Quantifying Backtest Overfitting in Alternative Beta Strategies

Suhonen, Antti; Lennkh, Matthias; Perez, Fabrice

doi:10.3905/jpm.2017.43.2.090

Cited by 21 publications

(7 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In that regard, our results are in line with the ones obtained by Suhonen et al [2017] and the haircut that should be applied does not differ significantly. In their paper, median annualized Sharpe ratios before and after the launch date amount respectively to 1.20 and 0.31, hence translating into a median haircut of 73%.…”

Section: Selection Biasessupporting

confidence: 91%

See 1 more Smart Citation

Alternative Risk Premia: Is the Selection Process Important?

Naya¹,

Tuchschmid²

2019

JWM

View full text Add to dashboard Cite

Alternative Risk Premia (ARP) are rule-based strategies. They should reward investors exposed to non-traditional systematic risk factors. Yet, allocation to ARP is not straightforward. First, there are many ARP indices proposed by different providers that claim to capture the same underlying risk premia. Second, a proposed index may not automatically mimic an existing risk premium whose performance is sustainable or persistent. Our findings confirm these suspicions.If some categories of indices show risk-return characteristics that are rather homogeneous, others are highly heterogeneous. Stated otherwise, performance is provider dependent making the choice of an index an important component of the allocation process. Differences between simulated past results and live data are then calculated. Results are indisputable. There is a significant overfitting bias. Once launched, the performance of ARP indices dropped significantly. To summarize, this research paper shows that investors should take no short cuts.When it comes to allocating capital to ARP, an extensive due diligence process is required.Electronic copy available at: https://ssrn.com/abstract=3045057 3 Alternative Risk Premia (ARP) are better understood when compared to traditional long-only risk factors. Everyone investing in financial securities as equities or bonds is indeed aware that capital is at risk and that it is this same risk that calls for the existence of "risk premia". The idea behind ARP is the same. Investors must be rewarded for the risk they take. The difference comes from the way the exposure is achieved. ARP are non-traditional risk factors and roughly speaking correspond to any long-short strategies or styles whose risk is a priori not diversifiable. Examples could be FX carry, Interest Rate (IR) carry or equity size.ARP have experienced an increase in popularity recently, among both researchers and investors.Size-wise, the ARP market still looks small compared to the hedge fund industry whose assets under management (AuM) estimates are close to $3 trillion, but new indices are continuously created and AuM are increasing. 1 According to CitiGroup, the ARP market accounted for a tiny $15 billion in 2011. By the end of 2015 it had risen to $241 billion. 2 Yet, there is nothing new in the idea that returns of risk assets can be expressed with a set of limited factors, whatever the name we want to give them. Single-factor or multi-factor return generating processes can be found in articles published as early as the sixties and seventies. One can cite, for example, the paper of Sharpe [1964] and the use of a diagonal model or the seminal article of Ross [1976] and the contribution of his Arbitrage Pricing Theory to modern finance. However, when it comes to multi-factor pricing models, one often points to the article of Fama and French [1993] in which size and book-to-market were added to a "market" factor to explain returns on equities. Since then, other factors have been proposed. The work of Jegadeesh and Titman [1993] led Cahart [19...

show abstract

Section: Selection Biasessupporting

confidence: 91%

“…One could cite here the recently published article of Suhonen et al [2017] that analyses the persistence of ARP performance after the live date. Our paper belongs to this field of research.…”

mentioning

confidence: 99%

Alternative Risk Premia: Is the Selection Process Important?

Naya¹,

Tuchschmid²

2019

JWM

View full text Add to dashboard Cite

show abstract

“…As with nonbackfilled data for hedge funds, bank product returns are likely to be more realistic after the live date. Suhonen, Lennkh, and Perez (2017) found that alphas are reduced by half in the live period. 13 Second, products that perform poorly are likely to be dropped from the bank's current lineup and perhaps replaced by better-performing versions.…”

Section: Bank Risk Premia Indexesmentioning

confidence: 99%

Hedge Funds vs. Alternative Risk Premia

Jorion

2021

Financial Analysts Journal

View full text Add to dashboard Cite

show abstract

“…We can outline, chronologically, three distinct approaches in the literature to evaluate and deal with backtesting overfitting: Data Snooping, Overestimated Performance, and Cross-Validation Evaluation. The problem of overfitting cannot be understated, and innumerable references highlight the issues with phacking which has been an issue for considerable periods but making headlines more recently (see e.g., [35], [26], [7], [13]), and [24]) although it may not always be willful ( [16]), and is pernicious in finance (see for instance, [44]).…”

Section: Literature Reviewmentioning

confidence: 99%

Avoiding Backtesting Overfitting by Covariance-Penalties: An Empirical Investigation of the Ordinary and Total Least Squares Cases

Koshiyama

Firoozye

2018

SSRN Journal

View full text Add to dashboard Cite

Systematic trading strategies are rule-based procedures which choose portfolios and allocate assets. In order to attain certain desired return profiles, quantitative strategists must determine a large array of trading parameters. Backtesting, the attempt to identify the appropriate parameters using historical data available, has been highly criticized due to the abundance of misleading results. Hence, there is an increasing interest in devising procedures for the assessment and comparison of strategies, that is, devising schemes for preventing what is known as backtesting overfitting. So far, many financial researchers have proposed different ways to tackle this problem that can be broadly categorised in three types: Data Snooping, Overestimated Performance, and Cross-Validation Evaluation. In this paper we propose a new approach to dealing with financial overfitting, a Covariance-Penalty Correction, in which a risk metric is lowered given the amount of parameters and data used to underpins a trading strategy. We outlined the foundation and main results behind the Covariance-Penalty correction for trading strategies. After that, we pursue an empirical investigation, comparing its performance with some other approaches in the realm of Covariance-Penalties across more than 1300 assets, using Ordinary and Total Least Squares. Our results suggest that Covariance-Penalties are a suitable procedure to avoid Backtesting Overfitting, and Total Least Squares provides superior performance when compared to Ordinary Least Squares.We can outline, chronologically, three distinct approaches in the literature to evaluate and deal with backtesting overfitting: Data Snooping, Overestimated Performance, and Cross-Validation Evaluation. The problem of overfitting cannot be understated, and innumerable references highlight the issues with phacking which has been an issue for considerable periods but making headlines more recently (see e.g., [35], [26], [7], [13]), and [24]) although it may not always be willful ([16]), and is pernicious in finance (see for instance, [44]).

show abstract

Quantifying Backtest Overfitting in Alternative Beta Strategies

Cited by 21 publications

References 14 publications

Alternative Risk Premia: Is the Selection Process Important?

Alternative Risk Premia: Is the Selection Process Important?

Hedge Funds vs. Alternative Risk Premia

Avoiding Backtesting Overfitting by Covariance-Penalties: An Empirical Investigation of the Ordinary and Total Least Squares Cases

Contact Info

Product

Resources

About