Scotland Act 1998 (after the 74% Yes referendum of 1997), Scotland Act 2012 (borrowing and partial income tax), Scotland Act 2016 (permanence in statute, full income tax rates and bands, welfare powers). The Scottish Parliament of 129 MSPs is elected by AMS; primary law-making runs across health, education, justice, agriculture, environment and transport.
The 2014 independence referendum (45% Yes, 55% No) and the SNP's long hold on government 2007-2024 make Scotland the settlement that carries the Union question.
The Scotland Act 1998 devolved health, education, justice, agriculture, environment and transport; the 2012 Act added borrowing and partial income tax; the 2016 Act transferred full income tax rates and bands plus substantial welfare powers such as the Scottish Child Payment.
Lead any real-power paragraph with the tax transfer - nobody can dismiss full control of income tax rates as administrative devolution.
The 1997 referendum returned 74% Yes - the strongest founding mandate of any settlement - and Holyrood's 129 MSPs are elected by the proportional AMS system.
Contrast with Wales's 50.3% in the same year: mandate strength explains why Scotland's settlement started deepest and grew fastest.
Scottish students pay no university tuition while English students do, and Scotland introduced minimum unit alcohol pricing - policies visibly different from England's in areas Holyrood controls.
Divergence cuts both ways in essays: proof the devolved power is real, and fuel for the funding-fairness complaint that citizens in different nations get different deals.
The SNP won every Scottish election from 2007 to 2024 and took independence to 45% in the 2014 referendum - devolution built the platform from which independence became a permanent live question.
The counter sits in the same row: 2014 was a clear No, and Labour's 2024 recovery to 35 Scottish seats showed unionist parties can win the ground back.
The settlement has expanded by statute three times (1998, 2012, 2016) without suspension or reversal, and the Scotland Act 2016 declared the Parliament permanent in statute.
Contrast with Northern Ireland's repeated collapses: Scotland is the stability benchmark the other settlements are measured against.
Scotland holds powers no other settlement has: full income tax rates and bands, a separate legal system that pre-dates devolution, and substantial devolved welfare. Wales lacks justice and full tax powers; Northern Ireland has no income tax devolution at all.
Use the comparative powers table - the exam rewards knowing exactly who has what.
The Internal Market Act 2020 was passed without Holyrood's consent, constraining devolved regulation after Brexit - and Miller 1 (2017) had already confirmed the Sewel Convention is not legally enforceable.
However deep the settlement, the legal hierarchy still runs downhill from Westminster, and 2020 proved the override gets used. Compare the US federalism grid: in a federal system the states' powers are constitutionally entrenched and Washington has no equivalent override - the contrast is what makes devolution devolution.
Government of Wales Act 1998 created an Assembly with secondary powers only, after a 50.3% Yes on a 50% turnout. The 2011 referendum (63% Yes) granted primary law-making power. The Wales Act 2017 moved Wales to the reserved-powers model; Welsh rates of income tax followed in 2019. The Senedd expands to 96 members at the 2026 election.
Wales is the consent story: every expansion was approved by referendum or built on one.
Wales started with secondary powers only under the Government of Wales Act 1998, gained primary law-making after the 2011 referendum, moved to the reserved-powers model under the Wales Act 2017, and has set Welsh rates of income tax since 2019.
Use Wales as the trajectory example: the settlement that proves devolved power deepens over time rather than standing still.
The 1997 referendum passed by just 50.3% on a 50% turnout - the weakest founding consent of any settlement. The repair came in 2011, when 63% voted Yes to primary law-making powers.
The two votes together make the consent argument: each expansion of Welsh devolution was approved or legitimised by referendum.
Wales abolished prescription charges and runs its own curriculum and exam system through the WJEC - a distinct path in health and education that voters can see.
Divergence on this scale, sustained for two decades, is the evidence that even the settlement that started shallowest now exercises real policy choice.
Welsh devolution grew by referendum at each stage and produced no serious independence movement - Welsh independence polling sits around 20%, against 40-50% in Scotland.
Wales is the strengthened-the-Union side's best nation-level evidence: self-government inside the UK, expanding with consent, without separatist pressure.
The settlement has expanded steadily - 1998, the 2011 referendum, the Wales Act 2017, income tax powers in 2019 - and the Senedd grows to 96 members elected by a more proportional system at the 2026 election.
No suspensions, no reversals: alongside Scotland, the proof that devolution can be a one-way ratchet of orderly growth.
Wales now shares Scotland's reserved-powers model but not its scope: justice and policing remain reserved to Westminster (the notable exception among the three settlements), and only 10p of each income tax band is devolved against Scotland's full rates and bands.
Same model, smaller settlement - asymmetry persisting even as Wales catches up.
The reserved-powers model sounds generous, but Westminster writes the reserved list - and the Internal Market Act 2020 passed without Welsh consent just as it passed without Scottish consent.
Whatever the model, the legal last word stays at Westminster, which is the definition of devolution rather than federalism.
The Northern Ireland Act 1998 implemented the Good Friday Agreement: a mandatory power-sharing executive with cross-community consent rules, approved by referendum on both sides of the border.
The institutions have been suspended repeatedly - 2002-07, 2017-20, 2022-24 - and Westminster has legislated directly for Northern Ireland during suspensions. Post-Brexit, the Irish Sea border added a new strain the settlement was never designed to carry.
Stormont holds primary law-making power across health, education and (since the 2010 Hillsborough Agreement) justice and policing - but exercises it through a mandatory coalition in which the First Minister and deputy First Minister hold a joint office across the unionist and nationalist blocs.
Real power, conditional on the two communities agreeing to operate it - which is the next column's problem.
The Good Friday Agreement was approved by referendum in 1998 on both sides of the border - the only UK settlement endorsed by voters in two states, with cross-border bodies linking Stormont to the Republic built into the design.
That dual consent gives the settlement a legitimacy no Westminster statute alone could supply.
The institutions were suspended 2002-07, 2017-20 and 2022-24, and during suspensions Westminster legislated directly - including decriminalising abortion in 2019 while Stormont was down.
Years without a sitting Assembly mean years without devolved policy choices: collapse, not lack of powers, is what limits Northern Irish divergence.
Power-sharing manages the constitutional dispute rather than settling it, and Brexit added a strain the 1998 design never anticipated: the Irish Sea trade border under the Protocol triggered the 2022-24 suspension.
The counter: the institutions returned in February 2024 with Michelle O'Neill as First Minister - the framework held even under post-Brexit pressure.
Three suspensions - 2002-07, 2017-20 and 2022-24, the last over the DUP's objections to the Protocol - make Stormont the settlement that keeps collapsing. Yet February 2024 saw power-sharing restored after the longest suspension.
Run both halves: fragile by design, but the design has so far always recovered.
No other settlement looks like this: mandatory coalition between the largest unionist and nationalist parties, cross-community consent rules, STV elections locked by the Good Friday Agreement, and cross-border bodies with another sovereign state.
Use Northern Ireland to show asymmetry is deliberate design, not accident - the structure exists because of the conflict it ended.
Every suspension proved the point: when Stormont collapsed, Westminster legislated directly for Northern Ireland - including the 2019 abortion intervention - because the legal power never left.
Devolution paused, sovereignty operated. No other row demonstrates the statutory nature of devolution so repeatedly.
England has no devolved parliament: English laws are made at Westminster by the UK Parliament. The 2004 North East referendum rejected a regional assembly by 78%. English Votes for English Laws ran from 2015 to its abolition in 2021. Metro mayors have budgets and profile, not law-making power.
The English Question - who speaks for England inside an asymmetric Union - is the settlement's unsolved problem.
England is the only nation without a legislature: English laws are made at Westminster, where Scottish, Welsh and Northern Irish MPs can vote on them. Metro mayors - Burnham in Greater Manchester since 2017, Khan in London since 2016 - hold budgets over transport, skills and housing, but no law-making power.
A mayoral patchwork is not a parliament, and the gap is the row's whole story. For a comparative point, the US federalism grid shows the opposite design: every state has a legislature with the same constitutional status, so a gap like England's could not exist there.
The one attempt to build English regional devolution was put to voters and crushed: the 2004 North East referendum rejected a regional assembly by 78%.
Use this to complicate the asymmetry complaint - England lacks devolution partly because English voters, when asked, said no.
There is no separate English policy voice: NHS England, English schools and English transport are run by the UK government, so English policy is whatever the Westminster government of the day decides.
The divergence runs the other way - Scotland and Wales diverge from England, and English voters watching free Scottish tuition and free Welsh prescriptions feed the fairness argument.
The West Lothian question - why can Scottish MPs vote on English-only laws when English MPs cannot vote on Scottish ones? - remains unresolved, and the Barnett Formula gives Scotland higher per-capita spending than England.
Reform UK and parts of the Conservative right now campaign on exactly these imbalances, making the English Question a live electoral issue, not just a seminar point.
England's one structural experiment, English Votes for English Laws, ran from 2015 to its abolition in July 2021 - introduced by one government and scrapped by the next without a referendum either way.
Six years and out: contrast with Scotland and Wales, where every change has stuck and built.
Around 84% of the UK population lives in the one nation with no devolved legislature - asymmetry at its sharpest, since the largest nation by far is the one left out of the settlement.
This single statistic carries an essay paragraph: whatever devolution solved, it created the English Question in the same stroke.
English laws are made directly by the sovereign UK Parliament - no devolution, no consent convention, no Sewel question even arises. England is governed by the unmediated Westminster system the other three nations devolved away from.
Use this cell to define the baseline the other rows are measured against.
Authorised by Westminster through the Edinburgh Agreement 2012, the referendum asked whether Scotland should be an independent country: 45% Yes, 55% No, on an 85% turnout. The pre-vote Vow led to the Smith Commission and the Scotland Act 2016.
In essays the referendum cuts both ways: the Union survived its sternest test, and independence was normalised as a live political option.
The pre-vote Vow promising further powers led to the Smith Commission and then the Scotland Act 2016: permanence in statute, full income tax rates and bands, substantial welfare devolution.
The referendum Scotland lost delivered the biggest single transfer of powers since 1998 - a result worth a sentence in any real-power paragraph.
Turnout reached 85% - the strongest democratic exercise in modern UK history - on the most fundamental question a polity can ask itself.
Whatever side of the Union debate an essay takes, the referendum demonstrated devolved Scotland could conduct an existential vote peacefully and at extraordinary participation levels.
The referendum decided constitutional status, not policy - no school, hospital or tax changed because of the vote itself. The policy consequences came indirectly, through the 2016 Act the result triggered.
Keep this row for the Union, mandate and asymmetry columns where it scores heavily.
The Union survived its sternest test at 55-45 - the strengthened side reads this as devolution working as a constitutional safety valve, giving Scotland a route to decide without rupture. The weakened side reads 45% Yes as independence normalised: support has polled around 40-50% ever since.
Both readings are defensible; the examiner rewards choosing one and weighing the other.
The vote was billed as once in a generation, yet the SNP returned to demanding a second referendum within years, and independence support has stayed around 40-50% in polling.
One referendum did not settle the question - which is the instability the weakened-the-Union side builds on.
The post-referendum Scotland Act 2016 pushed Scotland's settlement further ahead of Wales and Northern Ireland - full income tax devolution and welfare powers neither of the others holds.
Each crisis in one nation deepens the overall unevenness: the referendum is the clearest example of asymmetry compounding.
The vote happened because Westminster authorised it: the Edinburgh Agreement 2012 transferred the power to hold the referendum, on terms the UK government negotiated.
Even Scotland's independence question ran through Westminster's legal gate - the sharpest illustration that devolved Scotland cannot decide its own constitutional future unilaterally.
Passed after Brexit to preserve a single UK market, the Act applies mutual-recognition rules that constrain how far devolved regulation can diverge. It was passed over the objection of all three devolved legislatures - the clearest modern demonstration that the Sewel Convention is a convention, not a rule.
For the devolved governments it is a Westminster power grab; for Westminster it is the sovereignty devolution always preserved.
The Act's mutual-recognition rules mean goods and services acceptable in one part of the UK must be accepted in the others - which caps how far devolved regulation can practically diverge, whatever powers the devolution Acts confer on paper.
A constraint on the use of devolved power, imposed without devolved consent: the row reads minus almost straight down.
All three devolved legislatures refused consent and the Act passed anyway - the clearest modern demonstration that the Sewel Convention, confirmed legally unenforceable in Miller 1 (2017), yields when Westminster decides it must.
No referendum, no devolved agreement: the contrast with every settlement row above is the point.
Mutual recognition undercuts divergence in practice: a devolved standard on goods can be set, but products meeting English standards must still be accepted, hollowing out the regulation's effect.
Use this to sharpen the divergence column - the Act attacks the mechanism by which devolved policy difference becomes real.
The devolved governments treated the Act as a Westminster power grab arriving on top of a Brexit that Scotland and Northern Ireland had voted against - post-Brexit strain layered on constitutional grievance.
Westminster's defence is the next column: preserving a single UK market is what a union government is for. The disagreement is the essay.
The Act reopened a settlement Brexit had already strained: it arrived alongside the Protocol dispute that suspended Stormont from 2022 to 2024, and it converted the Sewel Convention from working norm into open grievance.
Twenty years of settlement-building, then one statute that set the governments against each other.
Unusually for this grid, the Act treats the settlements alike: mutual-recognition rules apply uniformly, cutting across Scotland, Wales and Northern Ireland regardless of how deep each settlement runs.
A uniform ceiling over asymmetric floors - worth a sentence when an essay needs to show not everything in the territorial constitution is uneven.
Passed over the objection of all three devolved legislatures, the Act is the working proof of what Miller 1 (2017) established in principle: Sewel is a convention, and Westminster can legislate on devolved matters without consent when it chooses.
This is the cell to quote in any sovereignty-intact judgement - the override is not theoretical, it happened in 2020.