The correlation between spending and good outcomes actually tells you very little about cost-effectiveness or about whether spending “works” at achieving good outcomes. I learned this from an older article by Abhijit Banerjee that I’m reading for class:
…one can easily imagine one country choosing one of these [aid programs], spending a lot, and getting the same results as another that spent very little. If both projects were aid- financed, someone comparing them would conclude that spending does not correlate with success in development projects, which is what one finds when one compares aid and growth across countries.
The key is that if two countries are getting the same outcomes, but using different methods, with different costs, then both methods might very well be effective in improving outcomes, but because they cost different amounts there is no correlation between spending and good outcomes.
This is generalizable to all sorts of other spending, both public and private. For example, I spend a lot more on food than some of my friends, yet we all eat pretty much the same amount of calories. Does that mean money can’t buy calories? Of course not; it just means that I eat out more than they do. I would certainly not advise either of us to stop buying food altogether.
PS — There is a great idea for a paper-abstract-as-April-Fool-joke somewhere in here. It will have to wait until 2015 I suppose. Consider yourselves warned.