Quote:
Originally Posted by Hrafn
No, they do not have "a whole lot more information" they have information that is a whole lot more aggregated, and therefore is a whole lot less useful for estimating elasticities. Is the fact that we sold less books this holiday season for higher unit prices because of the price, or because of an economic downturn? Is the fact that people are buying more $20 Harry Potter books than $25 Stephen King books because they're cheaper, or because people like Harry Potter more?
Statistically, heavily aggregated data means very few 'degrees of freedom' and a model that is nowhere near robust, because so many contributory factors end up being conflated.
I know all this because (i) I've got a degree in Statistics & (ii) because I've actually worked in the field of Regulatory Economics/Competition Law, including on calculating price elasticities. |
There is continuing disparagement of Amazon's figures on this forum. Some of it is valid, including your own criticism of "heavily aggregated data". This, of course, can only be put to rest by Amazon releasing the detailed data which has been aggregated. This is not going to happen any more than Hachette or the other BPH will release their detailed data. In fact, what Amazon has released, aggregated as it is, is far more than I have seen on this topic from any member of the BPH.
Your criticism is of aggregated data in general. Recongnising that we are not going to get the detailed data you demand, I would have thought that the most productive course was to attempt to evaluate the data we do have. If it is that bad, I would expect that there would be some data which contradicted it in major respects. If there is, I have not yet seen it.
Yet we have in this topic seen many efforts to evaluate the efficacy of the information. These include comparable data from Smashwords, speculation on Amazon's reasons for lying and for taking the risk of the consequences thereof, and importantly, the very fact that Amazon's business practices accord with its position. They did prefer the $9.99 price point, just as they profess. If they make more money at $14.99, why on earth would they not set that price? If any of the BPH wish to claim that they suffer at $9.99 they are at liberty to provide their own data and indeed their own explanation. But looking at what explanatins we have seen so far are troubling. Retarding the EBook market in an attempt to bolster their preferred print book market. Subsidising loss-making publications which are nevertheless required to be published for the Public Good? Encouraging the local industry. And avoiding cheapening the image of books in the public mind. Really?
Using what seems to be the best data available, and evaluating that data against the real world and the relevant explanations provided seems to me to favour Amazon's position. All the criticisms of methedologies and errors that they might lead to are moot. That an error or errors might have occurred does not mean that such error or errors did occur, particularly in these circumstances. Please look to the particular case.