Asset allocation choices are the most important decisions institutional investors make, shaping long-term returns, risk exposures, and ultimately the success of their portfolios. Whether the question is how much risk to take, whether to hold sovereign bonds or credit, or how to position between the US and Europe, finding answers is never simple. Answers to these questions vary dramatically between investors with different liabilities, governance structures, and objectives.
This complexity is why institutions such as pension funds, sovereign wealth funds, endowments, private banks, and wealth managers, devote so many resources to their allocation process. CIOs, asset allocation teams, and portfolio managers can spend months debating, modelling, and stress-testing decisions before portfolios are adjusted. The challenge is not just "what should we invest in?" but how to create a rigorous, repeatable decision-making process that produces robust portfolios.
Defining the Analytics Stack
Behind every well-run allocation process sits a constellation of data, models, and analytical tools. Together, they form what we call the Analytics Stack — the analytics suite required to tackle asset allocation in a highly-informed way.
Institutions build or procure a wide range of tools to help guide their thinking. The stack typically includes forward-looking return assumptions (Capital Market Assumptions), risk engines (a.k.a. Economic Scenario Generators) to simulate portfolios under different macro conditions and regimes, factor models to explain returns and manage exposures, and scenario tools to stress-test asset allocations. Each layer involves extensive development and maintenance time, and are not easy to build. Building powerful and state-of-the-art models requires deep expertise and real-world testing, with years of iteration and refinement.
There are also other underappreciated and time-consuming layers involved in building up analytics. A significant example is the procurement, onboarding, and maintenance of the raw data inputs that feed into analytics. Every asset allocation analyst knows first-hand how painful this data handling can be, and how things can go wrong with the data infrastructure at the worst possible moment. In addition, the data landscape is exploding in size, and it is more confusing than ever when trying to understand which datasets are relevant for asset allocation. At the other end of the workflow is the distribution of analytics and their insights to decision-makers — in the form of dashboards or model portfolios – that are equally underappreciated and resource-intensive.
Each layer of the stack is essential. By covering the whole stack, the benefits multiply and asset allocation processes improve exponentially. Miss one component, and you leave a blind spot. But building and maintaining a complete stack is neither cheap nor easy.
The Build-Versus-Buy Dilemma
Institutional investors typically face stark choices in deciding whether to build or procure each layer of the analytics stack. Both paths have challenges, and across the industry we typically observe a mix of in-house development and procurement from external providers.
Building in-house is only an option for the largest investors. These institutions can hire and maintain teams focused on asset allocation. They can often construct a bespoke, high-fidelity solution — but it is typically slow, expensive, and operationally fragile.
Onboarding datasets can take months, and building useful models needs decades of experience. Analysts spend large amounts of time just making data usable, fixing broken pipelines, and maintaining back-end infrastructure. Data and model development projects hoover up resources, and are necessary before even starting to answer difficult asset allocation questions. It can take years to build a competent team that adds value to asset allocation decisions. A single departure or change in priorities with an organisation can stall or even reverse this progress. Even the largest and most sophisticated institutions will still source analytics externally.
Buying from providers is no easy fix, either. Coverage across the analytics stack is fragmented: return assumptions from one vendor do not line up with factor models from another, and scenario tools often use outdated methodologies. Delivery technology lags behind modern expectations, with clunky onboarding, weak APIs, and poor user experience. Many providers have been using the same intellectual frameworks for decades, with little incentive to innovate so long as clients remain in their ecosystem.
Both approaches share the same first-order problem: it is extremely difficult to build a fully integrated, consistent, and cutting-edge analytics platform for asset allocation. Any gaps in the stack expose investors: expected return estimates are poor signals, asset classes cannot be compared when building portfolios, and portfolios are exposed to risks that should have been anticipated.
This is especially problematic as we see political, geopolitical, economic, and market landscapes all shift before our eyes. The investment environment is being overturned, and we need fresh and dynamic approaches to tackling asset allocation problems more than ever.
A Different Approach
At Allocation Strategy, our mission is to solve this problem by delivering a state-of-the-art and comprehensive analytics stack. We are not here to replace CIOs, analysts, or other investment professionals. Rather, we aim to empower them with the best possible tools, with robust data and tech underpinnings.
Our vision is to remove the long and resource-intensive journeys that so many investors embark upon in trying to solve difficult asset allocation problems. Our platform removes the years-long journey many teams face when building from scratch, while avoiding the compromises of legacy vendor solutions. This lets asset allocation functions spend less time reconciling data and models and more time focusing on the ultimate questions for the asset allocation process: what is the right allocation today, and how should it evolve if conditions change?
Better Portfolios, Better Outcomes
When institutional investors have a complete, coherent analytics stack, asset allocation decisions become highly informed and more resilient. Return expectations are consistent with risk models, scenario analysis is directly tied to other analytics, and decisions can be made with confidence. The result is better performing portfolios, and better alignment between portfolios and objectives. This holds true whether meeting pension obligations, financing universities through endowments, supporting sovereign fiscal plans, or securing retirement outcomes. Efficient and powerful asset allocation tools will be essential for navigating the major global shifts that lie ahead for investors.