It's a complex issue. There are some calculations and they can go both ways, depending in what and how you calculate. E.g. you could consider training costs - b

It's a complex issue. There are some calculations and they can go both ways, depending in what and how you calculate. E.g. you could consider training costs - but training costs are one-off for a change, not eternal. Or you can consider coding costs, but they might vary massively from region to region and depending on how (in)competent the procurement goes and how the local market works. Estonia is a country that seems to have figures this out to an immense degree, creating most of the infrastructure themselves and then tendering small-scale procurements for small parts of the system and sometimes allow private competition - but always with government-set data standards and often more on interfaces etc. To add one other angle: what is cheaper, paying 5 million to a company in Redmond yp provide licenses for opersting systems and officd software, or 10 million to a number of local companies to set up and adapt a slightly less refined open source tool for the local needs? In one scenario the money leaves the local economy, in other the city might spend more money but the money stays local and is used to buy houses or sandwiches, pay handymen or bus tickets or taxes. Where is the long term cost/benefit/ratio higher?