Research Review | 8 March 2024 | Combination Model Forecasting

Market Risk Premium Expectation: Combining Option Theory with Traditional Predictors
Hong Liu (Washington University in St. Louis), et al.
December 2022
In general, the slackness between the Martin lower bound (solely based on option prices) and the market risk premium depends on economic state variables. Empirically, we find that combining information from option prices and economic state variables yields forecasts of the market risk premium with greater out-of-sample performance compared to forecasts using option prices alone or economic state variables alone. Additionally, these combination-based forecasts can significantly increase investors’ utility by improving their portfolios’ Sharpe ratios. Our findings suggest the importance of combining information from option prices and economic state variables.

Flexible Global Forecast Combinations
Yilin Qian (The University of Sydney), et al.
April 2023
Abstract Forecast combination—the aggregation of individual forecasts from multiple experts or models—is a proven approach to economic forecasting. To date, research on economic forecasting has concentrated on local combination methods, which handle separate but related forecasting tasks in isolation. Yet, it has been known for over two decades in the machine learning community that global methods, which exploit taskrelatedness, can improve on local methods that ignore it. Motivated by the possibility for improvement, this paper introduces a framework for globally combining forecasts while being flexible to the level of taskrelatedness. Through our framework, we develop global versions of several existing forecast combinations. To evaluate the efficacy of these new global forecast combinations, we conduct extensive comparisons using synthetic and real data. Our real data comparisons, which involve forecasts of core economic indicators in the Eurozone, provide empirical evidence that the accuracy of global combinations of economic forecasts can surpass local combinations.

Empirical Asset Pricing with Probability Forecasts
Songrun He (Washington University in St. Louis), et al.
February 2024
We study probability forecasts in the context of cross-sectional asset pricing with a large number of firm characteristics. Empirically, we find that a simple probability forecast model can surprisingly perform as well as a sophisticated probability forecast model, and all of which deliver long-short portfolios whose Sharpe ratios are comparable to those of the widely used return forecasts. Moreover, we show that combining probability forecasts with return forecasts yields superior portfolio performance versus using each type of forecast individually, suggesting that probability forecasts provide valuable information beyond return forecasts for our understanding of the cross-section of stock returns.

Forecast Combination in the Frequency Domain
Gonçalo Faria (Catholic University of Portugal and Fabio Verona (Bank of Finland)
February 2023
We propose a new forecasting method – forecast combination in the frequency domain – that takes into account the fact that predictability is time and frequency dependent. We use this method to forecast the equity premium and real GDP growth rate. Combining forecasts in the frequency domain produces markedly more accurate predictions relative to the standard forecast combination in the time domain, both in terms of statistical and economic measures of out-of-sample predictability. In a real-time forecasting exercise, the flexibility of this method allows to capture remarkably well the sudden and abrupt drops associated with recessions and further improve predictability.

Forecasting the Risk of Cryptocurrencies: Comparison and Combination of GARCH and Stochastic Volatility Models
Jan Prüser (Technical University of Dortmund)
January 2023
We provide a comparison of several GARCH and stochastic volatility models for forecasting the risk of cryptocurrencies. It turns out that the widely used GARCH(1,1) does not provide accurate risk predictions. In contrast, adding t-distributed innovations or allowing for regime changes improves the accuracy in both model classes. Moreover, we find that Markov switching stochastic volatility models perform in particular well. Finally, we consider a Bayesian decision-guided approach with discount learning to combine the different models and provide robust evidence that combining the model predictions leads to accurate combined risk.

A Structural Approach to Combining External and DSGE Model Forecasts
Thorsten Drautzburg (Federal Reserve Bank of Philadelphia)
June 2023
This note shows that combining external forecasts such as the Survey of Professional Fore casters can significantly increase DSGE forecast accuracy while preserving the interpretability in terms of structural shocks. Applied to pseudo real-time from 1997q2 onward, the canonical Smets and Wouters (2007) model has significantly smaller forecast errors when giving a high weight to the SPF forecasts. Incorporating the SPF forecast gives a larger role to risk premium shocks during the global financial crisis. A model with financial frictions favors a larger weight on the DSGE model forecast.

Wisdom of the Silicon Crowd: LLM Ensemble Prediction Capabilities Rival Human Crowd Accuracy
Philipp Schoenegger (London School of Economics and Political Science), et al.
February 2024
Human forecasting accuracy in practice relies on the ‘wisdom of the crowd’ effect, in which predictions about future events are significantly improved by aggregating across a crowd of individual forecasters. Past work on the forecasting ability of large language models (LLMs) suggests that frontier LLMs, as individual forecasters, underperform compared to the gold standard of a human crowd forecasting tournament aggregate. In Study 1, we expand this research by using an LLM ensemble approach consisting of a crowd of twelve LLMs. We compare the aggregated LLM predictions on 31 binary questions to that of a crowd of 925 human forecasters from a three-month forecasting tournament. Our preregistered main analysis shows that the LLM crowd outperforms a simple no-information benchmark and is not statistically different from the human crowd. In exploratory analyses, we find that these two approaches are equivalent with respect to medium-effect-size equivalence bounds. We also observe an acquiescence effect, with mean model predictions being significantly above 50%, despite an almost even split of positive and negative resolutions. Moreover, in Study 2, we test whether LLM predictions (of GPT-4 and Claude 2) can be improved by drawing on human cognitive output. We find that both models’ forecasting accuracy benefits from exposure to the median human prediction as information, improving accuracy by between 17% and 28%: though this leads to less accurate predictions than simply averaging human and machine forecasts. Our results suggest that LLMs can achieve forecasting accuracy rivaling that of human crowd forecasting tournaments: via the simple, practically applicable method of forecast aggregation. This replicates the ‘wisdom of the crowd’ effect for LLMs, and opens up their use for a variety of applications throughout society.


Learn To Use R For Portfolio Analysis
Quantitative Investment Portfolio Analytics In R:
An Introduction To R For Modeling Portfolio Risk and Return

By James Picerno


One thought on “Research Review | 8 March 2024 | Combination Model Forecasting

  1. Pingback: Quantocracy's Daily Wrap for 03/11/2024 - Quantocracy

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.