[arin-ppml] Spartan allocation proposal for transfers

[Owen DeLong]

On Oct 10, 2011, at 3:38 PM, Ben Edelman wrote:

> Thanks for your interest.  A couple quick thoughts --
>  
> As John says, this paper is not work ARIN requested of me; it's a paper I wrote on my own.  Incidentally I wrote an earlier paper on v4 scarcity, available here: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1337075, also independent of my work with ARIN. 
>  
> On the complexity of the paper and mathematical approach: Indeed.  The economics profession typically uses a certain formality, which is sometimes helpful and sometimes a distraction.  I write in that genre in order to fit the norms of those who expect it.  But I have a healthy skepticism for how useful it is.  At the very least, I'll always do my best to help clarify wherever folks flag the problem.
>  
> Owen wrote: "I do find the assumption that the price of IPv4 addresses declines post exhaustion rather startling vs. the reality that everyone expects IPv4 addressing prices will increase until IPv4 is obsoleted." 
>  
> My intuition here: If a seller thought the price was about to jump, why would the seller sell?  Perhaps the seller badly needs the money, but then the seller might be better off getting a bank loan and repaying that loan with the IP addresses to be sold at a later date.  Clearly not everyone can get loans, for various important reasons.  But if most folks expect prices to rise, it's not clear why anyone would sell.  That's the basic intuition for why prices shouldn't rise slowly over time, but rather should quickly rise to a level that reflects expectations of future value.
>  

I can see that might be the case, but, I think that this ignores the potential for a feedback loop where the rapid rise creates new expectations among those that haven't freed up their resources yet and they come to the market later with even higher expectations.

> Of course if the v6 transition goes more slowly than expected -- bearing in mind that it may already be expected to take quite some time -- then demand for v4 may exceed prior expectations, and prices may increase further. 

Which, in spite of my level of enthusiasm and desire for expediting IPv6, is, IMHO, a very likely scenario.

>  
> Mike, in prior discussion with ARIN members and others, I did occasionally hear casual mention of approaches that sound like the spartan allocation rule.  For example, some suggested that if a seller is selling an entire allocation, it doesn't matter whether the buyer needs to buy from several such sellers because, on net, no new routes are being created.  Of course there might still be a downside of permitting those transactions, e.g. if that buyer would previously have bought just one block, but now ends up with several, and then some big sellers disaggregates to satisfy several small buyers.  Our paper is a step towards evaluating the pros and cons of policies in this area.

That's true, but, what is of concern here isn't the buyer buying multiple blocks to meet a larger need so much as the following (likely) scenarios:

1.	A buyer with a large need buys portions of several other blocks.
2.	Several buyers with smaller needs are willing to pay more for smaller pieces
	of a given block than the smaller number of buyers with larger needs.

>  
> I'll be in Philadelphia beginning Wednesday evening (unfortunately can't make it before then).  I read this list and am always reachable by email .

Cool. I don't know whether there is any time available in the ARIN agenda, but, could you at least consider discussing this at the Open Policy Hour if not elsewhere in the agenda?

Owen

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20111010/ae7ebced/attachment.htm>