Munich Personal RePEc Archive

Multi-Index Evaluation of Alternative Assets Funds. Time Lagged Effects and Linear Factors Capturing Non-linear Effects

Scorbureanu, Alexandrina Ioana (2013): Multi-Index Evaluation of Alternative Assets Funds. Time Lagged Effects and Linear Factors Capturing Non-linear Effects.

[img]
Preview
PDF
MPRA_paper_50208.pdf

Download (196kB) | Preview

Abstract

Investments such as venture capital, buyouts, distressed debt or assimilated, have the peculiarity of being difficult to value due to their illiquid nature on the market. The lack of transparency is determined by the market value being either determined infrequently or estimated through an "appraisal" process. Both methods of evaluation lead to the "smoothing" of returns, implying that the reported performance metrics are biased: in particular, the estimated volatility is lower than the true volatility of the investment. We propose and test two multi-index methods to evaluate performance of mezzanine, distressed debt and hedge funds, aiming at overcoming the existing gaps in the current traditional portfolio analysis. The proposed models are able to capture non-linear market effects - asset class exposure and style factor model - with the advantage of remaining in the simple framework of the linear regression analysis. We discuss the results obtained and compare them with the outcomes of a traditional regression analysis.

The paper is organized as follows: section 1 provides additional arguments to support the current methodology; section 2 presents the two types of methods used in the analysis and introduces the extensions to these models to incorporate market-lagged effects; section 3 describes the data used for the analysis, namely series of returns for three funds: mezzanine, distressed debt, and hedge fund; section 4 provides interpretation and comparisons of the results obtained when using different specifications of the first method based on the asset class exposure models whereas section 5 discusses the results obtained from different specifications of the style factor model. Finally, section 6 summarizes the main achievements.

UB_LMU-Logo
MPRA is a RePEc service hosted by
the Munich University Library in Germany.