Thanks.

So I think you all for coming and
I may thank Danny in particular for

inviting me and arranging for

me to be here as well as that lovely
generous introduction Goodman for

who I met at breakfast today in Great
talk about the program and Michael Murphy

has been handling much of the details
that are necessary to make this possible.

Bernie.

Thank you.

We met.

On the way over.

So my background is history
from historian by training

who also has had a longstanding interest
in economics and over the past decade or

so but more than that maybe fifteen
years in interest in technology and

in particular information technologies and
so this is a part of a new project

that I have sort of been playing with
the past few years on surveillance and

thinking about surveillance
in various forms and

guises that it takes in modern society and

then as a historian thinking historically
essentially how did we get here and

what can we learn by examining the past
that has brought us to this current moment

so my talk today it is the economics
of sort of surveillance or

really the strange economics of
surveillance as I hope to make clear

why information does not always do
what economic theory predicts that

it would do so a lot of this talk will
be about some of the well darker but

also perhaps less obvious or unforeseen

consequences of building a world through
information and information technology.

And I'm going to basically
give you a story a story about

one particular form that surveillance
has taken after some political.

Remarks about the larger sort
of theoretical and policy

framework that study surveillance or at
least as I understand it takes place in.

In his nine hundred forty eight paper
a mathematical theory of communication

Bell Labs engineer Claude Shannon
determined that information could best be

understood as a transmission that reduced
uncertainty and resolved ambiguous States.

Everything else should be treated
as either noise to be filtered out.

Or redundancy to the main signal.

Shannon's idea was influential enough
to enter popular culture in the one

nine hundred fifty S. and sixty's.

Designers Charles and great Eames gave it
a puckish twist in their film a primmer

on communication where they pointed out
that telling your lover you loved her

had no informational value.

Shall not have thus reduced diverse
expressive forms such as letters and

telegrams or today email as well
as novels plays paintings and

sculptures to mere
signals carrying messages

the semantic content of those
messages was not his concern.

By this way of thinking there was no way
to distinguish good information from bad

information truth from lawyers democratic
discourse from to tell a terror in

propaganda or slander gossip and emotional
appeals from reason and knowledge.

Much riding on information and the economy
as follow the same line of thought

information signals are presumed to move
only in a positive direction informing and

enlightening the receiving parties.

Information as an economic
resource expands agent knowledge

removes uncertainty
increases transparency and

thus contributes to economic
efficiency and innovation.

But what if information creates
unintended wrong wanted consequences

bad results that make for inefficiencies
stifle change and innovation.

But I don't mean here the well
known problem of information over.

Or load which could be resolved by
better filtering mechanisms and

the organizing of data.

I mean that information itself may be
a social bad as well as a social good.

And this is my issue for today.

If we shift our idiom from
the neutral wand of information

to a more strategic language then all
sorts of new possibilities problems and

potential conflicts of interest arise.

Let's take for
example information as surveillance.

Surveillance is the creation in
circulation of information but

it adds the connotations of power and
control.

Surveillance is a one way
rather than reciprocal at

someone who does surveillance on
someone else with some intent or

purpose in mind it's not just
the gathering of raw data but

the recording organizing classifying and
sorting of that data.

Now these are not neutral process these
there's no one best way to organize

information as librarians here can
tell you they reflect the needs and

goals of the surveillance.

Of course the word surveillance
immediately brings to mind the state

particularly the totality Tarion
state as in George Orwell's novel

Nineteen Eighty-Four Big Brother.

But surveillance also has a market mode.

It's the link in exchange
relationships and

the Meet the link between exchange
relationships and the means and method of

enforcement between contracts compliance
and the regulation of agent behavior.

Without surveillance contracts or just
pieces of paper filled with high flown or

obscure language.

Surveillance trust is nothing
more than a pure act of faith.

The dependence of modern business
transactions on surveillance

just makes the story much more complicated
than in the case with the state and

state surveillance States and post
surveillance with clear objectives in mind

capturing knowledge of citizens to
build armies gather taxes regular.

Populations were perhaps more benign
early to improve public health

require forms of surveillance of
people's behavior and location.

The age of surveillance here is clear but
in the marketplace.

No one wields sovereign power

mixed motivations rather
than clear objectives rain.

All reliance on surveillance in the
marketplace is also a relatively new one.

Now people clearly have always watched and
policed each other but

until about two hundred years ago they did
so in face to face or in small communities

only occasionally as with conscription
into the military did the state even come

to know much about its citizens or
more accurately the king his subjects.

In the economic realm most activity took
place among people living side by side

even long distance trade
depended on personal knowledge or

effective contacts for
example the religiously or

ethnically based trade D.S.
Bros or the small communities of merchants

who met regularly in London's coffee
houses in the seventeenth century.

And modern society by contrast we
trust people whom we do not know and

may never meet and extending
connections beyond physical proximity

the tools of surveillance permit and

the classic formulation of
George Zimmerman a society of strangers in

other words we know very little personal
into individual about our fellow citizens.

So in place of that we use
lots of in personal data.

Surveillance in the marketplace is
important to how the world functions but

it's far from the impartial process
depicted in information theory.

It operates through assemblages of
technology that may interact in

unexpected ways it's continually inflected
by unintended or incompletely foreseen

outcomes it functions in context shaped
by politics culture and social relations.

Now for the rest my talk I want
to bring this down to a more.

Create level by looking at one
of the most ubiquitous and

powerful surveillance techniques of
the modern economy credit reporting

and economic theory credit reporting
is all about increasing transparency

establishing trust regulating moral
hazard and disciplining opportunism.

With better information lenders more
easily evaluate borrowers permitting them

to assess risk and lower the price of
credit indeed with true transparency.

Everyone wins.

We see when all see all competing
lenders seek out the good risks

offering them favorable interest rates and
all that although the poor credit risks

are segregated out lenders can
correctly price credit to them as well.

Higher interest rates but
better than no credit at all.

This compelling narrative portrays
surveillance as offering up good

information that eliminates strategizing
but what if people can strategize

over information as much as
any other economic resource.

Credit reporting in fact has been a
continual struggle over representation or

how people are seen in
the credit mechanism

by what information on whose terms.

And practices that involve information
like credit assessment work they do so

through conventions that
deal with self interest and

strategizing over representation.

And these conventions may be quite durable

though like any institution if not
properly maintained they will break down.

Unlike any institution as they spread and

become normalized in accepted
practice they cease to be questioned.

We don't look inside them anymore.

They start to take over the thinking for
us.

When this happens in this is to
anticipate my conclusion a little

bit when this happens once they
become normalized and accepted and

unquestioned it becomes important
becomes possible again for

actors to game the system to engage in
informational strategizing In other words.

In the ways of information
the evolution of new strategic

viruses is always a possibility.

So to understand this process.

Let me look at the evolution of
consumer credit reporting and

how it produced the fight go score.

I'm sure you're all familiar with
the FICA score your credit score which

no one else sell back to you for
a profit if you're interested.

Today almost all consumer
credit serve Eylandt

is done by a mathematical algorithm based
on statistical relationships between

the behavior of the borrower and
the likelihood of repayment.

By the business credit market I should
note is much less formalized than this

even though credit reporting in trade
has a longer history than it does with

consumers' trade credit reporting has
not devolved down to a single predictive

scoring methodology something I discussed
in another paper but today I want to

tell the story of why scoring came
to dominate the consumer market.

Let me just note though with the contrast
between consumer credit reporting and

trade credit reporting that in
both cases the relevant causes and

effects are the same that is they reflect
the problems of information sharing and

the aligning of interests and
incentives over representation.

Information markets like
credit reporting or

subject to the classic chicken in a or
network externality problem

users of information will only
subscribe to an information network.

If coverage is sufficient
to make it useful to them.

Sellers of information can only afford
to collect the data if there is a clear

market for selling it including
the ability to exclude non payers.

The subjects of surveillance need to
be seen in order to obtain credit but

they resist too much of or the wrong sort
of at least in their view the wrong sort

of surveillance that might cut
off their financial life blood.

Thus the actors in the system
have both divergent and

convergent interests resulting
in a continual Strug.

Over what constitutes
valid information and

how that information is represented
the chicken and egg meaning that you need

people to buy into the system to make it
worthwhile to sell the information about

people and buy into the system if they
believe it's widely accepted and useful.

How do we overcome this problem
the problem of aligning interests and

sharing information.

Networks will only operate when incentives
are carefully aligned to permit

information be shared
among participants and

accepted as legitimate by
the subjects of surveillance.

The market for consumer credit expanded
in part because credit reporting there.

Evolved such mechanisms credit to the
consumer change from a traditional form or

lenders such as tradesmen and storekeepers
only gave credit to those well known in

their local community now into
a global system of instantly available

open line revolving credit for
almost any consumer purchase credit

in other words shifted from face to face
practice to a highly surveilled one.

This this change took place in
a series of steps beginning in

the one nine hundred twenty S..

The one hundred twenty S. were a consumer
decade the sole widespread use of

installment purchases producer provided
loans notably for automobiles which became

mass items of mass consumption in the one
nine hundred twenty S. and vastly expanded

use of credit in retail establishments
especially department stores.

In all by nine hundred twenty nine
twenty percent of all retail sales were

made on credit.

Information serving the consumer credit
market at first consisted of diverse

forms of data assembled in
a file constructing a life

narrative investigators thought
close familiarity with that hers.

Chattel mortgage lenders for
example took careful inventory

of the homes of borrowers to recording
detail the property securing the mortgage

making repossession easier
should they default and

also to assess the character of
the bar in the most personal settings.

Since most people bought
primarily from the nearby Murcia.

It's thousands of local credit bureaus
sprung up across the nation to collect and

file this information in
addition to the in-house

files kept in retailers
credit departments.

As consumer credit data expanded
in the one nine hundred twenty S.

resellers were able to more closely track
their customers department stores used

Corning's tokens and charge cards that
linked the purchase of each customer

to the back of the house credit
ledgers perhaps the most notable and

effective of these tracking
systems was the charger plate.

First adopted by filings department store
in one nine hundred twenty nine and

then spreading to other stores in
the one nine hundred thirty S.

and one nine hundred forty S..

A metal square embossed with
the customer's name address and

account number charge plates were
run through a machine at the counter

automatically imprinting the identifying
information on the on the sales slip.

By filing the sales slips the credit part
was assured of an accurate report on

the credit use at the individual
level stores also overcame the self

interested in have Bishan that normally
would prevent them from sharing this data.

The Boston retail trade board
designed a common plate with notches

that would fit in
the machines of each store

board members agreed to
treat each other equally so

that one store did not use the commonality
of credit to steal customers from another.

The system also disciplined
freeriding by mandating that

each time a customer opened a new account
that store would conduct an independent

investigation of credit worthiness
to be shared with the other members.

And here's an example I'm talking about.

You got a little thing to put it in so

it didn't make a hole in your
pocket down charge plates and

related practices overcame the chicken and
egg problem of credit reporting

stores were willing to share information
because their customers were part of

a single monitoring network customers
had incentive to join this network.

Because it was simple and
convenient to use the same charge plate

in multiple locations the shared data
was also sufficiently rich that stores

could capture the behavioral
patterns of the customers and

able in them to construct basic
ratings of credit worthiness.

And this is the little thing you've got
right so that each charge plate notice is

useful in a number of different locations
like actually if you look up here you

can see the notches line on the chart
plate where it would go into the machine

that you were authorized the store that
you were authorized to have credit with.

When the Great Depression hit in
the one nine hundred thirty S.

stores became especially concerned about
maintaining their control of credit.

At the same time customers became
more dependent on credit for

purchases than ever before.

By nine hundred thirty eight credit
sales had risen to twenty one credit

sales had risen to twenty
five percent of all retail.

Department stores that's expanded their
cooperative information system to cast

a wider net for
this they relied on the thousands of

independent credit bureaus located in
towns and cities across the nation.

How to perform the labor intensive work of
gathering information and maintaining and

updating files files which included
information on the individual's

purchases and credit routines as
well as details of identity and

lifestyle race marital older or
divorced status drinking habits.

The highly localized the credit
bureaus exchange their data since

each of them served a home market and
did not compete outside of that territory.

Meanwhile retailers wary
about sharing customer

information that could fall into the hands
of a competitor were willing to cooperate

through the independent credit bureaus
which in turn maintained a national

network through their trade association
the associated credit bureaus of America.

It was a system that gave credence
to the trade associations motto.

Your credit bureau follows
you like a shadow.

Through the month.

Model through the one
nine hundred fifty S.

consumer credit depended on this highly
that highly diverse interconnected system

of information collection and
sharing over the next thirty years.

That would change into a single
method of standardized data

feeding into a predictive school or
why did this happen.

The idea of using statistics
in predictive fashion for

credit was actually not that new.

In the early one nine hundred fifty S.
credit managers and

professionals recognize the potential
of shifting their monitoring from

the characters to behavior or as they put
it from what a man is to what a man does.

What they need it.

However was a method of gathering the data
which depended heavily on cooperation and

sharing among the different parties.

The first part of this new system came
when consumer credit swung from open book

to revolving credit.

A store charge plates were not true.

Credit cards they were monitoring devices
that linked the customer to a line

of credit sent monthly under the stern
gaze of the store credit manager.

Accounts were to be paid off each month.

In the late one nine hundred fifty S.
However banks sense in the consumer credit

was now a profitable and
safe market develop the first open line of

revolving credit cards gradually
supplanting store based credit.

Revolving credit from Universal cards
severed the old link between information

and credit and required a new system
of information sharing and exchange.

As with the earlier credit system
the credit card confronted a chicken and

a dilemma.

Retailers would only be going to accept
credit cards if they were popular and

in general use but customers would only.

Sign up for cards if they were
widely accepted by retailers.

Credit cards had advantages
to stores particularly small

stores since they eliminated most
of the record keeping financing and

monitoring costs associated
with store based credit.

But in outsourcing credit retailers lost
control of the credit monitoring function

and hence the ability to control risk and
promote sales.

Given how much retail merchants had
invested in gaining this control and

insight into their credit
customers the bank

issued cards had to come
with some inducements.

Want to do some it was
protection against loss or

fraud by credit card users that is
the store wouldn't be liable if

a fraudulent card was used to buy
something the bank would absorb the loss.

Another was data on sales and
purchases by their customers.

Well those inducements
took care of the chicken.

But what of the a.

The answer mass mailing.

In one nine hundred fifty seven Bank of
America introduced the bank America hard

by sending it to some sixty
thousand California customers and

foregoing the usual credit checks.

It worked.

The card gain wide acceptance
by merchants and customers both

nonetheless Bank of America had to sustain
heavy losses in the early years and

other banks move cautiously
to follow the leader.

When Chase Manhattan Bank unveiled its
charge card a few years later it steered

clear of the mass mailings and
screened customers the traditional way.

As a result it failed to convince
merchants to accept the card.

The Divergent experience of the two
pioneering banks showed that credit cards

would work only if one moved aggressively
in signing up customers and merchants.

But to do so.

Banks needed in effect of system
a credit monitoring and fraud control

the capacity to monitor credit
continuously was Key Banc of America

had designed its credit card strategy
following the installation of its.

E R. and A or
electronic recording method of accounting

system a few years before to speed
check processing and account

balancing by nine hundred fifty seven
Erma and related information technology

investments yielded more sophisticated
databases that could relate that could

connect different forms of information to
paint a composite picture of the borrower.

As credit cards.

Spread outside of a single bank
customer base some means was

needed to track customers across
regions and among different financial

institutions this two required
a new form of information sharing.

Consumer market already had
a means of credit information

sharing through retailer cooperation and
credit bureaus.

When Sears Roebuck computerized its
accounting systems in the one nine hundred

fifty S. for example the trade credit
Association provided assistance to

independent bureaus in accessing this data
obtaining the monthly computer tapes and

distributing them to members after
converting their various Ledger

setups into a single format.

In the one nine hundred sixty S. the small
scattered local credit bureaus also

began to merge creating larger
databases of information on consumers.

The larger credit bureaus adopted
computer and electronic data

processing which aligned their information
structure with out of the banks so

that information could be shared
by exchanging computer tapes.

Especially important in the credit
card business with the entry of

new actors from the electronics industry
in the one nine hundred sixty S..

Taking advantage of electronic
processing power they designed credit

verification systems that allowed
merchants to query charges in real time

which meant that credit cards
could be used securely for

store purchases with little weighting
by nine hundred sixty four a credit

transaction could be confirmed or
denied within ninety seconds.

In one nine hundred sixty nine T.R.W.

acquired credit data
credit data had started.

In the one nine hundred thirty S.

as the Michigan credit association but
it also been a pioneer in the shift from

manual to computerized methods in
the one nine hundred sixty S..

T.R.W. completed this process operating
a large mainframe in Anaheim California

with fifty million names selling
credit profiles to request lenders.

But those who availed themselves of T.R.W.
as credit reporting service

were also required to share their
data on customers and borrowers.

In other cases independent credit bureaus
and regional systems manage the shift to

new methods and technologies notably
retail credit company of Georgia.

Which transformed itself into Equifax.

Computerized credit bureaus by
the one nine hundred sixty S.

had clearly moved credit reporting
through consolidation and

the sharing of information and
large computer data banks and

of course this is the structure that
pretty much evolved today although.

Believe some of them have changed their
names at this point can't keep track of

called anymore but the big three have
emerged by the one nine hundred seventy S.

dominating computerized data bank form
of credit consumer credit reporting.

Financial institutions had initially
set up their credit card approval and

authorization authorization systems
to cover their own customers but

by nine hundred sixty six two nationwide
card associations were in operation

National Bank of America card
incorporation incorporated or N B I and

into a bank card association and beyond in
one thousand nine hundred six changed its

name to Visa interbank as you can
imagine evolved into Master Card.

But their most important early
achievement was an information sharing

allowing the construction of and
rapid nationwide verification system while

also accumulating vast panels of data
that helped underwrite credit marketing.

By the mid one nine hundred seventy S.

with information on consumer credit
standardized and centralize.

In this way it became easier for
credit reporting to be done

with quantitative data and probability
based algorithms those that had been first

outlined theoretically in
the one nine hundred fifty S..

Before we got the fight go however.

One final factor should
be noted in this shift

politics particularly
the politics of representation.

Let's remember my original point
that surveillance in the in

the market place is always
a struggle over the questions

of how people are represented and
the case of credit card and

related consumer credit
reporting this was no less true.

In the nineteenth century credit reporting
had had to walk a fine line between

collecting collecting data for a lender
clients and protecting the confidentiality

of information that could be
potentially damaging to borrowers.

This was one reason that business credit
reporting was much less standardized and

used much less share data than did
consumer credit fear of running afoul of

the law and being sued because because a
credit report had caused business to fail.

In the twentieth century.

However the politics of credit shifted.

Politics and
the law pushed consumer credit

in nearly the opposite direction
as law had affected train credit

putting a premium on Wired availability
sharing and standardization.

Reacting to charges that individuals
were denied credit based on innuendo

hearsay and lifestyle choices
that were irrelevant to risk.

Congress passed the Fair Credit Reporting
Act in one nine hundred seventy.

There is not required reporting agencies
to maintain accurate files purge and

verifiable information and
permit credit seekers to examine and

correct their records.

In one nine hundred seventy four
responding to pressure by women's rights

organizations Congress added the equal
credit Opportunity Act which for

a day for bait.

Credit discrimination on
the basis of sex or marital status

undercutting the longstanding tradition of
rating women for credit lower than men.

The Act also banned the use of certain
categories pertaining to race and

ethnic status.

And finally in one thousand nine
hundred six Congress opened hearings

on information privacy
in the private sector.

While the upshot of all this
legislation in these investigations

was to make puter data banks and
behavioral or

transactional information not what
a man is but what a man does or not.

What a woman is but what a woman does
the answer to concerns about privacy and

prejudice.

If women are African-Americans were
denied credit because of credit.

Of course credit evaluators took race or
marital status or

personal lifestyle into account.

Well then credit scoring based
strictly on patterns of credit use and

behavior provided to a defense
against charges of discrimination.

If the variable records kept on
individuals by thousands of small credit

bureaus were considered unsecure and

liable to misuse that highly
standardized data available only to

those with authorized access
overseen by a small number of large

corporate entities provided greater
security and superior accuracy.

Fairness would arise by processing
more information more cheaply for

more people to lower the cost
of credit to the poor and

to minorities while increasing
transparency to protect the borrower all

of these outcomes it was argued required
to shift to new high powered computerized

credit reporting systems and the move
to quantifiable standardized forms of

information to what looked like an
objective credit score and thus we get to

the five moment which I think you'll see
is actually the tail rather than the door.

The dog of the story.

It's the outcome.

It's not the call Soulforce.

It was the availability
of standardized data

that enabled the Fair Isaac
company to develop a credit

scoring methodology you probably know
FICA was just Fair Isaac company.

Name.

If I see.

At first.

This required a labor intensive process
of sorting through the records and

ledgers of an individual company but
as more data became available in digital

form the cost of the data input fell and
the scoring models developed.

The FI Coast score also fit the political
and legal legal temper of the times

based as it was on behavior rather
than on observed characteristics.

So to its supporters the credit score
revealed much more about the credit

worthiness and about the credit
worthiness and risk that had been hard to

its supporters the credit score revealed
much about the credit worthiness and

risk that had been hidden by partial
information and rule of thumb guesswork.

But this vast abstraction also sealed its
own ignorance behind the veil of data.

Indeed the FICA score worked in part
because it was a proprietary formula

that used a common data platform but
could be sold for profit.

In this way it aligned the interests
incentives in the market for

information it eliminated the danger
of strategizing by competing

lenders that represented consumers in
a way that accorded with legal and

political mandates and
earned sellers of information a profit.

It did these things by means of
a black box an institutional

arrangement the took over
the thinking about risk in credit and

as a black box it was subject to
the laws of unintended outcomes and

unexpected consequences in the one
nine hundred ninety S. general

acceptance of the fight those score led
federal and private lenders to use it for

mortgages as well with credit ratings
attached individual mortgages

bond rating agencies were then able to
score packages of asset backed securities.

Indeed without find go Standard and
Poor's and

the other bond traders would have
lacked a methodology to rate individual

mortgages comprising the tranches
of debt obligations.

Reduction of complex individual did.

To a single universally recognized score

made securitization of mortgages easier
but it was a method that was only valid to

the extent that past behavior predicted
future behavior based as it well

was on shallow information compressed
into the FICA score it worked best when

conditions in the future were
very much like those of the past

became instead unstable in the face of
sudden changes in conditions or behavior.

The history and indeed the assumptions
behind history in the limitations of FI Co

were either lost or ignored
the consequences of ignoring that history.

Well we're living through them today.

So let me conclude in theory information
is subject to the same laws of supply and

demand as is every other commodity.

At some point the cost of
knowing outweighs the value

of what is known as sort
of equilibrium state.

But historically forces have
conspired to change this calculus.

Some of this is technological reduction in
the cost of gathering holding sorting and

deploying information.

Some of its political the uses of
information to serve certain social ends

much of it is institutional.

Information is a strategic resource and
how and where it's

provided depends not on the efficiency of
markets upon the interests of actors and

the ability of entrepreneurs to align or

more precisely to serve certain
interests and generate a profit.

The extra analogies of too much
information not enough privacy or

even unforseen disasters.

These are not their concerns.

The danger is that as we gather ever
larger quantities of information

we create black boxes that we no
longer question or interrogate.

The tendency is inherent in the process
of acquiring information for use or

what I've term surveillance we
see we have observed we sense

all sorts of things all of the time but
all of this little is remembered and.

What is remembered little is used.

No matter how delicate
the sensory mechanism.

No matter how powerful
the information processing software.

It is impossible to eliminate
the opportunity cost of attention.

Beyond a certain point there are simply
too many things to pay attention to.

And here is where black boxes operate.

They impose an order grid or

classification scheme that tells us what
is interesting and what is irrelevant.

Surveillance operates by calling
attention to some things.

While ignoring others but

this is also why surveillance can
never yield simple transparency

way in which the economic theory of
information would predict or rather why

transparency a purely instrumental value
should never be mistaken for knowledge.

Nearly a century ago the philosopher
Alfred North Whitehead identified

the fallacy of transparency.

There is a danger in clarity.

He wrote the danger of overlooking
the subtleties of truth.

More recently sociologist Charles Perot
is called attention to the inherent

a piece to be of complex systems and
black boxes.

No matter how well designed
such systems will fail.

Predictably but it unpredictable moments.

What he termed normal
accidents the slow but

steady spread of the FICA score is
an example of a normal accident at work.

The response to such breakdowns is
frequently to assume that more information

will solve the informational problem.

If the mortgage market collapses
through vendor opportunism or

companies like Enron hide their operations
behind the thicket of misinformation.

Then we automatically assume
greater surveillance and

accountability are called for.

But by enabling further surveillance such
a response creates opportunities down

the road for yet
more strategic information behavior.

My argument in short is that
the economic logic is for

surveillance to grow
even when it fails to.

Reduce the results that
it was intended for.

By this process we are creating a world
where information draws more people into

an ever tighter grid not due to it
to tell a Tarion state but do to but

but by the logic that says information
makes for transparency and

transparency promotes efficiency and
exchange beneath the surface

this information creep may be having
debilitating consequences critic.

Philip a Gray has argued that too much
information drawing people too close

together stifles economically
valuable diversity diversity

thrives whenever behavior in action do not
have to be explained and accounted for

immediately exactly the opposite of
information and surveillance systems

to try to control or regulate behavior
by means of accountability and

statistical regularity and
you can think of an example from the case.

I just gave you the way in which the FICA
score induces a form of personal self

editing so that you can engage
in the sort of behavior and

be the sort of person that then has
a higher FICA score deep diversity

biologists No require independent
choirs independent and

unrelated evolution
the sort of a diversity

that arises in distinct environments
without contact with each other.

The creative possibilities with this
sort of pagan isolation pagan for

into the Galapagos Islands where Darwin
first developed his theory of evolution to

creativity possible with this
sort of pagan isolation.

Tim Harford argues is being undercut
by nervous corporate hierarchies that

are unable to innovate stifled by the very
information that identifies each potential

risk and demands that every outcome have
a clear accountable measurable result.

Preferably sooner rather than later.

Too much standardisation too many
protocols reduce the variety of contexts

wiping out the cultural knishes and
recesses where diversity and can.

And creativity can occur.

Futile attempts to complete
information total transparency or

full calculate billeted and
accountability.

The only cut down the sort of
rich embedded life that makes for

creative innovation and change

while making inevitable the sort of system
breakdowns and strategic opportunities

that are entirely to be expected in
a world awash in too much information.

Thank you.

With some new approach.

So I'm in it so
I suggest you take two questions

and answer these are on
The Factor discussion of

one that's the development
track the development

other than your question or
was it was you know safely away.

So why write what was it about
America you know the question.

I'm so thick to so you have to split that

person here right

for being a.

Right right.

Not perfect and
I mean I thought of exactly of that.

Why was I was reading something
on the plane coming over.

So yes you know there's a there's a great
story which I start out another version of

this paper with about a couple
that were applying for

a mortgage and this was at the sort
of bottom of the of the financial.

So lenders were very very wary and

they were actually asked to provide a sort
of narrative history of why did they want

you know very old fashioned sort
of why do you want a mortgage.

Anyway you're so.

So right I mean I think the answer
is that the instability

of these kinds of algorithmic
quantitative methods does

uncertainty that's inherent I think
in trying to you know never perfect

information what you know what's what
matters what doesn't matter what plans

work with what if the context shifts all
that good stuff might become irrelevant.

So there's always this kind of the sort
of nervous logic about gathering more

and more information.

I don't think the current
Fair Credit Reporting Act and

related things necessarily provide much
of a barrier because member they were

actually they were sort of intended to
regularize what type of information.

Now clearly you know if they
thought you were discriminating

the basis of gender race you probably
would be you know run afoul of it.

One problem is you know one problem
the development immediately

was that you know lots of things
are correlated with undesirable.

You know like zip codes and so
you can collect zip code information

they actually did have an addendum
to limit use of zip code but

it just happens that the zip code that you
don't want people to get you want to give

credit to people with his ninety percent.

You know minority or something.

And so.
So this is a continual problem right.

So I don't think that the the laws
would do much to your question.

I mean you know big question for so I
don't know as much about other places but

I do know that in general credit reporting
was less developed and I think in general.

Also America has always been sort of at
the forefront I mean there may be other

some other countries but compared to say
the main European nations America has

always been sort of the forefront of
consumer credit right I mean that's our

that's the knock against American style
capitalism or such a credit driven system.

So I start.

Well I mean not in the one
nine hundred twenty S.

certainly right I mean that really
is driven by diet by businesses why

automobile makers are trying to find
a way to deal with the cost of right.

If there's an argument and then in the one
nine hundred sixty S. and seventy's.

Some of those some of
the Fair Credit acts.

You know they are kind of designed to get
credit into the hands of people who don't

have them and there's a there's
a actually a very good recent book.

Louis.
Hyman on the history of debt in credit in

America called debtor nation and you know
he actually spelled sorry about pretty

well about how there was a kind of
presumption that one way to address

poverty or one way address prejudice was
make credit more democratic rights and

it's a big assumption and but
that's I think the only moment.

I mean I think a lot of it has more
to do with business sector trying

to figure out ways to solve problems.

Interesting.

That I don't know I didn't realize
that I actually said you start over.

So none of your experience in Europe.

Yeah route from you is that so
I mean that's very interesting.

I didn't know that but I like the logical
way of thinking about that is OK What we

need to do is to integrate
these systems even better.

We do have a universal credit score
that will follow you everywhere you go.

Yeah yeah

I mean that you know I don't know whether
this is a policy driven thing or not but

your privacy issues vary widely.

And then odd ways right so so

France can actually be very strong
about protecting data privacy.

Whereas a country you might
think would be like suite.

Actually has a very open.

They believe well we're all part of
the same society why should we share.

You know we have to we
have to promote credit or

we have to promote social programs
we want lots of information.

So it might just be aligning different I
mean there are wide variations in laws and

that might be as much as anything that
effects of I'm not quite sure why going

from Europe to
the United States is an issue.

So at the end of your talk about her
correctly it sounded like you basically

are my two major problems that this credit
reporting system the finance was right

on the very good sort of the problem
which is you define your use

of interest you create a new area for
Lafayette's strategizing strategies.

But so you have more opportunities for

the other probably outline how
people shape their behaviors.

Perhaps homogenizing and
therefore detrimental to creativity which

of these two problems you see
is most significant and why.

What sort of options do we have
because every solution for

is to perpetuate what you saw
in one probably right right.

Take another step.

Yeah

really.

But you can't really separate
the two right because

your standards are partly determined
by how you evaluate information.

You know I mean can you

now.

Now.

Yeah right.

But right.

But you're set.

Right.
Your standards for creditor are.

I think this is a roundabout
way of getting your question.

I'll take it out more directly if I can.

So the standards for credit right
are determined by OK some sense

of being able to predict what you're
going to do if I give you a credit.

So partly my ability.

My sense of what I can predict
is dependent on how much

faith I have in my methodology and
my system and

how much you know how much information how
much faith I have and so you know I mean.

I don't have an answer and
it could be it could be that you like that

the most cynical interpretation would
be that all of this credit reporting.

You know figure leaf to cover

just you know the same way that
they that investment bankers right.

Employed all these X.

physicists to do quantitative
analysis to just write that that.

One way of reading that it was just
basically a way of covering up what

you say malfeasance or or lowering of
standards as they just saw an opportunity

to make money and that all this kind of
quantification has mostly a P.R. type.

I think that's probably a little
too cynical the other way would be

to say that they believe their own B.S.
You know they really believed.

We've got a law for get out.

We know what we're doing we're so

precise with our measurements that you
know we can now go into these lesser

credit markets because we have
good ways of providing prod right.

I mean it's only one part of the story
because obviously a lot of other things

going on but
you want me to take a report Yeah.

What was the question that

you in order to do with more information.

Right right.

So there too.

I mean there are two things One
is what Steve was getting at.

Right which is that which is that you
often end up with these black boxes that

you're over confident in or perhaps
even just become an opportunity for

people to pretend they
know what they're doing so

they can you know lower the standards or
whatever the other is

more what you're saying right
let's like we become obsessed.

Asked with measuring everything.

I mean I would I think that
most of the emphasis and

most of the work at a university probably
experienced this firsthand right.

The endless attempt of self
measurement and accountability.

I think is is that is
the danger of this not the sky.

It's the greater danger but
it's not the one that's in the way it's

not everyone's about we were transparency
and accountability and it's sort of like

taken as though there was no potential
negative draw consequences to that.

So I would say that I guess my
role is to is to bring to the fore

the downside of and
mostly trying to measure and account for

everything but I don't I don't
know it which is worse for me.

I mean it is as you say it is a trade off.

Yes

I would agree.

Right right right.

Yeah no I would agree totally say
that information is basically

the bottom line is not is not neutral in
the way people think of it as being true.

Yeah I mean you know it raises
the question like like you're saying.

Right.
What do people do

before they had this ability to and
listening gather information and

I mean some of it was what we
would have trouble accepting right

making decisions on the basis of
things that we think are illegitimate.

You know to make decisions on you
know on the other hand right.

I mean the type of I mean the belief
in measuring control everything.

I think is the fact first fallacy.

Right.

I mean.

Yeah but
I mean investment bankers apparently

do a lot of the same thing despite all
the quantitative analysis and all the X.

physicists say you know it's still very
much a relationship type of business.

But again it because it raises
the question well you know

all of human history.

From until fifty years ago was like that
now lots of horrible things happened in

history and lots of precious and all kinds
of other things were part of that system.

I mean societies.

You know societies have run without
trying to become information obsessed.

Or you could say that that's is that
is that what's promoted our well being.

The fact that we are so
good at information.

I don't mean that's.

Yeah.

At the very least if we're going to base
much of what we do in the corporate world

or the work world or the political world
on the idea of information will solve

all of our problems we should
have a kind of sensitivity and

awareness of you know what that means
in other words to open the black box or

to get away from the sort of Claude
Shannon notion that information is this

you know neutral process of receiving
signals that that direct your behavior and

recognize that it's information as much
a cultural product is that as everything

else that we made was

trying

to lead you where money.

You know

but.

Right.

You

are right.

Totally right.

Absolutely.

We have a long history.

Yeah absolutely.

That is definitely part of what's going on

the THE RIGHT RIGHT

RIGHT RIGHT.

Well that's you would expect.

So yeah the first point which
is in the longer paper I

do discuss that as part of
the development of the methodology.

Bear in mind that insurance companies
have discovered how to do that

you know fifty one hundred
years before discovered

you could you have to just
leave based risk this.

So actually the compliment you know if
you want to do the full history of this I

should give you a whole paper on
the history of insurance as well but so

the other thing is I mean I think right
now the answer is that they don't have

good alternatives that's I think
the upshot of that little article of

The New York Times about four
years ago about the bank like

you know just tell us you're
like they were trying to.

Back to some notion of character.

Maybe Social media is a way that
that will that will do that but

also they're just not lending.

Right.
I mean just basically you know I mean

that's the danger of of abandoning the
system is when you have nothing to really

replace it with that is you know accepted
legitimate to all parties right.

Then you do have a kind
of information freeze and

I think that's kind of the moment
we're in right now up and

I put my sorry your black ministers you

know that

house and you know something

must say yeah well bear in mind
that the score is not used so

much to make choices about individuals
borrowing but to rank the mortgages

in the collateral so that you have
the A.B.C. tranches So the side

it wasn't that wasn't they didn't know
that some of the lending was higher risk.

It's that they believe that they had a
good handle on that on what was a low risk

medium risk and high risk they
could package them together and

then while that gets in the financial
stuff that's not really my area but but

it was the idea that that not so
much they didn't know but

that they they believed that they
understood what the numbers meant and

how they could use them and
the black box saying it's not

it's not that Fair Isaac doesn't
know what's inside that box.

It's that the other actors in the system.

Treat it like a black box and they don't
have any penetration into it to know what

they're what they're using
a pleasure thank you.