" another idea I have is to pay department heads a percentage of every dollar saved each year." We can't reasonably measure how much money is 'saved' because t

" another idea I have is to pay department heads a percentage of every dollar saved each year." We can't reasonably measure how much money is 'saved' because there would have to be some rational benchmark agains which there would be 'savings'. What you would be inducing is merely 'less spending'. The 'spending' is not the problem, it's the efficiency. We want governments to spend wherein the can be a high ROI on that spend, but not where there is less obviously. As the article hints, if SF had an amazing new subway system for it's squandered billions, that would probably be a net win. In other words what we want is 'surplus' to be generated from government spending, the 'more surplus for the dollar', the better. Because most government services are free, we don't have any direct way of measuring said surpluses (we don't in the private sector either, but assuming any degree of market efficiency than price can be a proxy for surplus). So either we have to 'guess' at the value of services created (easier for subways than for say, having 'judges'), or create some kind of direct measurement. Or perhaps levels of spending and baseline measures of output could be measured, and then held to account against other cities in the region/country and have cities more comparably 'benchmark' against one another. And finally, just transparency: if people see how much money is going where and it's popularized, published, widely disseminated, it can have a populist effect.