Technocracy and fixed incentives: The centralised doom for ... · Monetary supply, Means of production, Governance The centralised doom of Bitcoin | Outlook 2 Bitcoin is a cryptocurrency,

Technocracy and fixed incentives:The centralised doom for BitcoinClaudio J. TessoneURPP Social Networks

WeCos - Wien | 26.10.17

Outlook 2

� Bitcoin is a cryptocurrency, a complex techno-economicsystem designed top-down

� For its correct functioning, it requires decentralisation

z From a technical point of view, which is the working state thesystem is converging to?

z Which is the economic outcome of the design put in place?

� Wrongly placed incentives at different levels foster itscentralisation

� Monetary supply, Means of production, Governance

The centralised doom of Bitcoin |

Outlook 2








Outlook 2








Economic transactions in the digital era 3

� Economic transactions haveregressed since electroniccommerce

� Credit cards� Debit transfers� Remi¤ances

� All require a central authority thatasserts identity and validates thetransactions before they occur

Centralised payment systems propagate trustworthy informa-tion because of trusted communication channels

The obvious

The centralised doom of Bitcoin | 1. Prelude






The obvious







The obvious







The obvious


Blockchain: A decentralised approach 4

In a decentralised payment network, there exist some immediatemisbehaviours by the users

� A user claims to own assets she actually does not own� A user a¤empts to use multiple times the same assets

� Risk of forging transactions / assets

How to propagate trustworthy information and create confi-dence among peers in a decentralised system with non-trustedcommunication channels?

The difficult part

The centralised doom of Bitcoin | 2. Introduction to Bitcoin

Blockchain: A decentralised approach 4

In a decentralised payment network, there exist some immediatemisbehaviours by the users

� A user claims to own assets she actually does not own� A user a¤empts to use multiple times the same assets� Risk of forging transactions / assets

How to propagate trustworthy information and create confi-dence among peers in a decentralised system with non-trustedcommunication channels?

The difficult part


Blocks 5

How to keep the information stored in the system consistent?

Transaction_ID Sender Receiver Amount Received

------------------------

2001193F04EA23 Alice Bob 5 BTC 1070BA782056FA

20020AC93670A6 Charles Dan, Alice 3 BTC, 1 BTC 1090945761BE90

2003620FB49204 Eva Faythe 2 BTC 10509587DAB3BA

20041957701BC3 Bob Sam 5 BTC 2001193F04EA23

2005AB30C94045 Wendy Oscar 5 BTC 135294594CB10EE

------------------------

� Bundle multiple verified,new transactions into ablock

� Each block buildsupon/continues theprevious one


Blocks 5



------------------------




20041957701BC3 Bob Sam 5 BTC 2001193F04EA23


------------------------




Blocks 5



------------------------




20041957701BC3 Bob Sam 5 BTC 2001193F04EA23


------------------------




Blocks 5



------------------------

20010BA4569035 MINE Gandalf 12.5 BTC



20030DAD029445 MINE Merlin 12.5 BTC


20041957701BC3 Bob Sam 5 BTC 2001193F04EA23


------------------------

Why should someone verify?

� Allow the creator of theblock (miner) to add atransaction issued to himwithout an actual sender

� Plus fees paid by alltransaction senders


Blocks 5



------------------------

20010BA4569035 MINE Gandalf 12.5 BTC



20030DAD029445 MINE Merlin 12.5 BTC


20041957701BC3 Bob Sam 5 BTC 2001193F04EA23


------------------------

This is how Bitcoins enterinto the system

Money supply


Blockchain 6

Block height

� An append-only sequential data structure

� New blocks can only be appended at the end of the chain� To change a block in the middle of the chain, all subsequent

blocks need to be changed� But the blockchain is replicated in all network nodes!� History redundancy makes it resilient to a¤acks


Blockchain 6

Block height

� An append-only sequential data structure� New blocks can only be appended at the end of the chain

� To change a block in the middle of the chain, all subsequentblocks need to be changed

� But the blockchain is replicated in all network nodes!� History redundancy makes it resilient to a¤acks


Blockchain 6

Block height

� An append-only sequential data structure� New blocks can only be appended at the end of the chain� To change a block in the middle of the chain, all subsequent

blocks need to be changed� But the blockchain is replicated in all network nodes!

� History redundancy makes it resilient to a¤acks


Blockchain 6

Block height

� An append-only sequential data structure� New blocks can only be appended at the end of the chain� To change a block in the middle of the chain, all subsequent

blocks need to be changed� But the blockchain is replicated in all network nodes!� History redundancy makes it resilient to a¤acks


Mining 7

f (prev block, verified tx, x) ≤ ∆

� ... Make it (computationally) costly to add new information,cheap to verify it

� Miners (verifiers) must solve by a difficult mathematicalproblem. The solution is by brute force

� All the miners do the same� The first that by chance finds a solution, broadcasts it to others� Why? He wants to ensure the creation transaction goes to him


Mining 7






Mining 7




� All the miners do the same� The first that by chance finds a solution, broadcasts it to others

� Why? He wants to ensure the creation transaction goes to him


Mining 7






Currency supply 8

2010 2011 2012 2013 2014 2015 2016Year

0

2

4

6

8

10

12

14

16

Tota

lSup

ply×

106

Bitc

oin

Complete SupplyCirculating Supply

� Bitcoin supply designed to increaselinearly with time

� However, circulating supply issizeable smaller

� Anyone can participate of the money supply, i.e. it isdecentralised

� Monetary supply policy: arbitrary decision

� Predictable: Circulating money (piece-wise) linearly over time� Every two years, supply rate halves� Reason? Incentivise early adoption even in absence of

transactions


Currency supply 8

2010 2011 2012 2013 2014 2015 2016Year

0

2

4

6

8

10

12

14

16

Tota

lSup

ply×

106

Bitc

oin





� Monetary supply policy: arbitrary decision

� Predictable: Circulating money (piece-wise) linearly over time� Every two years, supply rate halves� Reason? Incentivise early adoption even in absence of

transactions


Currency supply 8

2010 2011 2012 2013 2014 2015 2016Year

0

2

4

6

8

10

12

14

16

Tota

lSup

ply×

106

Bitc

oin





� Monetary supply policy: arbitrary decision� Predictable: Circulating money (piece-wise) linearly over time� Every two years, supply rate halves

� Reason? Incentivise early adoption even in absence oftransactions


Currency supply 8

2010 2011 2012 2013 2014 2015 2016Year

0

2

4

6

8

10

12

14

16

Tota

lSup

ply×

106

Bitc

oin





� Monetary supply policy: arbitrary decision� Predictable: Circulating money (piece-wise) linearly over time� Every two years, supply rate halves� Reason? Incentivise early adoption even in absence of

transactions


The effect of the monetary policy 9

2009 2010 2011 2012 2013 2014 2015 2016Year

104

106

108

1010

1012

1014

1016

1018

Has

hR

ate

[Has

h/se

c]

BTCTop 500

� This created an arms race to devote more resources to mining� Nowadays, the computational power devoted to mine bitcoins

is 104 the combined power of Top500


Incentive scheme 10

150000 200000 250000 300000 350000 400000 450000block height

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

Thei

l Ind

ex

0 10 20 30 40Rank

5

4

3

2

1

0

log(

1-P

(Ran

k) )

Mining Pool Participation47006141728236214330699425185

� The incentive scheme was tailored to foster early adoption� But the outcome is that everybody has an incentive to mine

now instead of in the future

� Only large-scale investors can afford nowadays Bitcoin mining


Incentive scheme 10

150000 200000 250000 300000 350000 400000 450000block height

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

Thei

l Ind

ex

0 10 20 30 40Rank

5

4

3

2

1

0

log(

1-P

(Ran

k) )

Mining Pool Participation47006141728236214330699425185

� The incentive scheme was tailored to foster early adoption� But the outcome is that everybody has an incentive to mine

now instead of in the future� Only large-scale investors can afford nowadays Bitcoin mining


Bitcoin users and miners 11

addr_03 1 BTCaddr_04 3 BTC addr_10 3 BTC

addr_11 0 BTCaddr_12 4 BTC

� Set of individuals running a client and connected to aPeer-to-Peer network

� Each user in the verification network has a copy of theblockchain

� Each user in the network has a wallet, with many addresses

Which wallet hosts which addresses is a priori unknownAnonymity in the network


Bitcoin users and miners 11

addr_03 1 BTCaddr_04 3 BTC addr_10 3 BTC


� Set of individuals running a client and connected to aPeer-to-Peer network

� Each user in the verification network has a copy of theblockchain

� Each user in the network has a wallet, with many addresses

Which wallet hosts which addresses is a priori unknownAnonymity in the network


The verification dynamics 12addr_01


� Users issue transactions, which build a transaction network

� The transaction is broadcasted to the network� A node in the P2P mines a new block and includes the

transaction into the blockchain� The block (with the TX) is added to all the blockchains in the

P2P network� The transaction occurred




� Users issue transactions, which build a transaction network

� The transaction is broadcasted to the network� A node in the P2P mines a new block and includes the






� Users issue transactions, which build a transaction network� The transaction is broadcasted to the network

� A node in the P2P mines a new block and includes thetransaction into the blockchain

� The block (with the TX) is added to all the blockchains in theP2P network

� The transaction occurred


















� Users issue transactions, which build a transaction network� The transaction is broadcasted to the network� A node in the P2P mines a new block and includes the

transaction into the blockchain








P2P network



The verification dynamics 12addr_01 8 BTC






Bitcoin user detection 13

� The detection is based on identifying addresses that belong tothe same user (wallets)

� Incorrect aggregation is reduced as much as possibleThe picture I will provide is much more egalitarian than realworld

The centralised doom of Bitcoin | 3. Economic flow






addr_01addr_02addr_03

addr_11addr_12

addr_13addr_14





100 101 102 103 104 105 106 107

No. Addresses

10−6

10−5

10−4

10−3

10−2

10−1

100

CC

DF

No.

Add

ress

es

power-law fit α = 2.14power-law fit α = 1.96power-law fit α = 2.22power-law fit α = 2.22power-law fit α = 2.32Year 2010Year 2011Year 2012Year 2013Year 2014

10−4 10−3 10−2 10−1 100 101 102 103 104

Bitcoin Balance

10−4

10−3

10−2

10−1

100

CC

DF

Bitc

oin

Bal

ance

UnfilteredMinersPure users

� Participants of the transactions can be: Miners, End-Users

� Very broad distribution in the number of addresses per user:Detection is fundamental to understand economic activity inthe Bitcoin economy

� Very broad wealth distribution: Much more unequal thanunequal contries. Gini index = 0.9 (2015)



100 101 102 103 104 105 106 107

No. Addresses

10−6

10−5

10−4

10−3

10−2

10−1

100

CC

DF

No.

Add

ress

es

power-law fit α = 2.14power-law fit α = 1.96power-law fit α = 2.22power-law fit α = 2.22power-law fit α = 2.32Year 2010Year 2011Year 2012Year 2013Year 2014

10−4 10−3 10−2 10−1 100 101 102 103 104

Bitcoin Balance

10−4

10−3

10−2

10−1

100

CC

DF

Bitc

oin

Bal

ance

UnfilteredMinersPure users

� Participants of the transactions can be: Miners, End-Users� Very broad distribution in the number of addresses per user:

Detection is fundamental to understand economic activity inthe Bitcoin economy

� Very broad wealth distribution: Much more unequal thanunequal contries. Gini index = 0.9 (2015)


Bitcoin adoption 14

2010 2011 2012 2013 2014 2015 2016Year

1

10

102

103

104

105

106

107

N(t

),N

i(t)

End-usersMinersTotal

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

Use

rsha

re

End-UsersMiners

� User adoption shows two regimes of adoption: early adopters–miners–, end-users

� The relative proportion of users involved in supply becomesnegligible


Bitcoin adoption 14

2010 2011 2012 2013 2014 2015 2016Year

1

10

102

103

104

105

106

107

N(t

),N

i(t)

End-usersMinersTotal

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

Use

rsha

re

End-UsersMiners

� User adoption shows two regimes of adoption: early adopters–miners–, end-users

� The relative proportion of users involved in supply becomesnegligible


Bitcoin transaction network 15We constructed the network encompassing all transactions inthe Bitcoin economy in a given time window

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

N(1

)n/N

n

Bitcoin

� A�er users’ usage: a giant connected component emerges

� Miners show larger degree and lower clustering than end-users

Topologically, end-users and miners occupy different locationsin the network

Network location


Bitcoin transaction network 15We constructed the network encompassing all transactions inthe Bitcoin economy in a given time window

2010 2011 2012 2013 2014 2015 2016Year

100

101

102

103

104

105

k(o)

i,max

TotalMinerEnd-user

2010 2011 2012 2013 2014 2015 2016Year

0

10

20

30

40

50

〈k(o

)M〉−〈k

(o)

U〉

2010 2011 2012 2013 2014 2015 2016Year

−0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Ci

MinerEnd-user

� A�er users’ usage: a giant connected component emerges� Miners show larger degree and lower clustering than end-users

Topologically, end-users and miners occupy different locationsin the network

Network location


Description of the Bitcoin economy 16

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

π iR

elat

ive

Wea

lth

i = Useri = Miner

� Miners still hold a sizeable portionof the total amount of bitcoins incirculation

� Transition in the transactionbehaviour. Nowadays, mosttransactions are

� miner→ end-user� end-user→ end-user


Description of the Bitcoin economy 16

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

1.2

T i→

j/∑

j′T i→

j′

Miner to MinerMiner to end-userEnd-user to minerEnd-user to End-user

� Miners still hold a sizeable portionof the total amount of bitcoins incirculation

� Transition in the transactionbehaviour. Nowadays, mosttransactions are

� miner→ end-user� end-user→ end-user


Implications 17

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

Use

rsha

re

End-UsersMiners

2010 2011 2012 2013 2014 2015 2016Year

0.0

0.2

0.4

0.6

0.8

1.0

π iR

elat

ive

Wea

lth

i = Useri = Miner

2010 2011 2012 2013 2014 2015 2016Year

0

10

20

30

40

50

〈k(o

)M〉−〈k

(o)

U〉

2010 2011 2012 2013 2014 2015 2016Year

−0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Ci

MinerEnd-user


Implications 17

� Centralisation of supply� wealth inequality

Bitcoin has evolved into a state that works in a stark contrastwith its design principles

Because of misplaced incentives


The effect of Governance 18

2009 2010 2011 2012 2013 2014 2015 2016 2017Year

0

500

1000

1500

2000

2500

3000

Mar

ketP

rice

[USD

]

2011 2012 2013 2014 2015 2016Year

106

107

108

109

1010

Mar

ketC

ap[U

SD]

� The main developers, and other stakeholders have their ownincentive to maintain the status quo

� Changes of monetary policy would damage tremendously theprice

� But this is something that weakens the system!!!

The centralised doom of Bitcoin | 4. Governance


2009 2010 2011 2012 2013 2014 2015 2016 2017Year

0

500

1000

1500

2000

2500

3000

Mar

ketP

rice

[USD

]

2011 2012 2013 2014 2015 2016Year

106

107

108

109

1010

Mar

ketC

ap[U

SD]






2009 2010 2011 2012 2013 2014 2015 2016 2017Year

0

500

1000

1500

2000

2500

3000

Mar

ketP

rice

[USD

]

2011 2012 2013 2014 2015 2016Year

106

107

108

109

1010

Mar

ketC

ap[U

SD]





The blockchain mechanics 19

The centralised doom of Bitcoin | 5. Mining concentration




� Nodes mine searchingfor a new block

� Diffusion of blockswithin the network

� Consensus is unknownto nodes

� Diffusion does notoccur over nodeswhich have the sameheight


The importance of mining diversity 20

Diverse

Concentrated

� When mining becomes concentrated, large portions of theverification is performed by the same node

� It can decide which transactions are accepted and which not� Others lose incentive to participate of verification (no reward)� Vicious cycle


Results: Blockchain and diffusion 21

10−1 100 101 102 103 104 105

λnd

0.10.20.30.40.50.60.70.8

Nbr

anch

es/N

Bnum nodes=1000

10−1 100 101 102 103 104 105

λnd

0.2

0.4

0.6

0.8

1.0

Nor

ph/N

tot

num nodes=1000

nodes conn=3nodes conn=4nodes conn=8nodes conn=16nodes conn=32nodes conn=64

ηi ∼ Exponential(µ = 1). Network: ER

Blocktree properties:� Three regions: Full performance, intermediate performance,

congestion

� Number of branches decreases for large network delay� In the congested region, orphaned branches are longer

(security collapse)� Region of intermediate performance widens with system size



10−1 100 101 102 103 104 105

λnd

0.10.20.30.40.50.60.70.8

Nbr

anch

es/N

Bnum nodes=1000

10−1 100 101 102 103 104 105

λnd

0.2

0.4

0.6

0.8

1.0

Nor

ph/N

tot

num nodes=1000




congestion� Number of branches decreases for large network delay� In the congested region, orphaned branches are longer

(security collapse)

� Region of intermediate performance widens with system size



10−1 100 101 102 103 104 105

λnd

0.10.20.30.40.50.60.70.8

Nbr

anch

es/N

Bnum nodes=1000

10−1 100 101 102 103 104 105

λnd

0.2

0.4

0.6

0.8

1.0

Nor

ph/N

tot

num nodes=1000




congestion� Number of branches decreases for large network delay� In the congested region, orphaned branches are longer

(security collapse)� Region of intermediate performance widens with system size


Take-home messages 22

� Blockchain is indeed a technological breakthrough� Incentive placement is indeed crucial� Outcome far different from expected

� This is a complex socio-technical system. Design cannot betechnocratic, but co-evolutionary

The centralised doom of Bitcoin | 6. Take-home messages

Take-home messages 22

� Blockchain is indeed a technological breakthrough� Incentive placement is indeed crucial� Outcome far different from expected� This is a complex socio-technical system. Design cannot be

technocratic, but co-evolutionary

The centralised doom of Bitcoin | 6. Take-home messages

Claudio J. Tessone

URPP Social Networks

B [email protected]

m h¤p://www.socialnetworks.uzh.ch

Documents

Technocracy and fixed incentives: The centralised doom for ... · Monetary supply, Means of production, Governance The centralised doom of Bitcoin | Outlook 2 Bitcoin is a cryptocurrency,