A further source of unwanted volatility is the conversion of GNI to US dollars, since standard market rates can fluctuate in the short term. However the smoothing “Atlas” method works largely as intended. The use of purchasing power parity exchange rates would likely be a further improvement that would provide a better basis for crosscountry comparisons of GNI. But data availability and reliability continues to limit their use for this particular purpose the sizeable revisions at each “benchmark” round of the International Comparison Project make PPP estimates currently unsuitable for an annual classification system (see this selection of GNI per capita estimates using PPPs to see the impact of PPP revisions from the last three benchmark years of the ICP: 1993, 2005, and 2011.)