How should the economic performance of the Central Asian countries during the 1990s be measured? The most commonly used indicator is GDP, even though GDP estimates for transition economies are known to suffer from conceptual and methodological shortcomings. This paper combines national accounts estimates and household survey results to assess the biases in GDP measures as indicators of economic well-being. The main conclusion is that, although per capita consumption fell during the first half of the 1990s, the decline was not as drastic as GDP estimates suggest. Although the cardinal measures are subject to scepticism, the ordinal ranking of the five Central Asian countries' performance is clearer, at least to the extent of dividing them into better (Kazakhstan, the Kyrgyz Republic and Uzbekistan) and worse (Tajikistan and Turkmenistan) performers.