A statistical learning workflow for macro trading strategies

April 4, 2020

Statistical learning for macro trading involves model training, model validation and learning method testing. A simple workflow [1] determines form and parameters of trading models, [2] chooses the best of these models based on past out-of-sample performance, and [3] assesses the value of the deployed learning method based on further out-of-sample results. A convenient technology is the ‘list-column workflow’ based on the tidyverse packages in R. It stores all related objects in a single data table, including models and nested data sets, and implements statistical learning through functional programming on that table. Key steps are [1] the creation of point-in-time data sets that represent information available at a particular date in the past, [2] the estimation of different model types based on initial training sets prior to each point in time, [3] the evaluation of these different model types based on subsequent validation data just before each point in time, and [4] the testing of the overall learning method based on testing data at each point in time.

(more…)

The basics of low-risk strategies

March 28, 2020

Low-risk investment strategies prefer leveraged low-risk assets over high-risk assets. The measure of risk can be based on price statistics, such as volatility and market correlation, or fundamental features. The rationale for low-risk strategies is that leverage is not available for all investors (but required to increase the weight of low-risk longs) and that many investors pay over the odds for assets with lottery-like upwardly skewed return expectations. Popular versions of this strategy principle include “betting against beta”, “betting against correlation”, “stable minus risky” or “quality minus junk”. Empirical research suggests that low-risk strategies have delivered significant risk-adjusted returns for nearly a century and that this performance has not deteriorated over time.

(more…)

How loss aversion increases market volatility and predicts returns

March 21, 2020

Loss aversion means that people are more sensitive to losses than to gains. This asymmetry is backed by ample experimental evidence. Loss aversion is not the same as risk aversion, because the aversion is disproportionate towards drawdowns below a threshold. Importantly, loss aversion implies that risk aversion is changing with market prices. This means that the compensation an investor requires for holding a risky asset varies over time, giving rise to excessive price volatility (relative to the volatility of fundamentals), volatility clustering across time, and predictability of returns. All these phenomena are consistent with historical experience and form a useful basis for trading strategies, such as trend following.

(more…)

Reward-risk timing

March 14, 2020

Reward-risk timing refers to methods for allocating between a risky market index and a risk-free asset. It is a combination of reward timing, based on expected future risk asset returns, and volatility timing, based on recent price volatility. A new paper proposes to use machine learning with random forests for estimating both risk premia (return expectations) and optimal lookback windows for volatility estimates This method allows for non-linear prediction interaction and averages forecasts across a range of simplistic valid prediction functions. In an empirical analysis with data going back to 1952 the random forest method for reward-risk timing has outperformed other methods and earned significantly higher risk-adjusted returns than a buy-and-hold strategy.

(more…)

Lagged correlation between asset prices

March 7, 2020

Efficient market theory assumes that all market prices incorporate all information at the same time. Realistically, different market segments focus on different news flows, depending on the nature of the traded security and their research capacity. Such specialization makes it plausible that lagged correlations arise between securities prices, even though their specifics may change overtime. Indeed, there is empirical evidence for lagged correlation between the price trends of different U.S. stocks. Such lagged correlation can be identified and tested through a neural network. Academic research finds that price trends of some stocks have been predictable out-of-sample based on information about the price trends of others.

(more…)

Detecting market price distortions with neural networks

February 22, 2020

Detecting price deviations from fundamental value is challenging because the fundamental value itself is uncertain. A shortcut for doing so is to look at return time series alone and to detect “strict local martingales”, i.e. episodes when the risk-neutral return temporarily follows a random walk while medium-term return expectations decline with the forward horizon length. There is a test based on the instantaneous volatility to identify such strict local martingales. The difficulty is to model the functional form of volatility, which may vary over time. A new approach is to use a recurrent neural network for this purpose, specifically a long short-term memory network. Based on simulated data the neural network approach achieves much higher detection rates for strict local martingales than methods based on conventional volatility estimates.

(more…)

Systemic risk under non-conventional monetary policy

February 15, 2020

Central bank operations in the form of quantitative easing, qualitative easing, forward guidance and collateral policies wield great influence over market prices of risk. These policies reduce market volatility by design, compromising statistical assessment of risk and fostering leverage through endogenous market dynamics, such as collateral amplification. Also, current non-conventional monetary policies become less effective with increasing use. Yields, credit spreads, and term premia all have effective lower bounds and the more they are compressed, the less incremental economic support can be provided. Yet reversing such policies is challenging since leveraged institutions become dependent on cheap funding and credit market stabilization programs. In future economic downturns, another policy regime break is likely, possibly towards monetary financing of fiscal expansion. Hence, macro trading strategies need to consider the structural change in the monetary policy response function.

(more…)

Tracking investor expectations with ETF data

February 8, 2020

Retail investors’ return expectations affect market momentum and risk premia. The rise of ETFs with varying and inverse leverage offers an opportunity to estimate the distribution of such expectations based on actual transactions. A new paper shows how to do this through ETFs that track the S&P 500. The resulting estimates are correlated with investor sentiment surveys but more informative. An important empirical finding is that expectations are extrapolating past price actions. After a negative return shock, investor beliefs become more pessimistic on average, more dispersed, and more negatively skewed.

(more…)

Machine learning and macro trading strategies

February 1, 2020

Machine learning can improve macro trading strategies, mainly because it makes them more flexible and adaptable, and generalizes knowledge better than fixed rules or trial-and-error approaches. Within the constraints of pre-set hyperparameters machine learning is continuously and autonomously learning from new data, thereby challenging or refining prevalent beliefs. Machine learning and expert domain knowledge are not rivals but complementary. Domain expertise is critical for the quality of featurization, the choice of hyperparameters, the selection of training and test samples, and the choice of regularization methods. Modern macro strategists may not need to make predictions themselves but could provide great value by helping machine learning algorithms to find the best prediction functions.

(more…)

The q-factor model for equity returns

January 25, 2020

Investment-based capital asset pricing looks at equity returns from the angle of issuers, rather than investors. It is based on the cost of capital and the net present value rule of corporate finance. The q-factor model is an implementation of investment capital asset pricing that explains many empirical features of relative equity returns. In particular, the model proposes that the following factors support outperformance of stocks: low investment, high profitability, high expected growth, low valuation ratios, low long-term prior returns, and positive momentum. According to its proponents, the investment CAPM and q-factor model complement the classical consumption-based CAPM and explain why many so-called ‘anomalies’ are actually consistent with efficient markets.

(more…)

Tags

Tags
ABOUT US

Macrosynergy is a London based macroeconomic research and technology company whose founders have developed and employed macro quantamental investment strategies in liquid, tradable asset classes, across many markets and for a variety of different factors to generate competitive, uncorrelated investment returns for institutional investors for two decades. Our quantitative-fundamental (quantamental) computing system tracks a broad range of real-time macroeconomic trends in developed and emerging countries, transforming them into macro systematic quantamental investment strategies. In June 2020 Macrosynergy and J.P. Morgan started a collaboration to scale the quantamental system and to popularize tradable economics across financial markets.

FOLLOW US
Linkedin
LEGAL
Research
Quantamental Academy
About Us
Macro-quantamental investment handbook