Between 1998 and 2020 the Isle of Man's economy
nearly tripled, far outpacing the United Kingdom.
Then it stopped. Real output has fallen
twelve per cent since,
and what the Island actually earns — its GNP — is now
below its 2014 level.
This is a scroll-story of why.
Correction, 14 August 2026. An earlier version of this page
showed an erroneous Isle of Man growth chart, built from a Statistics Jersey series misfiled into the
Island's data folder. It has been corrected to use GDP and GNP from the Treasury's own National Income
Reports. The corrected data has since been re-verified under the Observatory's data governance standard:
an independent, cell-by-cell re-extraction of the underlying National Income tables — 2,678 cells
across 18 Isle of Man Government publications — matched with zero discrepancies.
Acts 2, 3 and 4 draw on separate sources and are unaffected; Act 5's output line has been
rebuilt, and the Act 6 simulator — withdrawn on 14 August — was restored on 17 August 2026,
running on the model recalibrated against the corrected, independently verified series. The full account
is in Act 1, under “What this replaces”.
Real output since 1998
Real output, adjusted for prices. 1998 = 100.
Isle of Man GDPIsle of Man GNPUnited Kingdom
Source: Isle of Man National Income Reports, Table 3 (growth in GDP and GNP at constant prices), chained from 1998; ONS, GDP output approach low-level aggregates, chained volume, Q1 2026 edition.
VerificationVerified the growth rates: the Treasury's 2008/09 split (GDP +4.7%, GNP −5.5%) is corroborated by the World Bank's constant-price series, which shows the same shape on its own basis (+4.9%, −5.9%); 2020/21–2023/24 rates confirmed by Manx media reporting the release; Moody's (Dec 2025) independently puts Manx growth at about 1% a year over 2014–2024. UK line matches ONS series ABMM exactly at 2022. Cannot verify the index levels: no institution outside the Isle of Man measures this economy. See the data-governance note at the foot of the page.
Scroll
Act 1
The growth that stopped
Two decades of headway, then five years of drift.
1998. Everything is set to 100. Real output, with inflation stripped out. One line for the United Kingdom, and two for the Isle of Man: GDP, what is produced here, and GNP, what Manx residents and companies actually earn.
2007. The boom was real, and it was extraordinary. The Island reached 194 against the UK's 128 — nearly double its 1998 output in nine years. For three decades the Isle of Man genuinely was one of the fastest-growing economies in the developed world.
2008. The banking crisis barely touched Manx GDP — it rose 4.7 per cent in 2008/09, while the UK fell six points. But GNP fell 5.5 per cent that same year. What is booked here and what is earned here began to part company, and the gap has never closed.
2019/20 was the peak. GDP reached 299, very nearly triple its 1998 level — though growth had already stalled the year before it: 2018/19 to 2019/20 added just 0.1 per cent, real GDP essentially flat before COVID ever reached the Island. E-gaming's share of the economy had itself been sliding since 2017/18 (Act 2); on this reading, the sector's earlier fade — not the pandemic — was already capping growth before the pandemic arrived.
It has fallen in every year but one since — −8.0 per cent, then a one-year rebound, then −5.0, then −2.5. The first fall and the rebound track COVID and its recovery, much like the UK's. The renewed falls in 2022/23 and 2023/24 come well after that recovery, and they land in the same years e-gaming's share of the economy kept sliding — 15.7 per cent to 13.6 (Act 2).
By 2024 Manx GNP was below its 2014 level. Measured over the last decade rather than the last quarter-century — fiscal 2013/14 to 2023/24 — the Island's real GDP grew 4.1% against the UK's 15.2%.
Do the Manx feel twice as rich as the UK? On paper, nearly: nominal output per head runs at roughly £69,000 here (GDP, 2023/24 national income accounts) against about £40,000 in the UK (ONS, 2023). But output per head measures what is booked on the Island, not what islanders take home: the same accounts put company income at 59.5 per cent of the economy, and personal income at about a third. This page is about the take-home number.
For the technically mindedThis section was rebuilt on 14 August 2026. The Manx lines are chained from the constant-price growth rates published in Table 3 of the Isle of Man National Income Reports (“Growth in GDP and GNP at Constant Prices”), cross-validated across successive editions; the UK line is ONS chained-volume GVA. Note that the Isle of Man publishes no gross value added series — its accounts run on GDP and GNP, and on Manx convention product taxes are added at the GNI line, so Manx GDP is already close to what the UK would call GVA. Manx figures are fiscal years (1998/99 onward) plotted against UK calendar years, so alignment is approximate to within a year, and a fiscal-year series for a small economy is volatile because one large sector can move the whole index. Which window you choose matters, and we say so plainly: from 1998 the Island is far ahead of the UK; from 2007 it is still ahead on GDP but behind on GNP; from 2014 GDP has grown at about a quarter of the UK's pace (+4.1% against +15.2%) and GNP has fallen outright. The point of this page is the turn in 2020, not the level.
What this replaces. Until 14 August 2026 this chart showed the Island reaching only 110 by 2022 against a UK 154. That series was Statistics Jersey's real GVA, downloaded from opendata.gov.je and misfiled into the Island's data folder without its je- prefix. It was a correctly-built chart of the wrong jurisdiction. Jersey's economy really was flat over this period, which is why nothing about it looked wrong. Acts 2, 3 and 4 draw on separate sources and are unaffected; Act 5's output line has been rebuilt, and the Act 6 simulator — withdrawn while this correction was made — was restored on 17 August 2026 on a model recalibrated against the corrected, independently verified series (calibration pass 6; its disclosures are in Act 6).
Real output: Isle of Man GDP and GNP v United Kingdom
Real output index, 1998 = 100
IoM GDPIoM GNPUnited Kingdom
Source: Isle of Man National Income Reports, Table 3; ONS. Real output, 1998 = 100.
VerificationVerified against third parties: the year-on-year growth rates. The Treasury's 2008/09 split (GDP +4.7%, GNP −5.5%) is corroborated by the World Bank's constant-price GDP and GNI series, which shows the same shape on its own basis (+4.9%, −5.9%). Moody's December 2025 affirmation independently describes growth as “approximately 1% annually between 2014 and 2024”. The UK line matches ONS series ABMM exactly at 2022 (153.8). Two separate 2008–11 shocks, not one: the 2008/09 GNP fall reads as the global financial crisis hitting investment and company income directly — that is inference, not a sourced causal claim. It is a different mechanism from the Customs & Excise rebasings of 2009 and 2011 (Act 4), which cut government revenue on its own timeline and did not, on this data, drive the GNP fall. Cannot verify: the index levels. No independent institution measures the Manx economy, so a reconstruction from a different source lands 6–10% lower on the same period. Read the shape and the turning points, not the level. The UK 2024 point uses the ONS Q1 2026 vintage (155.7); ONS series ABMM gives 156.1 on the current edition.Data table
Act 2
The churn
Behind the climb and the stall, the mix was changing completely.
1997. Finance is two-fifths of the economy — banking, insurance, funds. Everything else shares the rest.
Then come the flavours, each with its own promotion push. Film from 1995. E-gaming licensed from 2001; the big operators land from 2005. Space and digital follow. Watch the purple band.
By 2015/16 e-gaming and the wider ICT sector are around 26 per cent of the economy, from nothing in 1998 — e-gaming alone, about 17 per cent, rising to a peak near 21 per cent in 2017/18. The corrected real-terms accounts trace the same arc: on the consistent post-2007/08 basis, the ICT-and-e-gaming block multiplies more than sixfold in inflation-adjusted terms to its 2017/18 peak, then gives back 44 per cent of that by 2023/24 — still around three and a half times its 2007/08 size. Meanwhile everything that is not finance, digital or public services has shrunk from 56 per cent of the economy to 26.
Each new sector claimed a bigger slice. Until 2020 the pie was growing too — but hold Act 1 in mind: the sector that grew fastest is the one whose margins are booked here and earned largely elsewhere, which is exactly where GDP and GNP came apart.
And the steady-looking finance band is steady only in total. From 2007/08 — the first year the accounts split the block — banking and the other labour-heavy finance lines have drifted down in real terms, about −0.3 per cent a year, while insurance — margin-booked, employing comparatively few — grew about +3 per cent a year: sixty per cent larger by 2023/24, and up from roughly a third of the finance block to nearly half. Inside the flat band, the part of finance that employs people has been shrinking; the part that books margins has been growing.
The pie grew. Then it stopped. And the slices had already changed hands. Whether that was addition or replacement is a question shares of money cannot answer. Counting people can.
For the technically minded
Shares are of nominal national income by economic activity, by fiscal year — the Isle of Man publishes no gross value added series, so this is the sector breakdown of GDP from Table 1 of the National Income Reports. The bands are sums of the accounts' published sectors: the purple band is e-gaming plus the wider ICT sector (e-gaming alone peaked at 21 per cent in 2017/18); the green band is public administration plus education and health. The step in the green band at 2007 is a data boundary, not a hiring spree: the pre-2007/08 accounts do not separate education and health, so before then the band shows public administration only — which is roughly flat at 3½–5 per cent of the economy across the whole period. E-gaming's slice includes operating margins booked to the Island on business with customers largely elsewhere; its share overstates its footprint in jobs and local spending. The finance band is Banking, Insurance and Other Finance & Business Services; Corporate Service Providers are counted in “everything else”, so this is a narrower definition of finance than some readers will expect. Act 1, by contrast, measures real, price-adjusted output. Note that shares are shares: a sector whose share falls can still be growing in real terms while the whole economy grows faster.
On the real-terms figures quoted in the text. The ICT-and-e-gaming arc and the finance split come from the constant-price tables of the same accounts (National Income Table 2, extended back via the Digests of Economic and Social Statistics), assembled into a thirty-year sector series that passed a full independent cell-by-cell re-verification on 16 August 2026 (2,678 cells across 18 official publications, zero discrepancies). Two caveats bind those figures. First, the accounts were restated onto ESA10 at 2007/08, and the restatement was large and uneven — it moved measured insurance up by ×1.90 against roughly ×1.13 for banking and other finance, and the ICT-and-e-gaming block by ×1.91 — so no growth rate quoted here spans that boundary: the finance split is quoted from 2007/08 only, the first year the accounts publish it (it is deliberately not back-cast), and the ICT arc likewise runs from 2007/08. Second, the Manx constant-price tables deflate every sector by one economy-wide price index, so sector figures are real-income comparisons, not volume measures — relative sector movements are relative movements in money terms.
Who holds the slices
Share of nominal GDP, per cent, fiscal years 1997/98–2023/24 (labelled by opening year)
Source: Isle of Man National Income Reports, Table 1 — National Income by Economic Activity at current prices.
Data table
Act 3
A fixed pool of people
New sectors need workers. The Island stopped producing them.
1951: 54,024 people. The Island then grew for sixty years — through the tourist trade, then the finance boom, then everything after. 84,497 by 2011.
Then it stopped. 2011: 84,497. 2024: 84,756. Thirteen years, 259 more people. Flatlined.
Deaths have exceeded births every year since 2016. In 2024 natural change was −351 and net migration +651. Every extra worker must now be imported, housed, and kept.
So where did each new flavour find its workers? Not by retraining the old ones. The Island swaps people with the UK. Each new sector imports its specialists: older, experienced, mid-career. Meanwhile, between the 2011 and 2021 censuses — the flat-population years — transport lost 1,190 jobs. Banking lost 1,076. Insurance lost 605. Wholesale distribution lost 489. Manufacturing lost 236. Agriculture and fishing lost 186.
Mid-career specialists in; disproportionately younger, local people out. A flat population is not stillness. It is churn at equilibrium: specialists in, school-leavers out. The Government's own housing evidence base puts a number on the other half of that exchange: approximately 50% of new migrants leave the Island within 10 years (Objective Assessment of Housing Need, GD 2024/0041, May 2024, citing the Isle of Man Population Report). A census counts who is here, not who passed through.
Total employment barely moved either: 43,134 in 2011, 43,519 in 2021 — up under one per cent, tracking the population. The public sector traced a V that the decade’s totals hide: 10,042 → 9,144 → 9,934 across the three censuses — cut hard after the VAT settlement was rebased, rebuilt once the new revenue arrangement bedded in (Act 4 tells that story). What grew across the decade was construction, health services, IT — and the flavour of the decade, e-gaming: from 658 jobs in 2016 to 1,086 by 2021, about two and a half per cent of the workforce. But flavours fade. E-gaming’s share of the economy peaked at 21 per cent in 2017/18 and has slid since — 15.7 per cent by 2021/22 and 13.6 per cent by 2022/23, before edging back to 14.2 per cent in 2023/24. The fall is real, but it is not a straight line. The rise was fast; the fall began before the decade was out. A fixed pool of workers, exchanged rather than grown — and one employer able to keep bidding regardless of what the economy did. To see why, follow the money that paid for it.For the technically minded
Employment is from the census of population (residents in work, self-classified by industry), so it is a different lens from the national income accounts. Sector figures above are Census 2011 and Census 2021 published industry tables (Census 2011 Table 3.6; Census 2021 Part II Table 1.7), matched to the same 2011–2021 window as the population comparison above; e-gaming is not broken out as its own census category until 2016, so its earlier point uses that year instead of 2011. An earlier version of this section compared 2001 to 2021 using a different, non-census-report employment series that undercounted both years by a growing margin and implied a fall where the census reports themselves show resident employment rising 2001–2021; it was corrected after further fact-checking. Note that a census measures stocks, not flows: it shows where jobs stood, not the gross traffic in between, so the two-way churn cannot be read from it directly. The churn mechanism — specialists in, displaced younger workers out — is the reading of the 2010/11 systems-dynamics work, consistent with these stocks. One honest limitation: the simulator in Act 6 carries net migration only; it cannot represent the two-way flow. E-gaming employment counts licence-holder staff; its £801m of national income (2023/24) includes margins booked to the Island on business with customers largely elsewhere.
Population of the Isle of Man
Residents, 1951–2024
CensusAnnual estimate
Source: Isle of Man census reports 1951–2021; annual population estimates (experimental) 2022–24. Census counts are primary; 1951–2021 are direct counts, not estimates, and are not separately “verified” here. The natural-change run (deaths > births every year since 2016) is corroborated by UN Statistics Division data, a genuinely separate compilation from the Manx figures. The Q1 2025 estimate (84,975 — a record high, +478 since 2011 on that reading) is in the data table below rather than the chart, since it is a newer experimental quarterly figure not yet on the same annual basis as the rest of the series.
Data table
Where the jobs went, 2011 → 2021
Employment by sector, census counts
2011 census (e-gaming: 2016)2021 census
Source: Isle of Man census of population, 2011 and 2021 (Census 2011 Table 3.6; Census 2021 Part II Table 1.7). Residents in employment by industry. E-gaming's earlier point is 2016 (dashed line), the first year the census breaks it out as its own category.
VerificationChecked against primary source: both the 2011 and 2021 baselines have now been read directly from the Census Report tables (2011 Table 3.6; 2021 Part II Table 1.7) — not taken from a media summary. As a consistency check, the ten sectors itemised here sum to 26,102 workers in 2011 and 26,122 in 2021, against published total employment of 43,134 (2011) and 43,519 (2021); the 17,032 and 17,397 unitemised the totals imply are the remaining census sectors not broken out on this chart.Data table
Act 4
The engine in the basement
The economy grew. Government grew faster — and on a different engine.
The Island does not keep its own VAT takings directly. Under a revenue-sharing agreement with the UK it draws a share of a common pool. In 1998 that share brought in £307m in today's money. By 2006, the peak: £752m — up 145 per cent in eight years, against a real economy that itself grew fast over the same window (Act 1) but nowhere near that fast: this revenue stream outran even the genuine boom.
This one stream rose from 52 per cent to 72 per cent of Treasury receipts. Spending followed it up. So did public hiring. For a decade, the public sector's growth was paid for by a formula, not by the economy. The loop on the right span faster each year: spend, get measured, claim the share, receive, spend again.
2009 and 2011: the UK rebased the formula, twice. Receipts fell from that £752m peak to £403m in 2012, in real terms — a 46 per cent drop. The engine had been paying for the government — and now the government genuinely shrank: the workforce fell through to 2015, following the receipts down. Watch the two lines move together.
April 2015: the government suspended the Personnel Control Mechanism, the headcount cap that had governed staff numbers since 1991, replacing it with a limit on staffing costs only. Hiring resumed: government FTE climbed from 6,737 in 2016 to 7,533 by the end of 2024 — up 797 posts, +12%. Then in 2016 a new arrangement, FERSA, tied the Island's Customs & Excise share to its own measured expenditure, averaged over five years; a further true-up follows in 2025. The revenue base FERSA anchors to has not grown to match: real receipts were £465m in 2016 and £452m in 2023, essentially flat. Public-sector employment expanded against a flat income stream, not a growing one. By January 2025 the Council of Ministers felt the need for a new Recruitment Control Framework — which the Treasury Minister described as strengthening governance on hiring to “restrict headcount growth and drive value for money”.
For the technically minded
Under the 1979 Customs and Excise Agreement, shared duties are pooled and divided by formula. Before 2009 the Island's share keyed off an arrangement that outgrew its measured economy; the 2009 and 2011 revisions and the 2016 Final Expenditure Revenue Sharing Arrangement progressively re-anchored it to actual island final expenditure, with a five-year averaging lag. Receipts are Customs & Excise revenue from Treasury accounts, originally reported in nominal £m and deflated here to 2024/25 prices via the IoM CPI (Statistics IoM, Jan 2008=100, rebased to 2024/25=100) so the chart can be read directly against the real-output and real-population lines elsewhere on this page; nominal originals are kept in the underlying data for anyone who wants to reconcile against the published Treasury figures. The workforce series is whole-of-government full-time equivalents, excluding the Post Office and Manx Utilities, assembled from the Government Annual Reports (2004–08, 2011), Civil Service Commission reports (the 2013–14 points are derived from a stated Civil-Service-to-total ratio and shown hollow), an Office of Human Resources report (2015) and FoI/Cabinet Office figures (2016–24). The census confirms the same V independently: public-sector employment 10,042 (2011) → 9,144 (2016) → 9,934 (2021). On staff-number controls: the Personnel Control Mechanism (headcount targets, established by the Council of Ministers in 1991) was suspended in April 2015 and replaced by the Employment Costs Budgetary Control Mechanism, which limits staffing costs but not staff numbers; a Recruitment Control Framework was introduced alongside it, effective 20 January 2025, aiming to “restrict the growth of the public service” and “reduce the employee costs of Isle of Man Government” (Council of Ministers, Recruitment Control Framework, approved 9 January 2025).
Customs & Excise receipts — and the workforce they paid for
Receipts: £ million per year, real terms (2024/25 prices), 1998–2023 · Government workforce (FTE), right axis, 2004–2024
Sources: Isle of Man Treasury accounts (receipts, originally nominal, deflated here via IoM CPI, Statistics IoM, Jan 2008=100 rebased to 2024/25=100); Isle of Man Government Annual Reports 2009 & 2011, Civil Service Commission reports, OHR and Cabinet Office figures (workforce FTE — excludes Post Office & Manx Utilities; 2013–14 points are ratio-derived estimates, shown hollow).
VerificationAdjusted to real terms, 15 August 2026: this chart previously showed nominal receipts; it has been converted to real (2024/25-price) terms throughout, using the IoM CPI. Verified against audited accounts: the 2009 and 2011 rebasings and FERSA's 2025 true-up, from the HMRC account audited by the National Audit Office — the only externally audited source anywhere on this page. Cannot verify independently: the receipts levels — no institution outside Treasury measures this series, and the only third-party reports of it are very likely republishing the same release rather than checking it separately.Data table (real, 2024/25 £m)
What just happened, in general form. The formula was an administrative
difference from the neighbours, and it was closed from outside — twice, on somebody else's timetable. That is
not bad luck. It is the mechanism. An island economy has a real ceiling: a fixed, small population and a fixed
footprint of land. Growth this fast, this far past that ceiling, was not conjured from nowhere — it ran on
administrative differentiation from the neighbours the Island competes with: lighter tax, a VAT-sharing formula,
an e-gaming licence regime none of them offered. That is what a haven is, and it is how the Island grew. The same
mechanism is why it could not last. Each differentiation eventually gets closed, capped or copied, and when it is,
the sector or the revenue stream built on it goes with it, and the next one has to be found. Film, then e-gaming,
then the VAT formula itself — the story told across these four acts is that mechanism working itself out.
Act 5
Three claimants, one island
Put the three stories on one axis and the shape of the problem appears.
Start with the people. Everything on this chart is indexed to 1998, as in Act 1. Population climbed through the boom years, then hit the plateau of Act 3: up 15 per cent over the whole quarter-century — and almost none of it after 2011.
Add output. The real GDP line of Act 1, on the same axis: it nearly triples by 2019/20, then gives back twelve per cent in four years. Up, then down — on a population that barely moved.
Now houses — priced not in pounds, but in years of pay. The amber line is the official affordability ratio — median house price divided by median full-time earnings — on its own scale, right. 5.1× earnings in 1998. 10.4× at the 2008 peak. Never below 8× since. 9.1× in 2025 (Housing Market Report 2025, Appendix Four).
The sale price alone hides the sharper truth. After 2011 the nominal median fell, and did not regain its 2011 level until 2018 — from 2012 to 2013 it dropped from £244,300 to £239,500, yet the ratio worsened, 9.00× to 9.12×. Prices drifting down while affordability gets worse can mean only one thing: earnings fell further behind. In real terms a median house costs roughly what it did in 2011 — the deflated median is about 1 per cent lower than in 2011 (see the technical note); the ground lost since is in pay.
Three claimants bid for one fixed pool of workers and homes: the public sector, the existing economy, and this year’s flavour. A new sector does not conjure new people. It outbids an old employer for staff, and outbids a neighbour for a house.
Our competition isn’t with the globe. It’s internal.
And this is the relationship with migration, concluded: on a full island, every migrant in is a migrant out. Attraction policy decides who is swapped — never how many we are. The one lever that adds instead of swaps is supply: homes first, then people, then growth. That is why the supply side is the priority — and the simulator below lets you test it.
There is a second lever, on the demand side: stop targeting individual sectors for promotion. On a fixed pool, promotion does not add a bidder to the market — it hands one bidder an advantage over the others, and the gain capitalises into wages and house prices rather than into output. Ending sector-by-sector promotion removes that source of internal competition.
And there is a third option: keep playing the haven game — knowingly. It is the strategy of the last five decades — the lighter-tax regime, the 1979 Agreement’s formula, the e-gaming licence — and it can keep working, on the condition Act 4 laid bare: every differentiation is eventually closed, capped or copied on somebody else’s timetable, so the next one must always be found before the current one goes. That is a treadmill, and running it well is a real skill; the only dishonest version is pretending it is a growth policy. Build capacity, end the internal competition, or keep the treadmill — but choose knowingly.
For the technically mindedOn the public sector as a competing bidder. Until 16 August 2026 this act argued for a further demand-side lever: narrowing the pay and job-security gap that lets government outbid other employers for the same fixed pool. The pay half of that does not survive the published evidence. The Isle of Man Earnings Survey does show public-sector median weekly earnings above private (£828 against £720 in 2025), but Statistics Isle of Man disclaims that comparison in the same table every year: the two workforces “are composed quite differently” and the difference “does not reveal differences in rates of pay for comparable jobs”. Cut by industry rather than by ownership, Public Administration averages £936 a week against a whole-economy £920, while Other Finance & Business Services runs at £1,391 and e-gaming at £1,271. On the published numbers the highest bidders for the fixed pool are the promoted sectors themselves. What is genuinely unmeasured is total reward and conditions rather than pay: the employer future service rate — the actuarial cost of benefits being earned — is 17.3% of pay across all PSPA schemes (16.9% for the Unified Scheme, 33.5% for Police), against roughly 14.2% of basic pay actually charged to departments in Pink Book Table 14B, with no published private-sector comparator for either, and public-sector turnover is published while no private-sector figure exists at all. Statistics Isle of Man already collects what would settle it — occupation, grade and contract permanence are on the Earnings Survey questionnaire every year and have never been published. Until they are, this page makes no claim either way.
The affordability line is the “Median House Price to Median Earnings” column of Appendix Four of the Isle of Man Housing Market Report 2025, re-extracted directly from the published table (37 annual values, 1989–2025; the chart starts at 1998 to match the page’s base year — the data table carries the full series). It is deflator-independent: numerator and denominator are both nominal, so no price-index choice touches it. The nominal medians quoted in the text are the official houses-only median (Appendix One, annual average of quarters), kept in the data table; deflated by the IoM CPI (Statistics IoM, Jan 2008=100, rebased to 2024/25=100) they show a median house roughly flat in real terms since 2011 — about a 1 per cent real fall to 2025 — which is why the post-2011 worsening of the ratio must be an earnings story rather than a house-price one. Population between censuses is interpolated linearly; output is the real GDP index of Act 1 (1998 = 100, latest year 2024, rebuilt 14 August 2026 after the misfiled-Jersey-series correction). In a supply-constrained labour and housing market, sector promotion largely reallocates inputs and capitalises into prices — the mechanism the systems-dynamics model behind this page makes explicit. An earlier version of this chart plotted the house-price level instead: first a median of all land-registry transactions (flats and land included, running ~12% below the houses-only median — a reader’s cross-check prompted the switch to the official series), then the official median deflated to 2024/25 prices. It was replaced by the affordability ratio on 17 August 2026 so the housing line no longer depends on any deflator and no longer mixes a nominal level into an otherwise real chart.
People, output — and the price of a home in years of pay
Population & real output, 1998 = 100 (left) · median house price ÷ median earnings (right, amber)
Sources: Isle of Man Housing Market Report 2025, Appendix Four — median house price to median earnings ratio (right axis; full series 1989–2025 in the data table); Isle of Man National Income Reports, Table 3 (real output, as Act 1); census reports and annual estimates (population, interpolated between counts). Left axis 1998 = 100.
VerificationVerified against the primary source: the affordability series is the “Median House Price to Median Earnings” column of Appendix Four, Housing Market Report 2025 (Isle of Man Government), re-extracted from the published table on 17 August 2026 — 37 values, 1989–2025 — and cross-checked against the report’s own body text, which quotes the same closing figure: “median house prices were 9.1 times the median full time salary”. Corroborated in shape: roughly 5× earnings in the late 1990s and around 9× now is supported by a Liberal Vannin freedom-of-information investigation and by Livingstone Surveyors. The ratio is deflator-independent — both sides are nominal — so it is immune to the price-index choices that complicate other charts on this page. Cannot verify independently: the underlying median price and median earnings levels — both are single-sourced from Isle of Man Government statistics, and third-party mentions very likely republish the same release. Corrected 16 August 2026: an earlier version of this note attributed housing-completions counts to the Housing Market Report; it does not carry them — completions are published annually in Isle of Man in Numbers (sourced to the Department of Infrastructure), with a continuous series for 1996–2008 and 2016–2024 and the years 2009–2015 charted but never printed, so they are not reproduced here.Data table
Three claimants, one island
One fixed pool — and the bidders competing for it
Workers: 43,519 residents in employment at the 2021 census (Act 3). Homes are shown qualitatively — the stock is fixed on any one year’s timescale. The door is Act 3’s churn: roughly 50% of new migrants leave within 10 years, on a population flat since 2011.
Adding, and swapping. Look back at what each chart in this story could
not see. Real output could not separate what is booked on the Island from what is earned on it.
Sector shares could not separate a new industry from a displaced one. The census could not separate churn from
growth. And when the headcount cap became a cost cap in 2015, the control on government stopped counting people
at all. Four instruments, one blind spot: not one of them can tell adding from
swapping. On an island with a fixed pool of workers and homes that is the whole question, because
a new claimant does not arrive alongside the others — it arrives instead of one of them. Which leaves
exactly one lever that adds rather than swaps.
Act 6
Run the Island
This page rests on a small calibrated model: five private-sector blocks, a public
sector, houses, people, and the revenue formula in its three historical regimes. Recalibrated on the
corrected, independently verified Manx series (16 August 2026), it is back. Run the Island forward
to 2035 yourself.
Read before running. Three things this simulator does not do,
disclosed up front:
1 · 2009–2015 revenue is historical data, not simulation.
The Customs & Excise rebasings of 2010 and 2012 were negotiated impositions on the UK’s timetable
(Act 4) — no island-side model can generate them, so this one does not pretend to. Observed receipts are
fed through unchanged for 2009–2015, drawn as fixed amber dots on the revenue chart. The sliders do
not affect that window.
2 · The model under-predicts the 1999–2007 boom.
It cannot yet produce the observed 7–8%-a-year real growth of that era and runs roughly 20% low
through the mid-2000s (endpoints fit; the boom’s pace does not). Do not use this simulator to
characterise the pre-2008 economy; that era’s story is told from primary data in Acts 1–5.
3 · The wage-cost drag’s strength is a floor, not a point
estimate. The crowding-out mechanism this page describes is statistically identified on the
corrected data — the fit actively demands it — but the preferred estimate sits at the top of the
range the calibration allowed. The data says “at least this much”, not “exactly this
much”: the displacement effects you will see are, if anything, understated.
Whether the Island actively courts a next sector from 2025.
The response is fitted on the one promotion episode the data contains — e-gaming — so read
this lever as “a sector that behaves as e-gaming did”. A different sector may respond quite
differently, and assuming the next flavour repeats the last one is itself a bet, not a law.
“None” switches active promotion off.
How hard the courting is, from 2025. 25% is roughly today’s
level (the baseline assumption); 100% is the 2005–15 push.
Building rate from 2025, as a percentage of the model’s
own baseline path — the readout above converts your setting into net new homes in 2025. The model
builds more when real prices rise, so a building surge partly cools its own market. For scale:
officially recorded completions were 199 (2023) and 150 (2024).
How freely workers move in when jobs pull, from 2025.
100% matches calibrated history; 0% closes the door.
A model, not a forecast. It is the story above made mechanical: use it to compare
policies, not to predict 2035. Controls apply from 2025; history to 2024 is the calibrated baseline and
does not move. Beyond 2026 the outside world is held steady (UK growth 1.4%/yr, inflation 2%, policy
rate 4%).
Real output 2035
–
index, 1998 = 100
Population 2035
–
residents
Median house 2035
–
nominal
Real output
Real output index, 1998 = 100
Your runBaseline modelObserved (Manx GDP)
Observed: verified Manx real GDP index chained from National Income Report
Table 3 growth rates, fiscal-year keyed (1998 = 1998/99). The model under-predicts the pre-2008
boom — see the notice above.
House prices
Median house price (houses only), £ thousand, nominal
Your runBaseline modelHistory (observed)
Observed: Isle of Man Housing Market Reports, Appendix One: median house price
(houses), annual average of quarters, 2011–2025. The pass-6 model is fitted to this official
houses-only basis directly — no display rescaling.
Observed: Treasury Customs & Excise outturns, nominal. Three regimes:
Common Purse share formula to 2008; 2009–15 observed receipts imposed (the rebasing window
— sliders cannot move it); FERSA retrospective share formula from 2016.
Why the amber dots do not move
The window the model refuses to simulate
The revenue-sharing
rebasings of 2010 and 2012 were renegotiations imposed from outside — a change in the formula
itself, not in island activity. A model of the Island can generate receipts from the economy it
simulates; it cannot generate a treaty renegotiation in London. So for 2009–2015 this simulator
plays back the observed receipts unchanged, whatever you do with the sliders, and its endogenous
revenue machinery only resumes with the FERSA formula in 2016. Treat those seven amber dots as
quotation, not prediction. The fit statistics quoted below exclude them for exactly that reason.
Five sectors under your run
Real output by model block, £ million (2023/24 prices), 1998–2035
Model output. The finance block splits in 2007/08 (first year the accounts split
it): the labour-heavy half feels the wage-cost drag; the margin-booked half is nearly immune. The
sheltered block (retail, construction, private services) sells into on-island income; the exposed
block (agriculture, manufacturing) is world-price-capped.
Data table — your run v baselineFor the technically minded
The simulator is a line-for-line JavaScript port of the calibration pass 6 systems-dynamics
model (16 August 2026; 72 free parameters, staged fit, final cost 7.5654) — the recalibration on
verified Manx data after the Jersey series contamination documented above, fitted against a target set
independently verified cell by cell (2,678 of 2,678 cells). Mechanics: an annual deterministic loop,
1998–2035, with five private-sector blocks (labour-heavy finance, margin-booked finance, e-gaming,
sheltered and exposed existing economy), migration responding to vacancy pressure, and house prices driven
by the people-to-dwellings ratio. The wage-cost mechanism is a persistence wedge: sectors
are dragged in proportion to the relative-cost index’s excess over its own trailing five-year mean
— cost pressure bites while it outruns the adapted norm, fades as contracts and bargains catch up
(about a five-year cycle), and a falling cost level yields only a transient, bounded
competitiveness gain, never a permanent subsidy (verified by a forced-fall test). Revenue runs in its
three historical regimes: the Common Purse share formula, endogenous, to 2008; observed
receipts imposed for the 2009–15 rebasing window (see the notice above — fit statistics are
reported excluding those seven years); and the FERSA retrospective rule — a fitted share of the
trailing five-year revenue base — from 2016. Beyond 2026, external inputs are held steady: UK growth
1.4%/yr, UK finance growth 1%/yr, inflation 2%, policy rate 4%, promotion 0.25. Levers replace their
baseline paths from 2025 only; at their defaults the run is bit-identical to the calibrated baseline.
Port parity: sim_p6.js reproduces the Python original (model_p6.py, params_p6.json) over
1998–2024 to a maximum relative deviation of 3×10−12 per cent across all 48
output series — float rounding only (harness: parity.js in the model bundle).