Before we all get carried away with the comforting certainty of the tax policy’s forecast figures, and while we wait for the Policy & Resources Committee’s ‘detailed analysis’ underpinning the tax policy letter to appear (whenever that may be), it’s probably worth starting with a slightly awkward question – how often have the States actually delivered the efficiency savings it has previously promised?
If you take a look back at the track record, a pattern starts to emerge that is difficult to ignore. A Bayesian reading of historical delivery suggests there is around a 71% probability that the £20m. of efficiency savings baked into the current proposals will not succeed. This is, of course, not a prophecy, just a reasonable reading of past behaviour. But if that pattern holds, then we have a problem. Any shortfall has to be plugged somewhere and somehow – higher taxes, deeper cuts, or yet more pressure on an already stretched system. And given how elusive both meaningful cost reduction and sustained economic growth have been, it doesn’t take much imagination to see which lever tends to get pulled.
Against that backdrop, we are told that a 3% GST will raise around £55m. a year, before costs. A neat number. A reassuring number. But also one that rests on foundations which are starting to look a little less solid the closer you examine them.
The modelling, we are told, is largely based on household consumption data from 2018/2019. At the time, that may have been perfectly reasonable. But we are nearly eight years, a pandemic, a surge in inflation, and a fair amount of behavioural change later. People don’t spend money the same way they did in 2018. They don’t shop the same way, travel the same way, or in many cases even live in quite the same way. Using that as the anchor for a forward-looking tax forecast inevitably introduces uncertainty – not just at the edges, but at the core of the calculation.
And that uncertainty doesn’t sit in isolation. It sits alongside a series of assumptions that can only really be tested once GST is actually in place – how people respond to higher prices, how much of the economy is carved out through exemptions, how compliance behaves in practice, and how the system settles once the initial adjustments have played through. None of these are fatal flaws. But they are real unknowns.
Which is why the comparison with Jersey becomes hard to avoid.
Jersey raises around £129m. a year from GST at 5%. If you scale that down to reflect Guernsey’s smaller economy and a 3% rate, you arrive at something closer to £41m. That is a fairly straightforward, back-of-the-envelope comparison – but it gives you a sense of scale. Because against that benchmark, the £55-57m. forecast looks not just optimistic, but noticeably so.
Especially when you remember that this figure is presented before you even account for categories of spending that sit outside the tax base altogether – smaller businesses, childcare providers, the States itself. On the figures provided, that alone may amount to around £25m. of annual expenditure that is effectively removed from the equation.
So the implication, whether intended or not, is that Guernsey will raise almost twice the GST revenue of Jersey once you adjust for size and rate. That is quite a claim for a new tax system in a smaller economy, drawing on older data, and introducing a structure that hasn’t yet been tested in practice.
A similar sensitivity runs through the international service entity scheme, expected to raise around £11m. Again, not insignificant. But also dependent on assumptions about participation and behaviour that could shift meaningfully with relatively small changes in the underlying commercial environment.
Then there is the 1.9% inflation effect. On paper, this feeds through into higher public sector pay costs of around £8.2m. In reality, that reflects a fairly basic principle – if prices rise, pay usually has to follow if you want to recruit and retain staff. The committee’s position, however, is that no corresponding pay adjustment will be made. Which, at the very least, is an interesting tension.
A further £1.7m. arises through the impact on public sector pensions. The argument here is that this is not a direct cost to general revenue because the scheme is 99.1% funded, and therefore has sufficient ‘headroom’ to absorb short-term inflation. In other words, it is treated not as a budget pressure, but as something happening safely inside a well-funded system.
But if that headroom genuinely exists in the way suggested, then it also implies something else. In the absence of GST, it could just as easily be reflected as lower contributions over time. Not a saving in any meaningful structural sense, but a timing shift — pressure eased in one direction only for it to reappear later.
And this is where the wider issue starts to show itself. Because the treatment of pension and contributory costs seems to rely on a distinction between what shows up directly in the annual budget, and what sits inside longer-term funding mechanisms. That distinction may be administratively neat, but economically it is harder to sustain. Costs can be smoothed, re-allocated, and re-labelled, as the policy letter attempts to do – but they do not disappear. They are still paid somewhere, somehow, either through contributions, benefit adjustments, or future calls on general revenue.
In other words, the taxpayer always ends up picking up the ever increasing tab.
But these flaws don’t stop at revenue. They run through the whole system.
Using the most recent States accounts, a proportional increase in contributory benefits alone would add around £4.6m. a year to expenditure. Not transformative in isolation, but still relevant when assessing the net effect of the reforms. And when combined with the pay and pension-related impacts, the policy letter is treating close to £16m. of costs directly relating to the implementation of GST as non-existent.
Yet these are not theoretical costs. They only exist because GST is being introduced. Without it, they simply would not arise.
None of this is to suggest the reforms are fundamentally misguided, or that the forecasts are necessarily wrong. Public finance modelling always involves judgement, assumptions, and uncertainty.
But that is precisely why transparency matters. If the forecasts are robust, there should be no hesitation in publicly publishing the underlying workings so they can be independently tested. That is not an obstacle to good policy — it is a prerequisite for it.
At present, however, the initial modelling underpinning the tax policy letter has not been published, and no clear explanation has been given for that omission.
And perhaps that is where the real question lies. Not whether GST is politically desirable or undesirable, but whether the assumptions supporting it are sufficiently strong, sufficiently current, and sufficiently transparent to justify the confidence being placed upon them.
Because in the end, even in public finance, arithmetic has a habit of reasserting itself.
