Page 119 - Contributed Paper Session (CPS) - Volume 7
P. 119

CPS2043 Adnan Dawood K. B. et al.
                            Table (1) Leakage and Under-coverage Rate calculations
               Predicted Welfare Level By            Actual Welfare Level
                      the Formula
                                                Poor                Non-Poor         Total
                                                  5
                                           Under-Coverage
                      Non-Poor             (Exclusion error)           65             70
                                                               Successful Targeting
                                            (Under-coverage
                                              rate=0.25)
                                                                       15
                                                 15                  Leakage
                         Poor                                                         30
                                         Successful Targeting    (Inclusion error)
                                                               (Leakage rate=0.50)
               Total                             20                    80            100
               There are no such predetermined rates for the leakage and under-coverage
               telling whether to accept the formula or to reject it, more reducing for these
               rates  more  efficiency  of  the  formula.  In  the  process  of  constructing  the
               regression formula for targeting the poor, actions are taken to reduce these
               rates,  but  when  reducing  the  leakage  rate  we  could  found  that  the  under
               coverage increased. Alternatively, when trying to reduce the under-coverage
               rate we found that the leakage rate increased.

               2.  Methodology
                   This  enhanced  methodology  of  the  PMTF  is  based  on  dividing  the
               population into four categories of households; the first is the High Welfare
               Level  (HWL)  which  don’t  have  poor  households,  the  second  is  the  Middle
               Welfare Level (MWL) which could include poor households, the third segment
               is the Lower Middle Welfare Level (LMWL) and could include poor households,
               and the fourth is the Bottom Welfare Level (BWL) in which most of it are poor
               households. Determining these four segments is mainly based on selecting
               three cutoff points on the natural logarithm of annual per capita expenditure.
               Each  cutoff  point  represent  a  step  on  the  scale  of  the  annual  per  capita
               expenditure consumption. To select these cutoff points we need knowledge of
               the poverty line and poverty incidence. The detailed methodology is explained
               through the following process:
                   Based on the poverty line and the poverty incidence in the society, we
               select a cutoff point (call it CP_1) on the natural logarithm curve for the annual
               per capita expenditure consumption. This CP_1 should be approximately ten
               per cent above the poverty line. This cutoff point divide the population into
               two welfare levels, High Welfare Level (HWL) which represent the non‐poor
               households, and the Lower Welfare Level (LWL) which represent the rest of the
               population (households).


                                                                  106 | I S I   W S C   2 0 1 9
   114   115   116   117   118   119   120   121   122   123   124