I wanted to create a regression model based on the following data: United States Crime Rates 1960 - 2009

For my model, I wanted to estimate the population of US based on reported crime (murder, theft, etc...), and I wanted to know if it's a good idea for a model because you can clearly see that as the population increases so does the crime rate (until early 90's when it starts to decrease).

Is there a better idea for a model?

Thank you in advance.