Forum
|
|
Forum
|
     | Monday, January 17, 2000 - 09:57 am  People don't test every portion of the model, but tests have and are run by people based on the "black box" viewpoint. That is the end result of the model is checked. A company will take 1,000 customers that score 620 and 1,000 customers that score 650 and determine which of these two sets has more delinquencies. When the people who score 620 have more delinquencies then the scoring model is considered to be generally effective. "...the Fed did undertake an analysis of credit scoring systems in the sense of trying to determine whether or not there were solid correlations between low scores and poor performance. And generally they found that there were...This was not an exhaustive study of each factor in a given scoring system and a relationship." So, basically, the scoring system does work as advertised. Groups of people who score low have more people not paying than groups of people who score high. This is off your own website, Greg -- I'm surprised you were unaware of it.
|
| | Monday, January 17, 2000 - 11:20 am Sean: I am perfectly aware of it. In fact, I obtained a copy of the report to which you referred. Did you? The report does not use FICO (credit bureau scores) scores in its analysis. It uses something Equifax fed the Fed (I was fed up when I read where the Fed was to head) called The Mortgage Score. The report does not validate FICO scores. Try again.
|
| | Monday, January 17, 2000 - 12:02 pm No, I did not obtain a copy of the report to which they referred. Nor did I obtain a copy of the FTC transcript. Rather I trust you to have posted it correctly on your website. Nor do I particularly feel the need to see the validation of the models. I feel confident that the office of the comptroller of the currency can handle that adequately. Would you like to read their guidelines that banks need to follow in order to ensure that their models are statistically sound? I thought that their 145 page document on the matter was pretty clear, but you should review it for yourself.
|
| | Monday, January 17, 2000 - 05:22 pm hahahahahahahaha that is hilarious! BTW, has Sean read the Transamerica Page? Can you tell me why the Cal. Dept. of Corporations didn't give a damn about Transamerica defrauding bankrupt widows? My trust disappeared a long time ago. MOST of the times I filed complaints the government did nothing. Once the Dept. of Banking in NY actually resolved a credit reporting dispute for me, in my favor. The other times I received anything other than a form letter was when I filed a complaint against a small company. The big guys are immune. I couldn't care less about any 145 page guidelines. Not only are they not enforced, but since nobody can calculate a Score violations can't be proven. Sean, what exactly do you think the "office of the comptroller of the currency" is doing to validate anything?
|
| | Monday, January 17, 2000 - 05:36 pm And here's another question: Both consumer A and B have a $1,000 take home pay. They're single and live in identical apartments next to each other and have the same job, car, and driving record. Credit Score for A: 645 for B: 680 They each have a car loan, and $5K in credit/charge card debt. A's monthly payments are $350. B's monthly payments are $250. Does everybody agree that A is much more likely to default? Isn't it obvious that Credit Scores are the CAUSE of A's default, NOT the prediction? Credit Scores are self fulfilling prophecies.
|
| | Tuesday, January 18, 2000 - 07:51 am Sean: Gosh. Do all 145 pages deal with ensuring "that their models are statistically sound"?
|
| | Tuesday, January 18, 2000 - 10:32 am I don't believe that all 145 pages deal with ensuring that the models are statistically sound, but it does seem clear to me that the OCC requires each bank under their regulatory control to verify that the scoring model they have purchased and are using is statistically sound not just at the beginning of use, but on an ongoing basis. Of course it should be noted that the OCC doesn't care if the models are fair. They are only looking to ensure that banks remain financially viable and do not need to be bailed out by FDIC. Accordingly the OCC will take the lender's perspective as to whether risk scoring is a good technique for loss mitigation.
|
| | Tuesday, January 18, 2000 - 12:29 pm Sean: The name of that document is Credit Card Lending. Which of those 145 pages deals with residential mortgage loans or automobile loans? You didn't mention the name of the document in this thread because it didn't serve your rhetorical needs. Start using your name instead of anonymous255@yahoo.com; it will increase your accountability and decrease your wild statements.
|
| | Wednesday, January 19, 2000 - 02:59 pm You sure are trying very hard to wiggle out of the fact that every bank is required by the Comptroller of the Currency to ensure that the scoring model they are using will predict with 95 percent accuracy the current performance of their accounts before it can be used. Let's review your previous words shall we? Greg: "...[N]obody but Fair, Isaac and the credit bureaus has checked them. I looked into it." I would be interested to know how many banks you personally contacted to determine if they used the generic FICO model for their credit card underwriting process. Could you please provide the names of a few banks you contacted, the names of the person you spoke to and their phone numbers so we can know you actually did call every bank in the United States to ensure that not a single one of them used a generic FICO model for their credit card underwriting process?
|
| | Saturday, January 22, 2000 - 11:17 am I didn't contact any banks to determine if they used the generic FICO model for their credit card underwriting process because that would be an invalid test of other uses of scores. I did, however contact the Office of the Comptroller of the Currency, asking if they regulate consumer reporting agencies with regard to the validity of credit bureau scores. They responded that they do not regulate credit bureaus. It is clear that the bureaus are regulated by the FTC, and they claim that there has been no study. http://www.creditscoring.com/pages/regulation.htm If you want to believe that individual lenders use credit scores to justify their decisions and make a buck doing it-- and that fact validates credit scoring as a fair, or even an accurate practice-- you are welcome to do so; naivete is not against the law. Because the OCC orders them to "validate" doesn't mean they do it. Because there is a law that forces credit bureaus to maintain toll-free numbers on which consumers can ask them questions, did that stop them from not answering their telephones? Do housing anti-discrimination laws mean that mortgage finance is pure? I think it is ridiculous that so many government agencies regulate banking. They were created at a particular point in time to solve a then-current problem. Now the OCC (formed in 1863 to coordinate Civil War financing), the FDIC, the Fed, the OTS, and the FTC all throw-in their two-cents and create a system of mush. Who should prove that credit scoring is valid? Those who sell it: the credit bureaus. But they don't even like to acknowledge its existence. How was scoring validated before anybody started using it so they knew is was valid to begin with? If any studies exist which prove validity, they are all locked up in the desk of some hack in San Rafael. What are they afraid of? Why don't they defend themselves? You started the thread with a reference to a report that I handily discredited. Had I not, you would have walked away thinking, and would have been able to lead others to believe it was a credible study. That incredible blunder leaves huge questions about your entire premise. You are losing this argument because you didn't do your homework. Try again.
|
| | Sunday, January 23, 2000 - 10:06 am Wait a minute, let me get this straight. The OCC requires **ALL** banks to validate credit scoring and now you, with no proof whatsoever, want me to believe that not a single bank out of the thousands that are regulated by OCC have actually taken the time to determine that it works? That's pure bullshit and you know it.
|
| | Sunday, January 23, 2000 - 05:13 pm Pretzel logic. You are the one making the case (poorly), not me. Let's see the numbers. Let's see the blue-ribbon panel of experts testifying FICO scoring is valid. It's never going to happen because the source code is top secret. This is a fine mess they've gotten themselves into. The OCC made some comments at the FTC forum in July-- here they are: http://creditscoring.com/pages/forumtranscript.htm#page198
|
| | Monday, January 24, 2000 - 04:08 am Washington Post casts a light on Scoring and CRA accountability. http://www.washingtonpost.com/wp-srv/WPlate/2000-01/23/059l-012300-idx.html
|
| | Monday, January 24, 2000 - 05:38 am Hey Sean, here's a good quote from your favorite agency from the article Denise listed: "The failure to report is unfair to customers, unfair to competitors who play by the rules, and ultimately inconsistent with the values of our national economy," John D. Hawke Jr., comptroller of the currency, which regulates the nation's banks, said in a speech last year.
|
| | Monday, January 24, 2000 - 08:14 am Ok first of all, we're splitting subjects here -- and I'm fine with that -- but maybe we should've started a new thread for that. My response will, therefore, be split. First of all let's talk about the credit reporting agency's failure to handle 1 million calls in a manner that was acceptable to the Federal Trade Commission. Let's furthermore cut through all the bull and go straight to the root of the problem. A sizeable percentage of the calls that a credit reporting agency has to process are from people who have accurate, derogatory information on their credit profile that they're trying to delete. These people either have heard or know that they can call up, and dispute an item repeatedly and eventually it will go away because the credit reporting agency doesn't have the manpower to deal with a person that's bound and determined to erase accurate information. And this article is either spin or commentary. If it's commentary, great -- let's just straight out label it an editorial and we'll all know that's the reporter's opinion. If it's supposed to be merely an article then phrases like "Clearly, the industry still needs more oversight" should not be there. This is a statement of opinion and not fact. This article is just another example of media bias. I note that people opposed to the credit reporting agencies got quoted but did anyone from any of the big three get quoted? Nope, they sure did not. Is that a fair and balanced coverage of a story that presents the facts and lets people make their own decision? No it is not. In fact, I came up with a good quote right here within seconds by searching the Experian website according to Ask Max 40 percent of all disputes they have to process come from credit repair clinics and most of those are attempts to delete accurate information. Of course the reporter isn't going to include that information. That wouldn't give the article the right spin, would it? ************************************************** But none of the above has anything to do with scoring, which was the original thread. Let's check out Greg's link and see what the OCC has to say about credit scoring. "[W]e think the odds ratio is that credit scoring systems make an underwriting system likely more fair." (this according to the OCC). Let's press further. When are scoring systems unfair? They give three examples: Changing the input information based on whether the applicant and co-applicant were married or unmarried so that unmarried persons half of their income not count. Age-split scorecards on a non-statistically validated system. In this case the scoring model is faulty because it wasn't properly created, maintained and tested. Spanish speakers needing a higher score than English speakers. Is this the fault of the scoring model? Nope, sure isn't. The fourth story was anecdotal so we are hard pressed to say whether such a situation would be fair or unfair based on a different scoring model used for people applying from Ebony magazine. It sounds like something that should be avoided and isn't cost effective. Why employ two statistically-validated scoring models when you can employ just one for half the cost? Except for your standard "garbage in garbage out" argument (which I agree with, to a limited extent) you have been able to produce no one that has studied scoring models and doesn't come up with the same message: That they work and that they're more fair than oldstyle underwriting methods. Fair Isaac studies show that their generic credit scoring is actually affirmative action scoring in that for any given score a person located in a minority zipcode poses "a slightly higher degree of risk" than their score indicates. Furthermore let's state that banks are required by the OCC to statistically validate scoring systems and the OCC has made referrals to the Department of Justice for at least one bank that wasn't doing it right.
|
| | Monday, January 24, 2000 - 09:37 am What is your definition of statistical validation? In your previous post, are you referring only to generic, broad-based risk credit bureau scores?
|
| | Monday, January 24, 2000 - 01:22 pm Sean, obviously you either did not read the article in its entirety or read it with your eyes half open in an attempt to only see what you want to see. I point you to the following Quote: "It's like taking the SAT and nobody will tell you your score," says Peggy Twohig, assistant director for financial practices for the FTC." Read the ENTIRE article before you make false representations the article only had to do with the issue of their not answering their phones. In your refusal to accept, even a slightly different view point, makes it sadly apparent that no matter what is pointed out to you, or what can be proven, you hear no truth...see no truth and therefore... can speak no truth...
|
| | Monday, January 24, 2000 - 01:29 pm By the way Sean, "Ask Max" is not someone I would trust by any strech of the imagination for truthful, honest and unbiased information. Give me a break! I could point out many, many discrepancies in what the law says and what "Ask Max" says but, I do not have the time, strength or desire to attempt to enlighten someone who refuses to see or hear the truth or accept any other viewpoint than their own.
|
| | Monday, January 24, 2000 - 02:33 pm What do I mean by statistically valid? Let's say I'm a credit card issuer and a Fair, Isaac representative shows up one day and offers to sell me their new, presto-chango Fair, Isaac Revolving Credit Model (Industry Option). It slices, it dices -- it makes the whole family happy. I'm open-minded, but skeptical. I'm from the "show me" state. How do I know that this scoring model works? Well, he's willing to show me. It so happens that I run account-review credit profiles on all my customers every 6 months and we run all of those through his credit model and out pops scores. The median score is, let's say, 680 -- so we divide my customer base into two groups ... those above 680 and those below 680. Low and behold those below 680 have more delinquencies than those above 680. Ok now he's got my attention, but I'm still not convinced. So we divide them further. We divide it into four groups: below 640, 640-679, 680-719 and 720 and up. Low and behold those below 640 have the worst record, those in 640-679 second, 680-719 third and 720+ best. At that point I'm convinced the scoring model (at 60 cents a score) is far better than my underwriters at $18/hr and that the scoring software is statistically valid. What's your definition of statistically valid? ************************************************** Denise: I did read the entire article and I found no one in the entire article that was had anything to say in favor of scoring, or even the credit industry's defense. Is this your definition of fair and unbiased reporting? Look at the words chosen by the reporter: "Of course, the three bureaus...admitted no wrongdoing" (emphasis added). Wow, isn't that a kind of mushy way of saying they firmly feel they are doing right by the consumer? The story claims that One million plus callers were turned away from Experian; One million plus callers were turned away from Trans Union and hundreds of thousands of "customers" trying to reach Equifax had similar problems (ever notice that Equifax does a better job of dealing with consumers?) Well, I dare say that at least 90% of those consumers who called aren't "customers." They're most likely calling to dispute something or calling for a free credit profile. Only a fraction of those people want to pay $8.00 to get their credit profile. It's funny how in a world where I have to pay money for water (and I need that to live) I can get my credit profile for free (which I don't need to live) yet everyone is upset when some people have difficulty getting through on the first try or gets put on hold for "too long."
|
| | Monday, January 24, 2000 - 03:04 pm I'll ask again: In your previous post, are you referring only to generic, broad-based risk credit bureau scores? Repeating: Which of those 145 pages deals with residential mortgage loans or automobile loans? One more time: Let's see the numbers. Let's see the blue-ribbon panel of experts testifying FICO scoring is valid. Your weak supporting quote: "'[W]e think the odds ratio is that credit scoring systems make an underwriting system likely more fair.' (this according to the OCC)." Now there's a real statement of fact. Could he have couched his words any more? We "think"? The "odds ratio"? "Likely"? Where's (even as in the simple example you gave) the study?
|
| | Monday, January 24, 2000 - 03:57 pm Sean...where do you get your info from? You prove my point! Where did it say or how did you derive at this quote of yours "Well, I dare say that at least 90% of those consumers who called aren't "customers."? Addtionally, its funny that you so heavily refer to how the reporter wasn't fair by not getting the opinions of the Big three, well, I don't see them refering or including consumer quotes in their reports or press releases. Also, have you EVER heard of one of the CRA's actually EVER admitting to WRONG DOING? Lastly, I can't believe you complain because we get copies of our credit reports for free!! We don't get paid to do the work we need to do in correcting and re-correcting their errors nor do we get any money from our personal information that they profit billions of dollars from....gee, I guess a free credit report once a year won't break them!!! See..It actually costs them less to provide us with a copy of our credit report so that we can clean up their messes than it would cost them to actually hire more employees and do their own work!!!
|
| | Tuesday, January 25, 2000 - 11:09 am As I said before, this topic has split. That's why my answer on the Washington Post article can be found elsewhere. Now let's again remind ourselves of Greg's comment that started this thread: "In other words, nobody but Fair, Isaac and the credit bureaus has checked them. I looked into it." Let's see what the OCC has to say on the matter: 'Credit scoring systems generally fall into one of two categories: those that are empirically derived, demonstrably and statistically sound (hereinafter referred to as "validated"), and those that are not.' The OCC includes the definitions of both, which you can check out in your spare time. As you can see from this document the OCC does perform "backtesting" to determine if the model performs as specified by the model designer. I have already requested a copy of any such backtesting done through the use of the Freedom of Information Act, and I would be delighted to mail a copy to you when it arrives.
|
| | Tuesday, January 25, 2000 - 03:52 pm You're fading fast. Your quote of "what the OCC has to say" is no study. The next one is equally irrelevant. Is this what you do when you get frustrated-- throw up a bunch of non sequiturs? Stop grasping for anything you can find with the word "validate." Anybody can do that; whoopee. Try to find something that has a remote semblance to the words: "we tested credit bureau scores, and they are valid-- here's our study" from a credible source. Now, run along.
|
| | Wednesday, January 26, 2000 - 09:45 am Since their studies aren't placed on the Internet their studies do not exist -- is that what I'm hearing, Greg? Well, since people's credit profiles aren't posted on the Internet they don't exist either so we're arguing over nothing.
|
| | Wednesday, January 26, 2000 - 10:21 am Your logic has failed you again-- comparing a government study to an individual credit report. One we have the right to see; one we don't. You started the thread, and lost in the first round by quoting an irrelevant study. Now, all that is left of your premise is baloney-- and you don't know when to quit. With your anonymous255@yahoo.com email address, you have just become another Internet flamer. Post something of value, or leave it alone.
|
| | Wednesday, January 26, 2000 - 12:05 pm I have already posted that I requested a study (any study) from the OCC via the Freedom of Information Act. I even tried calling them, but with the snowstorms in Washington DC they were closed all that day (and now, it's 2:00 pm, 5:00 pm where they are, again it's not possible). I completely reject your two contentions: that no one ever did a study (I seriously doubt this) and that you checked (How did you check? Can you offer any evidence at all that you checked?). Furthermore you seem to have a problem with my right to privacy and seem to feel that you're above or beyond privacy. Since that is the case, would you be so kind as to post your: Name Address City, State, Zip Phone Number Place of Business Business Phone Number Credit Report url of your jpeg. I'd sure appreciate it.
|
| | Wednesday, January 26, 2000 - 02:58 pm Sean: Now there's a post that's sure to shut me down quick. This is so corny an effort, it is laughable. If there is some study, I would be remiss in not posting it at creditscoring.com and Fair, Isaac and the repositories would certainly be hyping it. I checked by writing to the regulators (including the OCC). And when I finally got the famous Federal Reserve report, it didn't validate FICO scores. Again, I contacted the OCC myself, and they responded (http://www.creditscoring.com/pages/regulation.htm). Apparently, your comments about me not checking is your hokey idea of good spin. My credit report? Get real. You want my medical report, too? You forgot my birth certificate and dental records. Shoe size? My mother's maiden name? My name (LAST and first) and address are on my web site, as I'm sure you are aware (you refer to the site often). If anybody wants to identify me, they can. You, on the other hand are still some guy who may have the name of Sean at the address anonymous255@yahoo.com. Your last post is just another example of non-credible flame from an anonymous source. You have the right to privacy. And the readers have the right to see you as they wish, too. Your obsession with pictures is bizarre. Hack on, wayward flamer.
|
| | Wednesday, January 26, 2000 - 03:05 pm My name: Christine Baker Address: USA -- Arizona at this minute City: About 8 miles South of the intersection of 8 and 10, parked at the Flying J's NE parking lot Phone/Fax Number: (415) 449-3484 Place of Business: Bend, OR Credit Report URL: none -- Next time I see my credit report I'll probably go shoot somebody. It's been years since I saw my reports, why would YOU want to see it? I think most of the above is VERY irrelevant to anyone here. I'd like to know: Name: City/State (#of years): Profession (# of years): URL: Credit related experience (description and # of years):
|
| | Thursday, January 27, 2000 - 11:51 am Greg: Ahh, yes, the regulation page. I didn't recall reading anything off of that page, and now I remember why. The first 5 links are broken. Your page generates the impression that credit scoring is not regulated at all and that no studies have been done to determine if credit scoring is valid. DESPITE THAT you know that the Federal Reserve has done studies on credit scoring and found them to be valid. Yet you split legal hairs over the idea that "Well, it's not Fair, Isaac scores, per se." Then, ironically, you flame me over being a "flamer." Christine: I called your phone number and left you a message. All I got was voicemail that picked up on the first ring.
|
| | Thursday, January 27, 2000 - 01:56 pm It is unfortunate the FTC has moved the documents to which the hyperlinks used to lead. It's a big job-- keeping track of all the government agencies, credit reporting agencies and vendors involved. But don't worry, I'll find them. In the meantime, knock the site all you want; your piddling comments mean almost nothing in the larger realm of the issue, but also show me what the fringe might think (while amusing, inconsequential). While you have a manic tendancy toward the argumentative, it gets you nowhere because you have chosen a losing cause. It brings back the issue of your credibility, again, however. The many other links on that page are very informative, but you didn't mention them. "Your page generates the impression... "? What impression you get is your problem. I get the "impression" that, to you, the discussions are only a game. You said, "DESPITE THAT you know that the Federal Reserve has done studies on credit scoring and found them to be valid. Yet you split legal hairs over the idea that 'Well, it's not Fair, Isaac scores, per se.'" Just another day at the office for you with one of the usual tactics of the opposition: trying to confuse the issue by stating that there are any number of scoring systems. See the bald-faced, almost embarrassing-to-watch attempt by the credit bureaus' association at http://www.creditscoring.com/pages/forumtranscript.htm#page271. It's similar to your everything-has-everything to do-with-everything problem. You can't focus, and lash out at everything-- making your arguments turn into mush. Keep playing your debate tactic games-- arguing the presentation, not the issue-- and I'll answer in kind. With any luck, this thread will remain forever, and haunt you forever, too. Then again, as an anonymous voice, what do you have to lose? Go ahead-- post your irrelevant response.
|
Credit Forum CreditCourt Forum 2003 Credit Suit CreditFactors Order Credit Reports |