Price distortions are an important source of short-term trading profits, particularly in turbulent markets. Here price distortions mean apparent price-value gaps that arise from large inefficient flows. An inefficient flow is a transaction that is not motivated by rational risk-return optimization. One source of such inefficient flows is ‘rebalancing’, large-scale institutional transactions that align allocation with fixed targets. Rebalancing flows are detectable or even predictable if one understands their rules. Their motives include benchmarking of portfolios, benchmark changes, regulatory changes, ETF designs, equity parity, capital protection, and – to some extent – high-frequency trading algorithms.
The below is the first in a series of updates of this site’s summary on market price distortions.
What are price distortions?
Price distortions are apparent price-value gaps. They are defined as deviations of quoted prices from a level that would clear the market if all participants were trading for conventional risk-return optimization. In short, they are discrepancies between mark-to-market prices and a plausible range of economic values of a contract. Trading strategies that are based on price distortions rely less on information advantage than on consistent price monitoring, flexibility of trading, privileged market access, and rational discipline in turbulent times. Price distortions arise from inefficient flows and prevail as long as a sizable share of market participants is either unwilling or unable to respond to obvious dislocations.
What is rebalancing?
Rebalancing is the process of realigning the weights in a portfolio with the designed purpose of the investment vehicle or strategy. Rebalancing seeks to limit exposure to unwanted risk, regardless of whether that risk pays a high premium or not. The main rebalancing processes simply prescribe periodically buying or selling assets to maintain an original or desired allocation.
Rebalancing is a source of inefficient flows. Importantly, mechanical rebalancing rules are in fact active algorithmic strategies, even if they do not explicitly seek return optimization. For example, simple periodic reallocation to fixed asset weights means that winning assets are systematically sold and losers are bought. In markets with trends and relative price momentum this creates losses and slows trends at the market level.
It is important to understand the motivation behind specific rebalancing flows in order to detect potential price distortions. Academic research shows that the effect of rebalancing cascades on the net demand for individual assets will look like noise, even if the flows are fully rational (view post here). Predictions of aggregate flows become infeasible because of alternating buy and sell orders, feedback loops and threshold-based execution rules. This cautions against just dismissing seemingly non-fundamental market flows as irrational and betting against them.
The most common motive of rules-based rebalancing is benchmarking, i.e the use of pre-set standards for allocation, risk, and return. Many investment managers are formally or informally benchmarked against some market index and prefer to contain deviations of their returns from those of the benchmark index.
“Underweights” in volatile but outperforming assets are the main risk of violating these margins, because such assets simultaneously outperform and gain weight in the benchmark. Hence investment managers often find themselves compelled to buy overpriced and risky assets merely to contain streaks of underperformance (view post here). Profit maximizing traders can exploit the market’s proclivity to overvalue high-beta and high-volatility assets on these occasions. Empirical research has provided evidence for a “low risk effect” in financial markets, i.e. the recurrent outperformance of low-risk versus high-risk assets, once both are scaled by volatility (view post here).
Benchmarking effects should not be confused with benchmark effects. Benchmark effects arise from changes in global securities indices that are commonly tracked by investment managers. In particular, the surge in passive investment means that a large share of institutional investors is under obligation to buy and sell in accordance with the constituents and weights used by benchmark indices, regardless of assets’ fundamental values (view post here).
Benchmark companies revise indices regularly, causing re-weighting of sectors or countries that is not in proportion to market capitalization. Sovereign credit rating changes, for example, can establish or remove the eligibility of a country’s securities for inclusion in benchmark indices. There is empirical evidence that these changes induce sizeable portfolio re-allocations and international capital flows, entailing an outperformance of ‘upgraded’ assets at the time of announcement and the time of actual index adjustment (view post here). Upgrading here does not mean necessarily better asset quality, but rather the assets’ greater access to index-tracking capital allocations.
Regulatory changes can necessitate the strategic rebalancing of large segments of the market. This motive is particularly important for tightly regulated institutions such as insurance companies and pension funds (view post here). Thus, the EU reform of the regulation and supervision of insurance and reinsurance undertakings in 2016 introduced bias against assets with high market and liquidity risk, such as equity, and in favor of low-yielding sovereign bonds (view post here). Also, regulatory changes seem to be one of the key motivations behind herding in the pension industry. Greater complexity and policymaker discretion in the wake of the great regulatory reform of the 2010s means that investment managers must pay more attention to regulatory policies, not unlike the way they have monitored monetary policies (view post here). Since regulatory allocation changes are unrelated to risk-return optimization resultant flows are likely to be inefficient and conducive to price distortions.
Exchange-traded funds are hybrid investment vehicles that are continuously traded in a liquid market. The goal of a traditional ETF is to match the returns of its associated index or market sector. ETFs have been a major part of the passive investment boom since the 2000s, expanding in size, diversity, scope, and complexity (view post here).
All ETFs rebalance periodically. A regular passive ETF weights its holdings in a fashion that is similar to the underlying index but might rebalance only on an annual or semi-annual basis. Such rebalancing may lead to just subtle market price distortions. Rebalancing flows can become a stronger force when the arbitrage mechanism between ETFs and their constituent securities runs into troubles. ETF prices can deviate significantly from those of the constituent securities, especially at high frequencies, for illiquid assets and during periods of financial stress. Empirically, ETFs have been associated with greater co-movement of asset prices: stocks tend to co-move more with their respective indices once they are included in ETF portfolios. There is also evidence that ETFs are associated with increased price volatility of the constituent securities (paper here).
Rebalancing flows of equity ETFs that are leveraged can be particularly conducive to price distortions. The goal of leveraged ETFs is to realize returns that may be double or triple those of the underlying index or market sector. Leveraged ETFs use borrowed money to increase returns. Leveraged ETFs are subject to automatic rebalancing rules, requiring them to buy when prices rise and sell when they fall. As leveraged ETFs have become a significant factor in U.S. equity markets they can reinforce or even escalate large directional moves in the stock market, both through their own transactions and other market participants’ front running (view post here).
More subtle rebalancing flows arise from the so-called “uncovered equity parity“. This parity suggests that when foreign equity holdings outperform domestic U.S. holdings, USD-based investors are exposed to elevated exchange rate and country risk in form of higher USD notional in the foreign currency area. There should be a tendency to reduce or hedge the exposure subsequently. There is indeed empirical evidence for investors selling winning equity markets in 1990-2010 (view post here). However, other analyses suggest that this effect may not be dominant (paper here) overtime. Plausibly, equity parity flows introduce subtle short-lived distortions around rebalancing dates.
Constant Proportion Portfolio Insurance
Constant Proportion Portfolio Insurance or CPPI products are capital protection products. Rather than using options they deploy a dynamic asset allocation strategy. Put simply, a CPPI strategy allocates between a riskless asset and a risky asset, such as equity, hedge funds, or commodity indices. The manager defines the “cushion” or the percentage of the fund’s assets that may be put at risk, which is estimated based on the difference between the initial value of the product and the present value minimum necessary to provide the capital guarantee at maturity.
In rising markets, a CPPI strategy allocates more towards the risky asset. In falling market, it allocates more towards the safe asset. Since CPPI flows do not consider any aspects of assets’ fundamental value they are an example of inefficient flows that reinforce market trends.
High-frequency traders do not really rebalance, but also follow strict rules-based position management. High-frequency trading became a large-scale business during the 2000s and is managed mostly by independent proprietary traders. It executes large numbers of trades in less than one millisecond, powered by trading algorithms and based on fast-moving market data. The share of high-frequency trading across various markets has been estimated between 10% and 50%. Most high-frequency strategies seek to exploit tiny arbitrage opportunities in large volumes. For example in “slow-market arbitrage”, the high-frequency trader detects price moves on one exchange and picks off orders sitting on another before it can react.
High-frequency trading strategies respond at high speed to changes in prices by using relatively simple strategies. Speed, rather than reflection, is of the essence. In particular, high-frequency trading algorithms have no fundamental anchor and simply move too fast for humans to intervene with judgment. For example, when stocks drop, even if due only to a “fat finger”, the programs may decide to stop trading, withdrawing liquidity from the market, or even aggravate the sell-off.
While high-frequency trading can provide liquidity and efficiency on many occasions it can magnify volatility on others. In particular, it can make markets more prone to vanishing liquidity. Increases in trading speed, in conjunction with market concentration, and regulatory costs of market making, augment the probability of liquidity events (view post here). The May 2010 “flash crash” in the U.S. stock market exemplified that risk. Moreover, there is a non-zero probability of outright “glitches” that can escalate modest price changes toward systemically destabilizing events (view post here). Modern physics teaches that objects behave differently as they reach the speed of light. In particular, quantum physics suggests that ‘freak events’ that destabilize the markets are likely to occur.