Quantitative Investment Strategies Unwrapped

Figure 1: Three Interconnected Models Used to Build QIS
Source: GreySpark analysis

(Click to enlarge)

How QIS Works

QIS strategies are built utilising the common architecture of three interconnected models, as shown in Figure 1:

• A signal generation model utilises historical data to identify return-predictive factors or patterns.
• A portfolio construction model translates generated signals into target positions, accounting for risk constraints, transaction costs and capacity.
• An execution model then automates order generation and placement to minimise market impact.

Signals are generated using a wide range of data types, including price and return series (momentum, mean reversion), fundamental data (value factors, earnings quality), alternative data such as satellite imagery, credit card transactions and web scraping, as well as macroeconomic indicators including interest rates, inflation and economic surprise indices.

Developing and deploying a QIS strategy follows a disciplined research and production cycle. Quants undertake alpha research, developing hypotheses about return-predictive factors grounded in economic intuition or statistical observation. These hypotheses are rigorously backtested against historical data and performance data, accounting for look-ahead bias, data snooping and overfitting. Before capital is committed, strategies typically run in a simulated environment, known as paper trading, allowing strategies to be tested against real-world conditions, without financial exposure. Performance attribution is continuously used throughout a strategy’s lifecycle to assess whether the strategy is working as intended by identifying the sources of return and performance and flagging signs of model degradation. These findings help determine whether the strategy should be recalibrated, scaled back or retired.

In practice, a systematic equity strategy might operate as follows (see Figure 2):

1. Data vendors supply daily price, volume, fundamental and alternative data, which feeds a signal computation layer that calculates factor scores for each security in the investment universe. For example, 12-month momentum, earnings yield or analyst revision momentum.
2. A mean-variance or risk parity optimiser then selects target portfolio weights subject to constraints such as sector limits or factor neutrality targets.
3. Differences between current and target weights generate a list of orders, which are routed to market via algorithmic execution engines (VWAP, TWAP or implementation shortfall algorithms) or dark pools to minimise market impact. Throughout, real-time risk monitoring tracks factor exposures, value-at-risk and drawdown against predefined limits.

Figure 2: Typical Systematic Equity Strategy Process
Source: GreySpark analysis

(Click to enlarge)

QIS Versus Active and Passive Investment Strategies

Historically, active investing was the default form of investing, with fund managers and stockpickers picking securities and assets on behalf of investors. Skilled analysts were trusted to identify mispriced securities and generate positive returns. Passive investing is where a long-term strategy is employed, aiming to match the performance of a specific market index rather than beat it. This is a far more accessible investment approach for retail investors, and it now dwarfs active investing, with trillions of AuM globally.

QIS presents a further alternative for investors, utilising elements of both strategies. Unlike active management, QIS has no human discretion, relying on rules-based, systematic, factor-based investing. Like passive investment, QIS avoids targeting large market cap indexes and pursues specific risk premia or return targets based on the investors risk appetite.

Consequently, QIS has grown significantly among institutional investors, less so for retail, as a solution to find uncorrelated returns within traditional securities markets. Figure 3 draws out the differences between the three investment approaches.

Figure 3: QIS Comparison to Traditional Active Management and Passive Indexing
Source: GreySpark analysisthe

(Click to enlarge)

Key Players in this Space

Typically, QIS are created, manufactured and distributed to investors by large global investment banks. Early adopters included Goldman Sachs, Société Générale, CIBC and BNP Paribas. In fact, Goldman Sachs’ QIS team dates to 1989. A minority of Tier 2 and an even smaller minority of Tier 3 Investment Banks offer QIS products, generally distributing the white-labelled QIS products of larger banks as a standardised hedging solution. The investors in these products are predominantly institutional clients, including pension funds, asset managers and hedge funds. There is also a minority of retail investors accessing QIS via ETFs.

While Tier 1 investment banks are the core manufacturers of QIS, most global investment banks have dedicated QIS groups. In total, GreySpark knows of more than 40 investment banks which provide QIS to clients. These include banks that provide strategies for institutional investors and those that provide for retail and wealth investors. These banks (or their Tier 1 suppliers) design the underlying strategies, create indices and distribute the final product to clients. QIS are delivered as indices, notes or fund-linked solutions to institutional clients. Banks can monetise QIS via structuring fees and bid-offer spreads on the underlying derivatives. Offering QIS services provides a steady revenue stream for banks, dampening the potential volatility of other trading operations.

As the use of QIS has grown, a specialist group of platforms dedicated to providing the aggregation of strategy data, QIS analytics and risk capabilities has emerged. These platforms aggregate strategy data from multiple banks and index providers, develop and enhance performance and risk metrics and provide institutional investors with the ability to screen, filter, compare and select strategies without relying on bank-produced materials. They also support portfolio construction, back-testing validation and regulatory reporting. Additionally, there are providers, such as MSCI and S&P, that design, calculate and maintain QIS and factor strategies indices. These can be used to benchmark QIS strategies against and used for performance tracking (MSCI, Factor Investing) (S&P, Factor Index Solutions).

Historically, pension funds have been the primary investors in QIS. Attracted by the rules-based transparency and cheaper and liquid alternative to the riskier hedge funds. Typically, they use the QIS for hedging and diversification in portfolios. A more recent user group is the institutional asset manager, which integrates QIS either as portfolio building blocks or for hedging and risk purposes. Traditionally, asset managers have relied on their in-house models and stock picking capabilities.

Having previously been a competitor to banks offering QIS products, hedge funds have begun utilising QIS as a complement to their own strategies. QIS provides them with opportunities that they otherwise would not be able to identify.

Lastly, the newest and smallest category of QIS users is the retail investor. ETF trading has been the fastest growing investment product for retail investors in Europe since 2022. Retail ownership of ETFs in the European region alone has increased from USD 60 billion to USD 270 billion from 2020 to 2025 alone and is expected to rise to over USD 500 billion in the next 4 years. Leveraging QIS via ETF trading has become a viable alternative for investors to access moderate returns without resorting to the riskier hedge funds.

Figure 3: Participants in the QIS Market
Source: GreySpark analysis

(Click to enlarge)

AI-enhanced QIS

The utilisation of AI/ML capabilities has become a vital component in the creation of QIS. Enabling the discovery and processing speeds to far surpass the capabilities of a human. Previously, analysts would be tasked with scraping not only performance data, price data and financial statements, but also unstructured data in the form of news articles and company announcements. AI/ML tools have allowed quants to discover patterns across ranges of structured and unstructured data more quickly, through random forests, gradient boosting and neural networks. These techniques are being used alongside existing econometric models.

Additionally, large language models (LLM) have begun extracting structured signals from unstructured data. For example, they can determine whether news articles, company announcements, earnings calls, satellite images or credit card data preceded a price move of a security or not. As the availability and expertise around AI develop, quants will be able to continuously improve models to better utilise these signals and enable a quicker go-to-market for QIS strategies.

Regulatory Environment

QIS sits within a well-established regulatory framework, though the precise obligations depend on how a strategy is structured and distributed. In the US, QIS managers are typically regulated as investment advisers by the Securities and Exchange Commission (SEC) or as commodity trading advisers (CTAs) by the Commodities and Futures Trading Commission (CFTC), depending on the instruments and strategies involved. In Europe, QIS managers operating alternative investment funds fall under the Alternative Investment Fund Managers Directive (AIFMD), while those operating as investment firms are subject to the revised Markets in Financial Instruments Directive (MiFID II). Where QIS strategies involve automated order submission, as is increasingly the case, MiFID II Article 17 governs the algorithmic trading controls required, including pre-trade risk limits, kill-switch capabilities and annual self-assessment obligations.

Model risk is an area of growing regulatory scrutiny. Regulators expect firms to maintain robust model documentation, conduct independent model validation and operate governance frameworks that ensure models are reviewed regularly and retired when they no longer perform as expected. This is particularly relevant for QIS, where the risk that a model performs differently in live markets than in backtesting (known as model risk) can be significant, especially during market regime changes. From a product distribution perspective, QIS products delivered as structured notes or UCITS funds are also subject to PRIIPs KID requirements in Europe, which mandate standardised risk and cost disclosure to investors.

What the Future Looks Like

Assets under management (AuM) allocated to QIS was nearly USD 1 trillion in 2025, and GreySpark forecasts the trajectory for QIS to be one of continued growth. The user base is still broadening, retail investors, corporates and smaller institutional investors represent a meaningful market, yet to be fully tapped. QIS packaged as listed futures and ETFs provide opportunities for this wide trading demographic to access these markets.

A further area of growth comes in the form of AI and LLM technological improvement. The ability to process and scrape not only numerical information but also text, and to use them together, facilitates the automated incorporation of qualitative, unstructured information that was previously the domain of investment analysts.

As QIS becomes a more mainstream component of investment portfolios, the excess returns they were designed to capture may erode because of overcrowding and investor pile in. The opportunity QIS provides can be best taken advantage of by first movers.