number of
alternative approaches to measuring and communicating uncertainty as well
to propose new methods and metrics to measure and communicate
uncertainties for official statistics.

274 | I S I W S C 2 0 1 9