Why Evaluation Matters More Than Ever: Lessons from Recent Local Government Evidence

/ jlowther2015 / Leave a comment

Jason Lowther

For many researchers and analysts working in local government, evaluation can sometimes feel like yet another additional reporting requirement attached to programme delivery. Yet a series of recent evaluations (including homelessness, local growth and skills, and community ownership) demonstrates why evaluation is increasingly central to effective public policy. These studies also highlight several methodological lessons that should influence how local authorities design, commission and use evaluation in the future.

The first lesson is that evaluation is no longer primarily about proving impact. Increasingly, it is about understanding systems. Recent evaluations of homelessness and rough sleeping have moved beyond individual interventions to examine how housing, welfare, health, justice and local government interact. Rather than asking simply whether a programme worked, they explored why outcomes emerge within a complex environment. The findings suggest that fragmented funding, conflicting incentives and a persistent focus on crisis response often undermine preventative approaches, regardless of the quality of local services.

This reflects a broader shift towards complexity-aware evaluation. As councils have known for years, many local challenges, including homelessness, economic development and community regeneration, are influenced by multiple organisations and external factors. Traditional experimental approaches are often difficult to apply in such settings. Consequently, recent evaluations have increasingly used systems mapping, qualitative fieldwork, contribution analysis and mixed-method designs to understand causal pathways rather than attempting to isolate a single intervention effect. The emphasis is increasingly on understanding contribution rather than claiming attribution.  We may need to re-consider having systematic reviews of RCTs atop our methodological pyramids?

A second lesson concerns the value of process evaluation. Across multiple policy areas, some of the most useful findings relate not to outcomes but to implementation. Evaluations of local growth funds, skills programmes and place-based regeneration initiatives consistently show that local flexibility enables adaptation to local circumstances. Conversely, they also reveal how compressed delivery timescales can constrain innovation, limit stakeholder engagement and encourage delivery of readily implementable projects rather than longer-term transformation.

Third, the evaluations expose the continuing challenge of measuring long-term outcomes. In local growth programmes, evaluators highlight the difficulty of attributing changes in productivity, employment or economic resilience to specific interventions when impacts emerge over many years and are shaped by wider economic conditions. Similarly, evaluations of community ownership initiatives have already generated evidence of strengthened community participation, local pride and volunteering, but economic sustainability and growth impacts remain less certain because many projects remain at an early stage of development.

This raises an important methodological point. Public sector evaluation often demands evidence before outcomes have had sufficient time to emerge. Analysts should therefore distinguish carefully between outputs, intermediate outcomes and longer-term impacts. Developing robust theories of change and identifying realistic milestones along a causal pathway may provide more meaningful evidence than premature attempts to demonstrate definitive impact. In many cases, demonstrating credible progress towards expected outcomes may be a more realistic and useful objective than proving final impacts.

A fourth lesson concerns the importance of context. Recent place-based evaluations show that outcomes depend heavily on local conditions, governance maturity, partnership networks and existing organisational capacity. Similarly, community ownership projects appear most successful where community organisations already possess strong leadership, governance and project management capabilities. These findings reinforce a recurring conclusion from evaluation research: “what works” is often contingent rather than universal.

For managers commissioning evaluations, this means moving beyond simple questions of effectiveness. The more useful questions are often: what works, for whom, in which circumstances, and through what mechanisms? Evaluations that address these questions may generate practical learning that is far more valuable than a single overall success measure.

Perhaps most importantly, these evaluations demonstrate that good evaluation supports organisational learning rather than accountability alone. While demonstrating value for money remains important, evaluation also provides a mechanism for understanding complexity, adapting delivery, strengthening partnerships and improving future decision-making. In an era of fiscal constraint, rising demand and increasing policy complexity, this learning function may be its greatest contribution.

Taken together, these recent evaluations point towards a more mature understanding of evidence in local public services. They show that evaluation is at its most valuable when it helps organisations understand systems, implementation, context and causation. For local government researchers and analysts, the challenge is not simply to generate more evidence, but to generate evidence that helps decision-makers navigate complexity. In an environment where public services face growing pressures and finite resources, that capability is becoming an essential component of effective public policy and organisational improvement.

Discussion of the available evaluations mentioned are available here.

This article was first published in the September 2026 edition of the LARIA (Local Area Research and Intelligence Association) newsletter.


Jason is Director of the Institute of Local Government Studies at the University of Birmingham. Email [email protected]

Fairness: the most ill-defined term

/ jlowther2015 / 2 Comments
Photo by cottonbro studio on Pexels.com

Clive Stevens

“My God, I never realised writing up a PhD was so difficult”, and may I add, it is frustrating, stressful and sleep depriving too. Perhaps my subject area is worse, when the meaning of nearly every word is contested and often used with imprecision. Take the word equity, for example, this has four meanings. One meaning is used to describe restorative justice when the law results in an unfair outcome, it is the legal principle of Equity in English Law. Equity also describes the share capital of a business with its related meaning of proportionality, what you get out is proportional to what you put in. Equity can also mean the opposite; in situations of disadvantage, sometimes it is desirable to put in more resource and support to level up the playing field, towards ‘equality of opportunity’,  please don’t set me off on the word equality! And finally, equity can be used as a synonym for any fairness principle as in, “this is equitable”.

Equity means all of those things, it depends on context, so a report saying “we will provide an equitable solution” means different things to lawyers, economists, business persons, teachers, doctors, and politicians. Councillors have taken different career paths; if equity is proposed by an officer, one councillor might understand it as, “we will help the disadvantaged” and another, sitting opposite, might understand, “we will help those who work harder” and that is before we ask the lawyers. “The subject cries out for conceptual clarity”, those are not my words, although I do agree, I have seven systematic review papers from different disciplines saying just that. Hence my write up is progressing one word at a time, all the way to the 80,000!

Although the PhD write up is slow, my research has already had some practical use in a Council, not Bristol, but one in the South West. You might recall Marcus Rashford’s successful campaign in 2020, he received an MBE; it resulted in the provision of holiday food vouchers for those children in receipt of free school meals, so they could afford to eat during the holidays. That is being withdrawn this year, nationally, and is being replaced by locally agreed systems relying on some holiday activities that provide a free lunch plus Community Fridge or Larder memberships. The Government Guidance says this has to be done fairly without explaining what they mean by fairly. They want each family to apply and be assessed to access these free activities, bearing in mind the family has already been assessed for free school meal provision. And this change is combined with budget cuts.

I was asked by some councillors to use my newly developed fairness template to analyse this and assist them because they were concerned about their Council’s proposals. The fairness template has eight different fairness principles, called realms; I applied each realm to the proposals and it was clear that certain aspects of fairness had not been taken into account. Fair process had been considered, not surprising as this was written by officers, but there was little assessment of outcome fairness. I will give two examples, firstly geographical equity, there was no consideration of families living in areas where there are no Community Larders. Secondly, there was no assessment of proportionality by family size; the food vouchers, those being withdrawn, would have fed a family with three children at lunchtimes for all the summer holidays. However, under the new scheme, free lunches run out when the allocation of 18 days of activities has been met, but there are about 30 weekdays in a summer holiday so 12 days are not covered. The councillors made their statements and the upshot was that the officers were tasked to work out more detailed proposals. It is scheduled to come back to scrutiny and decision making later this Autumn.

So, whilst I debate the meaning of many words in my PhD and the write-up is taking longer than planned, the conceptual clarity it should bring will not only reduce confusion but may improve the quality of reports, debate and even decision making. I can see uses for my work already, and that makes it all worthwhile.

Clive is a former councillor in Bristol and author of the book on Local Government, After the Revolution, Clive followed up on politicians’ conceptions of fairness. He is now in his final year of a PhD at the University of Bristol, interviews are complete and is writing it up. He is hoping Local Government will take heed and improve the clarity of fairness and related concepts.

Learning from the 2026 elections

/ jlowther2015 / Leave a comment
Photo credit: https://www.southoxon.gov.uk/south-oxfordshire-district-council/local-democracy-and-elections/elections-and-voting/first-time-voters-guide/what-happens-at-an-election-count/

Jason Lowther

The Electoral Commission’s new report on the 2026 English local government and mayoral elections provides largely positive picture of local democracy. Turnout was the highest seen at comparable local elections for a decade, public confidence in the electoral process remained strong, and the overwhelming majority of voters reported satisfaction with both registration and voting. At the same time, the report highlights several challenges that local authorities, elected members and policymakers will need to address if they are to further strengthen local democracy.

Perhaps the most striking finding is the increase in voter turnout. Participation reached 41.2% in metropolitan, London borough, unitary and district council elections and 45.2% in county council elections. The latter represents the highest county council turnout for three decades outside elections held alongside a general election or referendum. These figures suggest that local government increasingly matters to citizens and that many voters remain willing to engage when they feel local issues have real consequences for their communities.

The report also points to generally high levels of public confidence in election administration. Around 94% of voters were satisfied with the voting process.  While 73% of eligible voters expressed confidence that the elections were well run, 9% were not satisfied – reasons for this include the perceived lack of useful information about candidates and/or parties, the voting system being seen as unfair, or the belief that the process of voting is not secure. Local authorities and electoral administrators should take credit for delivering elections effectively despite continuing operational pressures and resource constraints.

However, the report identifies several areas where citizen engagement could be improved. One concerns voter registration. Although most people were satisfied with the registration process, younger voters were less positive, and a significant proportion of dissatisfied respondents argued that registration should be automatic. The Electoral Commission reiterates its support for a consistent automated voter registration system and urges governments across the UK to coordinate reforms.

Accessibility is another continuing concern. Despite very significant efforts by returning officers to provide equipment and support in polling stations, disabled voters still face barriers to participation. Awareness of available assistance remains relatively low, even among people who could benefit from it. This finding highlights the importance of ongoing public communication and inclusive service design.

The report also raises important questions about information quality and democratic engagement. While awareness of the elections was high, many voters remained concerned about misinformation. More than half of respondents reported seeing misinformation during the campaign and around one in ten reported encountering deepfake material. Although the Commission’s pilot work found no evidence that political deepfakes were widespread, the findings underline growing risks associated with digital campaigning.

A major challenge concerns the experience of candidates. Abuse, intimidation and harassment, especially online, continue to affect many people standing for election. Candidates who were female, from an ethnic minority background, had a disability or were standing for Reform UK were more likely to experience abuse and intimidation.  The report warns that these experiences are influencing how some individuals campaign, for example avoiding canvassing alone, and one third of respondents considered not standing for election because of harassment, intimidation or concerns about future abuse.

The report also highlights a wider issue for local democracy: the availability of reliable information about candidates and parties. Although large majorities found it easy to obtain information about voter ID requirements and voting methods, fewer found it easy to access information about candidates. This suggests that there remains a gap between procedural information and substantive information about political choices.

Finally, the report notes the pressures facing electoral administrators. Recruitment challenges, supplier issues and uncertainty surrounding potential electoral reforms all created difficulties for those delivering the elections. While these problems were largely managed successfully, councils and government need to address the long-term capacity and resilience of local electoral services. Effective citizen engagement ultimately depends on electoral systems that remain adequately resourced and capable of adapting to future reforms.

Overall, the 2026 elections offer grounds for cautious optimism. Higher turnout, strong voter satisfaction and confidence in election administration suggest that local democracy remains resilient. Yet the report also shows that participation cannot be taken for granted. Improving accessibility, modernising voter registration, combating misinformation and protecting candidates from abuse will all be central to sustaining citizen engagement and local democracy.

Counting the cost of talking to residents — a question for reorganisation

/ jlowther2015 / Leave a comment

Eren Bahadir Pehlivan

Between July and September 2026, I sent Freedom of Information requests to eighty-nine district, borough and city councils in England. Thirty-two went out in July; a further fifty-seven in August, adding questions about the annual electoral canvass. The full wording is in the appendix, and readers should judge it for themselves.

The questions were deliberately dull. How much did the council spend printing and posting correspondence to residents last year? How many items did it send? What did the canvass cost, and how did people respond to it?

As at 7 September 2026, seventy-three councils had replied. Thirty-six gave substantive answers. Twenty-one produced figures complete and consistent enough to compare with each other. Responses arriving after that date are not counted here. Excluding one council that restricts commercial re-use, the remaining twenty spent more than £4.4 million in a year on print and post alone. The median was £205,890.

Eighteen recorded both spend and volume. Divide one by the other and the cost of a single item runs from 74 pence to £15.01.

That spread is the first thing to be careful about, because it is not a finding about efficiency.

Councils count different things. Some totals include overheads, machine leases, paper and staff time; others are franking alone. Where a service is outsourced, the largest mailing may sit inside a contract price and never appear as postage at all. Shared services split costs between partners on conventions of their own. Year-on-year movements can reflect a restructure, a change of supplier, or a re-coded nominal ledger rather than anything operational. Geography, service mix and the degree of outsourcing all move the number. None of that is bad practice, and none of it can be read off a single figure.

What it does mean is that these figures are not comparable in the way a reader might assume, and that most councils cannot say which of those conventions their own number reflects.

Some answers, though, are not explained by accounting convention. A city council serving around 138,000 people replied that its total print and postage spend was “not held”, and that the number of items posted was “not held” as well. A district put its whole-council outgoing post at “£500k approx” and does not record how many items that buys. One council asked ten of its own departments and they disagreed with each other about whether the authority uses a digital notification platform.

A borough in the east of England explained that it had “recently wiped our 2025 canvassing data to commence canvassing for the current year”, rebuilt its answer from invoices and an Electoral Commission feedback sheet, and warned that it “may be incomplete.”

These councils answered courteously and within the deadline. The information was not there to give.

The canvass makes the point more precisely, because every authority is doing the same statutory job. Among four councils that answered in detail, one made no household visits at all; the others paid for 7,254, 8,839 and 12,306. Cost per property, where it could be isolated, clustered between 67 and 73 pence — a similarity that conceals four different operating models.

It also unsettles an assumption worth naming. In two of those councils about three-quarters of canvass responses arrived over the internet. In the third, which paid canvassers to call at the door, the internet share was 36 per cent and nearly half of all responses were collected in person. The digital share of responses appears to track the council’s chosen method rather than residents’ preferences. That is a caution for anyone reading channel-shift statistics as evidence of public appetite.

All of this now matters more than it did. Reorganisation will move these budgets into new and larger unitary authorities from April 2028. A successor authority inherits whatever its predecessors could measure. Where a district cannot state its own communication spend, the new organisation begins without a baseline for it; where a dozen districts each used a different convention, aggregating them produces a figure that is precise and meaningless.

Which raises a governance question rather than a spending one. Two councils told me, in different words, that nobody owns this. One wrote that it has “no officer responsible for customer contact strategy”. Another said all decisions on communications and digital engagement for the canvass are taken through the elections office. Correspondence with residents is not a service with a director. It is a cost distributed across every service, which is precisely why it falls between owners.

Three modest arrangements would help, and none requires new spending: a named officer accountable for the whole channel; a single nominal code that records volume alongside expenditure; and a canvass costed as one activity rather than split between an elections budget and a general print budget.

Reorganisation is an expensive way to find out what you were spending. It is also, for the next eighteen months, the cheapest moment to start counting.

Figure

“What one item costs, according to councils’ own figures.” Produced by the author from the FOI responses; councils anonymised. Copyright held by the author; free to reproduce with the article.

Eren Bahadir Pehlivan is the founder of ADORASEC Ltd, working on notification between institutions and the people they serve. The dataset here is his own. Email: [email protected]

Appendix — the Freedom of Information request

Reproduced in full, as sent. This is the wording used for the second round (fifty-seven councils, August 2026). The first round (thirty-two councils, July 2026) used the same Part A and Part C questions but did not include Part B, and asked for “the most recently completed financial year” rather than naming the period. That relative wording proved ambiguous: several councils answered without stating which year the figures covered, and their responses are excluded from the comparisons above. It was replaced with absolute dates for the second round. The change is noted here because it bears directly on how far the two rounds can be read together.

This is a request under the Freedom of Information Act 2000. I have tried to keep it as light as possible to answer, and I have set out below what would make it lighter still.

Please provide the following for the 2025/26 financial year only — that is, 1 April 2025 to 31 March 2026.

If your figures for that period are not available, or if the council’s records are held to a different reporting period, please answer for whichever full year is most readily available and simply state in your response which period the figures cover. Knowing the period matters more to me than which year it is.

Four things before the questions, each intended to reduce the work involved.

First, we do not require a departmental breakdown. A single whole-organisation total is sufficient for every figure below. These figures are typically held against nominal codes in the finance ledger, and a total taken from there is exactly what we are asking for.

Second, if a figure is not separately recorded, please simply say so. That is a useful answer to us and we would much prefer it to an estimate. A number of authorities have already told us that certain figures are not held, and we have recorded that as a finding rather than a failure.

Third, the questions are grouped so that each group can be answered by one team: questions 1 to 3 by finance, questions 4 to 6 by the elections team, and questions 7 and 8 by whoever holds the staffing information.

I should be explicit about how those first two groups relate, because it has caused confusion elsewhere. Part A is not about elections. It concerns the organisation as a whole — every department, and all correspondence sent to residents whatever the service area, including council tax, benefits, waste, housing and planning. Part B is the election-related part and concerns only the annual canvass. The two do not overlap, and a figure covering electoral registration alone would not answer Part A.

Fourth, if any single question would take the request over the appropriate cost limit under section 12, please tell me which one and I will withdraw it, rather than the request being refused as a whole. I would be glad to agree a narrower wording by email before any work is done on it, in the spirit of the duty to advise and assist under section 16.

PART A — CORRESPONDENCE WITH RESIDENTS (ALL DEPARTMENTS — NOT ELECTIONS)

1. The total amount spent by the council on printing and posting correspondence to residents during the period, across the organisation as a whole. This is not limited to electoral or canvass mail. A single whole-organisation total is sufficient.

2. The number of items posted to residents during the same period, if this is recorded.

3. The total amount spent on SMS or text messaging to residents during the same period, and the number of messages sent, if these are recorded.

PART B — THE ANNUAL CANVASS

4. The total cost to the council of the 2025 annual canvass — that is, the most recently completed canvass, whose revised register was published on 1 December 2025 — including printing, postage, staffing and any personal canvasser visits. If the canvass cost is not recorded separately from the council’s general print and postage budget, please say so.

5. For that same canvass: the number of residential properties to which a canvass communication was sent; the number that did not respond to the first communication; the number of reminder communications sent; and the number of household visits carried out.

6. For that same canvass, the number of responses received by each route — paper, online, telephone and SMS — if this breakdown is recorded.

PART C — WHO HOLDS THIS WORK

7. The job title and, if the council is willing to share it, the direct contact details of the officer responsible for resident communications, customer contact strategy or digital transformation.

8. The job title and, if the council is willing to share it, the direct contact details of the Electoral Registration Officer and of the officer who manages the annual canvass day to day.

I am content to receive the response electronically, and I understand the statutory period is twenty working days.

One last thing, which is an offer rather than a request. The same questions have gone to a number of district councils in England. Any council that responds is welcome to a copy of the comparative results, showing where it sits against the others, with the case reference against every figure so that anything in it can be checked at source. It is provided at no cost and on no conditions. You do not need to do anything to receive it beyond answering this request, and you are under no obligation to take it.

Creative Bureaucracy Festival 2026: why the future of public services is about systems, not services

/ jlowther2015 / Leave a comment

Claire Dhami

I was fortunate enough to attend the Creative Bureaucracy Festival in Berlin this June, bringing together public servants, leaders and practitioners from across the world who are grappling with the same fundamental challenge: how do we make public services work better for people in increasingly complex and uncertain times.

The festival itself has an interesting origin. It was founded by Charles Landry and others as a provocation. The term ‘creative bureaucracy’ deliberately challenges the stereotype of bureaucracy as rigid, slow and rule bound. Instead, it puts forward a different idea that public institutions, at their best, can be imaginative, humane and capable of solving complex problems.

What began as a provocation has now become a global movement. The festival has grown into a space where thousands of public servants come together each year to share what is actually working in practice, often in spite of system constraints, and to learn from each other.

What struck me most, both from the festival and from our work in the West Midlands, is that the issues we face are remarkably consistent across countries. There is a growing recognition that many of the challenges in public services are not about marginal service gaps. They are about how the system itself is designed.

We have built systems that are logically organised around institutions, but in practice expect people to navigate multiple disconnected pathways. Responsibility is too often passed between organisations, rather than shared. The result is fragmented, transactional experiences, particularly for those with the most complex needs.

This is not a capability problem. It is a design problem.

One of the most powerful shifts discussed at the festival is moving from improving individual services to improving life outcomes. That may sound like a subtle change, but it fundamentally alters how we think about reform.

Instead of asking how to optimise what already exists, we start by asking what people actually need to thrive and then work backwards to shape the system around that. In our own work, this means focusing less on creating new interventions and more on redesigning the pathways between education, health, employment and community support, particularly at the points where people are most at risk of falling through the gaps.

This is where relational public services become essential. When support is organised around trusted relationships and continuity, rather than organisational boundaries and eligibility criteria, the experience for residents changes completely. It becomes earlier, more consistent and more human.

But perhaps the most important insight for me was how we think about innovation.

There is a tendency in the public sector to equate innovation with new programmes or pilots. The festival challenged that directly. A more effective model is to see innovation as a small but deliberate part of the system, around 10 percent of activity, designed to influence the other 90 percent.

The pilots themselves are not the outcome. They may succeed or fail. The real value lies in the learning they generate and whether that learning changes how the system behaves. That means influencing mainstream commissioning, policy decisions and how resources are actually deployed.

This requires discipline. Without a clear link between experimentation and system change, innovation remains interesting but peripheral. With it, innovation becomes a lever for reform.

It also clarifies the role of organisations like the Combined Authority. Our role is not to direct delivery or create new services. It is to influence how the system behaves. That means convening partners, aligning incentives, building shared ownership of outcomes and creating the conditions for change.

At the same time, the festival highlighted a fundamental contradiction that we see locally. We talk about partnership, prevention and place-based working, but continue to fund services in ways that reinforce fragmentation. If we are serious about system outcomes, we have to fund them differently, aligning resources around shared cohorts and goals rather than organisational silos.

The message I took away from Berlin is that the future of public services will not be defined by new initiatives. It will be defined by our ability to redesign systems so that good people are able to do the right thing, consistently.

When we get that right, innovation is no longer something separate. It becomes part of how the system learns, adapts and improves over time. And that is where meaningful, lasting change happens.

Claire Dhami is Head of systems change and inclusion at West Midlands Combined Authority.

When mayors hold the pencil: gerrymandering inside a territorial reform

/ jlowther2015 / Leave a comment

Miguel Ângelo Vilela Rodrigues (University of Minho), with Oliver Meza (CIDE) and Carmen Navarro (Universidad Autónoma de Madrid)

Gerrymandering is usually filed away as an American problem — a decennial ritual of oddly shaped congressional districts and litigation. But the practice does not require an American ballot. It requires only two things: a boundary that has to be redrawn, and a politician with an interest in where the line falls. Territorial reform supplies both, and it does so in countries that would never describe themselves as having a redistricting problem.

Portugal is a case in point. In our recent article in Local Government Studies, we look at the 2013 reform that amalgamated Portuguese sub-municipal units — the freguesias, roughly comparable to parishes — and ask a question that the amalgamation literature has largely left aside: not whether mergers deliver economies of scale or damage local democracy, but who decides which units merge with which, and what they get out of it.

A reform designed to be captured

The reform came out of the 2011 Memorandum of Understanding with the troika. Cutting the number of local entities was one of the commitments made, and central government initially tried to impose specific merger criteria from above. Local government rejected that outright. So the government softened its approach: it kept the targets — urban units to be reduced by 50–55 per cent, rural ones by 25–35 per cent — and handed the map itself to the mayors. Across 278 municipalities, the number of sub-municipal units fell from 4,269 to 3,092.

The design has an unusual feature. The units that had to merge were not the units that decided. Mayors drew the proposals; municipal assemblies approved them. And Portuguese municipal assemblies are hybrid bodies: alongside directly elected members, every sub-municipal president in the jurisdiction sits with a full vote. Sub-municipal presidents are also, in practice, part of a mayor’s electoral machinery — the people who campaign, mobilise and deliver a vote in their own patch.

That makes an amalgamation reform something more than an administrative exercise. Reducing the number of sub-municipal units aligned with the mayor’s party reduces the mayor’s support in the assembly that approves the budget and the investment plan, and thins out the voices in the field who will be needed at the next election. Conversely, a well-drawn map can pack opponents into a single unit or dilute them inside a larger friendly one. The pencil was handed to precisely the actors with most to gain from how it moved.

Vulnerability and feasibility

We modelled the mayor’s decision on two dimensions. The first is vulnerability: does the mayor have a majority of politically aligned sub-municipal units, or not? The second is feasibility: does the mayor command a majority of the elected seats in the municipal assembly, and so the ability to get a map approved without concessions?

Crossing them yields four situations. Where mayors are not vulnerable, redrawing is either irrelevant or indifferent — there is nothing to gain. Where they are vulnerable but lack an assembly majority, redrawing is crucial to survival but has to be negotiated, which means sharing the spoils. Where they are vulnerable and hold the assembly, redrawing is appealing: the incentive and the means coincide.

Using the 2009 sub-municipal election results projected onto the new 2013 map — the only information mayors actually had when drawing it — we then estimated multinomial logit models on all 278 affected municipalities.

The pattern is clear. Across the country, 51 per cent of municipalities ended up with more aligned sub-municipal units after a reform whose entire purpose was to reduce their number. In the appealing quadrant, the odds of an increase in aligned units rise more than nine-fold; in the crucial quadrant, where the map has to survive a coalition, they rise just over four-fold. The gap between the two is itself the finding: opposition parties will not sign off on a map that benefits the incumbent alone, so the gains get shared and diluted. Results hold when we substitute a vote-efficiency measure of the kind used in the US gerrymandering literature.

Why this should interest anyone outside Portugal

Two implications travel. The first is procedural. Structural reform is often delegated downwards on the sensible-sounding grounds that local actors know their own territory. They do — and they also know their own electoral arithmetic. Where the reform reshapes the boundaries that determine who sits in the body that scrutinises the reformer, delegation is not neutral. Anyone currently watching local government reorganisation in England, where boundaries, wards and the composition of new authorities are all in play, may find the logic familiar.

The second is about what happens next. Since 2013 there has been sustained pressure in Portugal to demerge some units, and the initiative for those petitions again rests with mayors. If the same rationality holds, demergers will be permitted where they are politically harmless and blocked where they are not — meaning the same territories get gerrymandered twice.

None of this requires bad faith or illegality. Every map was lawful and approved through proper procedure. That is rather the point: an effective opposition, not a rule about compactness, was the only thing that reliably constrained the outcome.

Miguel Ângelo Vilela Rodrigues is Associate Professor with Habilitation in the Department of Political Science, School of Economics and Management, University of Minho, and a researcher at CICP.

The full article — ‘Gerrymandering to survive: an explanation of the political conditions that shaped mayors’ decisions over an amalgamation process in Portugal’, by Miguel Rodrigues, Oliver Meza and Carmen Navarro — is published in Local Government Studies, 51(5), 993–1015. DOI: 10.1080/03003930.2024.2404104