How could independence cost more to provide just the same?

Last week the FT reported that Owen Kelly, the chief executive of Scotland’s financial services industry trade body, Scottish Financial Enterprise, said: "A yes vote would require the creation of an additional financial regulator with hundreds of staff. The cost would run into millions and have to be paid for by the industry in Scotland".

Why would having a regulator in Edinburgh as well as London cause, or not cause, increased costs? Increased costs would arise if the work of the regulator involved a large percentage of fixed costs (say writing regulations) rather than variable costs (e.g. dealing with individual cases). I don't know much about the detail of financial service regulation, but I would have thought that checking the compliance of individual companies (i.e. variable costs) would account for much of the total costs and that any additional fixed costs could be covered by the savings made through reduced office rents and salaries for operation from Edinburgh rather than London.

Suppose there are no additional costs, but no savings, then would this be positive for or against independence? Clearly For. No additional costs so increased activity in Edinburgh must be at the expense of reduced activity in London (they have fewer cases to process because some firms now under jurisdiction of Scottish regulator). 2 minutes of speculation leads me to think that a Scottish regulator is likely to be a net positive for Scottish economy, and net neutral for Scottish financial service firms. But as usual the scream of 'there might be some changes' goes up and it's a massive blow for the independence campaign.

It does highlight a wider issue that many people possibly get wrong: there are public services provided for the UK as a whole that are not currently provided in Scotland - mainly central government services like HMT etc; surely it will cost Scottish taxpayers money to provide these services post independence? Well, no: unless there is evidence for fixed costs in public good provision, then reducing the amount of services required by UK by 8.4% and creating this level of service provision in Scotland, has no net cost to taxpayers. It may be a loss to the London economy and a boost to Scotland's economy to have the high status employment of providing these services located here.

But are there fixed costs in public service provision? You might have thought so, but perhaps surprisingly Deaner & Phillips (2013) (*) do not find any evidence of scale economies when looking across Europe at the costs of providing public services in countries of different sizes.

(*) Section 5.1 ‘Spending by service area – how might it change?’ on page 62 and 63