The Digital Privacy Paradox Small Money Small Costs Small Talk

Susan Athey Stanford

Christian Catalini MIT

Catherine Tucker MIT

PrivacyCon February 28th 2018

Why Do People Say They Care About Privacy and Government Surveillancehellip

= lnbox

- 991SCUrgteebillnctticali

- ~middotmiddotarY11r~lllnotflcAII

=vamprrouce E-Bill Schedlled Pamffll - Jur EbullBill CuampioTW tlt1ur r-KulTirg plffri t

tiilH1UJ PijfflA frL P11iliJI - --middotmiddotmiddotmiddotmiddotmiddot fcnbullbullthHI mw~~ ------ Frtar 3 t1W ~ micro

e orepl~middot

- KlfFi)I)

J ii-9middotfapruso CapsuiH

=)~tee by Feb 22 middot S99~ ~ lielmiddototler

0-r sjan Ca12 l i bull yc1middotmcbil~ bearding pass f or US Air as fl ih1 f-orr )CA IC BOS is

rgtfpe111nii Rmlnrlltffor eHR STIANCATltI It-JI -ConfrmRton mrl (

bull

JS A ayFl~ht2101

os-oce Ja 22 553AA- 725 Al1

But Then Share Personal Data with Firmshellip

Categories Social Networking

Free Paid Top Grossing

1 Messenger Social Networking

(903)

Signal - Private Messenger Social Networking

(166)

[+ OPEN I

OPEN

hellipand Knowingly Use Technologies That Do Not Safeguard their Privacy

Unpacking The Privacy Paradox bull What can we learn from the MIT Digital Currency Experiment about these apparent privacy paradoxes bull This is not a paper about digital currency per sehellip although a keypromise of blockchain and cryptocurrencies is improved digital privacy bull Surprisingly little economics work on the malleability of privacypreferences (Acquisti is the exception) bull Background policy question - What are we trying to regulate - What are we trying to protect consumers from

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Why Do People Say They Care About Privacy and Government Surveillancehellip

= lnbox

- 991SCUrgteebillnctticali

- ~middotmiddotarY11r~lllnotflcAII

=vamprrouce E-Bill Schedlled Pamffll - Jur EbullBill CuampioTW tlt1ur r-KulTirg plffri t

tiilH1UJ PijfflA frL P11iliJI - --middotmiddotmiddotmiddotmiddotmiddot fcnbullbullthHI mw~~ ------ Frtar 3 t1W ~ micro

e orepl~middot

- KlfFi)I)

J ii-9middotfapruso CapsuiH

=)~tee by Feb 22 middot S99~ ~ lielmiddototler

0-r sjan Ca12 l i bull yc1middotmcbil~ bearding pass f or US Air as fl ih1 f-orr )CA IC BOS is

rgtfpe111nii Rmlnrlltffor eHR STIANCATltI It-JI -ConfrmRton mrl (

bull

JS A ayFl~ht2101

os-oce Ja 22 553AA- 725 Al1

But Then Share Personal Data with Firmshellip

Categories Social Networking

Free Paid Top Grossing

1 Messenger Social Networking

(903)

Signal - Private Messenger Social Networking

(166)

[+ OPEN I

OPEN

hellipand Knowingly Use Technologies That Do Not Safeguard their Privacy

Unpacking The Privacy Paradox bull What can we learn from the MIT Digital Currency Experiment about these apparent privacy paradoxes bull This is not a paper about digital currency per sehellip although a keypromise of blockchain and cryptocurrencies is improved digital privacy bull Surprisingly little economics work on the malleability of privacypreferences (Acquisti is the exception) bull Background policy question - What are we trying to regulate - What are we trying to protect consumers from

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

= lnbox

- 991SCUrgteebillnctticali

- ~middotmiddotarY11r~lllnotflcAII

=vamprrouce E-Bill Schedlled Pamffll - Jur EbullBill CuampioTW tlt1ur r-KulTirg plffri t

tiilH1UJ PijfflA frL P11iliJI - --middotmiddotmiddotmiddotmiddotmiddot fcnbullbullthHI mw~~ ------ Frtar 3 t1W ~ micro

e orepl~middot

- KlfFi)I)

J ii-9middotfapruso CapsuiH

=)~tee by Feb 22 middot S99~ ~ lielmiddototler

0-r sjan Ca12 l i bull yc1middotmcbil~ bearding pass f or US Air as fl ih1 f-orr )CA IC BOS is

rgtfpe111nii Rmlnrlltffor eHR STIANCATltI It-JI -ConfrmRton mrl (

bull

JS A ayFl~ht2101

os-oce Ja 22 553AA- 725 Al1

But Then Share Personal Data with Firmshellip

Categories Social Networking

Free Paid Top Grossing

1 Messenger Social Networking

(903)

Signal - Private Messenger Social Networking

(166)

[+ OPEN I

OPEN

hellipand Knowingly Use Technologies That Do Not Safeguard their Privacy

Unpacking The Privacy Paradox bull What can we learn from the MIT Digital Currency Experiment about these apparent privacy paradoxes bull This is not a paper about digital currency per sehellip although a keypromise of blockchain and cryptocurrencies is improved digital privacy bull Surprisingly little economics work on the malleability of privacypreferences (Acquisti is the exception) bull Background policy question - What are we trying to regulate - What are we trying to protect consumers from

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Categories Social Networking

Free Paid Top Grossing

1 Messenger Social Networking

(903)

Signal - Private Messenger Social Networking

(166)

[+ OPEN I

OPEN

hellipand Knowingly Use Technologies That Do Not Safeguard their Privacy

Unpacking The Privacy Paradox bull What can we learn from the MIT Digital Currency Experiment about these apparent privacy paradoxes bull This is not a paper about digital currency per sehellip although a keypromise of blockchain and cryptocurrencies is improved digital privacy bull Surprisingly little economics work on the malleability of privacypreferences (Acquisti is the exception) bull Background policy question - What are we trying to regulate - What are we trying to protect consumers from

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Unpacking The Privacy Paradox bull What can we learn from the MIT Digital Currency Experiment about these apparent privacy paradoxes bull This is not a paper about digital currency per sehellip although a keypromise of blockchain and cryptocurrencies is improved digital privacy bull Surprisingly little economics work on the malleability of privacypreferences (Acquisti is the exception) bull Background policy question - What are we trying to regulate - What are we trying to protect consumers from

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

The MIT Digital Currency Experiment bull ~5000 students eligible Survey on privacy preferences digital payments etc

bull 3 randomizations and related privacy choices and outcomes Students had to selectgenerate a digital wallet learn about encryption

bull Key Findings small incentives small costs and small talk lead participants to ignore their privacy preferences (both stated and past revealed)

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

1 Small Money Whereas people say they care about privacy they are willing to relinquish private data quite easilyhellip

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

We asked for the emails of the studentsrsquo closest friends However it turns out that this is some of the personal data that is

considered most privatehellip

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

bull Most Sometimes I

Your Social Security number ~

lt Your list of contacts gt 10oo I

28 68

27 GE The content of your emails 10oo 33

40

52

The names of your children Vil 11

Your pholosvideos 15

44

54 Yoocorrem loooli whM ~~= ~~~ ~ 15

The websites you go to what you look at and what you ~~-2= ----

48

46

Your location

The hislorv of what youve purchased

Your phone number

Your home address

Your household Income

Your full name

Your age or date of birth

Your email address

Your gender

r 17

~ ~ I

18

23

26

53oo

53

62

36

60oo

57

53

50

47

ca

36 1

At Least in Surveyshellip

Source National Cyber Security Alliance (2014) N=493

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Tl News Video Event s Cru nchBase Search

BldiCJr Cu Fonin AIU COO f~1tulJJ lt u~un To 5fcJk At D1~rupl lo iJu11

bullbull The Wrong Way Path Uploads iOS Users Address Books Without Permission foampted eb 7 01

tyctiris Vefazco r ltttrimmiddotoucr

Wta started as a bit of aimless 1inke~ing for de eloper Arun Thampl ultlrrael) uneartha-0 simtl hing very surprbing about persormiddotal lifeshy

~r i1ring ~cr1-rbull Path tr a fan of hf 1pp ThrTlf)i ook it tprn hiM~P f rn kgtok it the API

cal s that 1he app made tc Paths erlce middot d otmo thc1t hi~ em ire c1ddr~s lgtook including

ull r lmc~ emails nd phine numb~rs) wJs

h Pini sltlt =i plis t to io=irr

Linkedln will pay $13M for sending those awful emails by J At1 John RohAt1~ ct)AfflohnmhPrl~ OCTOBERS 2015 823 AM EDT ----IIEII

Linked(ffl

hellip in Print and Lawsuits

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Random 50 of the Sample Pizza In Exchange For Their Closest Friendsrsquo Emails

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Ask Ask+ Incentive

Do You Protect the Privacy of Your Friends

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

(1) (2) (3) (4) (5) (6) (7) VARIABLES All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid All Invalid

Ask + Incentive (AI) -00285 -00268 -00224 -00249 -00245 -00327 -00332 (00059) (00066) (00076) (00068) (00074) (00060) (00066)

AI x Above Median -00045 Privacy Public (00079)

AI x Above Median -00110 Privacy Intermediary (00081)

AI x Above Median -00085 Privacy Government (00078)

AI x Above Median -00080 Trust Government (00079)

Al x Above Median 00153 Trust Startup (00100)

AI x Above Median 00105 Trust Retailer (00081)

Constant 00531 00531 00531 00531 00531 00531 00531 (00057) (00057) (00057) (00057) (00057) (00057) (00057)

Observations 3086 3086 3086 3086 3086 3086 3086 R-squared 0005 0006 0006 0006 0006 0006 0006

54 decrease

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

_ __ _ Others Privacy Sensitive Dorms

Ask Ask + Incentive

Heterogeneous Effects None bull Gender bull Race bull Citizenship Status bull Year of Study bull Coding Ability bull Digital Wallet Preference bull Will the Bitcoin price go down bull Are you a Mac or a PC bull Privacy Sensitive

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

2 Small Costs Small frictions in the process of selecting a new technology have large effects on privacy choices

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

motflla w- Google -Firefox Internet pPera- Safari_Q Explorers browser chrome

Your online security is Internet Explorer is the The powerfLI and easy- Google Chrome A fast Safari for Windows from Firefoxs top priority worlds most w iaely a-use Web browser Try new browser Made for Apple the worlds most Firefox is free and made used browser designed he only browser with everyone innovative browser

to help you get the moSl by Microsoft with yoi in Opera Turbo technology out of the web mine and speed up your

Internet connection

Install I ln5tall

Tell me more Tell me more I I Tell me more

Inspired by the Microsoft Windows ldquoBrowser Ballotrdquo Screen

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

MrT 8tcoln Pmjecl

To receive your bitcoin you will need to have a Bitcoin wallet set up and share the public address with

us

~

II -m

--middot

+ rtkatwdtor

-to-tfyyau wti)OUNndN -HaleocNtlOalYOI -

+ nony __

NCUitydlhelrlCCOlllhl --middot-middotdeg- Che~ aiecany lnfraslNcrure ___ ---+ OlY __ +--middot--degObullfN

110 11 )Qa-1ranUCtOII -sworaa n NCIft

-Egtn-- -Ofyyau-yau ---

HMacxalObullYQ -Elll9P11Mecoulel _y_ ---+bullmiddot-shyPlrtill10 aentity you when yal Nod cu Hu~toal)Ol -

+---- IOngMfN PIMWOfdlartMCUlI HIie ~ ~you aixxu1l lhuugh bmwM ~

+no___ NCUntyOltntilIOOOUltt

- Vcu dla b 9Klft a Ieng a anyaf _ __ _ __ _

+ Wyou tire lockedout of ~Wllllel

)OU rT9) t IO wbtt k by

---IYIOh-Y - lMccmpenyOOIJld tfeetofOAJI 6CCOllll0tOOOIAolbLMleN

How cat1 I acceu US -

+ ampy-10ono ltolIIUSdalwl(_hm - -- --

_

+----- OlY-middot- +No ff you ntockea OUI Ol)QJlWIIM Ot ~ ot ~ loee~IOJOIJrQClftClllteflifthOlII

I OfOOlf baclclCl It w I be lllIQ0MIIUt

IOr)OUIO~IOCelt

+ 10ng __ __ _ IIOONll~OfyltNlwdtt

-tr)IDU locudoulal)IDIJr~it may be krlpoattl for OLI io rga +~ --shy)OJtfirbtatiltlO~llbt profflg yot lmntity to Noompawy

+llNbllCOlftbullmlblCIUNd a

Dtlrldlamp1ClildlNrdlQOllt

lnluranco ffWlfcoYetJOU kleo Tho _ __

8CCCU4CWfilOOJtoC~

+fffgt-deg deg U8 OOMlt l O deg -ccxnaedilCllrd --

bull Wallets differ in terms of how much they expose a userrsquos transactions to an intermediary the government and the public

bull Random Order bull Increased Transparency 50 randomly exposed to additional information on key trade-offs

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

(1) (2) (3) (4) (5) (6) Maximized Maximized Maximized Maximized Maximized Maximized

Privacy Privacy Privacy Privacy Privacy Privacy from the from the from the from the from the from the

VARIABLES Public Public Intermediary Intermediary Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Iransparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082 -00391 lransparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observat ions 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

-17 -46 -32

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

VARIABLES

(1) Maximized

Privacy from the Public

(2) Maximized

Privacy from the Public

(3) (4) Maximized Maximized

Privacy Privacy from the from the

Intermediary Intermediary

(5) (6) Maximized Maximized

Privacy Privacy from the from the

Government Government

Best Wallet Not 1st -01301 -01761 -01320 -01839 -00379 -00164 (00161) (00240) (00147) (00225) (00130) (00165)

Increased Transparency 00902 -01845 00644 (00210) (00225) (00231)

Best Not 1st x Increased 00966 01082_ -00391 Transparency (00318) (00290) (00257)

Constant 07777 07313 02867 03779 01199 00859 (00105) (00161) (00115) (00174) (00117) (00148)

Observations 3108 3108 3108 3108 3108 3108 R-squared 0021 0047 0025 0054 0003 0008

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

unrollme Instant y see

Clean up your inbox of- o rbullpt ma be eas y from

whiltCW )Ou don t wint

Were trusted by millions of happy users

CBSreg TIME llfehacker II

3 Small Talk

orthogonal talkhellip Can easily distract people from escaping surveillance with

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

PGP is a program that gives your electronic mail something that it otherwise doesnt have Privacy It does this by encrypUng your mail so that nobody but the intended person can read it When encrypted the message looks like a meaningless jumble of random characters PGP has proven itself quite capable of resisting even the most sophisticated forms of analysis aimed at reading the encrypted text

PGP can also be used to apply a digital signature to a message without encrypting It This Is normally used In public postings where you dont want to hide what you are saying but rather want to allow others to confirm that the message actually came from you Once a digital signature Is created it is impossible for anyone to mOdify either the message or the signature without the modification being detected by PGP

PGP makes sure that any communication between you and someone else can only be read by the sender and the receiver

End-to-end encryption creates a sort of digital tunnel between the senders and receivers of e-mails - helping to llteep the prying eyes of everyone from governments to Internet service providers and mail providers themselves from seeing the content of messagesmiddot (source

bttoJlwwwwashjngtonoostcombloosthe-swtchfwp20140807vahoo-to-roJe-outsectnd-t o-end-encrvotion-ootion-for-all-yahoo-roail-users-10-20150

Although the technology has been available for a while It Is catching traction among those concerned about privacy and security Both Yahool and Google have recently announced plans to integrate PGP into their email services (For more information

httpJenwikipediaorqfwlkVPcettv Good Pnvacvl 1

EncryptionRandomization

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

1) (2) (3) Escaping Escaping Escaping

Surveillance Surveillance Surveillance from the from the from the

VARIABLES Public Intermediary Government

Encryption -00105 -00361 -00331 Randomization (00061) (00250) (00182)

Constant 00229 06921 08820 (00049) (00175) (00122)

0 bservations 1882 1410 1410 R-squared 0002 0001 0002

Effect of Small Talk on Privacy

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Conclusions (I) bull Although we study a very specific setting and technologyhellipwe are able to document possibly more general digital privacy paradoxes using field experiment data - People say they care about privacy but in the face of lsquosmall moneyrsquo they are willing to share very sensitive data

- People say they want privacy-protecting technologies but are put offfrom using them by very lsquosmall costsrsquo such as scrolling down a list

- People worry about government surveillance but stop worrying when receive reassurance lsquosmall talkrsquo about something completely orthogonal

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Conclusions (II) bull Debate in the US and Europe about the regulation of personal data bull Sensitive data is becoming more valuable because of machine learning and prediction bull Survey-based privacy research may not be best grounding for policy bull Policy focus has been on restricting data access through use of consent mechanismshellipbut this approach has implicit in it assumptions about the lack of malleability of usersrsquo preferences bull We document multiple potential market failures in digital privacy

Thank you

Christian Catalini blockchainmitedu catalinimitedu

Thank you

Christian Catalini blockchainmitedu catalinimitedu