Kiwi Mag

Getty Images

The government has framed its NZ$503 million budget spending on prisons as necessary to maintain public safety and manage a growing prison population, forecast to increase by 36% from the current 10,000 to 14,000 by 2035.

The appeal to public safety is tied to the goal of reducing violent crime, which most voters will understandably support.

But this broad messaging obscures two crucial facts. Most assaults in New Zealand happen inside private homes, not in public spaces. And the increase in the number of people in prison comes from an excessive remand population (people awaiting trial), not from an increase in serious offending.

More than half of those who are remanded will not receive a prison sentence once their case is heard. This shows they don’t present a risk to public safety.

As of March 2026, there were 4,537 people held on remand. Each is costing the taxpayer about $414 per day. This means the large share of the remand population that poses no threat to public safety is costing taxpayers nearly $1 million every single day.

From a fiscal perspective, it is a striking decision to build new prisons for the growing remand population instead of changing the law to release those who pose no risk on bail.

Why the remand population is growing

The remand population currently accounts for 41% of the prison population, up from 13% in 2000.

Over the past 25 years, a series of legislative changes has steadily increased the number of people on remand.

The most consequential change came when the previous National-led government amended the Bail Act in 2013 to tighten bail eligibility. Until then, most defendants were granted bail automatically.

It was largely on the prosecutor to prove the defendant posed a flight risk, might reoffend or interfere with justice – by intimidating witnesses for example – while on bail.

This amendment shifted the burden of proof onto defendants. Instead of bail, remand became the new norm, because it is harder to prove something will not happen. For example, how can you prove you will not intimidate witnesses?

As a result, more people are being detained, not because more people pose a proven risk, but because the legal threshold for release is now higher.

For the men and women held on remand, the consequences are often severe. People lose jobs, housing and family connections, all of which increase the likelihood of offending. Remand has become a costly and counterproductive system that harms both individuals and the public purse.

Because Māori are already more likely to be arrested, charged and have prior convictions due to long-standing systemic inequalities, a bail rule that forces defendants to prove they are not a risk perpetuates existing disparities.

The $1 million daily cost for keeping people unnecessarily on remand would be better divested into early intervention. We know a large number of people behind bars live with learning disabilities, fetal alcohol spectrum disorder, traumatic brain injuries and severe unresolved childhood trauma.

Many end up self-medicating with alcohol or drugs because they have never received proper support. Prisons are not designed to treat these underlying issues. Often they can make them worse.

The misleading appeal to public safety

Bail laws were tightened to protect public safety and the 2026 budget decision leans on that same argument.

The idea of public safety invokes images of physical threats in communal spaces. This overlooks that most violence in New Zealand happens at home. In 2025, around 10,500 people were convicted of a violent offence, of which more than half occured within families.

Family violence is a complex and often inter-generational problem.

Prison cannot teach emotional regulation, repair relationships or undo childhood trauma. Instead, it disrupts employment, isolates people from whānau and increases the stress factors that fuel further violence.

What works to address family harm is stable housing, addiction treatment, early intervention by specialist services and long-term therapeutic support.

These approaches strengthen families, reduce harm and cost far less than imprisonment. If we want safer homes and communities, we need to invest in solutions that change behaviour, not just contain it.

It is an old line, but it remains true: every dollar invested in early childhood support saves around 13 dollars in criminal justice costs later on.

A society that keeps expanding its prisons is admitting its social policies are not working. If we are serious about reducing crime, the government needs to invest earlier.

Antje Deckert does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

It costs a million dollars a day to keep low-risk defendants on remand. More prisons aren’t the answer Budget 2026 spends $503 million on expanding prisons. But 41% of the prison population are people on remand and more than half are released once their case is heard.

Even before receiving public funding, Wegovy has drawn much media attention in New Zealand for the dramatic difference it can make for weight loss.

But with its rise in popularity has come debate about what these next-generation drugs really mean for a worsening obesity crisis and its driving causes.

Last month, New Zealand’s drug-funding agency Pharmac added Wegovy to its list of medicines suitable for future public funding. If that happens – and it could quickly – the drug would initially be targeted at people with severe obesity, or who are overweight with related health conditions.

Right now, the drug’s private prescription costs – upwards of NZD$400 per month – places it beyond the reach of many New Zealanders, particularly those disproportionately affected by obesity. This has strengthened arguments that public funding could improve equity while reducing long-term health-care costs.

All the while, New Zealand continues to report some of the developed world’s highest obesity rates. Around one in three adults and one in eight children are today classified as obese, while roughly two thirds of adults are either overweight or obese.

That may make the prospect of public funding all the more attractive for the country and its health system. But it should also be asked: can these drugs really be expected to tip the scales against an epidemic rooted in complex social and environmental factors?

Why we’re hearing about Wegovy

Wegovy works through semaglutide, a drug that helps regulate appetite and blood sugar, making people feel fuller for longer.

Clinical trials have shown that, when the drug is taken in tandem with lifestyle change, effects can be striking. In one landmark study, participants lost around 15% of their body weight over 68 weeks, which was far more than those who instead took a placebo.

For some people, such weight-loss can be transformative – reducing the risk of diabetes, cardiovascular disease and other long-term conditions.

Still, the drug has drawbacks. People generally need to keep taking it to maintain weight loss, with many regaining weight once treatment stops. Side effects are common, and the long-term impacts remain uncertain.

NZ’s food ‘swamps and havens’

Debate around new drugs such as Wegovy can sometimes reduce obesity to a question of personal choice and responsibility.

But obesity in New Zealand has been rising for decades, particularly among children and more deprived communities, reflecting drivers that extend well beyond individual behaviour.

Research shows the environments people live in, for instance, strongly reflect what and how they eat. Highly processed, energy-dense foods are today widely available, aggressively marketed and often cheaper than healthier options.

But some sections of society are much more exposed than others. Māori and Pacific communities experience significantly higher rates of obesity, reflecting broader inequities in income, housing and access to healthy food.

Fast-food outlets are disproportionately concentrated in more deprived areas and these “food swamps” dominated by unhealthy options are common across New Zealand. In many neighbourhoods, unhealthy food is often the easiest and most accessible choice.

Children’s everyday environments also play a part. Studies suggest many New Zealand schools still make unhealthy food easier to access, with healthy food policies unevenly applied.

Yet there are also signs of what works. So-called “food havens” – community spaces designed to make healthy food affordable, accessible and culturally appropriate – show how local initiatives can improve food environments.

An intervention, but not an answer

All of this reinforces that obesity is fundamentally a systems problem, not one that can be solved through pharmaceutical treatment alone. Framing these drugs as a silver bullet risks diverting attention from the broader preventive changes needed to address its root causes.

There is also a policy tension. Public funding for Wegovy might indeed help reduce future health-care costs by lowering rates of diabetes and heart disease. But unless the drivers of obesity are addressed, the number of people needing treatment will likely continue to grow.

Evidence suggests that reducing obesity at a population level requires action on the environments which shape daily life.

That includes improving access to affordable healthy food, restricting marketing of unhealthy products, strengthening school food environments and addressing those broader social and economic conditions that influence health.

These interventions are obviously more complex than prescribing a drug. But they are also more likely to produce a lasting solution to a crisis that is bringing a heavier toll for New Zealand each year.

I work with some of the academics whose research is referred to in this article.

Can Wegovy move the needle on NZ’s obesity crisis, or simply treat its symptoms? The popular new drug can bring dramatic weight loss for some patients. But for NZ, tackling a rising obesity epidemic is likely to require much broader change.

Un des défauts de l’entretien annuel est qu’il est principalement tourné vers le passé. Charles Deluvio/Unsplash, CC BY

Pensées pour un monde du travail stable et prévisible, les évaluations annuelles peinent à saisir la complexité des contributions actuelles. Résultat : elles mesurent mal ce qui compte vraiment.

Chaque année, les organisations remettent à jour leur stratégie, annoncent de nouveaux indicateurs de performance, les fameux KPI (pour Key Performance Indicators), et affichent des objectifs ambitieux pour l’année à venir. Mais malgré l’évolution rapide des modes de travail, des technologies et des besoins des équipes, un élément reste étonnamment figé : les évaluations individuelles de performance.

La plupart d’entre nous voient très bien à quoi elles ressemblent : le formulaire classique à cases à cocher, les échelles de notation de un à dix, et cet espace vide un peu gênant réservé aux « commentaires complémentaires ».

Nous savons que ces dispositifs sont dépassés. Des travaux de recherche montrent depuis des années qu’ils regardent vers le passé, peuvent biaiser les comportements et négligent la collaboration ainsi que l’apprentissage. Nous savons qu’ils reposent sur une évaluation rétrospective d’une « performance » individuelle définie de manière étroite. Et nous savons qu’ils reflètent souvent mal le travail réel – par opposition à ce qui est effectivement valorisé. Pourtant, année après année, ils perdurent. Alors pourquoi continue-t-on à les utiliser ? (NDT : En France, la loi exige qu’un entretien professionnel ait lieu tous les deux ans. Cet entretien est à distinguer des évaluations annuelles dont nous parlons ici).

Un décalage avec le travail réel

De nombreux travaux expliquent pourquoi les dispositifs classiques d’évaluation de la performance et les KPI sont décevants. Une partie du problème tient au fait qu’ils brouillent la frontière entre rémunération et performance. Les chercheurs en management distinguent pourtant depuis longtemps la mesure de la performance – utilisée pour les salaires et les promotions – de l’amélioration de la performance, qui relève de l’apprentissage et du développement.

Fusionner ces deux objectifs dans un même processus annualisé crée une tension inhérente entre deux logiques pourtant très différentes. Le calendrier pose également problème. Des retours tardifs, concentrés une fois par an, sont souvent déjà dépassés et d’une utilité limitée, manquant des occasions clés d’amélioration tout au long de l’année.

Un article largement cité conclut d’ailleurs que les systèmes formels, fortement axés sur les notations, sont « fastidieux et de faible valeur ». L’amélioration de la performance, souligne-t-il, passe plutôt par des attentes continues, des retours en temps réel et des opportunités de développement – et non par des rituels annuels.

Une enquête de 2024 sur la gestion de la performance menée par le cabinet de conseil Betterworks va dans le même sens. Si la plupart des dirigeants jugent leurs dispositifs efficaces, les salariés dressent un constat bien différent : 44 % estiment qu’il s’agit d’un « échec significatif ».

Plus largement, les employés sont 57 % moins nombreux que les dirigeants à considérer que ces systèmes fonctionnent correctement. D’où vient un tel décalage – et pourquoi persiste-t-il ?

L’illusion d’objectivité

Le Center for Advanced Human Resource Studies de l’Université Cornell (État de New York) apporte quelques éléments de réponse. Son analyse des tendances en matière de gestion de la performance, publiée en 2025, montre que les systèmes traditionnels perdurent non pas parce qu’ils sont efficaces, mais parce qu’ils sont :

profondément ancrés dans les organisations, généralement liés à la rémunération, aux promotions, à la conformité et aux cycles RH,

perçus comme objectifs alors qu’ils ne le sont pas,

longs et coûteux à réformer, ce qui freine les transformations,

décalés par rapport aux attentes des salariés, dont seul un sur cinq se dit motivé par les dispositifs actuels.

Les indicateurs classiques de performance – production horaire, nombre de tâches réalisées, objectifs de vente – ont été conçus pour une époque où le travail était prévisible et ancré dans un lieu. Ils peinent à rendre compte de ce qui fait aujourd’hui la réussite des organisations.

Les sciences sociales rappellent depuis longtemps qu’à partir du moment où un indicateur devient un objectif, il cesse d’être une bonne mesure : les comportements s’ajustent pour maximiser le score plutôt que la réalité qu’il est censé refléter.

Concrètement, la sur-optimisation d’un KPI pousse à « jouer avec les chiffres » ou encourage des comportements contre-productifs. Les salariés prennent des raccourcis ou travaillent excessivement, au lieu d’améliorer réellement le résultat de leur travail.

Les organisations observent le même phénomène à une échelle plus large : se focaliser sur les volumes produits et la poursuite d’objectifs chiffrés peut nuire à la qualité, à la création de valeur à long terme et à la collaboration.

C’est d’autant plus vrai lorsque l’automatisation prend en charge les tâches routinières et que la contribution humaine se déplace vers la créativité, la résolution de problèmes et la production de valeur sur le long terme – autant d’éléments difficiles à résumer en un seul indicateur. À l’inverse, lorsque les indicateurs de performance manquent de clarté, les jugements subjectifs des supérieurs tendent à se substituer aux données réelles.

Pourtant, de nombreuses entreprises continuent de s’appuyer sur ces indicateurs anciens, simplement parce qu’ils sont familiers, quantifiables et profondément ancrés dans leurs pratiques. Le paradoxe est frappant : elles disposent de plus de données que jamais sur leurs salariés, mais leurs systèmes d’évaluation reposent encore sur des instantanés dépassés et des indicateurs réducteurs.

Comme le souligne un expert, des réalités de performance complexes sont souvent simplifiées à l’extrême dans des cadres qui « fabriquent » des données chiffrées pour créer une illusion d’objectivité.

Et si l’on mesurait ce qui compte vraiment ?

Les recherches sur les approches modernes de management de la performance montrent un net basculement : les évaluations annuelles rigides cèdent la place à des échanges continus, accompagnés par les managers.

Parmi les évolutions qui favorisent davantage la motivation et l’engagement des salariés :

des retours continus, en temps réel,

des objectifs à court terme, ajustables,

des échanges informels et réguliers entre managers et collaborateurs,

des dispositifs à « 360 degrés », où les retours proviennent de plusieurs collègues, offrant une vision plus équilibrée du travail collectif,

une logique de développement tournée vers l’avenir, plutôt que la notation des performances passées.

Ces approches reflètent mieux la manière dont le travail de qualité se construit réellement : progressivement, de façon collaborative et souvent imprévisible. À l’heure où les organisations définissent leurs objectifs pour le nouvel exercice, le véritable indicateur à adopter pourrait bien être celui qui évalue… leurs propres dispositifs de mesure de la performance.

Favorisent-ils la progression ? Captent-ils la valeur réelle du travail ? Suscitent-ils l’engagement ? Reflètent-ils ce que font réellement les équipes ? Répondre à ces questions permet d’identifier ce qui ne fonctionne plus. Et, dans bien des organisations, les systèmes d’évaluation de la performance figurent sans doute parmi les chantiers les plus urgents à lancer.

Les auteurs ne travaillent pas, ne conseillent pas, ne possèdent pas de parts, ne reçoivent pas de fonds d'une organisation qui pourrait tirer profit de cet article, et n'ont déclaré aucune autre affiliation que leur organisme de recherche.

Tout le monde sait que les entretiens annuels sont inefficaces, alors pourquoi en fait-on encore ? Rituels annuels, grilles de notation et « Key Performance Indicators » : les évaluations de performance structurent encore la vie des entreprises. Pourtant, de plus en plus de travaux montrent qu’elles sont mal adaptées au travail actuel.

NurPhoto/Getty Images

We’ve heard a lot about the artificial intelligence (AI) boom and how enormous amounts of money are being poured into companies building ever more powerful technologies.

That boom is now taking a new turn as major AI players edge closer to becoming publicly-traded companies.

According to reports, OpenAI is preparing to file confidentially for a public listing that could value the ChatGPT maker at hundreds of billions of dollars.
Rivals including Anthropic (Claude) and Elon Musk’s SpaceX – which just absorbed xAI (Grok) – are also moving toward the stock market.

What many people may not realise, however, is that, through retirement funds, pensions and other managed investments, they could end up owning shares in these giants – whether they choose to or not.

And while people might have moral concerns with the AI companies they’re tied to, the greater issue at play is about money, risk and who ends up holding it.

Where AI’s billions go

Building a cutting-edge AI system requires vast numbers of specialised computer chips, running nonstop in data centres that consume enough electricity to power a small city.

OpenAI plans to spend around US$50 billion on computing power in 2026 alone. In 2017, that same company spent roughly US$30 million, a 1,600-fold increase in less than a decade. OpenAI is targeting roughly US$600 billion in compute spending – or that in areas such as processing power, data storage and cloud infrastructure – through to 2030.

It’s not just OpenAI. The big technology companies are collectively expected to invest around US$650 billion in AI infrastructure in 2026. That’s roughly one-third of Australia’s annual GDP – or two-and-a-half times New Zealand’s – being committed to one technology bet in a single year.

All these expenses must be covered before the companies earn consistent profits. This is why they keep raising money – and why the question of where that money will eventually come from is enormously important.

In 2025, total investment in AI companies reached US$217 billion. Then, in just the first three months of 2026, private AI companies raised a further US$226 billion, surpassing the entire 2025 total in a single quarter.

Much of this was concentrated in three transactions: US$122 billion for OpenAI, US$30 billion for Anthropic and US$7.5 billion for xAI.

Together, these three deals alone accounted for 71% of all AI funding that quarter. Mega-rounds above US$100 million now make up 94% of all AI investment by value.

The funding has mostly come from large institutions: venture capital firms, sovereign wealth funds, and technology giants that can afford to take the risk. The gains and the losses stay within a small, specialist group.

When the AI sector goes public

Once a company is publicly listed, anyone can buy its shares.

More importantly, large index-tracking funds – widely used in Australia’s Super system and New Zealand’s KiwiSaver funds – automatically gain exposure to companies once they become large enough to enter the indices those funds follow.

In other words, they don’t get to decide whether it looks like a good investment; the index simply decides for them.

That matters for ethical investors, too. The thought of AI raises concerns about privacy, labour, misinformation and security. But unlike to***co or gambling, it may prove difficult to exclude because it is being increasingly woven into the world’s largest listed companies.

There is, therefore, a case for asking whether these companies should trigger an opt-out mechanism for fund managers and regulators before the listings arrive.

Neither OpenAI, Anthropic, nor xAI has formally announced a stock market listing, and timelines remain uncertain. When that shift comes, the risk also shifts. The investors who funded the early stages of this race knew what they were getting into.

The people who will end up holding the shares through their pension funds or index trackers may not.

Index providers are rewriting the rules

Here is where the picture becomes more complex. Major index providers are changing their rules so newly listed mega-cap AI companies can enter key benchmarks much faster.

Nasdaq has already adopted a fast-track rule that allows a newly listed mega-cap company to join the Nasdaq-100 after just 15 trading days. S&P Dow Jones Indices is consulting on similar changes that would reduce the waiting period and waive profitability requirements for mega-caps.

These changes are reshaping the index system to funnel passive money into AI giants almost as soon as they list – and before most investors have had time to decide whether they belong in their portfolios at all.

So, what can ordinary investors do?

As they likely won’t be making the call themselves on whether to invest in an AI stock market float, they can put questions to the fund managers doing so on their behalf.

Those might be questions about whether the company is becoming more efficient, what their customer retention looks like, or how their leadership holds up under pressure.

OpenAI’s 2023 board crisis showed how unusual governance structures can create sudden instability.

There is no doubt the AI revolution is real and is changing economies in the same way it is changing our everyday lives.

But whether the AI boom will create lasting value for ordinary investors – or mainly provide an exit for early-stage insiders – is a question fund managers and regulators cannot afford to leave unanswered.

Before the listings arrive, they need to decide: should ordinary investors be automatically swept into the AI gamble, or should they have a choice?

Sara Ali does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

When AI giants go public, will ordinary investors know if they are along for the ride? Large index-tracking funds could soon gain automatic exposure to AI giants as companies such as OpenAI edge toward public markets.

Marty Melville/AFP via Getty Images

New Zealand has a voter turnout problem. Although enrolment is compulsory for all eligible voters, casting a ballot is optional. And the number of people voting has been falling in both general and local body elections since the 1980s.

Enrolment is high, however, hitting 94.7% at the 2023 general election, the highest rate since 2008. But only 77.5% of those enrolled cast a ballot, considerably below 20th century averages, which hovered around 85%.

Participation among younger voters was even lower: 74.17% of 18–24 year olds, and 69.1% of those aged 25–29. Turnout in Māori electorates was 68%.

Older voters, by contrast, are far more likely to participate. In 2023, 84.94% over 60s turned out – meaning they had a much greater influence on who formed the government.

When almost a quarter of enrolled voters – 22.49%, or 829,326 people – don’t vote, there is a real risk that governments don’t accurately reflect society.

Low voter turnout can skew other outcomes, too. For example, the Brexit referendum that led to Britain leaving the European Union was based on just a 72% turnout rate.

But while research suggests compulsory voting can be a “cost-efficient institutional remedy” to low turnout, we would need to carefully weigh up the costs and benefits before implementing it in New Zealand.

Voting as a community event

It’s compulsory to vote in around 13% of the world’s democracies, including ten of the 30 OECD countries.

Australia celebrated a century of compulsory voting in 2024, with its Commonwealth Electoral Act stating, “It shall be the duty of every elector to vote at each election.” Compulsory enrolment was introduced in 1911 but turnout remained voluntary.

Australian voter participation sat at 71% in 1919 and fell below 60% in 1922, but it jumped to 91% in 1925 after the introduction of compulsory voting. It has remained around 90% ever since.

At the 2025 federal election, Australia’s turnout was 90.7%, with enrolment (also compulsory) very high at around 98%.

New Zealand, by contrast, has not exceeded 90% turnout in a general election since 1984.

Compulsion helps, but enforcement amplifies the effect. Research suggests compulsory voting without penalties increases turnout by around 7.5 to 10 percentage points. Where sanctions are enforced, turnout rises by a further 14 to 18 points.

In Australia, failure to vote without a valid reason is an offence punishable by a modest fine (around A$20). The penalty is light, but apparently sufficient.

More importantly, voting has become culturally normalised in Australia as a form of civic participation and community connection, epitomised by the famous “democracy sausages” sold at polling booths. As one submitter to an Australian electoral reform inquiry observed in 2013:

Election day is a community event, a common civic enterprise – and one in which most people are happy to participate.

What are the objections?

New Zealand already accepts compulsory jury service and compulsory voter enrolment. Is there any principled reason not to follow Australia’s lead and extend compulsion to voting itself?

Four objections are usually raised, starting with the assertion that you can’t force people to vote.

In practice, compulsory voting doesn’t compel anyone to make a political choice. The obligation is to attend, be marked off the roll, and receive a ballot paper. Casting a blank or informal vote remains an option.

Second, compulsory voting requires the act of voting to be as easy and accessible as possible, which has raised questions about potential fraud. But digital cross-checking and identity verification processes have largely eliminated the risk in Australia.

Research over decades has shown countries with some form of compulsory system see increased voter turnout without corresponding increases in electoral fraud.

Read more:
Compulsory voting in Australia is 100 years old. We should celebrate how special it makes our democracy

Third, there is the question of people who genuinely can’t vote. Compulsory systems rely on maximising access and the acceptance of reasonable excuses for opting out.

Advance voting, mobile polling, phone voting for blind voters, and exemptions for illness or incapacity are standard (and already familiar to New Zealanders through jury service).

Fourth, and crucially in an era of declining trust in institutions generally, would
compulsion be a spark that ignited another culture war about government overreach?

The COVID pandemic demonstrated how social cohesion can suffer when governments mandate certain behaviours. And research has shown this can affect lower socioeconomic groups more.

But the same research also showed people with a “high sense of belonging” were equally likely to participate in civic and social activities, and those who feel isolated are less likely to vote in voluntary systems.

While Australia has seen a slight rise in spoiled and “donkey votes”, evidence from Latin America suggests compulsory voting disproportionately increases turnout among the less socioeconomically advantaged.

Better democracy

To be widely accepted, compulsory voting would need to be presented as a shared, nonpartisan civic duty. As the American academic Shane P Singh has written (in the context of very low turnouts in the United States):

The promise of compulsory voting is that it would incentivize people to turn out and, in turn, bolster the quality of democracy.

Closer to home, it seems people share that optimism. Across five different surveys, more than 70% of Australians have consistently said compulsory voting has been good for their country.

New Zealanders might be generally reluctant to acknowledge their nearest neighbour has got something right, but perhaps it’s time to make an exception.

Myra Williamson does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Compulsory voting has boosted turnout in Australia – is it time NZ tried it? Voter turnout in New Zealand elections has been sliding for decades. Research suggests compulsory voting could reverse the trend – so what are the objections?