Asset Class
An asset class is a grouping of investments that share economic characteristics and respond similarly to market conditions. The standard top level is equities, fixed income, cash, real estate, and alternatives, with each subdivided into subclasses (large-cap growth equities, investment-grade corporate bonds, etc.).
Asset-class taxonomy is the foundation of every portfolio-allocation conversation, every benchmarking comparison, and every risk model. Modern Portfolio Theory operates at the class and subclass level — covariance matrices are built between classes, optimization runs over class weights, and rebalancing rules trigger at class drift thresholds. A wealth platform that doesn't have a clean class taxonomy ships with a known accuracy ceiling on every downstream analytics feature.
The complication is that classification is not always obvious. A REIT is legally an equity but behaves partly like real estate; a high-yield bond fund holds bonds but has equity-like volatility; a master limited partnership trades like a stock but distributes K-1s like a partnership. Different platforms classify these inconsistently, which causes real disputes when household data is portable across institutions.
The practical contract is: every position has both a class (top-level) and a subclass (named segment within), and the subclass mapping is explicit. Common subclasses include US large-cap growth/value/blend, US small-cap, developed international, emerging markets, investment-grade corporates, high-yield bonds, US Treasuries, municipals, REITs, MLPs, commodities, and cash equivalents. The taxonomy is closed-set — every position must map to exactly one entry — and the mapping is stable across versions so that historical comparisons remain meaningful.
Synthetic households need realistic class-and-subclass mappings, not just ticker → 'equity'. The covariance structure between classes drives every Monte Carlo result and every rebalancing-engine test. A test dataset where 80% of households hold the same three ETFs cannot exercise diversification logic. Synthetic data should produce the empirical class-weight distributions of real wealth-tier segments — affluent households are 60% equity / 30% fixed / 10% alternative on average; UHNW households are 35% equity / 25% fixed / 40% alternative.
Common pitfalls
- Using a flat ticker → class table without subclass — kills any factor-tilt or style-drift analysis.
- Letting alternatives leak into the equity bucket because the position is held through a public-equity wrapper (BDC, listed PE).
- Failing to separate cash from cash-equivalents — money-market funds behave differently from sweep deposits at scale.
- Mapping the same ticker to different subclasses in different builds, breaking time-series consistency.
Examples
The skeleton most allocation engines operate on.
{
"asset_class": "equity",
"subclass": "us_large_cap_growth",
"geography": "us",
"factor_tilts": ["momentum", "quality"]
}