Page 166 - Special Topic Session (STS) - Volume 3

STS538 Ibrahim S Y.
(ii) Housing (H) - relates to rent and other housing-related costs for the
primary dwelling at the location of assignment, including costs for
maintenance, utilities, and other housing costs. Its weight is based on staff-
reported survey data. Its index is based on market rent data or staff reported
data.
(iii) Pension Contribution (PC) - relates to the amount of pension
contribution paid by staff, obtained from administrative sources. Its weight is
the fixed amount expressed as a percentage of a reference net remuneration
(that of an average staff member);
(iv) Medical Insurance (MI)- relates to the amount of insurance
premium paid by staff, obtained from administrative sources. Its weight is the
average medical insurance premium paid by staff, and the index is the ratio of
this average premium at the duty station versus New York;
(v) Out-of-Area (OA)- relates to expenditures outside the country of
assignment. Its weight is determined from out-of-area expenditures reported
by survey respondents, and the index is estimated as a weighted arithmetic
average of US dollarized CPIs of 26 selected countries.
The first methodological challenge is how to derive the overall weight of
the PAI and allocate it to its five major components. The accuracy of the weight
depends on the quality of expenditure data reported by staff, and this has
challenges of its own. Since the surveys are self-administered and voluntary,
with no requirement for diaries, there is a risk of both recall and telescoping
errors. In any case, due to a variety of reasons, including response burden,
recall problems, vested interest, and confidentiality, it is simply not practicable
to obtain good-quality information regarding all sources of household income
or all household expenditures. In view of all these considerations, and the
compensation context of the PAI, for which the measurement target is the
average household, the overall weight of the PAI is set equal to the net
remuneration of an average staff member, at the time of the survey.
Another methodological challenge is how to aggregate these five
components in a way that produces an index that simultaneously satisfies the
requirements of a cost-of-living index, and desirable compensation policy
objectives, including the overarching requirement that duty stations not be
disadvantaged in a uniform or systematic way. These policy requirements
effectively eliminate consideration of most of the superlative indices (see
Diewart (1976), for details), which, if applied in the spatial context of the PAI,
would lead to systematic decreases in PAIs for comparison duty stations, in
large part, due to the wide disparity in the weights and sub-indices between
the base and most of the comparison locations, as well as a certain lack of
substitution of consumption among the highest level components of the PAI.
The aggregation formula for the macro-components of the PAI, as well as
the subcomponents of the only two macro-components with an internal

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