1、Finding the Goldilocks data collection frequency for the Consumer Price IndexLUIGI PALUMBO1AND TIZIANA LAURETI21 BANK OF ITALY2 UNIVERSITY OF TUSCIAThe views expressed in this presentation are those of the authors only and do not necessarily represent the views of the Bank of Italy or the Eurosystem
2、.How often is“just right”?Assumptions:Timing of CPI price collection:prices often vary during the periodData collection,processing,and storage has a costObjectives:PALUMBO AND LAURETI-FINDING THE GOLDILOCKS DATA COLLECTION FREQUENCY FOR THE CPI2Estimate variability of CPI m-o-m variations at differe
3、nt sampling-in-time frequenciesPropose and empirically validate a framework to determine the optimal data collection frequencyElectricity and gas prices in ItalyMandatory publication of unregulated market offers for electricity and gas to guarantee transparency Used by ISTAT for official CPI price d
4、ata collectionSimplified data collection:9 cities National consumption profile Electricity consumption:2700kWh/year Gas consumption:1400 m3/year Selection of rates for electricity Providers weighted by national market shares Type of contracts(fixed or variable prices)weighted according to official r
5、eportsPALUMBO AND LAURETI-FINDING THE GOLDILOCKS DATA COLLECTION FREQUENCY FOR THE CPI3Daily Time-Product-Dummy indexWeighted Time-Product Dummy index:pit:average price of operator i in month tDt:dummy equal to 1 if month is equal to t and zero otherwiseDi:dummy equal to 1 for prices of operator i a
6、nd zero otherwise.Weighted Least Squares using operator i market share as weight for each observation.Aggregate daily price level:PALUMBO AND LAURETI-FINDING THE GOLDILOCKS DATA COLLECTION FREQUENCY FOR THE CPI4ln=1+=1+=Month-on-month CPI variationsMonth-on-month rate of change:Number of potential c