We used a longitudinal design to investigate the hypotheses that the components of the Job Demands-Control-Support model and changes in their levels over time predict subsequent changes in levels of positive affect of vigor over time, and vice versa. Our study was conducted on a sample of adults working in a variety of occupations (N = 909, 68% men) at three points in time (T1, T2, and T3), over a period of about four years, controlling for neuroticism and other potential confounding variables. Job control at T1 and increase in its levels from T1 to T2 predicted an increase from T2 to T3 in the levels of vigor, whereas for social support, only its level at T1 predicted an increase from T2 to T3 in levels of vigor. An increase from T1 to T2 in levels of job demands predicted an increase from T2 to T3 in levels of vigor only for those rated low on neuroticism. Vigor at T1 predicted an increase from T2 to T3 in levels of job control and social support, but not changes from T2 to T3 in levels of job demands. The reciprocal causal relationship between job resources and vigor exists regardless of the demands of the work environment.