An exogenously defined poverty line yields poverty headcounts between any two points in time that are a net outcome of the two-way traffic into and out of poverty. This paper argues that, for the rural Indian context, where housing is too lumpy and illiquid to be used for consumption smoothing, transitions in housing quality in cross-sectional data sets can provide revealed evidence of household perceptions of downside risk to their current consumption levels. Using the two most recent National Sample Survey housing surveys (the 58th round for 2002 and the 65th round for 2008-09), composite housing quality classifications are unbundled, and binary wall quality is selected from cross-quartile behaviour as the feature most responsive to rising household consumption levels. In both rounds, the incremental move to better quality declines beyond the consumption level at which half of all households are in better quality structures. The threshold consumption level at which this happens was lower in 2008-09 than in 2002 and reflects an improvement in housing conditions over the period. However, this effective saturation of the demand for the most basic element of better housing, much before actual saturation, provides a quantitative measure of the percentage of households even in the topmost quartile that fears downside consumption risk.