DARTS uses information from a representative sample of the Russian population taken from the Russia Longitudinal Monitoring Survey (RLMS) for the year 2000. The survey covers just over 3000 households and records a great deal of information about each household and about its individual members: incomes from different sources, employment history, home production, spending on different types of goods, housing and so on. DARTS uses this information to calculate the likely effects of tax and benefit changes on each household, and then aggregates this information over all households to produce a picture of how the changes affect the Russian population as a whole. population as a whole.
The richness of the RLMS data allows a wide variety of interesting and informative output to be produced. DARTS automatically generates tables and graphs that break down the results of alternative tax and benefit scenarios by income decile (or, more precisely, deciles of living standards), age group, region and family type. For each of these subgroups, the results report the numbers of gainers and losers, average gains and losses in roubles, and changes in the numbers of individuals in poverty.
DARTS was developed at UNU-WIDER by André Decoster, Tony Shorrocks and Inna Verbina. This online version was produced by Graham Stark and Jocelyn Paine of the Virtual Worlds Group in collaboration with the original authors and Bruck Tadesse of UNU-WIDER.
The tables and charts showing gainers and losers indicate the numbers of Russian households in thousands. These figures are calculated by "grossing up" the households in the RLMS data to represent the Russian population as a whole. Since there are just over 52 million households in Russia, and just over 3,200 households in the sample, each RLMS household is deemed on average to represent around 16,500 household in the Russian population, but the actual weights for each household type differs depending on whether we think there are too many of too few of that type of household in the population compared to the sample.