Executive function (EF) and mathematics abilities are consistently found to be related in early childhood, but several potential explanations for their longitudinal associations exist.  Proposed theoretical explanations include: 1) early EF skills lead to later mathematics abilities, 2) early mathematics skills and training lead to later EF abilities, 3) the effects between EF and mathematics are bidirectional, or 4) longitudinal associations between EF and mathematics are due to factors that similarly influence the development of both skills across time.  In a longitudinal study of children’s EF and mathematics development from the fall of prekindergarten to the spring of kindergarten, Duncan estimated a model that allowed for the contribution of early skills in both constructs to later skills in the same construct and to the other construct, and for relatively stable factors to influence the development of EF and mathematics across time.  Results suggest that EF and mathematics associations are largely due to the combination of characteristics with a stable influence on these constructs across this developmental period, r = .84, and not to immediately preceding time-varying skill level on the other construct.  Implications for theories of the co-development of EF and mathematics achievement in early childhood are discussed.