Do patents matter for commercialisation

End-point royalties and the performance of
Australian wheat breeders
Russell Thomson
Melbourne Institute of Applied Economic and Social Research and the
Intellectual Property Research Institute of Australia
THE UNIVERSITY OF MELBOURNE
The ‘performance’ of wheat breeders
The Innovation Pathway
• Not about farmers adoption of new varieties by farmers in practice
• Not about actual commercial on-farm productivity
• Not about choice between end-point and up-front royalties
CONTEXT: WHEAT BREEDING IN AUSTRALIA
• Historically, (almost) exclusively public breeding – State government (VIC, QLD, NSW, WA) + U. Adelaide & U.Syd. • 1996 First variety attracting ‘End Point Royalties’ released Did the shift toward royalty funded breeding affect the performance of Australian
wheat breeders?
LITERATURE
Few related studies using crop trial data: Babcock and Foster (1991); Alson and Venner (2002); Kolady and Canonical estimating equation, variety i in region j.: yield = α RFB + β X + ε Value created by new variety is the yield advantage. Varieties not tested in all regions (sites) = selection bias 248 varieties (hard and soft); 23 regions; 10 breeders Yield data from scientific variety trials. • National Variety Trials & old state government sowing guides • Does the breeder ‘look like’ it undertakes royalty funded breeding – Data: The share of varieties released in the previous 5 years that attract end 1 Thanks to John Brennan, John Sheppard, Rob Wheeler, Alan Bedggood and Neale Sutton & National Variety Trials, Peter Martin, and Colin Wellings for providing data and assistance in interpretation and use. ISSUE (1)‘VALUE’ = YIELD ADVANTAGE (NOT JUST YIELD)
Issue: What do we do with negative estimated yield advantage?
Proposed solution: Transform E(advantage)
If E(advantage) < 0, still some probability that they will outperform the frontier in some describes the yield advantage of variety i at any randomly selected location: POTENTIAL ADDITIONAL OUTPUT (Q)
ISSUE (2) VALUE DEPENDS ON AREA OF ADAPTATION
Issue: 23 regions differ in size
Proposed solution: Scale the dependent variable.
New data on area of wheat cultivation in each NVT ‘region’: geo-coded regions around ISSUE (3): VARIETIES NOT TESTED IN ALL REGIONS
Solution: Type II Tobit (Heckman selection model).
• Are the breeders recent releases included in the trial region? • (Robust to other candidates and identification through non-linearity of IMR). DATA SUMMARY
Measures of performance
Explanatory variables
Exclusion restrictions
Estimator OLS
Selection
Dependent
Pot. Add. Output
Pot. Add. Output
variable:
Yield advantage
(area. Weighted)
(area. Weighted)
Selection
CONCLUSION: HOW DO ROYALTY FUNDED BREEDERS PERFORM?
• New data covering wheat trials in Australia 1978-2011 • New method for treating sub-frontier yielding varieties • Accommodating selection in data & weighting by region size • Royalty funded breeding associated with varieties included in more trials. • Royalty funded breeding associated with less ‘valuable’; i.e., lower contribution to total potential output of Australian wheat farms. THANK­YOU 

Source: http://www.aares.org.au/aares/documents/2013AC/Presentations/Thomson.pdf