In producing the Supplemental Poverty Measure (SPM) thresholds and subsequent statistics, it is assumed that all consumer units, regardless of size and composition, share the same fraction of the thresholds on housing (shelter and utilities).  The implication of this assumption is that the implicit economies of scale for housing are the same as those for the thresholds as a whole.  If, on the other hand, one assumes that housing expenditures are subject to greater economies of scale than the food and clothing parts of the thresholds, it would be reasonable to use a larger percent to identify the housing portion of the thresholds for smaller families. This would have two consequences for SPM poverty statistics.  First, the portion of the SPM thresholds subject to the geographic adjustment would be larger for smaller families --- increasing thresholds for those who live in areas with housing costs greater than the national median and decreasing thresholds for those who live in areas with lower housing costs  Second, since the values of housing subsidies in SPM resources are capped at the housing portion of the thresholds, this would increase the value of housing subsidies for some smaller consumer units and could reduce their poverty rates. In this paper we investigate the impact of varying the housing share of the SPM poverty thresholds directly by changing housing directly first and then indirectly by applying differing equivalence scales by consumer unit size.  American Community Survey (ACS) and U.S. Consumer Expenditure Interview Survey (CE) data are used to explore how housing expenditures as a share of income and expenditures on food, clothing shelter and utilities (FCSU), respectively, vary by consumer unit size. Data from the CE are also used to estimate equivalence scales; these scales result in indirect adjustment to the housing shares. The Current Population Survey Annual Social and Economic Supplement data are used to analyze the impact of allowing the housing share of the thresholds to vary on SPM poverty rates. Results suggest that choice of the housing shares (and equivalence scales) has very little impact on either overall poverty rates or the impact of housing assistance on poverty.