The Commerce Commission’s retail grocery competition inquiries always had the wrong focus.  

If there are enormous profits to be made in groceries, why is nobody else trying to steal them away? The world has many supermarket chains. Kiwis can also build new ones. If there are $100 notes on the sidewalk, why is nobody picking them up? 

Zoning rules seemed to be the underlying problem. We argued that large-scale entry was de facto impossible.  

Our submissions and columns could point to some of the rules in a piecemeal fashion. But we could not put together a comprehensive map. There are literally hundreds of bits of zoning rules that need to be examined across the main centres. Our shop is tiny, and we have no GIS experts on staff.  

Anthropic’s Claude Fable 5 changed everything.  

Over the past couple of weeks, I’ve been working with Fable to do the work we could not do during the market studies process. Fable classified roughly 250 zones across nine plans, plus hundreds of precinct documents.  

We have comprehensive maps for Auckland, Hamilton, Wellington (and environs), Christchurch and Dunedin. The maps show where large-footprint supermarkets are permitted, where getting permission will be more difficult, and the places where cities have very literally banned competition.
  
In cities with mapped water infrastructure constraints, we present those as overlays.  

I went through several rounds with Fable, catching errors and steering improvements. We put up a test version of the site for urban-economist and urban-planning friends to play with and provide advice. The high-test version of GPT 5.6 provided a hard adversarial review, looking for errors and making suggestions. After a few iterations, we asked the high-test version of Perplexity Compute to provide additional peer review.  

The site is now live, as a running experiment. Help us improve it. If you find errors or opportunities for improvement, please let us know via the site’s feedback form. Fable will sweep feedback weekly, provide me with options for updates, and we’ll make the site better. 

Currently, the maps show where a 4000-square-metre supermarket could be built. I would like users to be able to select different-sized stores and have the maps update. It is possible when my Fable token quota resets.  

Watch for more of these experimental sites to come.  
 
We live in an age of wonders and our regulators are stuck in the dark ages.  

Explore the supermarket zoning maps and share your feedback here.

The Opportunity Party wants to be known as the party of ideas. Now polling near five percent, I decided to test one.

The party’s website offers a calculator. You enter your income and land value, and it reveals how much better off you will be under its radical tax proposal. Their so-called ‘Tax Reset’ promises a cash payment of $19,400 a year to every adult, the ‘Citizen’s Income’, funded by a 1.75 percent annual tax on land.

I tried to invent someone who would pay more net tax under the scheme. I failed.

Tenants would be way better off. They would receive the free cash but pay no land tax. That is one way of cooling the housing market.

Retirees could defer the tax until they die.

Farmers would pay just 0.5 percent, with conservation exemptions and weather deferrals.

And homeowners? Most would be better off, TOP explains, once their Citizen’s Income is factored in.

But I was determined to find a net loser. I invented a couple TOP could not possibly love, each earning $500,000, with their house on a million dollars of urban land.

Still, the calculator awarded them a net tax cut of $5,455 a year.

Why? It had charged them more income tax (because TOP wants to increase it for top earners), yet their land tax of $17,500 was no match for their $38,800 of free cash.

The tax falls only on land, not on buildings. By TOP’s own rule of thumb, land makes up 70 percent of an average freestanding home’s value. That means few couples’ land tax would outrun the free cash until their house passes around $3 million. And once introduced, land prices would fall, so the new land tax raises even less revenue.

Yet, the scheme must somehow raise $24 billion a year to pay for the ‘Citizen’s Income’. The party estimates one in ten New Zealanders will pay more.

The scheme seems to punish families with multiple earners or high land value, while rewarding those who shift their resources elsewhere. But given how seldom the calculator produces a tax increase, even that seems a stretch.

Milton Friedman warned there is no such thing as a free lunch. Indeed, the oldest lesson in economics is that somebody, somewhere, always pays. Still TOP, the party of ideas, has built a shiny policy and calculator that cheerfully disagree with him.

Perhaps they know something that Friedman missed. Or maybe it is just a policy that is not properly costed?