Using Spatial Econometric Techniques to Analyze the Joint Employment Decisions of Spouses

Studies of the joint time-use decisions of spouses have relied on joint estimation of time-use equations, sometimes assuming correlated errors across spouses’ equations and sometimes directly examining the effects of one spouse’s time use on another’s, relying on panel data or instrumental variables techniques to account for endogeneity. However, panel data often are not available and available instruments often are not satisfactory, making examination of the direct relationship between spouses’ time use difficult. Spatial econometric techniques applied to cross-sectional data do not require instrumental variables. This study estimates both a Spatial Autoregressive (SAR) Model and a Spatial Durbin Error Model (SDEM) to examine the labor hours of husbands and wives in dual-earner couples using the 2012 Annual Social and Economic Supplement to the Current Population Survey (ASEC). In these models, each spouse is treated as a direct “neighbor” of the other in a spatial weight matrix and non-spouses are treated as non-neighbors. Estimates of both the own- and cross-wage effects on labor hours and an estimate of the direct relationship between spouses’ labor hours are obtained.