Well we know this isn't true, as reporting by furnishers varies between CRAs. Secondly, its not the CRAs that have ever used FICO. 3. FICO 08 could bring legal action against Fair Isaac. Anybody know what he's talking about? We have all been so focused on the AU aspect... Well even the non-experts have been saying this...
Interesting read. We are still operating on second- or third-hand information, and the views of a pundit. It is what F-I themselves say that is important, not what some who report what they say. Primary source materials are always the most valuable.
The new model has 12 population segments as opposed to 10 in the old models. 8 segments for consumers with good credit, and 4 for those with bad credit. The theory is that more segments will make the model more accurate. This could result in a consumer losing only a few points, but may put them in a whole new segment causing them higher interest rates, etc.
No one is arguing, I think, that the model won't be different, and that different credit profiles may see adjustments (perhaps major) to their scores. What one might see as discrimination may be viewed as accuracy on from the other. Straw men can always be blown away by strong winds. The speculation on the part of many seems to be creating a fair number of straw men. Enigma's point early in this thread is that F-I will do what their masters say. Too early to divine what that conversation may conclude.
i hope i am not stupid but does anybody have a link to this "fico '08" plan? i mean, the darned thing and not some newspaper report!! it scares me to NOT know what they want to do to us i have been a stay at home wife my dh has me as au on OUR credit cards so if he dies I can get credit on OUR PERFECT credit thats what he told me artifice if i die ya' can get credit. aargh if this au thing won't count, i will be SO upset..!!
F-I 's algorithms are proprietary and unavailable for public review. They limit their public statements regarding their techniques, too, to PR releases. That is at the core of the discussions here; much of it is based on speculation.
Hell, Oracle it's all based on speculation at this point, lol. That's what we're talking about. Again, I've said my peace and you are correct about one thing which is no one knows what F-I is going to do. I don't think anything is going to come of this, you do but, no one knows. One thing is for certain . . . if something happens, we will know when it does. Of course, we agree on this. In any event, this is a very good thread and if nothing else I think we've shed some information on this. However, to answer the poster a few above, we don't know if this is going to happen or when. We do know that if it does, it will take the AU accounts out of consideration in terms of credit scores. Oracle and I disagree on this but, I don't think the model will stay viable too long thereafter however, that is my opinion. So, who knows exactly what is going to happen. Anyway, happy holidays!
Well finally, agreement. Seriously though, I see Apex's thought as F-I withdrawing FIC0-08 much more likely than a full-tilt at the subject by F-I as predicated by at least one other strident voice. My guess is that F-I's approach may be far more subtle, and as a consequence, much less attackable. I grant that with release of FICO-08 and a perception of discrimination on the AU issue, a suit will ensue. Never a disagreement there. I do not believe, however, it will be as readily attackable or as easily defeated. F-I has a significant proprietary interest in their product, and they seem to have the legal cover and litigation muscle available to keep it from being attacked. I see the odds of prevailing low. They have been successful so far. Something else over which there is little disagreement. I also don't see resolution in the near future. Court cases such as this might be have the habit of going on for years. As for injunctive relief for the consumer? I just don't think there would be a sufficiently convincing argument for the Court to tell F-I to stop doing business as they see fit. My assessment. Apex has his. We'll see who is closer when things wash out. December 25, 2012? Let's check back!
It comes down to what strategy is taken if it does go to court. I can't see why anyone would go after FI in the first place. The courts won't tell them not to do business as usual, because that is not the problem. The better strategy would be to concentrate on the possible violations of the lender and their failure to follow proper procedures and regulations. The OCC says they are required to submit any slight possibility of violations to the DOJ for investigation. The federal agencies have not had a problem in the past of bringing the industry back in line, and they never had to go after FI and their secret model. You can get from LA to Sacramento by driving to Miami, hopping a plane to Europe and then Siberia, crossing the Strait to Alaska, then heading South. It would be much easier to drive a few hours to the North. Will this type of suit be easy? Maybe not. But it would be easier with the right strategy and going after the proper defendant. You have one idea of what the strategy should be, and there are many consumer lawyers who have another. A note of interest. FI is already showing their model in some cases. The state of Texas requires them to make their models available, and can be viewed or purchased in Austin.
Well, then, why not go to Austin and purchase it? Then you can be heroic. Until you do, you're just piling more ifs into the equation. An iffy way of doing business. Sorry, but facts do sometimes get in the way. No violations, no case. You can also get from LA to Sacramento by flying to Antarctica and back, but you'd probably not be able to get a seat because a lack of permissible purpose; you have to have a reason to go. Sometimes the shortest way to a destination takes many detours. Do I hear 2015 for a status review?
Wow, do you enjoy making things so difficult for yourself in life? It must be sad that everything you look at is so difficult, lol. I don't think you will find anywhere that I promoted filing any type of suit without violations or damages. I'll leave that to the people that think they can get to FI with an unknown drop in a fako score. I haven't put any ifs in the equation. You are doing that all by yourself. If you think I am, you misread everything I posted. I see this as a very straight forward case if it ever comes about, and will not include FI at all. Just the lenders. I'll leave all the speculation and ifs to you, and continue this discussion with others who can see the issue without all the invented roadblocks. Sometimes the destination has detours, and sometimes it is just a short Sunday drive. No plans for Austin either, as there is no need for that info in this case, IMO. Just thought it was interesting that the secret model was already out and available. When the time is right, I know how to get to Austin without any detours. Interesting conversation anyways. Merry Christmas.
Every one of your posts is predicated on an if, an if that I am unwilling to grant at this time because of its highly speculative nature.. Everything said about FICO-08 remains speculation. Don't want to go to Austin? Your choice. Hero button gets to go to someone else. You seem to be the one that's flailing, since it is clear that you are unwilling or unable to overcome the objection. Your not-so-veiled attempt at casting aspersions my way clearly shows that you are willing to try the disparaging personal comment route to shake my argument because the rational ones haven't. A technique, for sure, but one that is wasted here. Tends also to have not-so-great consequences if it is used in court. But I digress. Point still remains, one has to have a cause of action to file a sustainable complaint, something which appears difficult to establish from what has been discussed here. A complaint such as is anticipated by the ifs, clearly will not go unopposed; of that I am sure. If there is a difficulty, it seems to be in gaining sufficient factual justification for a complaint. To date, I have seen none. Ifs just won't make the cut. Once a complaint is filed, it may be very difficult to make it stick. Such is the world of civil law. It can move at a glacial pace. There is never a guarantee of victory.
Wow! It looks like I missed out on an intense debate here! Quite the read, taken in entirety. My personal perspective is that "FICO 08" is already out there, and being used as a comparison score somewhere. The (legal) danger in its use is for the true "credit grantors", not Fair Isaac. FI's danger is in a business risk, as to the marketability (revenue producing ability)of the new model. My position is that FI has no legal culpability under the ECOA, since they do not grant or extend credit. The ECOA falls upon the entity reviewing for considering the extension of credit. The danger for these grantors is now in the "practices" the industry has developed. Legally the extension and/or granting of credit is based upon "all of the information" in an applicant's credit file and application. But...we are all well aware how that review process has often reduced to the FICO score number. The credit industry faces a problem of which number to use, do they use a number that qualifies, or disqualifies an applicant? The underlying profit motivation is a huge variable here. Similar to the "gun analogy", a FICO model that discounts AU accounts would not be illegal, the "illegal use" of the model, as with the gun", is the issue. In a like approach, no one charges that the "auto enhanced scores" discriminate against applicants who have never taken an auto loan. The "enhanced" models are used at the discretion of the reviewing party and credit grantor. The "rub" of the new FICO model comes into play at the "denial" of credit. If a credit grantor did deny an applicant credit based upon a FICO credit score which did not include any AU accounts, the grantor would have to prove that they in fact completed a manual review of the applicant's credit file, and "somehow considered the AU accounts" in evaluating the credit worthiness of the applicant. What is interesting here is that the Federal Reserve did a study of credit scoring models a few years ago, specifically for gauging any degree of discrimination (as implied under the ECOA). The study showed a distinct lack of discrimination in the models. The Fed even designed their own model for comparison. I maintain that the "FICO '08" model is purely a business risk for Fair Isaac, which I feel they developed in response to their customers' request. What is paramount for Fair Isaac is the underlying data of default rates tied to the "new scores". This is the core value of the FICO models, and I do not know how FI would pull this data together. FI would have to support the "market value" of the new model with hard data evidencing the default rates of applicants who had AU accounts versus "traditional" applicants. I know I sound like I am contradicting myslef here, but this is the only area FI could find itself culpable. IF, Fair Issac claimed the FICO '08 model as the "predictor" of default rates of credit applicants who have AU accounts, and the credit grantor industry merely passed that through in its evaluation process, then Fair Isaac could be exposed if they could not validate the integrity and robustness of the new model with hard data as evidence, and the impact were widespread enough. This is analagous to Fair Isaac claiming they made a "safe gun". Fair Isaac would probably need to add a disclaimer to their model that use of the FICO '08 model does not remove the need for a complete review of an applicants credit report and application. IMO, the new FICO '08 model will not be a legal issue, but rather a business issue. I personally do not see the new model being used by credit grantors in its current version, as I cannot see how its predicitve integrity could be proven right now, and the use of it by a credit reviewer places too much ECOA liability on the reviewer right now. In short, the risks of the FICO '08 model outweigh the benefits of it for the credit industry. I think it was an interesting exercise by Fair Isaac, but I cannot see it taking off as to common usage. As a footnote to this discussion, I have been conversing with an associate who is an "expert" with modeling and algorithms, and the potential for "errors" in coding and replication is quite large. "Lost in translation" is quite a factor when working with these types of models. Another factor, it was mentioned earlier in the thread about 8-12 "inputs", the actual number is quite larger. The referenced Federal Reserve model used 312 inputs to calculate a credit score. Unfortunately, the study did not reveal the amount used by the commonly used models, though it hinted at well over 200 inputs.
