Abstract
In this paper, we investigate empirically the effect of using higher moments in portfolio allocation when parametric and nonparametric models are used. The nonparametric model considered in this paper is the sample approach; the parametric model is constructed assuming multivariate variance gamma (MVG) joint distribution for asset returns.We consider the MVG models proposed by Madan and Seneta (1990), Semeraro (2008) and Wang (2009). We perform an out-of-sample analysis comparing the optimal portfolios obtained using the MVG models and the sample approach. Our portfolio is composed of 18 assets selected from the S&P500 Index and the dataset consists of daily returns observed from 01/04/2000 to 01/09/2011.
Similar content being viewed by others
Notes
The variance gamma model is implemented in Bloomberg.
Daily portfolio returns are given by:
$$\begin{aligned} R_{p}^{i}(t)=\sum _{j=1}^{N}w_{j,i}R_{j}^{i}(t). \end{aligned}$$where \(i\) indicates the \(i\mathrm{th}\) window (in or out-of-sample), \(j\) is the \(j\mathrm{th}\) asset in portfolio, \(w_{j,i}\) is the weight of the \(j\mathrm{th}\) asset obtained in the \(i\mathrm{th}\) sample period and \(R_{j}^{i}(t)\) is the daily return of asset \(j\) observed at time \(t\) in the \(i\mathrm{th}\) out-of-sample period.
The compound interest formula is given by:
$$\begin{aligned} r_{p}^{i}(t)=\left(1+R_{p}^{i}(t)\right)^{252}-1. \end{aligned}$$Portfolio A dominates B according to the mean–variance–skewness–kurtosis criterion if \(E(A)\ge E(B)\), \(\mathrm{Var}(A)\le \mathrm{Var}(B)\), \(\mathrm{Skew}(A)\ge \mathrm{Skew}(B)\) and \(\mathrm{Kurt}(A)\le \mathrm{Kurt}(B)\). At least one of these inequalities must be strict.
References
Ang, A., Bekaert, G.: International asset allocation with regime shifts. Rev. Financial Stud. 15(4), 1137–1187 (2002)
Athayde, G., Flores, R.G.: The portfolio frontier with higher moments: the undiscovered country. Computing in Economics and Finance 2002 209, Society for Computational Economics (2002)
Athayde, G., Flores, R.G.: On certain geometric aspects of portfolio optimisation with higher moments. In: Jurczenko J.F., Maillet, B.B. (eds.) Multi-Moment Asset Allocation and Pricing Model, pp. 37–50. Wiley, New York (2006)
Bertini, C., Lozza, S.O., Staino, A.: Discrete time portfolio selection with lévy processes. In: Proceedings of the 8th International Conference on Intelligent Data Engineering and Automated Learning, , IDEAL’07, pp. 1032–1041. Springer, Berlin (2007)
Billingsley, P.: Probability and Measure. Wiley, New York (1995)
Chan, L.K.C., Karceski, J., Lakonishok, J.: On portfolio optimization: forecasting covariances and choosing the risk model. Rev. Financial Stud. 12(5), 937–974 (1999)
Cont, R., Tankov, P.: Financial Modelling with Jump Processes. Chapman & Hall/CRC Financial Mathematics Series, London (2003)
Desmoulins-Lebeault, F.: Gram-charlier expansions and portfolio selection in non-gaussian universes. In: Jurczenko, J.F., Maillet, B.B. (eds.) Multi-Moment Asset Allocation and Pricing Model, pp. 79–112. Wiley, New York (2006)
Hitaj, A., Martellini, L., Zambruno, G.: Optimal hedge fund allocation with improved estimates for coskewness and cokurtosis parameters. J. Altern. Invest. 14(3), 6–16 (2012)
Jean, W.H.: More on multidimensional portfolio analysis. J. Financial Quant. Anal. 8(03), 475–490 (1973)
Johnson, N.L., Kotz, S.: Continuous Univariate Distributions, vol. 1. Wiley Series in Probability and Statistics, New York (1994)
Jondeau, E., Rockinger, M.: Optimal portfolio allocation under higher moments. Eur. Financial Manag. 12(01), 29–55 (2006)
Jondeau, E., Poon, S.H., Rockinger, M.: Financial Modeling Under Non-Gaussian Distributions. Springer Finance, Springer, Berlin (2007)
Kassberger, S., Kiesel, R.: A fully parametric approach to return modelling and risk management of hedge funds. Financial Mark. Portfolio Manag. 20(4), 472–491 (2006)
Loistl, O.: The erroneous approximation of expected utility by means of a taylor’s series expansion: analytic and computational results. Am. Econ. Rev. 66(5), 904–910 (1976)
Loregian, A., Mercuri, L., Rroji, E.: Approximation of the variance gamma model with a finite mixture of normals. Stat. Probab. Lett. 82(2), 217–224 (2011)
Madan, D.B., Seneta, E.: The variance-gamma (v. g.) model for share market returns. J. Business 63(4), 511–524 (1990)
Madan, D.B., Yen, J.J. Chapter 23 asset allocation with multivariate non-gaussian returns. In: Birge, J.R., Linetsky, V. (eds.) Financial Engineering, Handbooks in Operations Research and Management Science, vol. 15, pp. 949–969. Elsevier, Amsterdam (2007)
Markowitz, H.: Portfolio selection. J. Finance 7(1), 77–91 (1952)
Martellini, L., Ziemann, V.: Improved estimates of higher-order comoments and implications for portfolio selection. Rev. Financial Stud. 23(4), 1467–1502 (2010)
Mendes, B., Semeraro, M., Leal, R.: Pair-copulas modeling in finance. Financial Mark. Portfolio Manag. 24(2), 193–213 (2010)
Meucci, A.: Risk and Asset Allocation. Springer Finance, Springer, Berlin (2005)
Rubinstein, M.E.: The fundamental theorem of parameter-preference security valuation. J. Financial Quantit. Anal. 8(01), 61–69 (1973)
Semeraro, P.: A multivariate variance gamma model for financial applications. Int. J. Theor. Appl. Finance (IJTAF) 11(01), 1–18 (2008)
Wang, J.: The multivariate variance gamma process and its applications in multi-asset option pricing. PhD thesis, Faculty of the Graduate School of the University of Maryland (2009)
Acknowledgments
The authors thank Markus Schmid and the anonymous referee for helpful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Appendices
A. Appendix
The characteristic function (c.f.) \(\varphi (t)\) for N-dimensional random vector \(X\) is defined as:
where \(t \in R^{N}\).
For all MVG models, the expected CARA utility can be connected with the characteristic function as follows:
In the following, the c.f. for each MVG model is derived.
1.1 MVG with common mixing density
For the MVG with common mixing density the c.f. \(\varphi _{MVG}(t)\) can be easily obtained as follows:
Using the law of total expectation (see Billingsley 1995), we get:
since \(Z_{h}\) are independent standard normals, we have:
Being V a univariate gamma, the c.f. is
1.2 Semeraro model
For the Semeraro model, using our parameterization, the c.f. can be obtained as follows:
Since \(W_{p}\) are independent standard normals, we have:
substituting \(G_{p}\) we can write:
Being \(Y_{p}\) and \(Z\) independent for \(p=1,\ldots ,N\), we obtain:
Since \(\frac{a_{p}}{k}=\frac{1}{m_{p}}\), the c.f. of Semeraro model is
1.3 Wang model
In Wang model, the c.f. is
Since \(A_{p}\), \(Y_{p}\) are independent for \(p=1,\ldots ,N\), we have:
Being \(A_{p}\) the pth component of the MVG with common mixing density and \(Y_{p}\) an univariate variance gamma, we can write:
B. Appendix
The general formula of co-skewness is
The general formula of co-kurtosis is
1.1 MVG with common mixing density
We derive the co-skewness of the MVG with common mixing density using the definition (29) and we have:
where \(Y_{i}=\sum _{h=1}^{i}a_{ih}Z_{h}\) and \(Y_{i}\sim {N(0,\sigma ^{2}_{i})}.\) Developing the equation, we have the following elements and their permutations w.r.t. indices \((i,j,k)\):
and the single final terms:
The term \(E\left[Y_{i}Y_{j}Y_{k}\right]=0\) since it is the coskeness of a multivariate Normal with zero mean.
Putting together these terms into equation(31) we have the element in (11) Starting from definition (30) we derive the co-kurtosis as follows:
Following the previous procedure we obtain the elements and their permutations:
The relation \(E\left[(V-E(V))^{2}V\right]=\alpha ^2+2\alpha \) comes from the property of the gamma distribution, see Johnson and Kotz (1994).
and the single final terms are given by:
where \(E\left[Y_{i}Y_{j} Y_{k} Y_{l}\right]=3\sum _{h=1}^{min(i,j,k,l)}a_{ih}a_{jh}a_{kh}a_{lh}\) since it the co-kurtosis of the multivariate Normal with mean zero.
Putting together the elements into (32) we obtain the co-kurtosis elements given in (12).
1.2 Semeraro model
From the definition of skewness (29), we obtain:
Developing the equation we obtain the following elements:
using the independence between \(Y_{i},\) \(Y_{j},\) \(Y_{k},\) \(Z\) and the property of gamma (see Johnson and Kotz 1994), we have:
since \(W_{k}\) is a standard normal independent from \(G_{i},\) \(G_{j}\) and \(G_{k}.\)
we observe that if \(i\ne j= k\) or \( i= j= k \):
otherwise the \(\mathrm{term}^{c}_{ijk}=0.\)
The last term:
Using these terms, we obtain the element of the co-skewness as reported in (19).
As above, we have the following terms:
using the definition of \(G_{i}=Y_{i}+\frac{k}{m_{i}}Z\), we can write:
Applying the properties of gamma distribution and the independence between \(Y_{i},\) \(Y_{j},\) \(Y_{k},\) \(Y_{l}\) and \(Z\) we obtain:
The following term and its permutations are always zero
since \(W\) is independent from \(G\).
The next term (with its permutation) is given by:
This term is zero except when \(k=l\)
The following term and its permutation are zero:
since \(E[W_{j}W_{k}W_{l}]=0\) for all possible combination of indeces.
The last term is
and we have the following cases:
With these terms, we obtain the elements of co-kurtosis as reported in (20).
1.3 Wang model
The co-skewness and co-kurtosis elements in Wang model can be easily obtained (using the same procedure as previously), since the Wang model is the sum of the MVG with common mixing density and an independent component with VG distribution.
C. Appendix
Rights and permissions
About this article
Cite this article
Hitaj, A., Mercuri, L. Portfolio allocation using multivariate variance gamma models. Financ Mark Portf Manag 27, 65–99 (2013). https://doi.org/10.1007/s11408-012-0202-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11408-012-0202-5