We report the effects of framing settings in annuity demand after conducting a survey-based experiment with members of a Dutch occupational pension plan. We gave participants the option to allocate up to 20% of their projected pension accrual to a lump sum. In particular, we investigated the joint effects of consumption and investment frames and gain and loss frames. We present strong evidence for framing effects in annuity demand. Framing effects remain significant when we control for individual characteristics. We also find robust evidence of individual characteristics influencing annuity demand, highlighting the importance of heterogeneity among participants, for example risk aversion, time preference and trust in the pension fund. Gender and long-term debt positions have significant impact how one responds to framing. We conclude that Dutch plan members generally welcome the partial lump sum option over mandatory full annuitization. The application of frames appears to predictively steer annuity demand. The precise effect framing may have, would probably also depend on the institutional environment, which predefines the perspective through which individuals filter annuities.