dfms (version 1.0.0)

news: News Decomposition

Description

Compute the Banbura and Modugno (2014) news decomposition of forecast updates. Given an old vintage and an updated vintage, the function decomposes the forecast revision at t.fcst into contributions from new releases.

Usage

news(object, ...)
# S3 method for dfm
news(
  object,
  comparison,
  t.fcst = nrow(object$X_imp),
  target.vars = NULL,
  series = NULL,
  standardized = FALSE,
  ...
)
# S3 method for dfm_news
print(x, digits = 4L, ...)
# S3 method for dfm_news_list
print(x, digits = 4L, ...)
# S3 method for dfm_news_list
$(x, name)
# S3 method for dfm_news_list
[[(x, i)
# S3 method for dfm_news_list
[(x, i)
# S3 method for dfm_news_list
as.data.frame(x, ...)

Value

For a single target, a dfm_news object with elements:

• y_old: old forecast for the target variable at t.fcst.

• y_new: new forecast for the target variable at t.fcst.

• news_df: data frame with one row per series and columns:

  • series: series name.

  • actual: actual release (if any).

  • forecast: old-vintage forecast of the release.

  • news: total innovation for the series on the output scale. If there is a single release, news equals actual - forecast. With multiple releases, news aggregates those innovations for the series.

  • gain: effective weight on news such that impact = news * gain (on the output scale).

  • gain_std: effective weight on the standardized innovations.

  • impact: contribution of the series to the target revision.

If target.vars selects multiple targets, a dfm_news_list object is returned, where each element is a dfm_news object and list names correspond to targets.

Arguments

object

a dfm object for the old vintage.

...

not used.

comparison

a dfm object or a new dataset for the updated vintage.

t.fcst

integer. Forecast target time index.

target.vars

Integer or character identifying target variables. Defaults to all variables.

series

optional character vector for naming variables.

standardized

logical. Return results on standardized scale?

x

an object of class 'dfm_news' or 'dfm_news_list'.

digits

integer. Number of digits to print.

name

character. Element name.

i

index. Element position or name.

Details

Let \(y_t^{old}\) and \(y_t^{new}\) be the old and new forecasts of a target series at \(t = t_{fcst}\). For each new release \(i\) (a previously missing observation that becomes observed), the innovation is $$\nu_i = x_i^{new} - \hat{x}_i^{old},$$ where \(\hat{x}_i^{old}\) is the smoothed estimate from the old vintage. The revision is decomposed as $$y_t^{new} - y_t^{old} = \sum_i g_i \nu_i,$$ with gain weights computed from Kalman smoother covariances: $$g = \sigma_y C_y P_1 P_2^{-1}.$$ Here \(\sigma_y\) is the target series standard deviation, \(C_y\) is the loading row for the target series, \(P_1\) collects cross-covariances between the target and each news item, and \(P_2\) is the covariance matrix of the news items (including measurement error where appropriate). See Section 2.3 and Appendix D in Banbura and Modugno (2014).

The function uses the system matrices and scaling from the new vintage. The old data are re-standardized to the new-vintage scale before smoothing so that innovations and gains are computed on a consistent scale. Set standardized = FALSE to report results on the original data scale.

References

Banbura, M., & Modugno, M. (2014). Maximum likelihood estimation of factor models on datasets with arbitrary pattern of missing data. Journal of Applied Econometrics, 29(1), 133-160.

See Also

dfms-package

Examples

# \donttest{
# (1) Monthly DFM example
X <- collapse::qM(BM14_M)[, BM14_Models$medium[BM14_Models$freq == "M"]]
X_old <- X
# Creating earlier vintage
X_old[nrow(X) - 1, sample(which(is.finite(X[nrow(X) - 1, ]) & is.na(X[nrow(X), ])), 5)] <- NA
X_old[nrow(X), sample(which(is.finite(X[nrow(X), ])), 5)] <- NA
# Estimating DFM
dfm <- DFM(X_old, r = 2, p = 2, em.method = "none")
# News computation (second DFM fit internally with same settings and rows)
res <- news(dfm, X, target.vars = c("ip_tot_cstr", "orders", "urx"))
# See results
print(res)
head(res$news_df)

# (2) MQ nowcast of GDP (idio.ar1 = FALSE for speed)
library(magrittr)
library(xts)
# Creating MQ dataset
BM14 <- merge(BM14_M, BM14_Q)
BM14[, BM14_Models$log_trans] %<>% log()
BM14[, BM14_Models$freq == "M"] %<>% diff()
BM14[, BM14_Models$freq == "Q"] %<>% diff(3)
X <- BM14[-1, BM14_Models$small]
quarterly.vars <- BM14_Models$series[BM14_Models$small & BM14_Models$freq == "Q"]
# Creating earlier vintage
X_old <- X
X_old[355, "ip_tot_cstr"] <- NA
X_old[355, "new_cars"] <- NA
X_old[356, "new_cars"] <- NA
X_old[356, "pms_pmi"] <- NA
X_old[356, "euro325"] <- NA
X_old[356, "capacity"] <- NA
# Estimating DFM
dfm <- DFM(X_old, r = 2, p = 2, quarterly.vars = quarterly.vars, max.missing = 1)
# News computation (second DFM fit internally with same settings and rows)
res_mq <- news(dfm, X, t.fcst = 356, target.vars = "gdp")
# See results
print(res_mq)
head(res_mq$news_df)
# }