Issue 15 argued most NZ portfolios are over-built - two funds, chosen for a reason you can state out loud, beats six funds quietly holding the same market on repeat. That's still true for the core of a portfolio. It was never meant to rule out the satellite - the smaller, deliberate position you hold because you understand something about a trend or an industry that the broad market hasn't fully priced in yet.
Missed last week - on leave, and finishing an assignment for my Level 5 investment paper. Back to it.
There's plenty written already about one kind of thematic investing: ESG, sustainability, values-screened funds. Filter out what you don't want to own, filter in what you do, accept the provider's definition of "sustainable," done. Less gets written about the other kind - putting money behind a trend you actually understand or believe will compound over five to ten years, independent of any values screen. The AI infrastructure buildout is the example everyone reaches for right now, but the same logic applies to renewable energy, US reindustrialisation, a wider US energy rebuild, an industry you work in, or any shift you can see coming before the index has fully caught up to it.
Buying Nvidia because a headline said AI is the future is speculation. Buying a sized position in a semiconductor or automation fund because you work in tech, understand the capital cycle behind data centre buildouts, and have a specific view on which part of that chain is underpriced - that's conviction investing. The topic can be identical. What separates them is whether the position is bounded, sized in advance, and backed by something you could actually explain in two sentences without reaching for a buzzword.
A useful test before buying anything on a theme: could you explain why you own it to someone who works in that industry, and would they nod along or wince. If the honest answer is "I read a headline," that's not a thesis, and no amount of conviction language changes what it actually is.
The standard framework for holding both a diversified core and a handful of thematic positions is core-satellite: the core, built the way Issue 15 described, sits at somewhere around 80% of the portfolio, with the remaining satellite allocation split across a small number of individual conviction positions, often sized at somewhere in the 3-8% range each so no single bad call does structural damage to the whole plan. The exact split is a personal risk-tolerance decision, not a formula - the point of the framework isn't optimisation, it's making sure a satellite bet stays a bet and never quietly becomes the portfolio.
The failure mode is a satellite position that grows past its intended size because it's gone up, and nobody ever rebalances it back down. A theme that's performed well is exactly the one that needs checking against its original sizing, not left to compound its own risk unmonitored.
One email a week. Straight writing on investing from New Zealand.
Smart runs a "Thematic shares" category with five funds, all structured the same PIE way as the broad index funds covered in Issue 15 - meaning FIF never applies regardless of position size:
None of those is a pure AI fund. BOT is the closest available NZX proxy - automation and robotics companies overlap heavily with the AI infrastructure buildout, chips and data centre hardware among them, without being a direct bet on any single AI company's earnings.
If what you actually want is direct exposure to a handful of AI-specific companies, or a US-listed AI-focused ETF rather than the automation-adjacent NZX option, that means going direct through Hatch, Stake or Interactive Brokers. That's a different tax bucket to everything above it: FIF applies once your total overseas cost basis crosses $50,000, the same threshold covered in Issue 3 and Issue 15, regardless of how strong the conviction is. A $30,000 direct position in a handful of AI stocks and a $30,000 holding in BOT sit under completely different tax treatments for an almost identical underlying bet - know which bucket you're in before you size the position.
A short list to run through before sizing anything, not because it's a formula but because skipping it is how a satellite position turns into an accident:
Would you be equally comfortable holding this if it fell 50% next month, or does the thesis only survive while the price is going up. Is it sized so a total loss doesn't change your retirement date. Have you actually checked whether it sits inside or outside FIF, and sized around the real after-tax number rather than the headline return. And has it grown past its original allocation without you doing anything - because if it has, that's not a win sitting there, it's an unmanaged risk wearing one.
I'm running exactly this test on a position I'm building right now. More on that next issue.
One email a week. Straight writing on investing from New Zealand.
Subscribe for free