Saturday, 31 May 2014

Lies, Damn Lies & Cost Estimates

A brief post for a change, with some observations on the recent rather fraught debate about the "one-off costs" of independence.

Neither side has made (or at least shared) a robust analysis of the potential costs 

[Look out for "could be" and "up to" and "as little as" in the quotes below]
  • HM Treasury has thrown around numbers based (explicitly) on some extremely crude assumptions (below from treasury press release)1
    • "Independent research ... estimates that it could cost up to 1% of a country's GDP" = £1.5bn
    • 180 departments x £15m per department = "up to £2.7bn"  -- the sources of the £15m per department have distanced themselves from this way of using that figure
    • UK government "has already published analysis which calculates that a new benefit system could cost £400m alone, and setting up a new tax system could be as much as £562m."  (I haven't yet chased down this analysis)
  • The Scottish Government has not been forthcoming with any analysis of their own; ducking the question or leaning on others' estimates
    • SNP Finance Minster John Swinney has resolutely refused to offer any figures -- hear him fail to answer the direct question thirteen times here
    • In the recent kerfuffle - some say stramash - Lone Wolf2 Prof Dunleavy has been cited as "estimating instead that the one-off set-up costs would be £150 to £200m
    • The next day Alex Salmond claimed the costs "could be as little as £250m" citing "London School of Economics expert Patrick Dunleavy, who has put the cost at between £150m and £300m"
    • There has been much noise made about the £575 - £625m figure apparently attributable to Swinney per the "leaked paper" below.  To be fair it is simply a "would on this basis be" estimate -- so if there were subsequent robust analysis leading to a different answer it would be reasonable to discard this early estimate. If
  • ICAS are one of our RIBs (Respected Industry Bodies: See Who We Can Trust ... I rate RIBs above Lone Wolves) have said in Scotland's Tax Future: Taxes Explained
    • "Examples from other countries can point towards the scale of costs which might be incurred. In New Zealand, for example, changes to the tax system which are less complex than those for an independent Scotland are costing around £750m. The cost for an independent Scotland could be significantly greater, especially considering the scale and the complexity of the legacy systems which might be inherited from the UK. What it is really going to be, and how it is to be paid for is a question that still needs to be answered"
    • Note the quote above is referring to Tax systems only, not all government departments require
1. Those who accuse the treasury of "lies" are wide of the mark. If the Treasury had said "we've done thorough and detailed analysis" and then been discovered to have actually just multiplied 180 x £15m, well that would have been lying. In fact they showed their working - it was in the press release - the figure was presented with no pretence to being anything more than a "could be up to" estimate

2. By referring to Prof Dunleavy as a "Lone Wolf" I do not mean to imply he can't be trusted; just that his is one person's opinion and should be treated as such (rather than as "the truth" because he is "an expert"). I explain further here > Sources: Who Can We Trust

The lack of good analysis by HM Treasury is disappointing; but the failure of the Scottish Government to come up with any robustly substantiated figures this late in the day is, quite frankly, astounding.

Its worth remembering that we are discussing unknowns here, that there is no truth (or to paraphrase Foucault: truth is at best plural). All we can ask for is honestly realistic estimates; I doubt we will get that from either side.

It might inform the debate to consider how accurately these sort of costs estimates normally turn out in reality.  When was the last major public project you can remember that ended up costing less than the initial estimate?  I offer two (admittedly not strictly comparable) examples for contemplation ..

So my conclusion?  We don't know what the "one-off" costs of independence would be ... but we can safely assume it would be a lot more than the Scottish Government are admitting to.


Anonymous said...

Cant compare Parlaiment and Trams. Parlaiment was a a joint commitee headed up by labour. Trams are a council issue and the current govt were not in power when costs began to escalate. Dont twist history to try and support your own agenda!

Kevin Hague said...

Anonymous: to be fair I do say not strictly comparable -- its a bit of fun as a sign-off to the more serious points above. Do you have examples you can give me where the Scottish Government has brought in major projects *under* the initial estimate?

Unknown said...

The £40m figure for the Scottish Parliament was plucked out of the air by Donald Dewar for political reasons, it was never a sensible estimate. Using the old Royal High School on Calton Hill was rejected by DD, also for political reasons. If the Royal High School had been used, the cost would certainly have been a lot closer to £40m than £414m.

M74 extension was £15 - £20m under budget, and completed nine months early.

The new Forth Crossing is currently projected to cost £145m less than the previous upper estimate.

Of course, you can't use civil engineering projects to make estimates about what will be largely IT costs, but it seems to me that the suggested IT costs are being ludicrously overstated. The software - the computer code - to do everything we need to do on day one of independence already exists at a UK level. Buy some hardware, copy over the software and relevant data, and you're 90% done. Most of the cost of big IT projects is in the development, and once you've got your software working there's no further development required until you decide to change something.

The UK government has a very poor record of successfully carrying out big IT projects - various departments have wasted billions in taxes over the years, to the huge benefit of the London IT and consultancy industry. Hopefully a Scottish government would complete such projects less badly; and to the extent that money is needlessly wasted, at least our tax money will be subsidising our indigenous IT/consulting industry, and not London's.

Kevin Hague said...

Useful examples Ewan thanks ... as someone who knows a little if such things I can't share your analysis of the ease of IT system cut & paste though.

More to the point: plucking #s out of the air is clearly what SNP have done with their one-off cost figure (hence no back-up and conflicts with what impartial opinion there is). Do you disagree with that substantive point?

Unknown said...

I'm not impressed with the SNP's failure to provide any thought-through analysis of costs. It's a serious blunder from a government that doesn't blunder much (at least not compared to the competition), and gives some genuine ammunition to their opponents, who have hitherto been scrabbling around in the mud screaming about scary monsters.

J Drew said...

That's easy then. No data centres needed, no data cleansing, no staff, no licence costs. I wonder why the Scottish govt have no estimates if your list was all they needed to add up, even a 5 year old with calculator could do that. Or maybe it's a bit more complicated than you think.