[arin-ppml] Dynamics of a transfer environment without operational need assessment

[Mike Burns]

Hi John,

 

Given that all of our experience has been with needs-based transfer policies (which 

provide some back pressure to speculation, whether via the direct prohibition that is

implied or the convolutions that is necessary to work around same), it is rather unclear 

if financial speculation would have occurred over the past five years in absence of the

needs-based assessment model.   

 

On the contrary we now have a good deal of experience with the most vibrant and active market, which has no needs test. And the absence of speculation grows clearer every day in the RIPE region. I see more and smaller transfers happening over there, but to be sure there might be some noise in the data related to their final /8 policy. Still I don’t see evidence that any single agency is buying up addresses for speculation, and personally I have not been contacted by any speculator issuing a blanket buy order, which is what would be required to begin assembling a hoard of addresses. As a daily participant in this market I have to shake my head at any reasonable belief that the needs test is what is holding back speculation. We all know that speculation can occur already in many forms, and there is a back door to avoid the needs-test anyway. What is this paper shield of a needs-test that can hold back these mighty and insidious speculators?  Yet we argue over a 30 day utilization period that everybody knows is and has been a no-op?  Do we really believe it’s the utilization requirement and not the price that buyers are concerned with?  It’s silly, IMO.  Drop the needs test and finally drop these sorts of discussions. 

 

What’s more, ARIN seems to be depicted poorly in other regions as having failed to create a reserve pool for new entrants, like they have. They think the ARIN region is in a bad state, particularly for new entrants. What we should be doing is demonstrating the vitality of our market, which would enable new entrants to participate through the reliable purchase of IP addresses just like they purchase equipment, man-hours, and every other component of their new business.

Eliminating barriers to trade will help to create a vibrant and reliable market to benefit new and old companies, and demonstrate to the world that a reserve pool for new entrants is just another way of keeping addresses fallow for years, whereas the ARIN policy provides the incentive for trade that naturally brings addresses into productive use.

 

 

In an early message, you characterize it as such - 

 

Policy in ARIN has been held hostage for years to the boogeyman of speculators capable of harmfully manipulating the market.  Never has evidence supporting the existence of these people been offered. And certain unavoidable facts are avoided:

1.       Any such speculation will drive IPv6 and could render IPv4 worthless

2.       Supply has been fragmented through a quarter century of worldwide RIR allocations

3.       Every policy-compliant transfer is publicly recorded

4.       Prices have not changed dramatically in five years of trading

 

These conditions make it all but impossible for any market manipulation to occur, because due to fragmentation, the acquisition of enough blocks to manipulate the market would take years, in my educated opinion, and would have to involve a very rich speculator who was willing to engage in a series of transfers under assumed names, with the speculated asset marching towards a zero value.  It’s ludicrous to believe that any such speculator exists, willing to risk not only the loss of his investments, but worldwide opprobrium.

 

While the circumstances you describe does inhibit speculation, it is unclear if it they

would still be latent in a transfer market environment which allowed and effectively 

endorsed speculation.  For example, it is unclear if a short-term market squeeze in

IPv4 (e.g. 3 months) would materially impact deployment of IPv6 (which has many

other factors that businesses face in making such decisions.) 

 

There is no way for a bad actor to create any squeeze in the market due to the inability to acquire enough space to affect price. Remember as price rises, so does available supply, since CGN provide some supply elasticity and there are still hundreds of millions of un-routed addresses*. But granted IPv6 deployment is not just a market-squeeze away, whether it’s short or long-term. Still it’s the Sword of Damocles to a market manipulator and provides a huge disincentive for what would have to be a yearslong project for that would-be manipulator. *Yes I know un-routed doesn’t mean unused, but c’mon we’re talking 20% of the universe of ipv4 space!

 

If it is unclear that an IPv4 squeeze would materially impact IPv6 deployment, it is at least as unclear that speculators would be a detriment to the market, even if they did exist.  This community is one of network operators primarily. Maybe some education on the role of speculation in a market would be appropriate, since the fear of speculation seems always to be at the root of objections to removing the needs test.  

 

 Similarly, we’ve seen

greatly accelerated recovery of unused IPv4 blocks since the IPv4 free pool runout

in the ARIN region in September 2015 (effectively describing fragmentation of the

marketplace over time) and absence of operational need could bring many more 

actors into seeking underutilized IPv4 number resources blocks.  If the community

decides to make IPv4 address block rights transferable to any party, regardless of

operational need, then they are likely to be considered an investable asset class

and analyzed very differently by financial firms going forward.  Whether that would 

be a good outcome (or not( is a matter of judgement regarding the functioning of 

markets and resulting impact/benefit to this community.

 

Yes, maybe we should talk about the benefits of aligning policy with the real-world treatment of valuable assets.

Because in the real-world the general analysis that every business applies is really the same (largely Net Present Value), the only difference is their own judgements of things like how long IPv4 will last, how much CGN the Internet can take, as well as normal factors like their growth rates and risk aversion.  Why not let them use their own determination of these things in their decision to purchase addresses?  Must they subordinate their decisions to ARIN policy designed in a free-pool environment?

 

 

In any case, the proposed policy 2015-3 does not directly touch on these merits 

of needs-based transfers or otherwise, as it describes removal of a criteria whose 

definition is unclear to operational need assessment for transfers (and thus not 

operational today), and thus my reason for moving this more general discussion 

of a transfer market without operational need-assessment to a separate thread 

and subject.

 

That’s a good idea for anybody who wants to pursue the general discussion. I’m reasonably sure Professor Mueller would be willing to provide relevant economic theory and maybe I can add something from my perspective with direct knowledge of hundreds of transfers around the world. I know there are other listening brokers who can challenge or support my observations.  Personally I am not aware of any broker who supports the continuance of the needs test as a means to conserve space.

 

Maybe the community should avail themselves of the information that only brokers have. Put a handful of brokers together and you have information covering a thousand transfers.  What other handful of community members has experienced a thousand transfers/allocations?  Probably only ARIN staff and David Huberman. ;-)  If we brokers speak with a unified voice about our experience with (the lack of) evidence of speculation, perhaps the community will credit our perspective.

 

Regards,
Mike

 

 

Thanks!

/John

 

John Curran

President and CEO

ARIN

 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20160519/5f8deefb/attachment.htm>