Agglomeration deficit outside the main centres

New Zealand's economic geography is highly concentrated in Auckland and Wellington: most regions lack the labour market depth, supply chain density, and knowledge spillovers that enable higher-productivity economic activity.

Agglomeration mechanisms

Firms in larger labour markets benefit from: easier recruitment of specialised workers, proximity to customers and suppliers, knowledge spillovers between co-located firms, and shared infrastructure. These benefits are largely absent outside Auckland and Wellington.

Policy implication

Regional economic development cannot easily replicate agglomeration effects. Policies that strengthen the Auckland-region connection for smaller centres — transport, digital infrastructure, talent pipelines — may be more effective than trying to build alternative agglomeration from scratch.

Manifests in

northland, gisborne, west-coast, taranaki, nelson, tasman, marlborough, southland

Evidence

• claim.northland.economy.business_investment_claim1
• claim.west_coast.economy.economic_2_claim
• claim.gisborne.economy.economic_fragility_65
• claim.taranaki.economy.diversification_01
• claim.nelson.economy.economic_2_claim

AI productivity gains are likely to be captured asymmetrically

New Zealand's productivity gap with the OECD frontier is a long-tail problem: a small group of frontier firms operate near the international frontier while the median firm lags far behind, with within-industry dispersion larger than in most OECD comparators. AI tooling is most effective when complemented by data, capital, and managerial capacity — all of which are concentrated in the frontier tier. Without diffusion mechanisms, AI sharpens the gap rather than closing it.

Capture mechanics

Productivity gains from AI accrue first to firms that already have digitised data, capital to spend on tooling, and managers who can redesign workflows around new tools. Frontier firms in Auckland and Wellington (large banks, telcos, parts of professional services, a small group of tech firms) clear all three bars. The median NZ firm — small-to-medium, paper-based or lightly digitised, with constrained management capacity — clears none of them. AI tooling sharpens this gap before it closes it.

Distributional consequences

Where productivity gains land in the wage and capital share is determined by firm-level pricing power, contract structure, and labour-market institutions. Concentrated capture pushes value toward capital owners and a narrow professional tier; broad diffusion would distribute it across the labour share and the long tail of firms. New Zealand's mix — concentrated capital ownership, weak union density outside the public sector, no sectoral bargaining — favours the concentrated outcome unless counteracted.

Diffusion is the policy lever

The Productivity Commission's frontier-firms inquiry identified diffusion (not frontier performance) as the binding NZ constraint. The diffusion lever for AI is a combination of accessible tooling (sovereign or shared cloud and data infrastructure), management capability (formal training programmes targeting SMEs), and a competition-policy frame that prevents incumbent capture of foundation-model access.

Manifests in

auckland, wellington, canterbury, waikato, bay-of-plenty, otago, nz

Evidence

• claim.auckland.economy.frontier_firm_productivity_gap_2021
• claim.auckland.economy.ai_diffusion_lag_2021
• claim.auckland.economy.ai_capture_distributional_risk_2024
• claim.auckland.economy.nz_productivity_gap
• claim.auckland.economy.low_rd_investment

Primary sector concentration and productivity limits

Most New Zealand regions outside Auckland and Wellington have economies heavily concentrated in primary industries — pastoral farming, horticulture, forestry, aquaculture — with relatively low labour productivity and limited wage growth.

Sector structure

Primary sector concentration is not inherently a problem: New Zealand comparative advantage in pastoral and horticultural production is real. But sectors with low labour productivity set a ceiling on wages that constrains regional incomes and makes retention of skilled labour difficult.

Transition pathway

Value-added processing, technology services, and knowledge-economy expansion are often proposed as diversification pathways. Each requires investment in workforce capability, digital infrastructure, and market access that most primary-dominated regions cannot self-fund.

Manifests in

waikato, northland, gisborne, hawkes-bay, taranaki, marlborough, nelson, tasman, southland, west-coast

Evidence

• claim.waikato.economy.agri_gdp_share_28pct
• claim.northland.economy.business_investment_claim1
• claim.gisborne.economy.economic_fragility_65
• claim.southland.economy.agri_commodity_prevalence
• claim.manawatu_whanganui.economy.agri_food_1

Tourism sector concentration and COVID vulnerability

Regions highly dependent on international tourism experienced severe economic shock during COVID-19 and face structural vulnerability to future demand shocks from health events, currency movements, or climate-related disruption.

COVID shock

The COVID-19 border closure reduced international visitor arrivals by 99% and devastated internationally-dependent tourism economies in Queenstown, Rotorua, the West Coast, and smaller regions. The shock exposed the risks of a single-sector concentration strategy.

Structural vulnerability

Beyond COVID, international tourism demand is sensitive to exchange rates, air connectivity, and increasingly to the climate credentials of long-haul travel. Regions that have not diversified from tourism dependence remain exposed to these structural trends.

Manifests in

otago, west-coast, nelson, tasman, gisborne, northland, marlborough

Evidence

• claim.otago.economy.agri_primary_prevalence
• claim.west_coast.economy.economic_2_claim
• claim.nelson.economy.economic_2_claim
• claim.northland.economy.business_investment_claim1
• claim.gisborne.economy.economic_fragility_65

Ageing population, contracting workforce, rising demand

New Zealand's median age is projected to rise from 38 (2023) to around 43 by 2048, with the 65+ share rising from ~16% to 23-25%. The same demographic shift increases demand on the health, aged-care, and superannuation systems while reducing the working-age share that funds and staffs them. The mismatch is a structural macroeconomic pressure, not a cyclical labour market issue.

The double bind

Ageing is both a demand shock (more health, aged-care, and superannuation services per capita) and a supply shock (fewer working-age people relative to dependants, and a workforce losing senior staff faster than it can replace them). The two sides reinforce each other, particularly in the health system, where the workforce that delivers ageing-driven care is itself ageing out. The scale is set by demographic projections that have low forecast variance over a 25-year horizon.

Regional unevenness

The aged share of population is projected to be highest in the same regions that are losing working-age population — West Coast, Tasman, Marlborough, parts of Northland and Hawke's Bay — so the workforce-to-demand ratio deteriorates fastest where it is already worst. Auckland's relative youth (sustained by net migration) shifts the national mean but does not compensate for regional concentration.

Migration as a partial offset

Net migration to New Zealand is the lever that most directly moves the working-age share. Sustained net inflows of working-age migrants slow but do not reverse the ageing trend; the projected ageing happens under all reasonable migration variants. Migration policy is therefore a margin-shifter, not a structural solution, and creates parallel pressure on housing supply (see the housing affordability pattern) and infrastructure capacity.

Manifests in

nz, west-coast, tasman, marlborough, northland, hawkes-bay, gisborne, southland, waikato, auckland

Evidence

• claim.auckland.economy.median_age_trajectory_2048
• claim.auckland.economy.dependency_ratio_2048
• claim.auckland.economy.health_workforce_retirement_2024
• claim.auckland.economy.regional_aged_share_divergence_2048

AI and automation exposure in the New Zealand labour market

Around 10% of New Zealand jobs are at high risk of automation under the OECD's task-based assessment, with a further ~25% likely to undergo significant change. Generative AI extends the exposure upward into cognitive and clerical work that earlier waves of automation left untouched. The exposure is uneven across regions and demographics — concentrated in routine clerical, transport, manufacturing, and parts of the public sector — and the country has no settled adjustment policy for it.

Exposure profile

Routine clerical, transport, and manufacturing occupations carry the heaviest automation exposure under the standard OECD model. Generative AI shifts the front of the exposure curve toward higher-skill knowledge work — first-draft writing, summarisation, coding, basic legal and accounting tasks — which means white-collar cohorts in Auckland and Wellington are now exposed alongside blue-collar cohorts in provincial regions. The geography of risk is broader than 2018-era estimates suggest.

The adjustment problem

Exposure is not displacement. Whether exposed jobs are augmented, redesigned, or eliminated depends on capital cost, regulatory environment, and the adjustment institutions that re-employ displaced workers. New Zealand has limited active labour market policy spend by OECD standards, no national reskilling entitlement, and a fragmented vocational training system mid-reform. The institutional gap is the binding constraint, not the technology.

Regional incidence

Regions with concentrated routine-task employment (manufacturing, transport, agriculture) carry higher exposure on the OECD task model. Regions with concentrated public-sector and professional- service employment (Wellington, Auckland CBD) carry higher exposure on the generative-AI model. Both flows hit New Zealand; neither is offset by an existing policy framework.

Manifests in

auckland, wellington, waikato, canterbury, bay-of-plenty, manawatu-whanganui, hawkes-bay, otago, southland, northland

Evidence

• claim.auckland.economy.automation_exposure_share_2018
• claim.auckland.economy.ai_diffusion_lag_2021