Gold continues to set new highs and remains a strong focus for institutional allocators. Yet, despite its prominence, there is a lack of sophisticated tools and modelling frameworks for understanding what drives gold prices, and how gold should be incorporated in a well-constructed portfolio.
We previously outlined a framework for estimating expected returns on gold—a piece which has since become one of our most-read. Given the continued interest, we decided to write more on how to analyse gold and its role in a portfolio.
In our view, gold is under-analysed compared to more mainstream asset classes. The analytics, forecasting, and risk decomposition tools available for other asset classes are much more developed. Our flexible modelling approach aims to fill this gap—by allowing us to decompose gold’s drivers (real rates, inflation expectations, term premia, secular vs transitory shocks etc.) and assess expected returns over both short and long horizons.
An analogous discussion is emerging around crypto/digital assets: an asset class that shares some characteristics with gold and also has fewer established quantitative tools. In the coming weeks, we will publish some of our perspectives on this asset class.
Understanding Gold’s Drivers Over the Recent Past
As with any other asset class, institutional investors need a robust analytical framework for gold: a view on expected returns, risk, and how gold co-moves with the rest of the portfolio.
When we wrote about gold in May, we outlined how we estimate expected returns on gold. At that time, the 1-year expected return was at its highest level in several years — reflecting a combination of elevated real rates and volatile inflation expectations. While a full year has not yet passed, gold’s subsequent strong performance has broadly validated those signals – though it is of course only one data point. The chart below shows how our 1-year expected return lines up with forward returns.
Our framework distinguishes between transitory and persistent components of real interest rates and inflation expectations. This decomposition allows us to attribute gold’s realised performance over the past six months to its fundamental macro drivers. Our analytics show that a significant fraction of the return can be traced to falling real rates (easier monetary policy) and term premium. Shifting inflation expectations contributed very little over that period.
In our previous piece, we showed that even as prices have continued to rise through late-2024 and early-2025, the 1-year expected return on gold remained elevated. However, with prices now at new highs, we are starting to observe a gradual decline in expected return. This is consistent with our model’s logic: a decline real rates lowers the opportunity cost of holding gold as an asset without cash flows, which is reflected in the expected return.
Downside Risk After Recent Rally
Our analytical toolkit provides additional insight into gold price dynamics — including option-implied probabilities of future moves in gold futures. During 2025, there have been two pronounced rallies: one in April and another in the past few weeks, each seeing gold rise by around 10% within a two-week period.
However, the market context following these two episodes looks very different. In April, option markets reflected a meaningful probability of a reversal over the next 3 months. From 8 April to 22 April, the probability of a 5% or greater decline increased to around 30%. By contrast, as of 10 September, the equivalent value was around 17%, barely changed versus 2 weeks ago. In other words, today’s rally is accompanied by a much lower expectation of near-term pullback than was the case in April.
Option-implied Return Distribution for Gold Futures (3-month horizon)
Adding Gold to Portfolios
A central question for investors is not just whether to hold gold, but how much to allocate within a diversified portfolio. As always, the answer depends on the specific objectives and constraints of each investor. For institutions with liabilities indexed to inflation, gold’s inflation exposure — which we have previously highlighted — is a natural point of interest. At the same time, gold carries other macro exposures, such as sensitivity to real interest rates, that overlap with assets already commonly held.
While the relative expected return on gold (appropriately measured over the relevant horizon) provides a first-order indication of its potential contribution to portfolio performance, its impact on risk is more nuanced. Capturing these interactions correctly requires a robust portfolio simulation framework.
Our in-house simulation engine is designed for this purpose. It models portfolio properties — expected returns, volatility, tail risks, and other distributional measures — over any investment horizon, with and without gold included. The framework integrates horizon-specific expected returns, as well as gold’s macro risk exposures and their correlations with other asset classes. By aligning simulated portfolio outcomes with investor liabilities, the engine enables a disciplined approach to assessing the implications of adding gold to an existing asset mix.
Interested in finding out more? Use our contact form to continue the discussion.