AI investment helps offset rising interest rates
September brought the first Federal Reserve (the Fed) rate hike since 2023 and the highest 10-year Treasury yield since 2007. Amid policy tightening by the Fed, European Central Bank and Bank of Japan, global bond yields moved sharply higher. Elevated oil prices, above-target inflation, strong US growth, lax government fiscal policy and the inflationary consequences of the AI data center buildout all contributed to the rise in interest rates.
That said, global stocks remained relatively well supported despite volatility in the bond market and September’s historically weak seasonality. In our view, this stability was due to the strong macro tailwinds being created by the ongoing AI investment boom. With frontier model providers competing for AI supremacy and AI use cases rising by the day, ever-increasing amounts of infrastructure are being required to support this major technological shift.
Strong AI investment activity is creating ripple effects throughout economies and supporting growth around the world. It’s for this reason we don’t want to become too negative on risk assets despite headwinds like higher oil prices and rising interest rates. Earnings are being supported by strong investment showing little evidence of slowing, and it’s hard for us to get too bearish on stocks just because growth is so strong it’s pushing up interest rates. Yes, this might create some near-term stalling, but we are entering a very strong seasonal period for US stocks, and it tends to be even stronger during midterm election years. For now, we remain constructive on risk assets with a close eye on everything AI-related.
September model performance
Natixis tactical model performance was negative in absolute terms on both a gross and net basis during September. Models across all risk levels declined by similar amounts as both stocks and bonds struggled. Relative performance was also negative across the models, except for some tax-managed portfolios with less active equity exposure. Riskier models underperformed their benchmarks by more than conservative models.
In the moderate-risk models, asset allocation and style selection contributed positively to relative performance. An underweight to fixed income, an overweight to cash and a tactical allocation to Japan all added value. However, manager selection detracted as the Loomis Sayles Growth Fund (LSGRX) and Oakmark Fund (OANMX) underperformed when the AI trade regained leadership within the US stock market. The two strategies’ underweights to the AI trade and overweights to legacy software stocks weighed on the models’ performance.
Natixis model portfolio positioning
The team didn’t make any official model changes during September but submitted its first new model trade since May during the first few days of October. The trade added 1% to the overweight position in stocks, reflecting continued earnings strength and favorable seasonality through year-end. The higher equity allocation was funded by selling fixed income, including a portion of the models’ Bloomberg US Aggregate bond ETF exposure.
The trade also added a long-term US Treasury ETF, increasing the models’ interest-rate sensitivity even as their overall fixed income allocation declined. With the 10-year Treasury yield well above 5%, we want to start positioning for lower interest rates. Lower rates might be catalyzed by an unexpected slowdown in the economy, a retracement in oil prices if the US/Iran situation improves, or a decline in inflation for other reasons that could allow the Fed to stop raising interest rates. While we are not expecting a major near-term slowdown, we believe long-term fixed income offers an attractive risk-reward profile at current levels.
The team also established new positions in the Technology Select Sector SPDR Fund (XLK), and the Invesco PHLX Semiconductor ETF (SOXQ). These positions are intended to help neutralize the technology and semiconductor underweights created by the models’ active US equity strategies. The hedges will be actively managed as the underlying funds’ allocations and the team’s views on technology evolve.
At month-end, model positioning was characterized by:
- Overweight stocks versus bonds
- Overweight US stocks
- Tilted towards growth stocks versus value stocks
- Underweight developed market international stocks
- Slightly underweight emerging market stocks
- Slightly long duration
