returns               package:monomvn               R Documentation

_F_i_n_a_n_c_i_a_l _R_e_t_u_r_n_s _d_a_t_a _f_r_o_m _N_Y_S_E _a_n_d _A_M_E_X

_D_e_s_c_r_i_p_t_i_o_n:

     Monthly returns of common domestic stocks traded on the NYSE and
     the AMEX from April 1968 until 1998; also contains the return to
     the market

_U_s_a_g_e:

     data(returns)
     data(returns.test)
     data(market)
     data(market.test)

_F_o_r_m_a_t:

     The returns provided are collected in a 'data.frame' with 1168
     columns, and 360 rows in the case of 'returns' and 12 rows for
     'returns.test'. The columns are uniquely coded to identify the
     stock traded on NYSE or AMEX.  The market return is in two vectors
     'market' and 'market.test' of length 360 and 12, respectively

_D_e_t_a_i_l_s:

     The columns contain monthly returns of common domestic stocks
     traded on the NYSE and the AMEX from April 1968 until 1998.
     'returns' contains returns up until 1997, whereas 'returns.test'
     has the returns for 1997.  Both data sets have been cleaned in the
     following way.  All stocks have a share price greater than $5 and
     a market capitalization greater than 20% based on the size
     distribution of NYSE firms.  Stocks without completely observed
     return series in 1997 were also discarded.

     The market returns provided are essentially the monthly return on
     the S&P500 during the same period, which is highly correlated with
     the raw monthly returns weighted by their market capitalization

_S_o_u_r_c_e:

     This data is a subset of that originally used by Chan, Karceski,
     and Lakonishok (1999), and subsequently by several others; see the
     references below.  We use it as part of the 'monomvn' package as
     an example of a real world data set following a nearly monotone
     missingness pattern

_R_e_f_e_r_e_n_c_e_s:

     Louis K. Chan, Jason Karceski, and Josef Lakonishok (1999). _On
     Portfolio Optimization: Forecasting Covariances and Choosing the
     Risk Model_. The Review of Financial Studies. *12*(5), 937-974

     Ravi Jagannathan and Tongshu Ma (2003). _Risk Reduction in Large
     Portfolios: Why Imposing the Wrong Constraints Helps_. Journal of
     Finance, American Finance Association. *58*(4), 1641-1684

     Robert B. Gramacy, Joo Hee Lee, and Ricardo Silva (2008). _On
     estimating covariances between many assets with histories  of
     highly variable length_. 
      Preprint available on arXiv:0710.5837: <URL:
     http://arxiv.org/abs/0710.5837>

     <URL: http://www.statslab.cam.ac.uk/~ bobby/monomvn.html>

_S_e_e _A_l_s_o:

     'monomvn', 'bmonomvn'

_E_x_a_m_p_l_e_s:

     data(returns)

     ## investigate the monotone missingness pattern
     returns.na <- is.na(returns)
     image(1:ncol(returns), 1:nrow(returns), t(returns.na))

     ## for a portfolio balancing exercise, see
     ## the example in the bmonomvn help file

