What Statistical Data?
The Pulp and Paper Edge Strategic Review 2007 exposed some interesting statistics in both Australia and New Zealand which poses some questions and exposes weaknesses with Australian Customs, ABS, NZBS and domestic paper industry reporting.
Newsprint consumption was 663.000 tonnes down (73,000MT) due to high stock holding at Albury during mill upgrade and declining standard newsprint consumption.

Uncoated mechanicals 214.000 tonnes
Directories 52.000 tonnes. A total of 930,000 tonnes. 
Coated Mechanical consumption was 327.800 tonnes with coated woodfree 392.800.  A total of 720.600 tonnes. With uncoated woodfree 431.000 tonnes.  A grand total for Printing and Communication papers of 1,415,900 tonnes (imports 69.4%), with newsprint added back in then 2,078,900 tonnes somewhat higher than the Packaging papers and Paperboard of 1,546,400 tonnes.
The question remains as to what volume was consumed at home and how much at work (C & I).
With the absence of imports and exports of Stationery and Printed Matter data arriving at any accurate consumption figures remains far too inaccurate so how can anyone guesstimate recovery and how many people are employed doing just this.
The New Zealand figures better illustrate the contradictions:
Consumption of Newsprint 130,000 tonnes
Newsprint reported recovery 118,900 tonnes
Percentage recovered 91.5%

P & W Papers consumption 221,800 tonnes
Reported recovered 36,600 tonnes
Percentage recovered 17.2%
Obviously neither are correct by approximately 20% and make a mockery of industry data quoted to Governments and for parts of the recovery industry for them to make informed investment decisions.
Accurate data from single stream waste and mixed paper and cardboard collection is misleading as it is often a total temporarily diverted gross figure and not a nett figure utilised for recycling or secondary processing.  Governments and MRF owners do not mind this as there is no accountability for true end of life costs so why do we persist with capturing and publishing data if it is misleading.