[Edaily Marketin KIM SUNG-SOO Reporter] In the alternative investment market, the importance of adopting a “benchmark portfolio”—which views the entire portfolio from the perspective of “risk and return”—is growing.

This is because funds are flooding into the single theme of artificial intelligence (AI)—regardless of whether they are public or private offerings—raising red flags for institutional investors’ traditional “asset class diversification” strategies. Even if assets are divided among stocks, bonds, and alternative investments, it is difficult to expect a practical diversification effect if the underlying assets of each class are all oriented toward the same direction—AI.

However, there is a limitation to the benchmark portfolio: if it relies too heavily on the judgment of portfolio managers, it can actually exacerbate concentration in specific themes. Ultimately, analysts suggest that rather than choosing between strategic asset allocation (SAA) and the benchmark portfolio, institutions should appropriately combine the strengths of both strategies.

“Diversified, but All Pointing in the Same Direction”… AI Shakes Up Diversified Investing

According to the financial investment industry on the 24th, there is a growing awareness among institutional investors (LPs) that they must move beyond simply adjusting investment weights by asset class and instead make investment decisions based on the actual risk and expected return of the entire portfolio.

The “AI investment boom” is behind the view among institutional investors that risk management cannot be adequately achieved through asset class diversification alone. This is because funds are flooding into the same “AI” theme across not only the public market but also alternative investment markets such as private equity (PE), private credit, real estate, and infrastructure.

In the stock market, the weight of large-cap AI-related stocks is increasing, while in the alternative investment market, investments in AI-related companies, data centers, and AI infrastructure projects are emerging as major investment targets.

The problem is that while the “vessels” of these investments may differ, the actual risk factors could become similar. Even if funds are allocated across stocks, bonds, real estate, and infrastructure, if the underlying assets of these portfolios all share the same expectations regarding AI growth, the likelihood that a shock in one area will spread to other asset classes increases.

This means that the “low correlation between asset classes” and the “diversification effect”—traditionally cited as the core advantages of alternative investments—could weaken.

Viewing Assets as “Risk Assets” and “Safe Assets” Rather Than Separate Categories

In this context, the “benchmark portfolio” is gaining attention.

While traditional strategic asset allocation (SAA) sets target investment weights for each asset class—such as stocks, bonds, and alternative investments—the benchmark portfolio is an approach that designs the portfolio within the broader framework of “risky assets” and “safe assets,” rather than focusing on individual asset classes.

The key is to view assets based on the institution’s overall risk and expected return, rather than the performance of individual investment departments.

In SAA, equity, fixed-income, and alternative investment teams often manage their performance based on their respective predefined benchmarks. As a result, each department is likely to focus on defending its own performance and target weightings rather than the overall portfolio.

This is known as the “silo effect.” The silo effect refers to “departmental self-interest,” where departments within an organization prioritize their own interests or performance without communicating sufficiently with one another.

For example, the fixed-income team may be reluctant to increase its bond allocation out of concern for valuation losses resulting from rising interest rates, while the equity team may oppose reducing the equity allocation due to the potential for falling stock prices. While these decisions may seem rational from each department’s perspective, they can make it difficult for the Chief Investment Officer (CIO) to quickly reallocate assets while considering the overall portfolio’s risk and return.

The benchmark portfolio focuses on relatively breaking down these barriers between asset classes and organizational silos. Its advantage lies in the ability to nimbly reallocate assets based on the total risk and expected return of the entire portfolio, rather than being constrained by the requirement to hold a specific percentage of any given asset.

The ‘Benchmark Portfolio’ Is Not a Panacea… Concerns About Concentration Risk

However, some point out that the benchmark portfolio is not a universal solution for asset allocation.

With a benchmark portfolio, the judgment of investment professionals becomes relatively more important than with SAA. This is because there is significant room for subjective judgment by portfolio managers in the process of deciding which assets to classify as risky and how much risk to accept.

Particularly in situations where market expectations are heavily skewed in one direction—as with AI—this flexibility can actually become a risk factor.

If investment professionals become overly optimistic about AI’s growth potential and profitability, it cannot be ruled out that they will continue to increase the allocation to AI-related assets under the guise of the “benchmark portfolio.” In other words, breaking down the barriers between asset classes could actually result in further concentration on specific themes.

In contrast, SAA sets target weightings and upper and lower bounds for each asset class and manages them within a specific range. It is advantageous for risk management because it can limit the investment weighting of each asset class to a certain level even if excessive optimism regarding a specific theme arises.

Ultimately, it is difficult to say that either the benchmark portfolio or SAA is an absolutely superior strategy. Even when comparing long-term returns, the performance of the two strategies is generally considered to be about 50-50, with no clear winner emerging.

Organizational restructuring is also an issue… “We should blend them rather than choose one over the other”

How to restructure the organization to implement a benchmark portfolio is also a practical challenge.

In a structure where departments or teams are divided by asset class to perform specialized management, immediate asset reallocation—as required by the benchmark portfolio approach—is difficult. If the organizational culture is conservative or the system involves rotational assignments, it is also difficult to push for large-scale organizational restructuring or system changes.

Accordingly, analysts suggest that what institutional investors need is not to completely replace SAA with the benchmark portfolio, but rather a “sense of balance” that appropriately combines the strengths of both strategies.

This approach involves using SAA to set minimum and maximum weightings for each asset class to prevent excessive concentration, while simultaneously utilizing a benchmark portfolio to assess risk and expected returns at the overall portfolio level.

In particular, with the emergence of themes—such as AI—that simultaneously impact multiple asset classes, it has become difficult to assess a portfolio’s actual risk based solely on asset-class classifications such as “X% in stocks, Y% in bonds, and Z% in alternative investments.”

Going forward, the scope of risk management is likely to expand to include not only the names of the funds invested in or the asset classes themselves, but also a closer examination of the specific companies and industries within those funds.

An LP official stated, “Ultimately, institutional investors’ asset allocation strategies are shifting from focusing on ‘how much to allocate to each asset class’ to assessing ‘how much actual risk is actually being exposed to,’” adding, “Since both SAA and benchmark portfolios have their own pros and cons as specific asset allocation methodologies, it is necessary to adopt an approach that comprehensively incorporates the strengths of both.”