Skip to content

NEWS.md

Latest commit

 

History

History
94 lines (64 loc) · 5.43 KB

NEWS.md

File metadata and controls

94 lines (64 loc) · 5.43 KB
Raw

Package update 0.5.0 -> 0.6.0

Improvements

  • Package documentation updated to document the special sentinel "_PACKAGE"

Bug fixes

  • Fix minor issues with 'weekIndex' function around end of year (thanks to Alberto-Agudo, Issue #72)

Package update 0.4.0 -> 0.5.0

Improvements

  • groupIndexes to estimate group indexes using a group variable.
  • yearOverYearIndexes to estimate year-over-year indexes.
  • predictedShareDissimilarity to estimate the predicted share relative price dissimilarity measure. This has also been incorporated into priceIndex to allow estimation of indexes using similarity linking and the new measure.
  • imputeCarryPrices to fill in missing observations using carry forward and carry backward prices.
  • parameter solveMethod to GKIndex now allows the GK equations to be solved by the iterative method. The matrix inverse method remains the default.

Bug fixes

  • priceIndex now allows estimation of elementary indexes on a dataset that does not contain quantities. Previously a quantity variable was always required, but elementary indexes do not need it (in fact it may not exist).
  • fixed issue with basePeriod parameter of priceIndex.
  • changed the way quantities and prices are calculated for the Dominicks dataset so that they represent unit prices and the associated quantity, instead of 'bundle' prices and quantities.

Package update 0.3.0 -> 0.4.0

Improvements

  • Added the Palgrave, Drobish, Stuvel, Marshall-Edgeworth, Lowe and Young bilateral indexes.
  • Added the intersection GEKS (int-GEKS) method

Bug fixes

  • Fixed matching with bilateral fixed base indexes when products leave from, and later return to the sample.

Package update 0.2.0 -> 0.3.0

Improvements

  • Added the wisp, hasp and mean_pub splicing methods. These are the same methods as window, half and mean splice, but splice onto the published series instead of the previously calculated window.
  • Added the Walsh and Jevons indexes as bilateral indexes in the GEKSIndex function.
  • Added a function for retrieving data from the Chicago Booth School of Business Dominicks Finer Foods dataset.

Bug fixes

  • Fixed issue with the time-product-dummy regression when there are missing products
  • Bug in the FBEW splice for the WTPD multilateral method
  • Fixed an issue with the half-splice splicing method

Package update 0.1.3 -> 0.2.0

Improvements

  • Added the bilateral time-product-dummy index number method. This optionally uses the Kennedy (1981) bias adjustment.
  • Added the Geary-Khamis bilateral index number method.
  • The bilateral time-product-dummy method added as a valid bilateral index to use within the GEKS multilateral index number method.
  • Added the Geary-Khamis and weighted time-product-dummy multilateral methods.
  • Added the Fixed Base Expanding Window and Fixed Base Moving Window splicing methods.
  • Added the function CESData() that calculates sample datasets using any value of sigma, assuming the data and methodology in Diewert and Fox (2017)
  • Added basePeriod parameter to priceIndex() and quantityIndex() so it is now possible to choose the base period when using output = "fixedbase" for bilateral indices.
  • Added "new" and "leaving" columns to the count matrix as part of the evaluateMatched() function output.

Bug fixes

  • Fixed an issue with evaluateMatched() incorrectly calculating matched expenditures/counts for the output = "fixedbase" option.

Package update 0.1.2 -> 0.1.3

Improvements

  • The column types of the input dataframe (for the time period, prices and quantities) are checked and converted to numeric if they aren't already.
  • Price and volume indicators can now be computed using Laspeyres, Paasche, Bennet and Montgomery methods, via the functions priceIndicator() and volumeIndicator().
  • Price and volume indicators can be used to decompose value change via the function valueDecomposition().
  • A new function, values(), can now be used to compute the values (prices*quantities) in each period, with optional matched sample. Options are available for how the matching is conducted.
  • Two new bilateral indexes are now available, the Geometric Laspeyres and the Geometric Paasche.

Bug fixes

  • The input data frame is now sorted prior to calculation. This solves an error where the products are passed to the priceIndex function in a different order in each time period, resulting in incorrectly matched product data when computing the index.
  • The output data frame of the unitValues function no longer returns factors (i.e., stringsAsFactors = FALSE), which could break subsequent calculations.

Package update 0.1.1 -> 0.1.2

Improvements

  • Performance improvements to the monthIndex function for overlapWeeks = "majority" or "wholeOnly".

Bug fixes

  • When matched sample is requested and no matches are found between the two periods, then for a bilateral index the corresponding element of the index is set to NA, and for a GEKS index the corresponding element of the bilateral comparison matrix is set to NA. A warning message tells the user which periods were set to NA.

Package update 0.1.0 -> 0.1.1

Improvements

  • Added additional argument "overlapWeeks" to monthIndex to allow flexibility in how the index is created.
  • Added more checks for correct parameter names in functions.

Bug fixes

  • Fixed bug in GEKSIndex function which affected indices computed with products that were initially in the sample, disappear for one or more periods, and then reappear later.