We generalize a standard technology diffusion model by allowing for IPRs regimes to be endogenously defined by the development level of each country. Also we insert differences in the composition of human capital between North (leader) and South (followers) which shape the relative costs of innovation and imitation. Results show how an optimal growth trajectory is found for the follower country which initially imitates and that, once a "threshold development stage" is reached, optimally switches to innovation by fully enforcing IPRs achieving a higher proximity with the technology frontier in the long-run. Other scenarios, such as a premature increase in the enforcement of IPRs or a switch from imitation to innovation at early stages of development of the followers are found to be sub-optimal.