Tangible economic impacts from sports facilities often fail to materialise for a variety of reasons. These include:
1. A substantial proportion of the crowds at stadiums are local rather than visitors. Some estimates I've seen in the literature suggest that it ranges from 80 to 95% of attendance being local.
1a. Spending by locals within a city on attending games is usually substituted from elsewhere within the local economy, for example, movie theatres, video rental stores, and other entertainment venues. A game merely redistributes spending rather than generates it.
2. Spending within a city often leaks outside the local area, as not all goods and services purchased by event attendees are produced locally, so a proportion of the spending has to go out of the local economy to pay for imported goods and services.
3. Government spending on stadiums, contrary to popular opinion, is not costless. That is, the funding has opportunity cost that must be considered. Money spent on a stadium could have been spent elsewhere in the local economy, and as such alternative activity is forgone. A benefit is only observed if the stadium activity more than outweighs the lost activity elsewhere.
4. Stadiums are almost always underutilised. Westpac Stadium in Wellington has around 45-50 event days per year. That is around one day per week. Game days are usually a hotbed of activity, but six of the seven days there is nothing going on. Surrounding development feels this too. Are businesses located nearby dependent on stadium activity going to survive with more off days than game days? It is unlikely.
5. Much of the projected activity that a new facility attracts comes from within the city at the expense of other facilities. Things such as conferences, conventions, trade shows, etc would by and large have been hosted elsewhere within the city at another venue. Thus we see another form of substitution in action here, which works towards reducing the overall realised impact of a new facility.
6. A replacement facility can not realistically be expected to do a lot more than a pre existing facility. Research in the US has suggested that there is a short term honeymoon effect of up to ten years where attendances spike due to the novelty of the new facility, but beyond this the experience has been that attendance returns to pre facility levels.
What about the intangible benefits? Surely they matter?
Relevant intangible benefits include consumer surplus that locals enjoy from attending games at the facility as well as the public good aspects. They are recognised as benefits but there are weaknesses in their ability to justify government funding. Firstly, consumer benefits are often captured to a greater or lesser degree by event organizers through ticket pricing structures - season tickets, family/adult/children, concessions, etc. It is in the organizers interest to capture as much of this as possible so as to maximize event profits. Secondly, it isn't just within the stadium that these benefits are appropriated. To watch your team elsewhere, you pay for it via Sky TV subscriptions. To read about your team you pay for it via newspapers, magazines, internet access, etc. A lot of benefits can be captured privately. Thirdly, one can argue that just about any activity or enterprise has some intangible benefits, but this doesn't mean we should subsidise every activity that generates intangibles!
The bottom line is that if tangible benefits don't materialise, the intangible benefits have to be substantial and international evidence suggests that while they aren't insignificant, they are nowhere near the size of subsidies given to build sports facilities and/or attract sports franchises. Take, for instance, the estimated willingness to pay for the London Olympics, which was measured in one study at GBP1.9b (for the UK), was estimated at GBP480m for London in a 2008 paper (see gated link here). In light of the most recent estimates of costs (US$14b). They might exist, but they aren't likely to be deal breakers.