Towards 2020

Editorial

(1) Harmonize the 2020 housing types with the EU specifications so that there is some overlap of specifications between the two lists. Konijn is not optimistic about this approach because the EU uses national weights for each specification, whereas there is only implicit weighting in the other regions to the extent some countries will not have rents for some specifications.
(2) Continue to gather country information on sample coverage in terms of the population where the sample rentals were taken. And following on (1) above, ask countries for any information they have on stock of rental and total housing stock for the ICP specifications.
(3) Based on coverage information in (2) seek ways to incorporate that information to modify the urban rental CPD information so it can be extended to all urban housing in urban areas and perhaps to the whole country.
(4) An alternative to (3) is to treat the stock of rural housing separately by applying a rough version of user cost to impute the rental volume. The urban and rural volumes could then be simply summed to obtain a national total for each country.
Housing: 5 (5) Following the initiatives of the ABD, research ways to improve the quality of housing measures used in the direct volume approach, or perhaps modifying country quality measures in a downward direction as was suggested in ADB19. The reason for suggesting this is the large gap between rental volumes using the indirect method (lower) and the direct method (much higher).
(6) Explore further the proposal in ABD19 of combining the physical quantity measures like number of rooms and number of dwellings (p.8, CQI_2) when applying the direct volume approach.
(7) In regions other than the EUO, look into the efficacy of combining the indirect and direct housing quantity estimates for each country as in ADB19, rather than choosing one or the other. This approach is consistent with the idea of treating rental and OOH as separate basic headings for which independent estimates are made.

Regional Shares of World GGP
• There are GCI (Global Core Items) and the non-GCI BHs. My comments will not deal with non-GCI BHs because the GO treats them specially anyway. • What is done now? Each GCI price is divided the BH PPP of the reference country PPP for each region from the regional comparison. A region-item regression provides estimates of linking factors for each BH.
• A variant takes the geo-mean of the GCI and Regional positions for each country BHs.
The simplified GCI approach: what I always thought should have been done would begin with GCI prices only. A CPD would be run over all countries for each basic heading to yield price levels for each country. Fixity: Regional Shares (cont.) • Combining the expenditures and the GCI price levels and non-GCI BHs PLs and expenditures in an EKS aggregation yields country price levels and GDPs in a common reference currency. • This directly provides the regional shares of global GDP in the common reference currency…no regional data so far. • Fixity can now be achieved in each region by distributing the regional total among the countries according to their regional results.