The Goodhart Global Real Return Fund (“The Fund”) generated a positive return over the quarter. The beta of daily returns compared with the broad global equity market was notably low at 0.12 over both the quarter and since inception[1]. This means the performance of the Fund has not been driven by broad market direction. That can be good or bad, of course, depending on which direction broad markets are heading.

Low beta DNA

Building a portfolio that has a very different factor risk profile to its comparator is a key requirement for an equity strategy that is aiming for low beta. The Fund’s geographic and industry biases certainly meet this criteria and so, on that simplistic level, its low beta to date is understandable. However, betas are not static. The relationship between stocks, industries, countries and other risk/return factors in a portfolio can change significantly over time. Betas may be stable for long periods and then change suddenly, sometimes for transient sentiment-driven reasons and at other times driven by longer term fundamental developments. Understanding this and predicting how betas may change in different environments is central to effective risk management, and building a portfolio that can generate returns independent of market direction over the longer term. We believe there are three key ingredients:

1. Breadth. The broader the genuine investment universe from which ideas are sourced the more potential there is to identify opportunities that will perform differently to market aggregates. It is particularly important to not restrict research to one style of company like “value” or “growth” because this artificially restricts the universe and creates long term factor bias.

   
   

2. Concentration. Portfolios must have significantly different construction to their comparators at the stock, country, sector, and factor/style level.

   
   

3. Change. The beta and correlation of a static portfolio, even if it has concentrated stock and factor exposures, will tend towards market averages over the very long term. In order to generate low correlation and beta over protracted periods, a portfolio must therefore change composition periodically to “stay ahead” of the market’s gravitational pull. That does not mean constant turnover in portfolios is necessary at all – but significant periodic change in key exposures is.

Unsurprisingly, this is what we aspire to in managing the Fund. It is the reason we seek to source ideas from across all “stages of the company lifecycle” from growth to quality, classic value and others. But we keep the portfolio concentrated at both the stock and thematic level and with very little “common money” to market comparators. We know we must shift the focus of the strategy materially over time as opportunities evolve if the Fund is to sustain low beta long term, and strive to get the balance right. We try to invest ahead of positive change that we believe will impact betas, and strive to create positive return asymmetry.

The AI iceberg

The key reason the Fund has had low beta since inception has been not owning hyperscalers and the broader AI infrastructure food chain, most notably semiconductor stocks. Unfortunately, this has also proven to be a substantial headwind for performance so far in 2026. The proportion of stocks globally that has had low or negative beta to the broad global equity market has been rising over the last year. This reflects the dominance of AI as a theme and the narrowness of investors’ focus on it.

The graphic below shows the percentage of constituents of the iShares MSCI World index Fund that had low (less than 0.5) or negative beta over rolling 90-day periods ending 30 September 2026.

Ordinarily our primary focus is on identifying stocks with substantial upside over the longer term and building a portfolio that successfully diversifies fundamental risks without reference to market indices or peers. Low beta is an expected long-term outcome rather than an explicit short-term target, and when warranted we are happy to accept higher beta than the Fund has had over the last year. However, currently we are specifically targeting low correlation to the AI theme because we believe it is one of three significant and deeply interconnected risks that are facing equity markets today (alongside interest rates and geopolitical risk).

The graphic below is illustrative of just how much the AI theme has been propping up markets. A large proportion of stocks currently trade at significant discounts to their all-time highs, and many are deeply into “correction” territory.

Stocks

No new stocks were added to the Fund over the quarter.

We exited two positions:

1. GQG Partners (“GQG”) is an investment management business and was sold after it made a significant change in its portfolios (it is now overweight hyperscalers). Our thesis on the stock rested on the idea that in a market downturn it was positioned more defensively than other large asset management businesses. We reasoned that if its positioning was correct, it would gain assets quickly, particularly in the US where it is relatively under-represented in the mutual fund market. Given how obvious and high profile its positioning was, the real attraction to us was the potential (counterintuitively) for the stock to go up when markets went down. We do not know why GQG changed positioning given how strident its client letters were about AI valuations, but it doesn’t really matter. We believe it is now less likely the stock will perform well in adverse markets and whilst we continue to respect management and the stock continues to be extremely cheap it no longer fits with its thematic role in the portfolio i.e. to have material “countercyclical potential”.

   
   

2. Koninklijke Philips (“Philips”) was sold in favour of increasing exposure to GTT, in which the Fund initiated a position the preceding quarter. Our conviction in Philips was not high enough when put in context with the “range of outcomes” for the portfolio in different adverse scenarios. In particular, we believed sentiment around China had the potential to negatively impact the share price during a downturn.

   
   

Themes

Exiting GQG essentially made the “Countercyclical Potential” theme redundant. Within “Geopolitical Winners” this year there has been a marked divergence between the performance of the energy and defense stocks. The addition of GTT softens the direct correlation of the Fund’s energy stocks to the oil price. But we think energy warrants being a separate theme in the current environment. At the margin it encourages us to think more carefully about situations (like March this year) in which energy and defense stocks had low correlation with each other.

The table below shows the performance of the Fund’s themes over the quarter in context with its other assets and ranked from best to worst performance (absolute return not contribution to return).

Source: Caceis, Capital IQ

Strong performance from Energy, Japanese Reform and Cyclical Winners was notable. Returns within Cyclical Winners were pleasing given weakness earlier in the year (Cicor +6.8%, Jet2 +12.2%, Bakkafrost +12.2%). GQG (Countercyclical Potential) was painful as the shift in its portfolio positioning spooked the market (and us). Inevitable Growth hasn’t really started working yet, with Raysearch -17.6% pulling down performance of the wider theme. Returns within the Resilient theme were mixed, with performance largely driven by shorter term newsflow on specific stocks rather than by broader demand for the theme itself. This is perhaps a little bit of a quid pro quo for the fact that the Fund’s Resilient holdings have (we think) significant idiosyncratic upside rather than just being dull companies we hope will be defensive in a falling market. We think this is important when aiming to generate positive returns over a broader range of market environments.

Are we too bearish?

We would like to deploy more capital but are trying to stick to a very low absolute risk profile. It has been tempting to step into more “growth” risk given the number of really interesting stock ideas we have across the team, particularly in software but also across a much broader universe. We think this is “good friction” because we are not comfortable. That said, our confidence that a defensive posture is warranted is bolstered when we contemplate what we regard as the three key drivers of risk today:

1. Interest Rates. Inflationary pressures in Europe and the US seem to be entrenching as expectations adjust and the longer-term drivers become more obvious. A resolution of war in Iran and subsequent reduction in oil prices would not be enough. Some bigger forces are at work, from fiscal deficits to reshoring, sustainability to demographics. This is partly reflected in bifurcating political sentiment in many countries.

   
   

2. Geopolitics. Whilst individual conflicts may de-escalate or even end, heightened geopolitical tension seems here to stay. This both strains budgets directly and increases “event risk” for which most countries need to be better prepared. Reprioritising security of supply is not a positive for economic growth at least short term.

   
   

3. AI. The hyperscaler “land grab” makes sense on lots of levels but it seems very unlikely the vast amounts being spent on datacentres and model development will generate attractive returns on capital within reasonable timeframes (if ever). Talk of AI being an existential threat to humanity that warrants a slowdown in development likely has more to do with hyperscalers recognising that the math on sales, costs and profits doesn’t work than social responsibility.

   
   

Individually each of these sources of risk are significant and need to be managed carefully. But they are also deeply intertwined, which increases the potential for something reflexive and more dramatic. Timing these big cycles is incredibly hard. So, we are doing our best to avoid the potential blast zone whilst investing in compelling opportunities elsewhere. Of course, AI pervades everything nowadays and we are certainly not negative on the potential benefits it brings to many companies. But we are trying hard to not build a portfolio that accidentally has second derivative exposure to the AI infrastructure theme if the music stops. Likewise, if inflation really starts to accelerate, interest rates balloon on sovereign credit concerns, Putin or Iran escalate or Trump does something unfathomable.

When low beta really matters…

The table below looks at the big global equity market drawdowns since 1987. It highlights the proportion of stocks that managed to generate a positive absolute return in these bear markets and those that either had a positive return or didn’t lose much (less than 10%). Obviously in practice navigating a bear market successfully is about much more than just having a defensive portfolio when it starts. This analysis ignores the many opportunities to redeploy capital within bear markets and doesn’t look at the recovery periods after bear markets. But that is also why it is an interesting reference.

Source: UBS HOLT, Goodhart. Methodology: For each of nine bear markets since 1987, we took every developed-market stock in the HOLT database (US and Canada, Europe, Japan and developed Asia-Pacific) that was listed at the start of the sell-off. We compounded each stock's monthly total return from the start to the end of the sell-off, then counted how many finished the period flat or up (≥ 0%) or no worse than -10%. The data is developed markets only (NA, Europe and Asia files), which is 4,773 to 15,751 stocks per period depending on specific bear market. Emerging and frontier markets are excluded. Universe: all stocks with data at the start date of each period. Stocks that delisted before the end date are kept in using their return up to their last month of data. Return: compounded monthly return from the start date (excluded) to the end date (included). Returns are in local currency, not USD. Periods: 1987 Crash (Aug–Nov 1987), Early 1990s (Jul–Oct 1990), Asian/EM Crisis 98 (Jul–Sep 1998), Dot-com 00–02 (Aug 2000–Sep 2002), GFC 07–09 (Oct 2007–Feb 2009), Eurozone 11 (May–Sep 2011), 2018 (Sep–Dec 2018), COVID 20 (Jan–Mar 2020) and 2022 (Dec 2021–Oct 2022). Average: the simple average of the nine periods, not weighted by stock count.

On average around a third of stocks have either gone up or not gone down more than 10% in the historic bear markets summarised. But the figures vary significantly between different time periods. 2022 was relatively easy in that a larger than average proportion of the market appreciated in value whereas the financial crisis was much harder, with very few stocks appreciating. The pertinent question, perhaps, is what sort of bear market do we expect next time? Of course we cannot know in advance, but three things make us cautiously optimistic:

1. First, we are currently able to identify plenty of stocks that we think are attractive on a longer-term basis (5 years) albeit they tend to be smaller companies in niche areas. We do not think we are in an “everything bubble”.

   
   

2. Second, we have high conviction that certain specific parts of global equity markets are hugely over-valued (hyperscalers and AI infrastructure).

   
   

3. Third, the fact so many stocks have had negative beta to the AI theme may mean they will be less negatively impacted in a downturn and may already partly discount more challenging macro and geopolitical scenarios.

We seek to balance these three things in the portfolio. The Fund has plenty of exposure to stocks we believe can generate substantial upside over the longer term. We do not believe it is necessary to be more fully invested in equities in order to generate attractive absolute returns over the longer term whilst at the same time we believe the current portfolio is well positioned to cope with a range of different scenarios in which markets decline.

As ever, if you have any questions on this letter or just want to catch up please reach out.

[1] Source: Goodhart Partners LLP, as at 30 September 2026. These statistics are based on the Fund's limited operating history and may change materially.