This function estimates a corrected Dynamic Conditional Correlation GARCH model.

1 2 3 4 5 |

`inia` |
a vector of initial values in the GARCH part. |

`iniA` |
a square or diagonal matrix of initial values for the ARCH parameter matrix in the GARCH part. |

`iniB` |
a square or diagonal matrix of initial values for the GARCH parameter matrix in the GARCH part. |

`ini.dcc` |
a vector of length 2 initial parameters for the cDCC part. |

`data` |
a data frame or a matrix object containing the variables. |

`model` |
a chacacter string setting the GARCH part of the model. |

`x` |
an object of class |

`...` |
optional arguments. Currently not in use. |

This function estimates a corrected Dynamic Conditional Correlation (cDCC-) GARCH model of Aielli (2013), which is a modification of the original DCC-GARCH model of Engle (2002).

The extractor function `summary()`

is available for a `"cdcc"`

class
object displaying a table with estimates and inferencial statistics,
information criterion and some diagnostic results of the standardized
residuals. See `summary.cdcc`

for details.

Estimation of the model is carried out in two steps.
The optimization in both steps is implemented by `solnp`

function in
Rsolnp package and its outcome is save ‘as is’ in a list component
`f.stage`

and `s.stage`

, respectively.

`solnp`

uses a sequencial quadratic programming (SQP) technique
to optimize the objective function. See the manual and the references therein for
details. During the optimization, positivity and stationarity
restrictions are imposed on the parameters.

Rsonlp has a function called `gosolnp`

that generates
initial starting values randomly. A logical argument `random`

controlls which
function to use in the optimization. The default is `random = FALSE`

meaning that
`solnp`

of Rsolnp is used. If `random = TRUE`

, `gosolnp`

is used.
It should be noted that optimization takes longer time and may not always be successful
when `gosolnp`

is used.

This function returns an S3 class object `"cdcc"`

that is a list with the following components.

`f.stage` |
an output of the first stage optimization from Rsolnp. |

`s.stage` |
an output of the second stage optimization from Rsolnp. |

`model` |
a chacacter string setting the GARCH part of the model. |

`method` |
the optimization method used. |

`initial` |
a list with the initial parameter vectors/matrices. The included components are |

`data` |
the original data matrix. Returned as a zoo object. |

`CDCC` |
a matrix of the conditional correlations. Each row corresponds
to |

`h` |
a matrix of the conditional variances. Each row corresponds
to |

`z` |
a matrix of the standardized residuals. Each row corresponds
to |

Aielli, G.P. (2013),
“Dynamic Conditional Correlation: On Properties and Estimation.”
*Journal of Business and Economic Statistics*
**31**, 282–299.

Engle, R.F. (2002),
“Dynamic Conditional Correlation: A Simple Class of
Multivariate Generalized Autoregressive Conditional
Heteroskedasticity Models.”
*Journal of Business and Economic Statistics*
**20**, 339–350.

`summary.cdcc`

,
`simulateCDCC`

,
`plot.cdcc`

,
`optim`

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | ```
## Not run:
ndim <- 3
nobs <- 3000
## Setting parameters in the DCC part
a <- 0.05
b <- 0.8
## Setting a correlation matrix
Q <- diag(0, ndim)
Q[lower.tri(Q)] <- c(0.8, 0.3, 0.1)
Q <- Q + t(Q)
diag(Q) <- 1
## setting parameters in the GARCH part
a0 <- c(0.05, 0.07, 0.1)
A <- matrix(c(0.06, 0.0, 0.0, 0.07, 0.08, 0.002, 0.003, 0.0, 0.06),
ndim, ndim)
B <- matrix(c(0.75, 0.09, 0.03, 0.001, 0.81, 0.003, 0.001, 0.008, 0.84),
ndim, ndim)
## Simulating data
sim.data <- simulateCDCC(a, b, Q, a0, diag(diag(A)), diag(diag(B)), nobs)
## Estimating a cDCC model
garch_cdcc <- estimateCDCC(inia=a0, iniA=diag(diag(A)), iniB=diag(diag(B)),
ini.dcc=c(a, b), data=sim.data$eps, model="diagonal")
## Summarizing the results
summary_garch_cdcc <- summary(garch_cdcc)
## Plotting items
plot(garch_cdcc, item="correlation") # for dynamic conditional correlations
plot(garch_cdcc, item="volatility") # for volatility (square root of conditional variance)
plot(garch_cdcc, item="std.residuals") # for satandardized residuals
plot(garch_cdcc, item="return") # for return (original data)
## End(Not run)
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.