Skip to main content
All CollectionsFeature Deep Dives
Feature Overview: Desmoothing of Private Asset Allocations
Feature Overview: Desmoothing of Private Asset Allocations
Written by Kyle Reinhardt
Updated this week

Rev. Date July14, 2023

What is the purpose of desmoothing on Venn?

The returns of private assets tend to exhibit high autocorrelation, meaning they show a high degree of similarity between data points along the time-series. This would suggest that past returns of private assets may predict future returns, which violates the efficient market hypothesis. Additionally, these returns tend to lag those of publicly traded proxies. Desmoothing aims to decrease autocorrelation of private assets in order to better capture their true economic return. We believe that it also helps to increase the explanatory power of the Two Sigma Factor Lens on these assets.

What types of assets can be desmoothed?

Venn supports desmoothing of assets with monthly or quarterly returns data.

What are the minimum number of data points required for desmoothing on Venn?

Venn requires a minimum of 12 data points for both monthly and quarterly data for desmoothing, with one additional data point needed for each lag. The maximum number of lags is 12 for monthly data and 6 for quarterly.[1]

How can I desmooth my private asset returns on Venn?

When analyzing a monthly or quarterly return stream, you can select the “Proxy” menu in the upper right hand corner of the Analysis page. Alternatively, this menu can be accessed by clicking on “Manage Data”, then “Analysis Date Range”, then “Proxy” on the right hand side of the page.

From the proxy menu, you can select either of the options containing desmoothing:

  • Interpolation with desmoothing (*recommended for quarterly investments)

  • Desmoothing

Within either of these menus, you can select a Public Market Proxy from the private asset category breakdown, or select your own proxy from the Venn library. Venn will identify any recommended proxies.

If you know the number of lags that a manager implements when smoothing their returns, you can simply input the number. In cases where the number of lags is not known, Venn uses the Akaike Information Criterion (AIC) to evaluate multiple models with different numbers of lags, and chooses the model of best fit. The expected impact of desmoothing is that the autocorrelation of the returns is decreased, and the desmoothed returns more closely match the mark-to-market value of the Public Market Proxy.

Why does Venn recommend to desmooth with interpolation?

The goal of both desmoothing and interpolation is to develop what we believe to be a more accurate representation of the true risks associated with private assets.

Interpolation allows users to increase the return frequency of a private asset from quarterly to daily, which aims to produce a risk profile that is closer to that of the Public Market Proxy. Desmoothing aims to decrease autocorrelation to produce a risk profile that is also more closely aligned with that of the proxy.

We believe that combining the two produces a resulting time-series that more closely resembles a marked-to-market public return stream, thus yielding more explanatory power to the Two Sigma Factor Lens.

Where can I see desmoothing on Venn?

The desmoothed returns for a private asset can be viewed on the investment’s Cumulative Return chart. Additionally, one can compare the Autocorrelation before and after desmoothing to confirm the impact of the function.

How does this feature fit in with the rest of Venn’s Private Asset suite?

Venn offers several proxy options incorporating Interpolation and Desmoothing, as well as an autocorrelation metric to help identify investments that could benefit from these features:

  • Interpolation: Utilizes a public market proxy to create a daily return stream in order to better reflect the underlying economic risk of assets with infrequent data.

  • Desmoothing: Leverages public market proxies to create a more realistic return stream that decreases autocorrelation. Venn’s belief is this will provide a view that more faithfully reflects the timing and level of true economic risk.

  • Desmooth with Interpolation: Combines the two proxy methods to produce a resulting time series that aims to closely resemble a marked-to-market public return stream. Desmoothing and interpolation, both separately and when combined, may increase the explanatory power of the Two Sigma Factor Lens.

  • Autocorrelation within Historical Risk Statistics will show the level of correlation between returns data over time and can help identify investments to desmooth.

Why are certain categories not recommended for desmoothing?

Venn will warn a user if desmoothing is not recommended in cases where the desmoothed private asset series does not have a meaningful and reliably positive relationship with a Public Market Proxy.

In order for a category to be recommended, the desmoothed private asset series must meet the following conditions in relation to the underlying Public Market Proxy:

  • Correlation must be greater than 0.4

[1] If the selected proxy starts less than 1 year before the investment, the desmoothed return series will start 1 year from the proxy start date. To preserve the start date of the original investment series, users must select a proxy with 12 months or more of data before the investment’s inception.

Proxies are for estimation purposes only and have many inherent limitations. The methodology for calculating potential proxies was chosen in our professional judgement, and will not always yield the most accurate available proxy. Our potential proxy suggestions are not a recommendation as to any portfolio, allocation, strategy, or investment nor an offer to purchase or sell any security.

This document highlights certain aspects of this feature. As an overview, it does not discuss all material facts or assumptions. Please see Important Disclosure and Disclaimer Information.

Did this answer your question?