Path: utzoo!utgpu!watserv1!watmath!att!linac!pacific.mps.ohio-state.edu!zaphod.mps.ohio-state.edu!think.com!barmar
From: barmar@think.com (Barry Margolin)
Newsgroups: comp.org.eff.talk
Subject: Re: OUTLAW ALL DATABASES!!  Damn right!
Message-ID: <1990Nov29.220837.15166@Think.COM>
Date: 29 Nov 90 22:08:37 GMT
References: <7425@hub.ucsb.edu> <5140@rsiatl.UUCP>
Sender: news@Think.COM
Organization: Thinking Machines Corporation, Cambridge MA, USA
Lines: 176

In article <5140@rsiatl.UUCP> jgd@rsiatl.UUCP (John G. DeArmond) writes:
>*	The IRS buys qualified databases and computer models your lifestyle
>	and therefore your income from your buying habits.
...
>*	Your detail buying habits are increasingly being recorded and kept
>	in databases.
...
>*	Even for such innocuous things as telephone saleslime, you will be
>	increasingly targeted based on your personal information stored
>	in databases.  If you watched the Nova PBS program on the subject
>	of direct marketing tonight (11/28), you will know what I mean.
>	I've worked very hard to stay off these lists and still I get one
>	or 2 slime calls each day.

I watched the Nova program last night, and the most surprising part (to me)
was not the detailed personal information, but the detailed IMpersonal
information.  All the stuff they are able to discern about neighborhoods
and block groups, allowing them to target people without actually having
individual profiles.

What this means is that the direct marketing agencies, and other interested
parties, may be able to infer correctly quite a bit about you without ever
looking at information with your name on it.  For instance, neighborhood
purchasing profile databases can be built even if everyone pays cash at the
local stores.

The insurance industry has been using this method for a long time.  Rather
than looking at individual history, they define broad categories
(male-under-25, etc.).  In the auto insurance industry your accident rate
is dependent mostly on your category and the value of the car model, your
theft rate is based on the car model and the town you live in.

What we've got is a two-edged sword.  Many people object to being lumped
into categories, being punished because they happen to be in the same
category as many bad risks (but does anyone in the good risk category
complain about being tagged?).  On the other hand, we don't want all the
companies we do business with knowing all the details of our lives.  I
admit that it's a hard problem to solve.

>1984 is upon us and it won't even take view screens.  Your life will be
>modeled well enough that the computer will predict what you'll buy
>before you do it.  From what I've seen in my work, this will happen
>in the next couple of years.

The Nova show suggests that it is already happening.  "They" know when your
baby was born, so "they" know when you'll be reading bedtime stories, when
you'll be interested in a preschool, etc.

>truth is that credit databases tell you practically nothing about the
>creditworthiness of a person.  All it shows is the HISTORY which
>has NOTHING AT ALL to do with your ability or willingness to pay in
>the future.  

Ah, the old "past history is no indication of future performance" paradox.
Unfortunately, past history is frequently the only objective information
available.  Credit purchasers are generally not willing to wait the length
of time it would take for the provider to get to know the purchaser well
enough to decide based on other information.

>	      A classic example is mortgages.  The mortgage companies
>claim to have a formula to determine whether or not you can make the
>payments.  The problem is that this formula, which usually calculates
>a ratio of your debt load to your net income, is changed radically
>to meet business objectives.  For many many years, the rule was
>1/3 of your net income could go toward house payments.  But when
>the market gets soft (as in Atlanta right now), they'll let  you
>do maybe 50% of your net income.   And in crazy areas like LA, some
>lenders will let you commit as much as 65% of your income to debt.
>Understand what is going on here.  They clain out of one side of their
>mouth that a formula exists with which to evaluate your past history
>and predict your future performance while speaking a number seemingly
>at random from the other.

I don't see it quite as badly.  The output of the formula is presumably a
probability of default, not a yes/no.  In a soft market the banks are
willing to take bigger risks, so they change the probability cutoff.

>Secondly, having had large holdings in the rental real estate business
>and having had a storefront business, I'm quite familiar with personal
>credit.  What I got from the credit reporting agencies was useless.
>My collection percentages went UP when I trashed the reporting services
>and simply made a SWAG (scientific wild-assed guess).  Here's why:

There are two conclusions that can be reached from the inaccuracy of the
credit (and other) databases: 1) they shouldn't be used, or 2) they should
be made more accurate.  The pessimist in me doesn't see these databases
going away, but the optimist believes that as society comes to rely on
these databases more there will be incentive to improve the accuracy.

>Now consider what my credit record shows.  It shows that I've repeatedly
>run up large balances and then paid them off in a few months.  In other
>words, I'm an excellent credit risk (and the unsolicited credit cards
>prove it.).  WRONG!  At the time I'm maxing out these cards, I'm at my
>most vunerable.  Especially when I'm financing new development, the risks
>are high.  So what looks like excellent credit is actually high risk
>unsecured loans.  One wrong move (knock on wood so far) and poof!
>Instant LARGE bad debt.

Yes, the credit industry's method of defining "good credit history" is
simplistic and flawed.  But how could businesses with thousands of
customers possibly understand the nuances of each one.  Coming up with
simple ratings based on complicated patterns is difficult.  For example,
the problem tracking database at my previous employer rated developers and
departments based on the number of problems that took various lengths of
time to fix.  So if a developer went in and fixed all the old, low-priority
problems that had been assigned to him for years his rating would go down
because he would now have a large number of problems that took a long time
to fix (the rating didn't count unfixed problems, because that wouldn't
account for the fact that some are insoluble without major system redesign,
and it would be inappropriate to penalize the developer for that).

I also dispute your claim that you are a bad credit risk.  You have
repeatedly succeeded at the projects that required you to run up these
debts.  I would assume that the development projects you finance in this
manner are generally successful.  Since you seem like a financially savvy
person, I assume the fruits of these projects have a better return than the
exhorbitant interest rate the banks are charging you.  Sure, the loans are
unsecured, but are venture capital or stock offerings much safer?  With
your history, you should be able to 

>I'll explain again.  Ahem....  My proposal would ban the use of your
>personal information WITHOUT YOUR PERMISSION.  Got that?  You could
>give ANYONE permission to use your information as you see fit.

Wouldn't that add an intolerable bottleneck to many transactions?  Someone
who wants to look at information about you would first have to send you a
letter.  This would sit on your desk for a week or month before you mail it
back with your signature.  It would be streamlined in many industries, I'm
sure (I think that when I applied for my mortgage I signed something giving
them the right to look at my credit history).

>An enhancement to my proposal could requre that a data collector get
>your permission and approval on EACH piece of data before it is 
>added to the database.  That way you'd have veto power over any
>data you did not want to have added.  Sure this would kill adverse
>databases such as credit but so what?   They are worthless anyway.

Hmm.  What would this mean for things like supermarket databases?  Would I
have a line-item veto for each purchase (i.e. yes, I bought the meat, but
don't list the Penthouse magazine)?  What would the logistics of this be?
Would I veto the items right there on the checkout line, or would the store
send me a checklist in the mail every month?

Of course, you can't remove all vestiges of the database entry.  The fact
that someone purchased a Penthouse magazine has to be recorded for
inventory purposes, and has to be used for planning purposes.

>While we're on the subject, perhaps you could explain why it is 
>right for the insurance company to raise your rate based on 
>your falling victim to the random revenue collection process called
>speeding tickets?  Why should they be allowed to profit from the
>state's dubious taxation system?

If the process is truly random, then statistically it is expected to catch
the chronic offenders more than the rare cases.  Also, since you consider
speeding tickets merely a taxation system, would you be less offended by
the insurance industry's use of this data if the punishment were
non-monetary (or do you think speeding tickets would be abolished if they
didn't produce revenue)?

>Why cannot I not work at night and sleep in the day and still have a 
>phone without an answering machine to screen out the trash that solicits
>me all day long?  Why can't I leave it open in case there is a family
>emergency and not be awakened all morning long?  Why the hell does the
>system allow the bastards to start calling at 8:00 am and continue 
>through 10:00 pm as they did today?  Why can't I post my phone and
>mailbox the same way I can post my property against intruders?

The Nova program featured a group of people who have told telemarketers
that they would be charged for the use of their phone as part of the
caller's business.
--
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar