| This study examines why states adopt certain social security programs in the latter half of the twentieth century. Previous research is grounded within one of two perspectives that focus either on internal contingencies or diffusion/globalization processes as the factors behind program adoptions. Internal contingency perspectives propose that variations in particular internal state characteristics lead to the adoption of diverse social security programs. Diffusion/globalization perspectives propose that program diversity results from forces that are external to the state.;In the present study, the adoption of two distinct social security programs—savings system (both public and private) old-age security provisions and means-tested old-age programs—are examined using discrete-time logistic regression models on country-level data for more than 100 nations from 1960 to 1999. The unique approach of this study includes: a larger nation sample than usual; multivariate modeling representing competing perspectives; and a dynamic analysis of the timing of program adoption.;The analyses are conducted in three phases. First, models including single predictors that represent alternative theories from the internal contingency perspective to estimate program adoptions are examined. Second, the same analyses are conducted using single predictors representing the diffusion/globalization theories. Then, multiple predictors from each perspective are used to estimate program adoption. Finally, models incorporating both internal contingency and diffusion/globalization indicators simultaneously are examined.;Results from the separate internal contingency models show support for the gendered bases of entitlement theory for savings system adoptions. For the means-tested program adoptions, interest-group politics theory is supported, though it is qualified by national economic circumstances. Results from the separate diffusion/globalization models support the geographic proximity thesis and the social networks theory for savings system adoptions. For means-tested programs, only the geographic proximity approach is supported. The combined model results mirror those found in the separate perspective models with one exception: geographic proximity is not a significant predictor of savings system adoptions.;These results suggest three broad conclusions about research on social security and, more generally, the welfare state. First, the lack of significance for most of the internal contingency variables suggests that different processes are behind program adoption in developed and developing nations, respectively. Second, the significance of the diffusion results demonstrates that external factors must be included in welfare state research. Last, the lack of consistent predictors for both program-type adoptions suggests that the processes involved in these adoptions are themselves more diverse and complex than previous studies have allowed. |
