Absorption Ratio

The Hidden Connections Between Markets

Financial markets are not isolated islands. When panic spreads, seemingly unrelated assets suddenly begin moving in lockstep. Stocks, bonds, commodities, and currencies that normally provide diversification benefits start falling together. This phenomenon, where correlations spike during crises, has devastated portfolios throughout history. The Absorption Ratio provides a quantitative measure of this hidden fragility.

The concept emerged from research at State Street Associates, where Mark Kritzman, Yuanzhen Li, Sebastien Page, and Roberto Rigobon developed a novel application of principal component analysis to measure systemic risk. Their 2011 paper in the Journal of Portfolio Management demonstrated that when markets become tightly coupled, the variance explained by the first few principal components increases dramatically. This concentration of variance signals elevated systemic risk.

What the Absorption Ratio Measures

Principal component analysis, or PCA, is a statistical technique that identifies the underlying factors driving a set of variables. When applied to asset returns, the first principal component typically captures broad market movements. The second might capture sector rotations or risk-on/risk-off dynamics. Additional components capture increasingly idiosyncratic patterns.

The Absorption Ratio measures the fraction of total variance absorbed or explained by a fixed number of principal components. In the original research, Kritzman and colleagues used the first fifth of the eigenvectors. When this fraction is high, it means a small number of factors are driving most of the market movements. Assets are moving together, and diversification provides less protection than usual.

Consider an analogy: imagine a room full of people having independent conversations. Each person speaks at different times about different topics. The total "variance" of sound in the room comes from many independent sources. Now imagine a fire alarm goes off. Suddenly everyone is talking about the same thing, moving in the same direction. The variance is now dominated by a single factor. The Absorption Ratio captures this transition from diverse, independent behavior to unified, correlated movement.

The Implementation Approach

TradingView does not support matrix algebra required for true principal component analysis. This implementation uses a closely related proxy: the average absolute correlation across a universe of major asset classes. This approach captures the same underlying phenomenon because when assets are highly correlated, the first principal component explains more variance by mathematical necessity.

The asset universe includes eight ETFs representing major investable categories: SPY and QQQ for large cap US equities, IWM for small caps, EFA for developed international markets, EEM for emerging markets, TLT for long-term treasuries, GLD for gold, and USO for oil. This selection provides exposure to equities across geographies and market caps, plus traditional diversifying assets.

From eight assets, there are twenty-eight unique pairwise correlations. The indicator calculates each using a rolling window, takes the absolute value to measure coupling strength regardless of direction, and averages across all pairs. This average correlation is then transformed to match the typical range of published Absorption Ratio values.

The transformation maps zero average correlation to an AR of 0.50 and perfect correlation to an AR of 1.00. This scaling aligns with empirical observations that the AR typically fluctuates between 0.60 and 0.95 in practice.

Interpreting the Regimes

The indicator classifies systemic risk into four regimes based on AR levels.

The Extreme regime occurs when the AR exceeds 0.90. At this level, nearly all asset classes are moving together. Diversification has largely failed. Historically, this regime has coincided with major market dislocations: the 2008 financial crisis, the 2020 COVID crash, and significant correction periods. Portfolios constructed under normal correlation assumptions will experience larger drawdowns than expected.

The High regime, between 0.80 and 0.90, indicates elevated systemic risk. Correlations across asset classes are above normal. This often occurs during the build-up to stress events or during volatile periods where fear is spreading but has not reached panic levels. Risk management should be more conservative.

The Normal regime covers AR values between 0.60 and 0.80. This represents typical market conditions where some correlation exists between assets but diversification still provides meaningful benefits. Standard portfolio construction assumptions are reasonable.

The Low regime, below 0.60, indicates that assets are behaving relatively independently. Diversification is working well. Idiosyncratic factors dominate returns rather than systematic risk. This environment is favorable for active management and security selection strategies.

The Relationship to Portfolio Construction

The implications for portfolio management are significant. Modern portfolio theory assumes correlations are stable and uses historical estimates to construct efficient portfolios. The Absorption Ratio reveals that this assumption is violated precisely when it matters most.

When AR is elevated, the effective number of independent bets in a diversified portfolio shrinks. A portfolio holding stocks, bonds, commodities, and real estate might behave as if it holds only one or two positions during high AR periods. Position sizing based on normal correlation estimates will underestimate portfolio risk.

Conversely, when AR is low, true diversification opportunities expand. The same nominal portfolio provides more independent return streams. Risk can be deployed more aggressively while maintaining the same effective exposure.

Component Analysis

The indicator separately tracks equity correlations and cross-asset correlations. These components tell different stories about market structure.

Equity correlations measure coupling within the stock market. High equity correlation indicates broad risk-on or risk-off behavior where all stocks move together. This is common during both rallies and selloffs driven by macroeconomic factors. Stock pickers face headwinds when equity correlations are elevated because individual company fundamentals matter less than market beta.

Cross-asset correlations measure coupling between different asset classes. When stocks, bonds, and commodities start moving together, traditional hedges fail. The classic 60/40 stock/bond portfolio, for example, assumes negative or low correlation between equities and treasuries. When cross-asset correlation spikes, this assumption breaks down.

During the 2022 market environment, for instance, both stocks and bonds fell significantly as inflation and rate hikes affected all assets simultaneously. High cross-asset correlation warned that the usual defensive allocations would not provide their expected protection.

Mean Reversion Characteristics

Like most risk metrics, the Absorption Ratio tends to mean-revert over time. Extremely high AR readings eventually normalize as panic subsides and assets return to more independent behavior. Extremely low readings tend to rise as some level of systematic risk always reasserts itself.

The indicator tracks AR in statistical terms by calculating its Z-score relative to the trailing distribution. When AR reaches extreme Z-scores, the probability of normalization increases. This creates potential opportunities for strategies that bet on mean reversion in systemic risk.

A buy signal triggers when AR recovers from extremely elevated levels, suggesting the worst of the correlation spike may be over. A sell signal triggers when AR rises from unusually low levels, warning that complacency about diversification benefits may be excessive.

Momentum and Trend

The rate of change in AR carries information beyond the absolute level. Rapidly rising AR suggests correlations are increasing and systemic risk is building. Even if AR has not yet reached the high regime, acceleration in coupling should prompt increased vigilance.

Falling AR momentum indicates normalizing conditions. Correlations are decreasing and assets are returning to more independent behavior. This often occurs in the recovery phase following stress events.

Practical Application

For asset allocators, the AR provides guidance on how much diversification benefit to expect from a given allocation. During high AR periods, reducing overall portfolio risk makes sense because the usual diversifiers provide less protection. During low AR periods, standard or even aggressive allocations are more appropriate.

For risk managers, the AR serves as an early warning indicator. Rising AR often precedes large market moves and volatility spikes. Tightening risk limits before correlations reach extreme levels can protect capital.

For systematic traders, the AR provides a regime filter. Mean reversion strategies may work better during high AR periods when panics create overshooting. Momentum strategies may work better during low AR periods when trends can develop independently across assets.

Limitations and Considerations

The proxy methodology introduces some approximation error relative to true PCA-based AR calculations. The asset universe, while representative, does not include all possible diversifiers. Correlation estimates are inherently backward-looking and can change rapidly.

The transformation from average correlation to AR scale is calibrated to match typical published ranges but is not mathematically equivalent to the eigenvalue ratio. Users should interpret levels directionally rather than as precise measurements.

Correlation regimes can persist longer than expected. Mean reversion signals indicate elevated probability of normalization but do not guarantee timing. High AR can remain elevated throughout extended crisis periods.


References

Kritzman, M., Li, Y., Page, S., and Rigobon, R. (2011). Principal Components as a Measure of Systemic Risk. Journal of Portfolio Management, 37(4), 112-126.

Kritzman, M., and Li, Y. (2010). Skulls, Financial Turbulence, and Risk Management. Financial Analysts Journal, 66(5), 30-41.

Billio, M., Getmansky, M., Lo, A., and Pelizzon, L. (2012). Econometric Measures of Connectedness and Systemic Risk in the Finance and Insurance Sectors. Journal of Financial Economics, 104(3), 535-559.