Welcome to Dr Lincoln's blog

Welcome for visiting my blog. Hope you enjoy the visit and always welcome back again. Have a nice day!


Don't correct errors by making errors deliberately

Comments on John Daley, Chief Executive Officer, Grattan Institute and Danielle Wood,

Fellow, Australian Perspectives, Grattan Institute: “Why biased budget forecasts make poor politics”, 5/05/2017

While budget forecast biases are concerning, the authors made like it has been the making of politicians or the Treasurer, as opposed to be a product of the Treasury which are bureaucrats. Unless the authors implied that there has been a conspiracy between the Treasurers and the Treasury staff, the kind of confusing generated by the authors are quite unhelpful in addressing the problems they say wanting to be addressed. Maybe the authors think that the Treasurers should force the Treasuries to make unbiased forecasts, or take some punishing measures if biases are consistent!

Besides, proposing another approach, that is implied by the mentioned conservative approach, is hardly helpful, because that is purposefully making wrong forecasts, as opposed to unbiased forecasts that should be aimed and achieved. Why do you need to do that, purely to cheat yourselves?

Perhaps, the authors of the post should consider having a more independent assessment of Treasury forecasts to see if there have been systematic biases existing, even though there was one assessment done that I suspect may not necessarily completely independent of Treasury if it was commissioned by the latter, or at least some may argue that its independence may have been compromised to some degree given its funding/commission source.

On a personal note, I do have some suspect that it seems there were some systematic biases as the authors pointed out that it was under-estimation and over-estimation of revenues year after year during the Costello and the period followed that respectively. But it needs to be further tested using a robust methodology.

No comments:

Post a Comment