20-year Returns Outlook
Our Temasek Geometric Expected Return Model, or T-GEM, simulates the range of possible returns for our portfolio over the next 20 years. These simulations do not predict actual outcomes.
Economic Scenario-Based Approach
Equity returns are volatile over time, and influenced by macroeconomic or geopolitical events and shocks. Even on a 20-year rolling basis, returns varied significantly across time. Likewise, on a longer 30-year or 50-year rolling basis, returns still vary significantly. Hence, T-GEM uses a scenario-based approach to simulate our 20-year long-term expected returns. This approach takes into account our views of long-term macroeconomic fundamentals over the next 20 years, adjusted for current market valuations.
US Annual Equity Returns
(for year ended 31 March)
US Equity Rolling 20, 30, and 50-year Returns over Time (Compounded Annualised)
(for year ended 31 March)
1 Rolling 20-year returns include events of the preceding 20 years.
Projection Based on Economic Fundamentals
In projecting economic fundamentals, we make more granular year-to-year assumptions for the first five years, and transition to more general assumptions based on longer-term fundamentals as we go beyond 10 years. This “pathing” approach incorporates assumptions about changes in economic conditions over time. We do not assume an equilibrium return that remains unchanged over the 20-year period, nor do we assume a reversion to the historical mean.
Economic Scenario Pathing (Illustrative)
20-year Expected Returns for Various Temasek Scenarios
We simulate our 20-year expected returns under different scenarios.
Potential Scenarios for 2024 and Beyond
Potential Scenarios |
Description |
Central (Our most likely scenario) |
Our baseline expectations of growth and inflation, that also include longer-term challenges such as climate change as well as high and rising geopolitical tensions.
|
Differing Climate Change Pathways (Less likely scenarios) |
Our “High Ambition” scenario assumes greater mitigation efforts to slow the rise in temperatures. Our “Low Ambition” scenario assumes no further mitigation efforts, which results in a disastrously steep rise in temperatures by 2100.
|
Major Escalation of Global Strategic Rivalry (Less likely scenario) |
A significant intensification of tensions between the global superpowers accelerates bifurcation across various economic blocs. This would result in a global stagflationary environment amidst structurally lower levels of productivity and economic growth (from reduced trade and investment), with elevated prices as global supply chains remain constrained.
|
China Downside (Less likely scenario) |
A downside scenario where China becomes less growth-oriented relative to our baseline. In this scenario, foreign participation is more significantly reduced, while the allocation of resources is less efficient, and the risk of ineffective policy is heightened. This will lead to lower productivity growth, which hurts sales and earnings potential, and compounds the negative effects of the challenging geopolitical backdrop. |
20-year Expected Returns for Different Portfolio Mix
The charts below illustrate the simulated returns for a Global Bond Portfolio, a Global Equity Portfolio, and the Temasek Portfolio. The simulations are based on our Central Scenario.
The Global Bond Portfolio has the lowest upside potential, as compared to the Global Equity and Temasek Portfolios. It also has the least volatility, as shown by its narrower year-to-year annual returns distribution curve.
The Temasek Portfolio has the highest upside potential (see blue shaded) at the end of the 20-year period, but also the highest volatility.
Likelihood of Geometric Returns (Compounded Annualised) at the End of 20-year Period, by Portfolio Mix
(as at 31 March 2024)
1 The Global Equity Portfolio uses the MSCI ACWI as a proxy.
2 The Global Bond Portfolio uses government bonds within the Bloomberg Global Aggregate Index as a proxy.
Likelihood of Year-to-year Annual Returns during 20-year Period, by Portfolio Mix
.jpg "boys_playing_chess")
