Attack of the Killer Algorithms Part 13–Bank of America Style - Flawed Data With Credit Scoring Agencies–Dead Man Ba
Posted Feb 10 2012 2:18pm
Yes here we have one more story about flawed data which is on the rise in the US with data aggregation services. Certainly there have always been few errors and most have normally been corrected quickly in the past, but when you think about all the sources today that access the same data it’s a snowball effect as that correction may have to travel many paths to reach all of those using “and selling” that data.
If you read here often enough you know I talk privacy and data quite frequently as when you have servers running 24/7 making life impacting decision about us that you can’t see, talk to or touch, it is frustrating to say the least. Combine that with a customer service department that can only elaborate to you what is shown on their computer screen and yes it is frustrating. This situation is not limited to just banking as other transactional industries have incurred this as well.
Let’s go back a ways to a story I did about HeatlhGrades when I found my former doctor who had been dead for 8 years still listed and honoring HealthNet insurance. They rely on data sources that may not be updated and in this case we found a ton of errors, other dead doctors, doctors affiliated with hospitals where they had never set foot and in one instance a person who had never been a doctor but was an administrator with an IPA. Below is a post I made back in October of 2010 that has some further links and issues with some of this erroneous data in healthcare and crooks live on it too. HealthGrades has since this time merged with a marketing company so perhaps the data will get cleaned up or we may get better at marketing dead doctors for that matter <grin>.
Here’s another great example of flawed data from last year when the City of Buffalo paid out over $2 million with insurance premiums on many of their employees who had been dead for years. This stuff makes it tough for the living. The Social Security death index has 31k living people listed as “dead” too.
Here’s the Killer Algorithms Chapter 8 which is well worth reading when you see how data is used and sometimes abused. North Carolina kicked out CoreLogic from mining data as many state servers are slowing down to a crawl due to all the data mining activity and some states have had to install software to keep them out and worse yet, some of the folks mining this data don’t pay for the quarterly updates on you and I so as consumers we are now chasing flawed data we can’t fix as a result. That is what is sounds like is happening to this man at Bank of America.
Aggregated information contains flaws…this is the problem. Score would not agree to go on camera the video states. Again I said I do hope our new Consumer Financial Czar knows about math and the Attack of the Killer Algorithms.
So now in the credit area we have FICO mismatching data sources here combining credit scores with a new set of algorithms they call Medication Adherence Scores. How long before you get denied credit based on some data that they have scored you and said you will not take your prescriptions let’s say if you have a diabetes or heart condition? This is not part of a smart decision making process but rather one more attempt to sell flawed data as tech companies and credit companies make big bucks marketing mismatched data and will go all out and pull all strings to make a case for it with steroid marketing.
CoreLogic has formed a partnership with FICO who is already in the process of selling algorithms with mismatched data which connects public information about you and combines it with your credit score to tell if you as a medical patient will take your prescriptions.
Somebody needs to start calling some of these folks on “mis matched data” the discriminates as now we have“Discrimination by the Algorithm” and it’s showing in more places all the time. Last time I wrote about the FICO mismatched data it all ended up over at the Daily Kos via another publisher as an awareness. No wonder nobody can make a decision when mismatched data is used in such fashions.
In this situation one of the possible reasons listed is the man’s last name and again another big issue with connecting the wrong data sources as folks trying to find a job have problems here too with good folks being connected erroneously to others with similar names that may have a criminal record, like this story where this woman says you just don’t know where the next “attack” will come from next.
This man in this story with Bank of America is definitely under the “Attack of the Killer Algorithms”.
Companies make big dollars selling your data and Walgreens said not too long ago that their data selling business was valued at just under $800 million so one wonders are filling prescriptions a side line business for getting that data. We should license and tax these folks.
Banks and other financial entities buy and sell you as well and when you take a look at what is happening to this man, how many times have they sold his data and to who? Good question and only “the shadow” knows. In the recent mortgage settlement how much of our data was sold and created profits? This is another good question as we saw with all the mismatched data cases in the news, someone didn’t pay attention to connecting all the data properly. As a matter of fact Hedge Funds and and High Frequency traders are buying and selling consumer data at a rate faster than Facebook so again how much of our consumer data is floating around there too, whether it be actual data or “profiles” as some sell?
Some put games out there to entertain you while they scrape data to give the naïve consumer an enjoyable process while the algorithms work behind the scenes.
So how and when will this man’s data be corrected and how much time is it taking on his part to fix all of this and he did not input the data so the middle class when errors are created are becoming “data chasers” to fix what corporate America messes up. This is just one example but again time to ask about who’s collecting what data and are they held accountable to fix it when problems arise as it’s not always that “bad middle class consumer” at fault. We can spend days and hours working on this and we have little jobs to do as well and thus so how much of our time is consumed with fixing the “data needs” of corporate America I ask? Back in August of 2009 when the Madoff news hit, I made this post and do we need a department of algorithms to certify we are getting non flawed data and correct calculations?
The reason I mention this is due to a story like this Quant on Wall Street who for 3 years used his flawed data out there and didn’t fix it until the SEC came knocking and this HFT that was fined over their problems with their algorithms.
Not only this but let’s look at the hardware side of this with the servers being over clocked…maybe not a good idea today with the speeds of transactions taking place and when the rogue algorithms hit are you going end up frying your hardware too in the IT area? When I did training for Intel I used to see quite a few kids who melted down their processors on their PC and servers are just one big PC of sorts if you want to look at it that way.
So now, can we and how much of the data produced by banks is accurate? For that matter in view of the case settled this week, did MERS also add some cash to their stash selling our data in the midst of all their other record problems? It’s a good question to ponder and think about when you see all the flaws everywhere else in the credit areas and all those cashing in on selling our data. I used to write queries and code that match documents and data and if you hurry the job along for the pursuit of money, these things get messed up and what kind of audit trails are kept on all of this aggregated data?
This man is dead, many others on the internet have died and still live with aggregated data today so again, when are we going to demand accountability with mathematical formulas and get some algorithms that are accurate that we can trust running out there? Everyone talks about compliance and I think that word should be used hand in hand with audit and certification, especially with some of the formulas running out there that attack consumers, and thus I started this series called the Attack of the Killer Algorithms as that’s what it all boils down to. You can read the full summary at the link below and these are all public cases and information that has been written elsewhere before I connected some dots here, but let’s starting looking at this big picture and work towards accurate results instead of desired results so we all have peace of mind soon.
One insurance company is very well up there in the IT business and they make a lot of profits from their technology interests and I did a short summary on a few of them so don’t forget how with mergers and acquisitions, the company as you knew it a few years ago may have new business models today based on business intelligence algorithms and the selling and combining data for profit.
If you read the links here to my prior posts you may find some real eye openers here and I do this as an attempt to educate and help those and myself find answers as to the logic or illogic used today as to why things happen. The video at the link below is well worth watching as you will gain some huge insight as to how all of this works behind the scenes and all of it is not bad or wrong, but there’s enough of an element with spun data and algorithms for desired results floating around out there that make life miserable for the rest of us.
Corporate US will keep running the formulas and it makes money for them as all you have to do is see the stock reports and gains, meanwhile back at the ranch, inequality continues to grow as mere middle class humans don’t stand a chance and sadly until some of this is addressed it will continue to get worse and I hope this man gets back to a “living” status as he works through his “Attack of the Killer Algorithms”. BD
Arthur Livingston, of Prosperity, S.C., may feel prosperous and alive, but his credit report says, "File not scored because subject is deceased."
That's because Livingston's bank, Bank of America, has been reporting him as deceased to the three major credit agencies since May 2009, he said.
Bank of America has still not resolved the issue, even after media attention, causing headaches for Livingston, 39, and his family in South Carolina.
A regional manager of a chemical company, Livingston, discovered the dilemma when he tried to obtain a loan from a mortgage company in October. The problem may have begun when Livingston, who said he has been a Bank of America customer for 14 years, sold his home in May 2009.
Five months since he discovered the problem, Bank of America still does not have a solution, and his mortgage company has not been able to obtain his credit score to give him a loan for his new home, which he said is the "major problem" with being "deceased." He also fears the inactivity on his credit will negatively affect his credit score.