GIC CEO warns diversified portfolios may share common AI theme
Singapore sovereign wealth fund GIC's CEO says diversification requires scrutiny of actual asset exposure, as many assets may involve AI. Canadian pension fund CEO agrees, noting S&P 500 outperformed diversified portfolios in recent years. Both spoke at the Hong Kong investment summit.
Singapore sovereign wealth fund GIC's chief executive Lim Chow Kiat said on 3 November at the International Financial Leaders Investment Summit in Hong Kong that diversification requires investors to examine the actual exposure of assets. He noted that about 70 per cent of stocks in the MSCI index are US companies, but their revenue comes from globally, meaning they are not purely US exposure. He added that investors should assess asset quality, including management teams and competitive advantages, when allocating capital.
Lim said that even if investors cannot completely avoid a highly concentrated theme, they should maintain good liquidity to respond flexibly to market changes. He observed that patient capital historically earned returns through time premium and mean reversion, but technological developments have brought structural changes that may prevent mean reversion. As a result, many long-term investors seek higher valuations when taking on such risks.
John Graham, president and chief executive of the Canada Pension Plan Investment Board, speaking at the same event, said diversification is necessary because no one can accurately predict how the world will evolve. However, he warned that when markets concentrate on a few themes or companies, it poses a challenge for institutions. He cited the S&P 500's clear outperformance of any diversified portfolio over the past several years.
Graham said the role of diversification is to provide resilience to a portfolio during economic downturns. The goal is not to maximise investment returns but to build a portfolio that can withstand any shock. He cautioned that asset classes previously considered diversifying may share common themes, giving the example of real estate potentially involving data centres, and infrastructure, energy and credit all potentially being linked to artificial intelligence.