scientific article; zbMATH DE number 1474998

From MaRDI portal
Publication:4489592





Based on the discrete-time system control theory, this paper analyses the risk sensitivity in single-developer real estate investment process: \[ \left[{{x_{1,k+1}}\atop {x_{2,k+1}}} \right]= \left[ \begin{matrix} 1 &0\\ 0&\alpha \end{matrix}\right] \left[ {{x_{1k}} \atop {x_{2k}}}\right]+ \left[ {{\nu_{1k}} \atop {\nu_{2k}}} \right]+ \left[ \begin{matrix} b_{11} &b_{12}\\ b_{21} &b_{21} \end{matrix}\right] \left[ {{y_{1k}} \atop {y_{2k}}} \right], \] where \([x_{1k}, x_{2k}]^T= {\mathbf X}_k\) represents the self-own and lent capitals in the \(k\)th year, \([\nu_{1k}, \nu_{2k}]^T= {\mathbf V}_k\) and \([y_{1k}, y_{2k}]^T= {\mathbf Y}_k\) are respectively the control variables and risk variables. The profit \(U_k\) in the \(k\)th year is \[ U_k= [c_1,c_2]{\mathbf X}_k+ [d_1,d_2]{\mathbf V}_k+fZ_k, \] where \(c_1\) and \(c_2\) are respectively the profit rates of self-own and lent capitals, and \(f\) is the risk coefficient of profit. Assuming \({\mathbf Y}_k\) and \(Z_k\) are uncorrelated, the minimum mean-square-error estimation of \({\mathbf X}_k\) can be given by the Kalman filter: \[ \widehat{\mathbf X}_{k+1}= \text{diag} [1,\alpha] \widehat{\mathbf X}_k+ {\mathbf V}_k+ M_k (U_k- \widehat{U}_k), \qquad \widehat{U}_k= [c_1,c_2] \widehat{\mathbf X}_k+ [d_1,d_2]{\mathbf V}_k, \] where \(M_k= [m_{1k}, m_{2k}]^T\) is the time-varying gain matrix. When choosing exponential state feedback as \[ {\mathbf V}_k= \text{diag} [a_1^k,a_2^k] \widehat{\mathbf X}_k, \quad 1> a_1,a_2\geq 0, \] the sensitivity of profit \(\widehat{U}_k\) to risk \((U_k-\widehat{U}_k)\) is the norm of a matrix \(G\), i.e. \[ a=\| G\|, \quad G= [c_1I+ C_{a_1^{-1}} d_1, c_2I+ C_{a_26{-1}} d_2] (T_1\oplus T_2)^{-1} \left[ {{m_1I} \atop {m_2I}} \right], \] where \(T_i\), \(i=1,2\) have diagonal components \(-(1+a_i^{-j})\), \(j=1,2,\dots, N\) and \(C_{a_i^{-1}}\), \(i=1,2\) are the composition operators defined by \(C_{a_i^{-1}} X(z)= X(a_i^{-1}z)\), where \(X(z)\) is the \(z\)-transform of \({\mathbf X}_k\). The result shows that the risk sensitivity \(a\) can suitably describe the influence of a risk factor to the profit. It depends not only on the investment control coefficients \(a_1\) and \(a_2\), but also on the investment period \(N\). To decrease ris of loss, the investment should be controlled such that the risk sensitivity is suitably reduced.












This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4489592)