If you're looking for examples in the real world, there are going to be multiple variables mixed in, such as spending cuts. If you want pure evidence, you could

If you're looking for examples in the real world, there are going to be multiple variables mixed in, such as spending cuts. If you want pure evidence, you could write a basic software simulation incorporating just the variables you earnestly want to understand the effects of.