2017-01-17 Meeting notes
- Dennis Wisnosky
Date
Attendees
Agenda:
Proceedings:
20170117 FIBO DER FCT/CFTC PoC
Meeting starting with a side discussion of how to use deep learning along with the DER ontology to do scalable reasoning. This is based on research at RPI. So far the network is learning RDFS entailments. Could this be used to learn more elaborate FIBO structures? JB wonders if we could really use this for Swap validation; more for classification. On to the agenda.
Open Jira Issues: FCP = FIBO CFTC PoC: https://jira.edmcouncil.org/projects/FCP/issues/FCP-1?filter=allopenissues
Also in DER: https://jira.edmcouncil.org/projects/DER/issues FCP is specific to the CFTC stuff. Both are relevant to this meeting.
EK: there are some fixes in GitHub (FND that fixes ContractTerms which is now Commitment). Deprecated some terms. Pull from EDMC/Pink Securities. Issuance Terms changes are underway, exempt offerings. Current situation is that the other chapters (e.g., DER) still refer to deprecated terms. They will change them to refer to the new ones. DN confirms in Protégé the current reference to deprecated terms (this is what Elisa is working on). BE Partnerships refers to this stuff, and will continue to point to a deprecated terms (TermSet). The issue has to do with updating all of the other domains based on these changes. We'll put others into OMG in March. BE will (probably) not make it until June. EK: The commitment is more important than the terms (according to Bruce Kendall). EK: Debt Terms, Standardized Terms, Settlement Terms are all in FBC. hasTerms is being deprecated to hasContractualElement (maybe should be hasCommitment).
When is something a term in a contract (e.g., delivery terms), and when is it a formal commitment? This is something that the SMEs in the FCTs need to discuss. XXX TermSet lingo seems uniformly disliked. It is being turned into various other terms. The Jira issue that caused this discussion is FND-84. We need to link this DER-14.
EK: SEC-7AssetClassAssetClass added (in comment) of SEC-7. Does CFTC have their own definition of AssetClass? What do we do with AssetClass? Do we pun? DA is against this. We should have a special property. EK: e.g., "represents" "refers to "get the list from somewhere "denotes" "defines". Find the right one "has definition" vs. (e.g.,) skos:defintion"has defintion" comes from FND, refers to "a defining concept" E.g., A classification scheme defines a classifier. Check out "hasDenotion" (better definition of that". Annotation for hasDenotion and hasDefintion is controversial. Nowlin: CFTC asset classes: http://www.cftc.gov/MarketReports/SwapsReports/DataDictionary/index.htm. DA feels that punning is bad practice and should be written up as a recommendation / pattern in FIBO(not a topic for this meeting - maybe FLT? Or discussion with DA/EK/MU?).
Move to spreadsheet from Wes: We will walk through these elements, find them in FIBO. If they don't exist, create them. JN: This came out in 2012, these are the primary economic terms that the CFTC says is the minimum that we need to evaluate swaps. 40-odd fields in CFR Part 45Last year they put together a list of 120 proposed fieldsin the FR notice for comment, not accepted. Industry's response is pending.
PET (Primary Economic Terms)These are used in all of the CFTC's calculations. Wes: we went through the appx and tracked concepts/terms, and then found good definitions for them.
Nowlin: http://www.cftc.gov/idc/groups/public/@newsroom/documents/file/specificationsswapdata122215.pdf
Definitions come from various sources as shown in Column C17 CFR part 45 appx. 1. Fed Register 77 vol 9, 2112. CFTC was going to say, "these are the PET that we need". Left it open enough so that the four SDRs ("Swap Data Repository") give them enough leeway to be able to provide it by their back-end technology. If you look at the FR notice, it says what they meant about the PET. Every Swap Data Repository does it a bit different. But this is what they must report. It would have been nice to be more specific, e.g., an FPML tag. But they couldn't do that at the time (political forces, multiple standards). They did it in a broad statement. All the CFTC did was to say that here is a field and here is a description. JB: is it safe to say, "Asset Class" definition reflects a market context? JN: We regulate five kinds of swaps. JB: This definition is limited to CFTC scope, not intended to be general?
WM: In the appx, it just says, "AssetClass" is a category, comment says "values are ..." (gives a list). So in the spreadsheet, they tracked down a definition. CFTC definitions are specific to US and FAR (that was EK.) EK is suggesting a module for the US to specify details of AssetClass referring to the FAR.
What is the priority of this? We'll see. Now we are working through the spreadsheet- agent. EK: That maps to two things. ThirdPartyAgent (other is not added yet; Responsible Party). Two things is a contract third party, an organization that is a third party to an offering. acting in an agency capacity. other is a "registered agent". This is natural person vs. organization. Yet to be named, likely to be "RegisteredAgentChild of "RelationshipManager", "Responsible Party". WM: There is a concept of "agency" in the law. The firm may act as an agent (e.g., broker). The individual who is taking the trade is also an agent. At top level, we have AutonomousAgent. here is "Agency" Role. DN: There is a more general notion of agent in the legal sense, that isn't as general as AutonomousAgent. EK: in FND, we have a Contract Third Party, not specifically human or organization, indirectly involved to a transaction, etc. ContractThirdParty is the general notion of agent that is the legal one. In SEC, Agent for process. Agent for Service of Process vs. agent representing. WM: One of those is a special case of the other.
DN: Financial Service Provider is an agent, a subclass of a generic legally empowered agent. EK: Some financial service providers are acting as an agent, some are acting on their own behalf. DN: If you need an agent for a financial role, but is a subclass of financial service provider and agent. EK: that's what we did, subclass of three things. EK: The definition in the spreadsheet is the def of Contract Third Party – that’s what that is.
EK: make 3rdpartyAgent and registeringAgent children of a single agent, to show the general notion of legal agency. If we do this, then row 2 will correspond to that new common superclass. Nowlin: The term “registered entity” includes SEFs and DCMs. See CEA sections 1a(40) (A), (C) and (D), 7 U.S.C. 1a(40) (A), (C) and (D). The Commission stated in the RTR Adopting Release that SEFS and DCMs may enter into a contractual relationship with a third party service provider to transmit swap transaction and pricing data to an SDR; however, the SF or DCM will remain responsible for such reporting requirement pursuant to part 43. RTR Adopting Release at 1201.Footnote 23 from the draft tech spec for swap spec.
http://www.cftc.gov/idc/groups/public/@newsroom/documents/file/specificationsswapdata122215.pdf Footnote 23FBC-106. Footnote 23 is on p. 13. JN: Keep in mind that these are CFTC requests, we need to look at the whole industry. CFTC is intentionally not overly prescriptive. There is tug-of-war FXML and FPML. They are trying to work with both. CFTC is not saying that this is the last word. Not concrete, more like a "slurry" :)
Jeff Braswell: The FIX / FpML 'duality' was also dealt with by the SEC in their securities-based swap data definitions by saying that they would accept either :)
AllocatedSwap? WM: These happen later in the day on the books of the swaps dealer. JN: Different types of allocations" give ups"...post-trade they allocate it. WM: The key is that when you identify it as allocated, this isn't current; the time of the trade isn't at market time, later in the day normally on the books of the swap dealer, not info about the swap market at that time. Part of the swap lifecycle