Robo advice - Target Risk Asset Management Model published in Wikifolio

Persephone Team, August 2018

Persephone’s proprietary robotIQ© harbors, among others, various financial asset management models. robotIQ© manages our clients’ assets applying highly specific quantitative approaches. Two of the largest European banks rely on robotIQ© as the core algorithmic platform for their robo-advice offerings. Given the frequent inquiries and the high interest in Persephone’s developed models that drive these robo-advice solutions, we decided to disclose a set of real-life portfolios to the public.

In this particular case, robotIQ© applies the target risk approach to manage portfolios. The idea behind target risk asset management is to manage assets given a pre-defined risk level, which is denoted by the Value-at-Risk (VaR) indicator. In our case, the VaR has a time horizon of one year and a confidence level of 95%. Hence, the risk level represents one key parameter and might be contractually defined for a DPM mandate.

There are various other parameters that clients may specify when delegating the management of their assets. By this means, robotIQ© enables clients to customize a truly individual portfolio management approach that fits their personal needs and financial situation. For instance, clients may prioritize or restrict investments separately for single asset classes or foreign currencies. robotIQ© also offers clients to allow investments in crypto currencies (e.g. bitcoin). Moreover, clients may permit the use of leveraged investment vehicles to increase return potential, or for hedging purposes.

We have chosen Wikifolio as a neutral and objective platform for publishing a set of sample portfolios with different VaR levels. Our aim is to provide transparency and insights into the process, relevant risks and, most interestingly, into the performance behavior of one of the most advanced robo-advice systems.

Currently, the following portfolios are published on Wikifolio (Source: Persephone robotIQ© system)


A legitimate concern with the plain vanilla target risk approach is that asset management and allocation tend to be pro-cyclical and hence deliver weak performance.

However, Persephone does not rely on the plain-vanilla approach of applying statistical methods on historical data to assess future market returns. We believe that, if at all, only human understanding of financial markets and expert investment experience can reasonably assess future market developments. Therefore, we developed a proprietary algorithm to transform high-level human market views into machine-readable return expectations. Our asset management with robotIQ© is based on a discretionary strategic asset allocation of a broadly diversified investment fund that is managed independently for any robo-advice portfolio. Persephone’s asset-liability committee (ALCO), which is comprised of renowned capital markets experts, reviews and adjusts the strategic asset allocation on a regular and ad-hoc basis following in-depth market research, fundamental analyses and outstanding capital investment experience.

Starting from this strategic allocation, algorithms compute instrument return expectations for each asset class and foreign currency, which are applied for portfolio management and construction of robo-advice portfolios with arbitrary target risk levels.

The portfolios listed on wikifolio are realizations of a standardized implementation of our robotIQ© platform. To showcase the flexibility, the two features of additional asset classes (some portfolios allow crypto-assets) and leveraged products are included as examples. The target risk approach is maintained for all portfolios, meaning that bitcoin or stock options are included in the management without breaking the risk limits.

The achieved performance of the published robotIQ© portfolios is compared against the performance of other robo-advisory offerings with help of the tests by Brokervergleich (brokervergleich.de). Brokervergleich analyzes and updates the performance of the robo-advice solutions on a monthly basis. Consequently, we will publish an updated comparison on this site on a monthly basis as well [1].

Needless to say, the performance measurement for robotIQ© portfolios follows the test framework of Brokervergleich [2]. Particularly, we have included transaction costs, management fees and taxation [3] when calculating robotIQ© portfolio performance.

Since the tax module has only been added in June, a truly fair comparison of robotIQ© against the performance KPIs presented by Brokervergleich starts in July 2018.

Performance July 2018 Robo-Advisors [5] [6]


Note that the performance on wikifolio deviates from that indicated herein since we correct for capital gains tax for a fair comparison.

Overperformance of the robotIQ© portfolios in July 2018 can mainly be explained by the above-average growth of crypto-currency assets during this month. However, having observed bitcoin and other crypto-currencies dropping about 25% in the first weeks of August, it will be highly interesting to see how that might affect the robotIQ© portfolios compared to other offerings not investing in crypto-currencies. The updated performance comparison for August 2018 will be available on this website (www.persephone-finance.com) beginning of September.

[1] Source: Brokervergleich, Franke-Media.net, Leipzig

[2] Please see Brokervergleich Testverfahren for a detailed description of the test approach

[3] Equally to Brokervergleich, we assume no tax exemption order (Freistellungsauftrag) and that tax loss accumulation totals (allgemeiner Verlusttopf) were zero on account opening

[4] Performance of robo-advisory offerings calculated by Brokervergleich(see above), all calculations for robotIQ© portfolios based on data from Wikifolio, Ariva, Lang&Schwarz and own calculations.

[5] robotIQ© portfolios in solid color (VaR 13, 16, 18) have risk characteristics that are similar to the ones of other robo-advice portfolios Brokervergleich has chosen for comparison

Legal Disclaimer

The views, thoughts, and opinions expressed in this article belong solely to the author, and not necessarily to the author’s employer, organization, committee or other group or individual.

Data is collected from various sources, and may be modified over time without notice to improve quality and accuracy. Persephone disclaims any responsibility or liability for errors, omissions, completeness, accuracy, usefulness, frequency of provision or timeliness of the data and the content of this site and makes no warranty concerning it.

Any information about performance represented here is derived from historical data. Past performance however is not a reliable indicator of future results.

The information provided here are for illustrative purposes only. This website and the information contained herein is not intended to be a source of advice or with respect to the material presented, and the information and/or documents contained in this website do not constitute investment advice.