Accounting for Product Quality Differences in PPP Measurement
The ICP approach to international comparisons is founded on the basic principle of comparing prices of like with like across countries. Accordingly, considerable effort goes into the determination of the basket goods and services for price comparisons at the regional level; and in the calibration of global core product list for global linking. Achieving a balance between comparability of the goods priced and at the same time ensure that the products are representative or important in the participating countries is well recognized, and ultimately a judicious compromise between these two competing considerations is reached. In order to ensure pricing of identical/similar products across countries, the ICP in its 2005 benchmark comparisons introduced SPDs, which attach a list of price determining characteristics associated with each product or service priced. SPDs for products are more narrowly defined in the case of Eurostat and OECD and the SPDs are more flexible in other regions like Africa.
There is a growing realization that despite the care taken in product specification and at the time of price collection, it is still difficult to ensure that there are no differences in qualities of products actually priced in different countries. Often this aspect is noticed at the price validation workshops where some countries may report lower than expected prices and ensuing discussions raise the issue as to whether the observed differences are a reflection of a lower quality.
A related issue is that even when a product is priced according to exact product specifications, there may be certain service elements that may have influence on price recorded. When it comes to services like medical and educations services, differences in quality are clearly apparent. Similarly, when it comes to transport sector prices, prices for rail and bus transport, the ICP specifications do not account for quality differences in terminal services, frequency and reliability of services. Similar differences in quality differences are generally acknowledged but are placed in the too hard basket at this stage.
Once it is recognized that quality differences may be driving some of the price differences then the PPPs observed are biased as they fail to account for price differences. It is indeed quite possible that the price levels in low income countries may need to be adjusted upwards in order to account for the likely lower quality associated with goods and services. This in turn means real per capita expenditures in these countries would be adjusted downwards leading to increased magnitudes of measures of inequality and poverty. The relationship between quality in manufacturing and services and the Balassa–Samuelson relationship is explored in a recent paper by Zhang (2016).
At this stage, the effect of quality differences on PPPs and the flow-on effects is a matter of conjecture. When researchers use unit values from household expenditure surveys, invariably attempts are made to adjust for possible quality differences in goods and services consumed by households in different income groups and regions. Similar attempts need to be explored and it is important for the ICP to devote resources to devise methods to account for quality differences across countries. The following aspects could be researched:
- Use of extended SPD approach to capture quality differences;
- Explore the possibility of incorporating quality indicators in the CPD regressions – similar to hedonic regressions but using broadly based quality indicators appropriate for each basic heading; and
- Conduct a pilot study in one or two regions for selected basic headings to measure the extent of bias induced by quality differences.