Library Reference AFSI-LR RM2016-2/1 Preface Docket RM2016-2

Total Page:16

File Type:pdf, Size:1020Kb

Library Reference AFSI-LR RM2016-2/1 Preface Docket RM2016-2

Docket RM2016-2 AFSI-LR RM2016-2/1 Preface

a. Purpose and content

The purpose of Library Reference AFSI-LR RM2016-2/1 is to provide the supporting materials used in the Declaration of T. Scott Thompson dated January 25, 2016.

b. Setup

The scripts found in this library reference use intermediate files from Dr. Neels’ backup material submitted as UPS-RM2016-2/LR-NP1. Dr. Neels’ Stata script “0 - Run Code.do” runs Dr. Neels’ other Stata scripts. This script contains code that deletes all of Dr. Neels’ intermediate files. Those intermediate files are needed to run the “01 Create dataset.do” script in this Library Reference.

Therefore this Library Reference includes a Stata script file titled “Modified Neels 0 - Run Code.do”. This is the same as Dr. Neels’ “0 - Run Code.do” script except that the code deleting Dr. Neels’ intermediate files has been commented out. Once the path in this script is set, it will run all of Dr. Neels’ backup scripts but will not delete the intermediate files. The path in this script should be edited for the local file system before running.1

Furthermore, before running the other scripts in this Library Reference, a Q: drive must be mapped to the “2 - Fixed Cost Regressions/Inter” folder in Dr. Neels’ backup materials, where the intermediate files are saved.

c. Organization

Library Reference AFSI-LR RM2016-2/1 consists of a zip file containing three folders, this document, and the “Modified Neels 0 - Run Code.do” script. The contents of the three folders are described in Section IV. The three folders are:

Data – This folder contains the file “neelsdata.dta”, a Stata formatted dataset that was created by combining some of the intermediate files created when running scripts in Dr. Neels’ backup materials

1 The input data used by “Modified Neels 0 – Run Code.do” are datasets FY07 Public B Cost Segments.csv, FY08 Public B Cost Segments.csv, FY09 Public B Cost Segments.csv, FY10 Public B Cost Segments.csv, FY11 Public B Cost Segments.csv, FY12 Public B Cost Segments.csv, FY13 Public B Cost Segments.csv, FY14 Public B Cost Segments.csv, FY14 Non-Public B Cost Segments.csv, FY14 Non-Public B Cost Segments.xlsx, Inflation Indices.xlsx, McBride Case1_with 2014.csv, and Reported Vols Weighted by 2014 Unit Cost.xlsx from UPS-RM2016-2/LR-NP1. These input files are not reproduced in AFSI-LR RM2016-2/1 so that this library reference can be filed publicly.

1300 Eye Street NW, Suite 600, Washington, DC 20005 main 202.408.6110 fax 202.408.7838BATESWHITE.COM Library Reference AFSI-LR RM2016-2/1 Preface Docket RM2016-2

contained in Library Reference UPS-RM2016-2/LR-NP1. These data are used for various analyses. While some of the data in Dr. Neels’ backup materials are non-public, all of the data in “neelsdata.dta” can be disclosed publicly, and so they are included in this public library reference.

Report backup – This folder contains files supporting the figures and numbers in the Declaration of T. Scott Thompson.

Scripts – This folder contains the Stata scripts that create “neelsdata.dta” from Dr. Neels’ intermediate files as well as scripts that use “neelsdata.dta” for analyses.

d. Specific contents of files

Data folder neelsdata.dta – This Stata-formatted file contains time series data of costs and weighted mail volume (as calculated by Dr. Neels), by component. It is created by the “01 Create dataset.do” script. This data contains the following fields: comp_num – The USPS component number comp_name – The USPS component name cost_seg – The USPS cost segment number year – The fiscal year wv_ – Dr. Neels’ calculated weighted volume for the (component,year) pair fc_ – Dr. Neels’ calculated reported fixed cost for the (component,year) pair ic_ – Dr. Neels’ calculated inframarginal cost for the (component,year) pair oc_ – USPS “Other” (or institutional) cost for the (component,year) pair ac_ – USPS attributable cost for the (component,year) pair tc_ – USPS total cost for the (component,year) pair neels84 – A flag indicating if the component was one of the 84 components for which Dr. Neels estimated a component level regression model

Report backup folder

Figures based on Neels table 7.xlsx – This file contains the backup material for Figures 1-4 of Dr. Thompson’s declaration based on Dr. Neels’ Table 7. The input for this file is Dr. Neels’ Table 7, which was copied into this file. The regressions were run using Microsoft Excel’s Analysis Toolkit add-in Regression feature. The Analysis Toolkit in Excel must be activated in order to reproduce the regressions. Hidden fixed costs.xlsx – This file contains the backup for Figures 5 and 7 of Dr. Thompson’s declaration. The raw data are created by the “04 Hidden fixed cost regressions.do” script and outputted to the “r.total regression” sheet. Hypothetical variable and fixed cost.xlsx – This file contains the backup for Figure 6 of Dr. Thompson’s declaration. There is no input to this file as all numbers are hypothetical. Neels component scatterplots.xlsx – This file allows for visual inspection of scatterplots for each of Dr. Neels’

Page 2 of 3 Library Reference AFSI-LR RM2016-2/1 Preface Docket RM2016-2

84 components. An orange dropdown box is used to select the component. The raw data are created by the “03 Scatterplot data.do” script and outputted to the “r.data” sheet. Pooled 69 and 70 regression.txt – This file is a Stata log containing the results of a regression of fixed costs on weighted volume pooled across components 69 and 70. It is created by the “02 Pooled 69 and 70 regression.do” script.

Scripts folder

00 Run.do – This script sets global references and runs the scripts listed below. 01 Create dataset.do – This script takes as inputs various files from “2 - Fixed Cost Regressions/Inter” of Dr. Neels’ backup material. Specifically it uses “total, attrib, fixed cost_all.dta” and filenames of the form “weighted volume_component `num'.dta” where `num' is a variable representing a component number. It merges these files together to create the “neelsdata.dta” file contained in the “Data” folder. 02 Pooled 69 and 70 regression.do – This script uses “neelsdata.dta” as an input and runs a regression of fixed cost on weighted volume for components 69 and 70 pooled together. It outputs the log file “Pooled 69 and 70 regression.txt”. 03 Scatterplot data.do – This script uses “neelsdata.dta” as an input and outputs time series data of fixed cost and weighted volume for each of the 84 components Dr. Neels uses. These data are outputted to the “Neels component scatterplots.xlsx” workbook. 04 Hidden fixed cost regressions.do – This script uses “neelsdata.dta” as an input and runs regressions patterned after those reported by Dr. Neels, but estimating instead “hidden fixed costs” in attributable costs and inframarginal costs. The script then outputs the results to the file “Hidden fixed costs.xlsx”.

Page 3 of 3

Recommended publications